Home

HSPICE RF User Guide - RFIC Group @ Fudan University

1. HI HI_thresh LO_thresh 7 i LO 1 5ns lt t lt 2 2ns For syntax and description of this statement see CHECK RISE in the HSPICE and HSPICE RF Command Reference You use the CHECK FALL statement to verify that a fall time occurs within the specified window of time For example CHECK FALL min max nodel lt node2 gt lt hi lo hi_th lo _th gt For syntax and description of this statement see CHECK FALL in the HSPICE and HSPICE RF Command Reference Edge Timing Verification The edge condition verifies that a triggering event provokes an appropriate RISE or FALL action within the specified time window You use the CHECK EDGE statement to verify this condition For example CHECK EDGE ref RISE FALL min max RISE FALL nodel lt node2 gt lt hi lo hi_th low_th gt Figure 51 EDGE Example voutA CLK HI HI_thresh LO_thresh LO Ins lt t lt 3ns HSPICE RF User Guide 413 Z 2007 03 Chapter 16 Advanced Features Using CHECK Statements For syntax and description of this statement see CHECK EDGE in the HSPICE and HSPICE RF Command Reference Setup and Hold Verification You use the CHECK SETUP and CHECK HOLD statements to ensure that specified signals do not switch for a specified period of time For example CHECK SETUP ref RISE FALL duration RISE FALL nodel lt node2 gt lt hi lo hi_th low _th gt CHECK HOLD ref RISE F
2. 0000 Analytical Model Types 0 0000 cece eee eee Simulating Circuit and Model Temperatures Temperature Analysis 0 000 e eee eee eee TEMP Statement nananana 000 cee eee Worst Case Analysis 60000 cee eee eee Model Skew Parameters 00 000 cee euee Monte Carlo Analysis 0 00 0 cece eee eee ee Monte Carlo Setup 1 0 0 2 000 c eee ee Monte Carlo Output 0 000 eee ee PARAM Distribution Function 000005 Monte Carlo Parameter Distribution Monte Carlo Examples 000 0c eee eens Worst Case and Monte Carlo Sweep Example Transient Sigma Sweep Results 05 Monte Carlo Results 0 0 0 Simulating the Effects of Global and Local Variations with Monte Carlo Variations Specified on Geometrical Instance Parameters 315 317 317 319 322 324 326 327 333 344 346 346 347 349 349 349 350 351 352 354 354 354 359 360 362 362 364 365 371 373 374 381 381 Contents Variations Specified in the Context of Subcircuits Variations on a Model Parameter Using a Local Model in Subcircuit Indirect Variations on a Model Parameter 00 02s Variations Specified on Model Parameters 0000 ee aee Variations Specified Using DEV and LOT 00055 Combinations of Varia
3. Figure 46 superimposes the required part grades for product sales onto the Monte Carlo plot This example uses a 250 ps delay and 6 0 mW power dissipation to determine the four binning grades HSPICE RF User Guide 379 Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Worst Case and Monte Carlo Sweep Example Figure 46 Speed Power Yield Estimation Monte Carla Results i m_delayt m_powerm delay a Sorting the results from inv mt1 yields Bini 18 Bin2 30 Bin3 31 Bin4 21 If this circuit is representative of the entire chip then the present yield should be 18 for the premium Bin 1 parts assuming variations in process parameters as specified in the netlist Of course this example only shows the principle on how to analyze the Monte Carlo results there is no market for a device with two of these inverters 380 HSPICE RF User Guide Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Simulating the Effects of Global and Local Variations with Monte Carlo Simulating the Effects of Global and Local Variations with Monte Carlo Monte Carlo analysis is dependent on a method to describe variability Four different approaches are available in HSPICE RF specify distributions on parameters and apply these to instance parameters specify distributions on parameters and apply these to model parameters specify distributions on model parameters using DEV LOT construct
4. kRref kRref Y lt NMA stamp Port Element The port element identifies the ports used in LIN analysis Each port element requires a unique port number If your design uses N port elements your netlist must contain the sequential set of port numbers 1 through N For example in a design containing 512 ports you must number each port sequentially 1 to 512 Each port has an associated system impedance zo If you do not explicitly specify the system impedance the default is 50 ohms The port element behaves as either a noiseless impedance or a voltage source in series with the port impedance for all other analyses DC AC or TRAN You can use this element as a pure terminating resistance or as a voltage or power source You can use the RDC RAC RHB RHBAC and RTRAN values to override the port impedance value for a particular analysis Port Element Syntax Pxxx p n port portnumber HSPICE RF User Guide Z 2007 03 155 Chapter 6 Testbench Elements Port Element Voltage or Power Information lt DC mag gt lt AC lt mag lt phase gt gt gt lt HBAC lt mag lt phase gt gt gt lt HB lt mag lt phase lt harm lt tone lt modharm lt modtone gt gt gt gt gt gt gt lt transient_waveform gt lt TRANFORHB 0 1 gt lt DCOPEN 0 1 gt Source Impeda
5. 0 0000 ccc eee ee Generating Measurement Output Files 0 00000 cee eee OptimiZationss ces csr i siege diel tends Bintan AE e Oh tina bess dies Sad last Optimizing AC DC and TRAN Analyses 0 000ee eee Optimizing HB Analysis 0000 cece eee Optimizing HBOSC Analysis 0 000 cee Using CHECK Statements 0 00 c cece Setting Global Hi Lo Levels 0 000 cece eee Slew Rise and Fall Conditions 0 0 0 0 cee es Edge Timing Verification 0 0 cee ee Setup and Hold Verification 0 00 cece eee IR Drop Detection 0020 et ee eee POWER DC Analysis nasrin ti 005 depo aa bene dt alah pas Ghee s Power DC Analysis Output Format 0000 cee eee POWER Analysis spia aie pei na E a A t ete Setting Default Start and Stop Times nanasa anaua Controlling Power Analysis Waveform DUMPS a n ansann nannaa Detecting and Reporting Surge Currents sasaaa nauan INDEX oora Berd E E eed Betas R E ead 2a 405 407 407 409 409 410 411 412 412 413 414 415 415 416 417 418 418 418 421 About This Guide This manual contains detailed reference information application examples and design flow descriptions that show how HSPICE RF features can be used for RF circuit characterization The manual supplements the HSPICE user documentation by describing the additional features built on top of the standard HSPICE f
6. The source connects to the s2 node stage references the MOSFET model The length of the gate is 2 microns The width of the gate is 10 microns In the following Mdrive MOSFET element Mdrive driver in output bsim3v3 W 3u L 0 25u DTEMP 4 0 The drain connects to the driver node The gate connects to the in node The source connects to the output node bsim3v3 references the MOSFET model HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Steady State Voltage and Current Sources The length of the gate is 3 microns The width of the gate is 0 25 microns The device temperature is 4 degrees Celsius higher than the circuit temperature Steady State Voltage and Current Sources The current source and V voltage source elements include extensions that allow you to use them as sources of steady state sinusoidal signals for HB and HBAC analyses When you use a power parameter to specify the available power you can also use these elements as power sources For a general description of the and V elements see Power Sources in the HSPICE Simulation and Analysis User Guide land V Element Syntax VXXX pn e 4 4 Voltage or Power Information x x lt lt dc gt mag gt lt ac lt mag lt phase gt gt gt lt HBAC lt mag lt phase gt gt gt lt hb lt mag lt phase lt harm lt tone lt modharm lt modtone gt gt gt gt gt gt gt lt transient waveform gt
7. sample configuration file the next line of code changes the delimiter for subcircuit hierarchies from to hier delimiter the next line of code matches any groups of characters wildcard_match_all the next line of code matches one character wildcard_match_one the next line of code begins the range expression with the character wildcard_left_range the next line of code ends the range expression with the character wildcard_right_range Using Wildcards in HSPICE RF 402 You can use wildcards to match node names HSPICE RF uses wildcards somewhat differently than standard HSPICE HSPICE RF User Guide Z 2007 03 Chapter 16 Advanced Features Limiting Output Data Size Before using wildcards you must define the wildcard configuration in a hspicerf file For example you can define the following wildcards in a hspicerf file file hspicerf wildcard _match_one wildcard_match_all wildcard_left_range wildcard right range The PRINT PROBE LPRINT and CHECK statements support wildcards in HSPICE RF For more information about using wildcards in an HSPICE configuration file see Using Wildcards in PRINT and PROBE Statements in the HSPICE Simulation and Analysis User Guide Limiting Output Data Size For multi million transistor simulations an unrestricted waveform file can grow to several gigabytes in size The file becomes unreadable in some wavefor
8. 4 Transforms the state variables to the frequency domain by using a Fast Fourier Transform FFT to establish an initial guess for HB oscillator analysis 5 Starts the standard HB oscillator analysis Additional HBOSC Analysis Options Oscillator analysis will make use of all standard HB analysis options as listed in the following table In addition the following options are specifically for oscillator applications Table 20 HBOSC Analysis Options for Oscillator Applications Parameter Description HBFREQABSTOL An additional convergence criterion for oscillator analysis HBFREQABSTOL is the maximum absolute change in frequency between solver iterations for convergence Default is 1 Hz HBFREQRELTOL An additional convergence criterion for oscillator analysis HBFREQRELTOL is the maximum relative change in frequency between solver iterations for convergence Default is 1 e 9 HBPROBETOL HBOSC analysis tries to find a probe voltage at which the probe current is less than HBPROBETOL This option defaults to the value of HBTOL which defaults to 1 e 9 HBMAXOSCITER Maximum number of outer loop iterations for HBOSC analysis It defaults to 10000 HSPICE RF User Guide 235 Z 2007 03 Chapter 9 Oscillator and Phase Noise Analysis Oscillator Analysis Using Shooting Newton SNOSC HBOSC Output Syntax The output syntax for HBOSC analysis is identical to that for HB analysis see Chapter 7 Steady State Harmonic Ba
9. Name of the vendor such as Synopsys whose tools you used to generate the DSPF file optional Name of the program such as Star RCXT that generated the DSPF file optional Version number of the program that generated the DSPF file optional Character that divides levels of hierarchy in a circuit path optional If you do not define this parameter the default hierarchy divider is a slash For example X1 X2 indicates that X2 is a subcircuit of the X1 circuit Character used to separate the name of an instance anda pin in a concatenated instance pin name or a net name and a sub node number in a concatenated sub node name If you do not define this parameter the default delimiter is a colon Hierarchical path to a net instance or pin within a circuit Name of a net in a circuit or subcircuit 329 Chapter 13 Post Layout Analysis Post Layout Back Annotation Table 25 DSPF Parameters Continued Parameter Definition instance_name Name of an instance of a subcircuit pin_name Name of a pin on an instance of a subcircuit pinCap Capacitance of a pin pin_type input O output B bidirectional X don t care S switch J jumper resistance Resistance on a pin in ohms for input I output O or bidirectional B pins You can use resistance capacitance RC pairs to model pin characteristics by using a higher order equivalent RC ladder circuit than a single capacito
10. The first line creates a simple two terminal current noise source whose value is described in A2 Hz The output noise generated from this noise source is noise_equation H Where H is the transfer function from the terminal pair node1 node2 to the circuit output where HSPICE RF measures the output noise The second line produces a noise source correlation between the node1 node2 and node3 node4 terminal pairs The resulting output noise is calculated as noise_equation sqrt H1 H2 where H1 is the transfer function from node1 node2 to the output 2 is the transfer function from node3 node4 to the output HSPICE RF User Guide 175 Z 2007 03 Chapter 6 Testbench Elements Behavioral Noise Sources 176 The noise_equation expression can involve node voltages and currents through voltage sources For the PAC phasenoise simulation to evaluate the frequency dependent noise the frequency dependent noise factor in the phasenoise must be expressed in between the parentheses For example gname nodel node2 noise frequency dependent noise bias dependent _noise This is only true when the total noise can be expressed in this form and when the frequency dependent noise can be evaluated in the PAC phasenoise simulation You can also input the behavioral noise source as a noise table with the help of predefined Table function The Table function takes two formats Noise table can be input directly throu
11. 320 Standard Post Layout Flow Control Options The standard post layout flow options are SIM _DSPF and SIM_SPEF Include one of these options in your netlist For example OPTION SIM _DSPF scope dspf filename OPTION SIM _SPEF spec_ filename In the SIM_DSPF syntax scope can be a subcircuit definition or an instance If you do not specify scope it defaults to the top level definition HSPICE RF requires both a DSPF file and an ideal netlist Only flat DSPF files are supported hierarchy statements such as SUBCKT and x1 are ignored Very large circuits generate very large DSPF files this is when using either the SIM_DSPF or the SIM _DSPF_ ACTIVE option can really improve performance You can specify a DSPF file in the SIM_SPEF option or a SPEF file in the SIM_DSPF option The scope function is not supported in the SPEF format For descriptions and usage examples see OPTION SIM_DSPF and OPTION SIM_SPEF in the HSPICE and HSPICE RF Command Reference Example models MODEL p pmos MODEL n nmos INCLUDE add4 dspf OPTION SIM DSPF add4 dspf VEC dspf adder vec TRAN 1n 5u vdd vdd 0 3 3 OPTION POST END SIM_DSPF With SIM_LA Option The SIM_DSPF option accelerates the simulation by more than 100 By using the SIM_LA option at the same time you can further reduce the total CPU time models MODEL p pmos MODEL n nmos INCLUDE add4 dspf OPTION SIM _ DSPF add4 dspf OPTI
12. Chapter 16 Advanced Features Using CHECK Statements 412 Setup and Hold Verification IR Drop Detection The results of these statements appear in a file with an err extension To prevent creating unwieldy files HSPICE RF reports only the first 10 violations for a particular check in the err file Setting Global Hi Lo Levels You use the CHECK GLOBAL LEVEL statement to globally set the desired high and low definitions for all CHECK statements For example CHECK GLOBAL LEVEL hi lo hi_th lo th Values for hi lo and the thresholds are defined by using this statement For syntax and description of this statement see CHECK GLOBAL_LEVEL in the HSPICE and HSPICE RF Command Reference Slew Rise and Fall Conditions You use the CHECK SLEW statement to verify that a slew rate occurs within the specified window of time For example CHECK SLEW min max nodel lt node2 gt lt hi lo hi_th lo th Figure 49 SLEW Example 3 3 2 6 0 7 0 0 Ins lt t lt 3ns For syntax and description of this statement see CHECK SLEW in the HSPICE and HSPICE RF Command Reference You use the CHECK RISE statement to verify that a rise time occurs within the specified window of time For example HSPICE RF User Guide Z 2007 03 Chapter 16 Advanced Features Using CHECK Statements CHECK RISE min max nodel lt node2 gt lt hi lo hi_th lo th gt Figure 50 RISE Time Example
13. For BPSK the symbol rate is the same as the data rate BPSK For QPSK modulation two bits are used to create each symbol so the symbol rate is half the data rate R n i 5 b 2 The period for each symbol is computed as Equation 19 T S This value is necessary for establishing the characteristics of Nyquist filters HSPICE RF User Guide 185 Z 2007 03 Chapter 6 Testbench Elements Complex Signal Sources and Stimuli 186 The following equation calculates the raised cosine COS filter response l a lt if 2T T s a TT l a 1 a l a Equation 20 Hf J 7 T cos z a or VSIT S l l a A gt or The VMRF signal source is designed primarily for TRAN and HB analyses and can generate baseband signals You can also specify DC and AC values as with any other HSPICE signal source In DC analysis the VMRF source is a constant DC source In AC analysis the source is a short or an open unless you specify an AC value m In HB analysis you must specify OPTION TRANFORHB on the source statement line The TRANFORHB option supports the VMRF signal source as well as the SIN PULSE and PWL sources The VMRF quadrature signal source typically involves an HF carrier signal that is modulated with a baseband signal on a much different time scale You must set source and simulation control parameters appropriately to avoid time consuming simulations in both the time and frequency domains E
14. HSPICE RF User Guide 67 Z 2007 03 Chapter 4 Input Netlist and Data Entry Input Netlist File Guidelines 68 f you create multiple definitions for the same parameter or option HSPICE RF uses the last parameter definition or OPTION statement even if that definition occurs later in the input than a reference to the parameter or option HSPICE RF does not warn you when you redefine a parameter You must define a parameter before you use that parameter to define another parameter When you select design parameter names be careful to avoid conflicts with parameterized libraries To delimit expressions use single or double quotes Expressions cannot exceed 1024 characters For improved readability use a double slash at end of a line to continue the line You can nest functions up to three levels Any function that you define can contain up to two arguments Use the PAR expression or parameter function to evaluate expressions in output statements Input Netlist File Structure An input netlist file should consist of one main program and can contain one or more optional submodules HSPICE RF uses a submodule preceded by an ALTER statement to automatically change an input netlist file then rerun the simulation with different options netlist analysis statements and test vectors You can use several high level call statements INCLUDE and LIB to structure the input netlist file modul
15. Harmonic balance analysis HB is a frequency domain steady state analysis technique In HSPICE RF you can use this analysis technique on a circuit that is excited by DC and periodic sources of one or more fundamental tones The solution that HB finds is a set of phasors for each harmonic signal in the circuit You can think of this solution as a set of truncated Fourier series HSPICE RF allows you to specify the solution spectrum to use in an analysis HB analysis then finds the set of phasors at these frequencies that describes the circuit response The result is a set of complex valued Fourier series coefficients that represent the waveforms at each node in the circuit Linear circuit elements are evaluated in the frequency domain while nonlinear elements are evaluated in the time domain The nonlinear response is then transformed to the frequency domain where it is added to or balanced with the linear response The resulting composite response satisfies KCL and KVL Kirchoff s current and voltage laws when the circuit solution is found Typical applications include performing intermodulation analysis oscillator analysis and gain compression analysis on amplifiers and mixers HB analysis also serves as a Starting point for periodic AC and noise analyses HSPICE RF User Guide Z 2007 03 Chapter 7 Steady State Harmonic Balance Analysis Harmonic Balance Analysis Harmonic Balance Equations We can write Kirchoff s current la
16. Recommended option settings are HSPICE RF User Guide Z 2007 03 Chapter 12 Envelope Analysis Envelope Simulation For BE integration set OPTION SIM ORDER 1 For TRAP set OPTION SIM ORDER 2 default METHOD TRAP default For GEAR set OPTION SIM _ORDER 2 default METHOD GEAR Example env tones le9 nharms 6 env_step 10n env_stop 1lu Oscillator Analysis Form ENVOSC TONE f1 NHARMS hi1 ENV _STEP tstep ENV_STOP tstop PROBENODE n1 n2 vosc lt FSPTS num min max gt Parameter Description TONE Carrier frequencies in hertz NHARMS Number of harmonics ENV_STEP Envelope step size in seconds ENV_STOP Envelope stop time in seconds PROBENODE Defines the nodes used for oscillator conditions and the initial probe voltage value FSPTS Specifies the frequency search points used in the initial small signal frequency search Usage depends on oscillator type Description You use the ENVOSC command to do envelope simulation for oscillator startup or shutdown Oscillator startup or shutdown analysis with this command must be helped along by converting a bias source from a DC description to a PWL description that either Starts at a low value that supports oscillation and ramps up to a final value startup simulation Starts at the DC value and ramps down to zero shutdown simulation In addition to solving for the state variables at each envelope time point the ENVOSC command also s
17. WDB Output Format You can use the waveform database WDB output formatin OPTION POST It was developed for maximum efficiency The output file is wdb For example to output to a wdb file enter OPTION POST wdba Signals across multiple hierarchies that map to the same node are named together They also share the same waveform data You can also set up the database so that CosmosScope extracts one signal at a time This means that CosmosScope does not need to read the entire output file to display a single waveform The WDB format was designed to make accessing waveform data faster and more efficient It is a true database so the waveform browser does not have to load the complete waveform file for you to view a single signal This feature is especially useful if the size of the waveform file is several gigabytes HSPICE RF User Guide 393 Z 2007 03 Chapter 15 Using HSPICE with HSPICE RF RF Transient Analysis Output File Formats 394 Furthermore the WDB format is usually more compact than XP and NW described later in this section However if the NW file is already very small then WDB offers little advantage in size or speed You can compress WDB files For additional information see Compressing Analog Files on page 396 TR Output Format HSPICE RF stores simulation results for analysis by using the AvanWaves graphical interface method For example these commands output a tr file in TR format
18. rm cm valm lt SHORTALL yes no gt lt IGNORE_COUPLING yes no gt Reluctance External File Form Lxxx nlp nin nNp nNn RELUCTANCE FILE lt filenamels FILE lt filename2 gt HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Passive Elements lt SHORTALL yes no gt lt IGNORE_COUPLING yes no gt Parameter Description Lxxx nip nin nNp nNn RELUCTANCE r1 c1 vall r2 c2 val2 rm cm valm FILE lt filename1 gt HSPICE RF User Guide Z 2007 03 Name of a reluctor Must begin with L followed by up to 1023 alphanumeric characters Names of the connecting terminal nodes The number of terminals must be even Each pair of ports represents the location of an inductor Keyword to specify reluctance inverse inductance Reluctance matrix data In general K will be sparse and only non zero values in the matrix need be given Each matrix entry is represented by a triplet r c val The value r and c are integers referring to a pair of inductors from the list of terminal nodes If there are 2 N terminal nodes there will be N inductors and the r and c values must be in the range 1 N The val value is a reluctance value for the r c matrix location and the unit for reluctance is the inverse Henry H Only terms along and above the diagonal are specified for the reluctance_matrix The simulator fills in the lower triangle to ensure symme
19. Chapter 11 S parameter Extraction Frequency Translation S Parameter HBLIN Extraction 300 Equation 65 N N V w yaw Zoili w yn 1 j a w 1 N N gt Ny nae ol N N V w yaw Zoli w yaw b w l l LN Na Ny 2 JZ Ol Where w is the ith tone The frequency translate S parameters are calculated by applying different nj 1 N to different ports Limitations The HBLIN analysis has these known limitations m Noise parameters are not calculated for mixed mode operation Only the S parameters corresponding to the set of frequencies specified at each port are extracted m Multiple small signal tones are not supported The port P element impedance cannot be specified as complex HB Analysis An HB analysis is required prior to an HBLIN analysis To extract the frequency translation S parameters a sweep of the small signal tone is necessary You can identify the small signal tone sweep in the HBLIN command or in the HB command together with a SS_TONE specification For additional information regarding HB analysis see Harmonic Balance Analysis on page 198 HSPICE RF User Guide Z 2007 03 Chapter 11 S parameter Extraction Frequency Translation S Parameter HBLIN Extraction Port element You must use a port P element as the termination at each port of the system To indicate the frequency band that the S parameters are extracted from it is necessary to spec
20. HBNOISE command which is included in the mix_hbac sp netlist The HBNOISE command invokes noise analysis identifying an output node where the noise is measured an input noise source in this case rrf1 which serves as a reference for noise figure computation and a frequency sweep for the noise analysis The PRINT and PROBE hbnoise commands instruct HSPICE RF to save the output noise and noise figure at each frequency in the mix_hbac printonO and mix_hbac pn0 output files This ideal mixer is noiseless except for the resistors at the input and output The mix_hbac lis file contains detailed data on the individual noise source contributions of the resistors You can view mix_hbac printpnO to see the output noise and noise figure at each frequency In CosmosScope you can view mix_hbac pn0 to plot the output noise and noise figure data as a function of frequency HSPICE RF User Guide Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 7 Using Shooting Newton Analysis on a Driven Phase Frequency Circuit and a Ring Oscillator Example 7 Using Shooting Newton Analysis on a Driven Phase Frequency Circuit and a Ring Oscillator Introduction While the Harmonic Balance HB algorithm represents the circuit s voltage and current waveforms as a Fourier series a series of sinusoidal waveforms the Shooting Newton SN algorithm provides analysis capability for digital logic circuits and RF components that require steady state analysis
21. HSPICE RF User Guide 317 Z 2007 03 Chapter 13 Post Layout Analysis Post Layout Back Annotation Star RCXT generates a hierarchical Layout Versus Schematic LVS ideal netlist and flat information about RC parasitics in a DSPF or standard parasitic exchange format SPEF file HSPICE RF uses the hybrid flat hierarchical approach to back annotate the RC parasitics from the DSPF or SPEF file into the hierarchical LVS ideal netlist Using the hierarchical LVS ideal netlist cuts simulation runtime and CPU memory usage Because HSPICE RF uses the hierarchical LVS ideal netlist as the top level netlist you can fully control the netlist For example You can set different modes to different blocks for better accuracy and speed trade off You can run power analysis based on the hierarchical LVS ideal netlist to determine the power consumption of each block If you use the hierarchical LVS ideal netlist you can reuse all post processing statements from the pre layout simulation for the post layout simulation This saves time and the capacity of the verification tool is not stressed so reliability is higher HSPICE RF supports only the XREF COMPLETE flow and the XREF NO flow from Star RCXT Refer to the Star RCXT User Guide for more information about the XREF flow To generate a hierarchical LVS ideal netlist with Star RCXT include the following options in the Star RCXT command file for XREF NO flow NETLIST
22. In this example fqmode1 is declared in both the S element statement and the S model statement and they have different qgmodel names This is not allowed in HSPICE Example 8 sl nl n2 n3 n ref mname smodel fqmodel sfqmodel model smodel s tstonefile expl s3p In this example fqmode1 is already declared in the s1 statement and tstonefile is declared in the related smodel card This is a conflict when describing the frequency varying behavior of the network which is not allowed in HSPICE Frequency Table Model The frequency table model SP model is a generic model that you can use to describe frequency varying behavior Currently the S element and the LIN command use this model For a description of this model see section Small Signal Parameter Data Frequency Table Model SP Model in the HSPICE Signal Integrity User Guide Group Delay Handler in Time Domain Analysis The S element accepts a constant group delay matrix in time domain analysis You can also express a weak dependence of the delay matrix on the frequency as a combination of the constant delay matrix and the phase shift value at each frequency point To activate or deactivate this delay handler specify the DELAYHANDLE keyword in the S model statement The delay matrix is a constant matrix which HSPICE RF extracts using finite difference calculation at selected target frequency points HSPICE RF obtains the T p delay matrix component as HSPICE RF Use
23. V ts timedomain v m1d m2d 200p 400p 600p 800p 2 14n dB V f H2z v mid_b m2d_b vimid m2d E a st wa tf rk amp a wt 25g 5g 75g 10g 125g 15g 17 5g 20g ti Hz 36 HSPICE RF User Guide Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 5 Using HBOSC Analysis fora CMOS GPS VCO Figure 3 VCO Tuning Curves Output in CosmosScope GPS VCO Tuning Curve Magi vtune vimid m2d 1 Freq vtune _ At_begin freq f0 2 D td pos vtune a ee eo 20 25 S20 So 4 4 vtune _ HSPICE RF User Guide Z 2007 03 37 Chapter 3 HSPICE RF Tutorial Example 6 Using Multi Tone HB and HBAC Analyses for a Mixer Figure 4 VCO Phase Noise Response in CosmosScope GPS VCO Phase Noise dBc H2 f H2 PHNOISE th an 3 120 0 J 140 0 1 60 0 100 0 1 0k 10 0k 100 0k imeg t Hz Example 6 Using Multi Tone HB and HBAC Analyses for a Mixer 38 The example in this section shows how to use HSPICE RF to analyze a circuit driven by multiple input stimuli with different frequencies Mixer circuits provide a typical example of this scenario in this case there might be two input signals LO and RF which are mixed to produce an IF output signal In this case HSPICE RF offers two options m Multi tone HB analysis specify the LO and RF base frequencies as two separate tones on the HB command Periodic AC analysis HBAC if one of the inputs is a sm
24. V and I Element Syntax Vxxx n n VMRF lt gt AMP sa FREQ fc PHASE ph MOD MOD FILTER FIL FILCOEF filpar RATE Rb BITSTREAM data lt TRANFORHB 0 1 gt lt gt Ixxx n n VMRF lt gt AMP sa FREQ fc PHASE ph MOD MOD FILTER FIL FILCOEF filpar RATE Rb BITSTREAM data lt TRANFORHB 0 1 gt lt gt Parameter Description VXXX Independent voltage source HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Complex Signal Sources and Stimuli Parameter Description xxx Independent current source n n Positive and negative controlled source connecting nodes VMRF Keyword that identifies and activates the Vector Modulated RF signal source AMP Signal amplitude in volts or amps FREQ Carrier frequency in hertz Set fc 0 0 to generate baseband 1 Q signals For harmonic balance analysis the frequency spacing must coincide with the HB TONES settings PHASE Carrier phase in degrees If fc 0 0 ph 0 and baseband I t is generated ph 90 and baseband q t is generated Otherwise s t I t cos o Q t sin MOD One of the following keywords identifies the modulation method used to convert a digital stream of information to I t and Q t variations BPSK binary phase shift keying QPSK quadrature phase shift keying FILTER One of the following keywords identifies the method used to filter the and Q signals before modulating the RF carrier signal COS raised
25. gl 0 if cur 1 0 v lo v rf mixer element cl 0 if g 1 0e 9 v lo v rf mixer element rout if ifg 1 0 vetrl ifg 0 0 0 hl out 0 vetrl 1 0 rhi out 0 1 0 vrf rfl 0 snac 001 24 0 Small signal for SNAC with 1 tone SN Input sn tones 1 0g nharms 3 snac lin 1 0 8g 0 8g print sn v rf1 v lo v out print snfd v rf1 v lo v out print snac v rf1 v lo v out measure snac vouti find v out 1 1 at 0 8g measure snac vout2 find v out 0 1 at 0 8g measure snac vout3 find v out 1 1 at 0 8g measure sn vlol find v lo at 0 5n measure sn vlo2 find v lo at 1n measure snfd vlo3 find v lo 1 at 1 end Multitone Harmonic Balance Noise HBNOISE 268 An HBNOISE Harmonic Balance noise analysis simulates the noise behavior in periodic systems It uses a Periodic AC PAC algorithm to perform noise analysis of nonautonomous driven circuits under periodic steady state tone conditions This can be extended to quasi periodic systems having more than one periodic steady state tone One application for a multitone HBNOISE analysis is determining mixer noise figures under the influence of a strong interfering signal The PAC method simulates noise assuming that the stationary noise sources and or the transfer function from the noise source to a specific output are periodically modulated HSPICE RF User Guide Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Multitone H
26. 350 To model parametric and statistical variation in circuit behavior use PARAM statement to investigate the performance of a circuit as you change circuit parameters For details about the PARAM statement see the PARAM statement in the HSPICE and HSPICE RF Command Reference Temperature variation analysis to vary the circuit and component temperatures and compare the circuit responses You can study the temperature dependent effects of the circuit in detail Monte Carlo analysis when you know the statistical standard deviations of component values to center a design This provides maximum process yield and determines component tolerances Worst case corner analysis when you know the component value limit to automate quality assurance for e basic circuit function e process extremes e quick estimation of speed and power tradeoffs e best case and worst case model selection e parameter corners e library files Data driven analysis for cell characterization response surface or Taguchi analysis See Performing Digital Cell Characterization in the HSPICE Applications Manual Automates characterization of cells and calculates the coefficient of polynomial delay for timing simulation You can simultaneously vary any number of parameters and perform an unlimited number of analyses This analysis uses an ASCII file format so HSPICE RF can automatically generate parameter values This analysis can replace hundreds or thous
27. Example 2 In the following Qopamp1 BJT element Qopampl1 cl b3 e2 s lstagepnp AREA 1 5 AREAB 2 5 AREAC 3 0 The collector connects to the c1 node The base connects to the b3 node The emitter connects to the e2 node The substrate connects to the s node 1stagepnp references the BJT model The AREA area factor is 1 5 The AREAB area factor is 2 5 The AREAC area factor is 3 0 Example 3 In the Qdrive BUT element below Qdrive driver in output model npn 0 1 The collector connects to the driver node The base connects to the in node HSPICE RF User Guide 163 Z 2007 03 Chapter 6 Testbench Elements Active Elements 164 The emitter connects to the output node model_npn references the BUT model The area factor is 0 1 JFETs and MESFETs Jxxx nd ng ns lt nb gt mname lt lt lt AREA gt area lt W val gt lt L val gt gt lt OFF gt lt IC vdsval vgsval gt lt M val gt lt DTEMP val gt Jxxx nd ng ns lt nb gt mname lt lt lt AREA gt area gt lt W val gt lt L val gt gt lt OFF gt lt VDS vdsval gt lt VGS vgsval gt lt M val gt lt DTEMP val gt Parameter JXXX nd ng ns nb mname area AREA area W L OFF Description JFET or MESFET element name Must begin with J followed by up to 1023 alphanumeric characters Drain terminal node name Gate terminal node name Source terminal node name Bulk terminal node name which i
28. HSPICE RF begins running at the 10th iteration then continues from the 20th to the 30th at the 40th and finally from the 46th to 72nd Monte Carlo iteration The numbers after list can not be parameter HSPICE RF User Guide 361 Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Monte Carlo Analysis 362 tran 1n 10n sweep monte list 10 20 30 40 46 72 Monte Carlo Output MEASURE statements are the most convenient way to summarize the results PRINT statements generate tabular results and print the values of all Monte Carlo parameters MCBRIEF determines the output types of the random parameters during Monte Carlo analysis to improve output performance If one iteration is out of specification you can obtain the component values from the tabular listing A detailed resimulation of that iteration might help identify the problem AvanWaves superimposes all iterations as a single plot so you can analyze each iteration individually PARAM Distribution Function This section describes how to use assign a PARAM parameter in Monte Carlo analysis For a general description of the PARAM statement see the PARAM command in the HSPICE and HSPICE RF Command Reference You can assign a PARAM parameter to the keywords of elements and models and assign a distribution function to each PARAM parameter HSPICE RF recalculates the distribution function each time that and element or model keyword uses a
29. IEEE J Solid State Circuits vol 34 no 6 pp 790 804 June 1999 Jitter Analysis Techniques for High Data Rates Application Note 1432 Agilent Technologies Feb 2003 6 Characterization of Clocks and Oscillators NIST Technical Note 1337 National Institute of Standards and Technology HSPICE RF User Guide 195 Z 2007 03 Chapter 6 Testbench Elements References 196 HSPICE RF User Guide Z 2007 03 7 Steady State Harmonic Balance Analysis Describes how to use harmonic balance analysis for frequency driven steady state analysis HSPICE RF provides several analyses that support the simulation and analysis of radio frequency integrated circuits RFICs These analyses provide simulation capabilities that are either much more difficult to perform or are not practically possible by using standard HSPICE analyses The RF analyses include Harmonic Balance HB for frequency domain steady state analysis see Harmonic Balance Analysis on page 198 Shooting Newton SN for frequency or time domain steady state analysis see Chapter 8 Steady State Shooting Newton Analysis plus spectrum analysis specific to the SN analysis see Shooting Newton with Fourier Transform SNFT Harmonic Balance oscillator analysis HBOSC see Harmonic Balance Analysis for Frequency of Oscillation on page 229 Shooting Newton oscillator analysis SNOSC see Oscillator Analysis Using Shooting Newton SNOSC on page 236 Harmonic Bal
30. Im A2 Re p2 Im p2 Im A3 Re p3 Im p3 Re A3 In the above syntax parenthesis commas and slashes are separators they have the same meaning as a space A pole residue pair is represented by four numbers real and imaginary part of the residue then real and imaginary part of the pole You must make sure that Re pi lt 0 otherwise the simulations will certainly diverge Also it is a good idea to assure passivity of the model for an N port admittance matrix Y Re Y should be positive definite or the simulation is likely to diverge Example To represent a G s in the form Equation 16 G s 0 001 1x 10 s s 1x 10 0 001 70 006 s 1x 10 j1 8x 10 You would input G1 1 0 FOSTER 2 0 0 001 le 12 0 0004 0 1e10 0 0 001 HSPICE RF User Guide Z 2007 03 0 0008 0 001 70 006 htt s 1x 10 j1 8x 10 0 006 1e8 1 8e10 179 Chapter 6 Testbench Elements Complex Signal Sources and Stimuli Note In the case of a real poles half the residue value is entered because it s essentially applied twice In the above example the first pole residue pair is real but we still write it as A1 s p1 A1 s p1 therefore 0 0004 is entered rather than 0 0008 Complex Signal Sources and Stimuli 180 To predict radio frequency integrated circuit RFIC performance some analyses require simulations that use representative RF signal sources Am
31. MODEL RF WB NPN NPN IS 1 32873E 015 BF 1 02000H 002 NF 1 00025E 000 VAF 5 19033E 001 EG 1 11000E 000 XTI 3 00000H 000 CJE 2 03216E 012 VJE 6 00000E 001 MJE 2 90076E 001 TF 6 55790E 012 XTF 3 89752E 001 VTF 1 09308E 001 ITF 5 21078E 001 CJC 1 00353E 012 VJC 3 40808E 001 MJC 1 94223E 001 Example 4 Using HBOSC Analysis for a Colpitts Oscillator 28 This section demonstrates HSPICE RF oscillator analysis using a single transistor oscillator circuit Oscillator analysis is an extension of Harmonic Balance in which the base frequency itself is an unknown to be solved for In oscillator analysis the user supplies a guess at the base frequency and no voltage or current source stimulus is needed To activate oscillator analysis include a HBOSC command with The TONE parameter set to a guess of the oscillation frequency The PROBENODE parameter set to identify an oscillating node or pair of nodes Always specify a pair of nodes if only one node oscillates specify ground as the second node To speed up the simulation also supply a guess at the magnitude of the oscillating voltage across these nodes The FSPTS parameter set to a frequency range and number of search points When you set FSPTS HSPICE RF precedes the HBOSC analysis with a frequency search in the specified range to obtain an optimal initial guess for the oscillation frequency This can accelerate the HB oscillator converge
32. RF User Guide Z 2007 03 Chapter 15 Using HSPICE with HSPICE RF RF Transient Analysis Accuracy Control same as first order Gear Also HSPICE RF supports a hybrid algorithm TRAPGEAR which is a mixture of the three basic algorithms HSPICE RF contains an algorithm for auto detection of numerical oscillations commonly encountered with trapezoidal integration If HSPICE RF detects such oscillations it inserts BE steps but not more than one BE step for every 10 time steps To turn off auto detection use the PURETP option The TRAPGEAR method combining 90 trapezoidal with 10 Gear 2 HSPICE RF inserts BE steps when the simulator encounters a breakpoint or when the auto detection algorithm finds numerical oscillations For the syntax and description of this control option see OPTION METHOD in the HSPICE and HSPICE RF Command Reference OPTION MAXORD You use the MAXORD option to select the maximum order of integration for the GEAR method Either the first order Gear Backward Euler or the second order Gear Gear 2 integration method For the syntax and description of this control option see OPTION MAXORD in the HSPICE and HSPICE RF Command Reference OPTION SIM_ORDER You use the SIM ORDER option to control the amount of Backward Euler BE to mix with the Trapezoidal method for hybrid integration This option affects time stepping when you set OPTION METHOD to TRAP or TRAPGEAR For the syntax and description of th
33. Scaling factor Initial capacitor voltage Capacitance can be a function of any node voltage and any branch current but not a function of time frequency or temperature HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Passive Elements Parameter Description CTYPE Determines the calculation mode for elements that use capacitance equations Set this parameter carefully to ensure correct simulation results HSPICE RF extends the definition and values of CTYPE relative to HSPICE 0 if C depends only on its own terminal voltages that is a function of V n1 lt n2 gt This is consistent with HSPICE 1 if C depends only on outside voltages or currents This is consistent with HSPICE 2 if C depends on both its own terminal and outside voltages default for HSPICE RF HSPICE does not use CTYPE 2 You can specify capacitance as a numeric value in units of farads as an equation or as a polynomial of the voltage The only required fields are the two nodes and the capacitance or model name If you use the parameter labels the nodes and model name must precede the labels Other arguments can follow in any order If you specify a capacitor model see the Passive Device Models chapter in the HSPICE Elements and Device Models Manual the capacitance value is optional If you use an equation to specify capacitance the CTYPE parameter determines how HSPICE calculates the capacitance charge
34. The calculation is different depending on whether the equation uses a self referential voltage that is the voltage across the capacitor whose capacitance is determined by the equation To avoid syntax conflicts if a capacitor model has the same name as a capacitance parameter HSPICE or HSPICE RF uses the model name Example 1 In the following example C1 assumes its capacitance value from the model not the parameter PARAMETER CAPXX 1 C1 1 2 CAPXX MODEL CAPXX C CAP 1 Example 2 In the following example the C1 capacitors connect from node 1 to node 2 with a capacitance of 20 picofarads HSPICE RF User Guide 121 Z 2007 03 Chapter 6 Testbench Elements Passive Elements 122 C1 1 2 20p In this next example Cshunt refers to three capacitors in parallel connected from the node output to ground each with a capacitance of 100 femtofarads Cshunt output gnd C 100f M 3 The Cload capacitor connects from the driver node to the output node The capacitance is determined by the voltage on the capcontrol node times 1E 6 The initial voltage across the capacitor is O volts Cload driver output C lu v capcontrol CTYPE 1 IC 0v The C99 capacitor connects from the in node to the out node The capacitance is determined by the polynomial C cO c1 v c2 v v where v is the voltage across the capacitor C99 in out POLY 2 0 0 5 0 01 Example 1 Cbhypass 1 0 10PF C1 2 3 CBX MODEL CBX C CB B 0 10P IC 4V
35. These three methods are still supported in HSPICE RF In the following sections the three methods are described The description relies on test cases which can be found in the tar file monte_test tar in directory lt installdir gt demo hspice apps Variations Specified on Geometrical Instance Parameters This method consists of defining parameters with variation using the distribution functions UNIF AUINF GAUSS AGAUSS and LIMIT These parameters are then used to generate dependent parameters or in the place of instance parameters In a Monte Carlo simulation at the beginning of each sample new random values are calculated for these parameters For each reference a new random value is generated however no new value is generated for a derived parameter Therefore it is possible to apply independent variations to parameters of different devices as well as the same variation to parameters of a group of devices Parameters that describe distributions can be used in expressions thus it is possible to create combinations of variations correlations These concepts are best explained with circuit examples In the three following examples variation is defined on the width of a physical resistor which has a model If this device was a polysilicon resistor for example then the variations describe essentially the effects of photoresist exposure and etching on the width of the poly layer test1 sp has a distribution parameter defined c
36. Vendor Libraries INCLUDE lt path gt T2N2211 inc This method requires you to store each model in its own inc file so it is not generally useful However you can use it to debug new models when you test only a small number of models Vendor Libraries The vendor library is the interface between commercial parts and circuit or system simulation ASIC vendors provide comprehensive cells corresponding to inverters gates latches and output buffers Memory and microprocessor vendors supply input and output buffers Interface vendors supply complete cells for simple functions and output buffers to use in generic family output Analog vendors supply behavioral models To avoid name and parameter conflicts models in vendor cell libraries should be within the subcircuit definitions Figure 15 Vendor Library Usage x1 in out vdd vss buffer_f __ OPTION search usr lib vendor usr lib vendor buffer_f inc usr lib vendor skew dat macro buffer_f in out vdd vss lib usr lib vendor skew dat ff inc usr lib vendor buffer inc lib ff fast model param vendor_xl 1u a 3 eom inc usr lib vendor model dat endl ff L usr lib vendor buffer inc L usr lib vendor model dat macro buffer in out vdd vss m1 out in vdd vdd nch w 10 l 1 model nch nmos level 28 xl vendor_x 92 HSPICE RF User Guide Z 2007 03 Chapte
37. automatically analyzed to search for periodic behavior near the TONE or PERIOD value specified HSPICE RF User Guide Z 2007 03 Chapter 9 Oscillator and Phase Noise Analysis Oscillator Analysis Using Shooting Newton SNOSC Parameter Description TRINIT MAXTRINITCYCLES TRES PERIOD SWEEP This the transient initialization time If not specified the transient initialization time will be equal to the period for Syntax 1 or the reciprocal of the tone for Syntax 2 For oscillators we recommend specifying a transient initialization time since the default initialization time is usually too short to effectively stabilize the circuit Stops SN stabilization simulation and frequency detection when the simulator detects that MAXTRINITCYCLES have been reached in the oscnode signal or when time trinit whichever comes first Minimum cycles is 1 The MAXTRINITCYCLES parmaeter is optional TRES Is the time resolution to be computed for the steady state waveforms in seconds The period of the steady state waveform may be entered either as PERIOD or its reciprocal TONE PERIOD is the expected period T seconds of the steady state waveforms Enter an approximate value when using for oscillator analysis Specifies the type of sweep You can sweep up to three variables You can specify either LIN DEC OCT POI SWEEPBLOCK DATA OPTIMIZE or MONTE SWEEP is an optional parameter Specify the nsteps start and s
38. but operate with waveforms that tend to be square instead of sinusoidal Simple examples of using the Shooting Newton analysis functions are presented in this section Driven Phase Frequency Example Ring Oscillator Example Shooting Newton Analysis Setup To set up a time domain steady state analysis the HSPICE input netlist must contain An SN command to activate the analysis The SN command specifies e The expected period of the steady state waveforms which must match the period of any input waveforms The period is specified in time domain units Seconds Alternatively you may specify a frequency in Hz e A time resolution which is analogous to the transient analysis TRAN command s TSTEP parameter and will affect the time step size selection It also affects the number of frequency terms used in small signal analyses such as periodic AC or noise analysis The time resolution is typically specified in seconds but alternatively may be specified in the frequency domain as a number of harmonics e A transient initialization time that is used by HSPICE RF to run a basic transient simulation of this length before attempting Newton Raphson iterations to converge on a steady state solution This parameter is optional If it is not specified the specified period is used as the HSPICE RF User Guide 43 Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 7 Using Shooting Newton Analysis on a Driven Phase Frequency
39. lt TRANFORHB 1 0 gt S kkxkxk k Dower Switch KKKKKKKK lt power 0 1 w dbm gt lt z0 val gt lt rdc val gt lt rac val gt lt RHBAC val gt lt rhb val gt lt rtran val gt Ixxx p n Current or Power Information x lt lt dc gt mag gt lt ac lt mag lt phase gt gt gt lt HBAC lt mag lt phase gt gt gt lt hb lt mag lt phase lt harm lt tone lt modharm lt modtone gt gt gt gt gt gt gt lt transient waveform gt lt TRANFORHB 1 0 gt S kkkxk Power Switch kkkkkkkxk lt power 0 1 W dbm gt lt z0 val gt lt rdc val gt lt rac val gt lt RHBAC val gt lt rhb val gt lt rtran val gt Parameter Description lt lt dc gt mag gt DC voltage or power source value You don t need to specify DC explicitly default 0 lt ac lt mag lt phase gt gt gt AC voltage or power source value HSPICE RF User Guide 169 Z 2007 03 Chapter 6 Testbench Elements Steady State Voltage and Current Sources 170 Parameter Description lt HBAC lt mag lt phase gt gt gt lt hb lt mag lt phase lt harm lt tone lt modharm lt modtone gt gt gt gt gt gt gt lt transient waveform gt lt power 0 1 W dbm gt lt z0 val gt lt rdc val gt HSPICE RF HBAC voltage or power source value HSPICE RF HB voltage current or power source value Multiple HB specifications with different harm tone modharm and modtone values are all
40. n If you specify only one node V n then HSPICE RF assumes that the second node is ground You can also specify a 2 terminal element A sweep of type LIN OCT DEC POI or SWEEPBLOCK Specify the type nsteps and start and stop time for each sweep type where type Frequency sweep type such as OCT DEC or LIN nsteps Number of steps per decade or total number of steps start Starting frequency stop Ending frequency The four parameters determine the offset frequency sweep about the carrier used for the phase noise analysis LIN type nsteps start stopOCT type nsteps start stopDEC type nsteps start stopPOl type nsteps start stopSWEEPBLOCK freq1 freq2 freqn METHODS 0 default selects the Nonlinear Perturbation NLP algorithm which is used for low offset frequencies METHOD 1 selects the Periodic AC PAC algorithm which is used for high offset frequencies METHOD 2 selects the Broadband Phase Noise BPN algorithm which you can use to span low and high offset frequencies You can use METHOD to specify any single method See the section on Phasenoise Algorithms below for a more detailed discussion on using the METHOD parameter HSPICE RF User Guide Z 2007 03 Chapter 9 Oscillator and Phase Noise Analysis Phase Noise Analysis Parameter Description carrierindex listfreq listcount listfloor listsources Optional Specifies the harmonic index of the carrier at
41. o 9 HSPICE RF User Guide 249 Z 2007 03 Chapter 9 Oscillator and Phase Noise Analysis Jitter Analysis 250 to enable currently supported HSPICE RF jitter measurements to be written as Equation 45 ns Gy Q J L f df RMS Phase Jitter 0 Orrp t a J L f sin nft Timing Time Interval Error Jitter Gi 0 From these definitions several other key jitter measurements can be derived including Period Jitter Tracking Jitter Long Term Jitter and Cycle to Cycle Jitter Period Jitter is equivalent to the value for Timing Jitter for a one period interval We therefore have co 2 Equation 46 Opper Org T a J L f sin nft Period Phase Jitter 0 Tracking Jitter is equivalent to the value in units of seconds for RMS Phase Jitter or oo Equation 47 6 6 Pas a 2 Lifdf Tracking Jitter 0 Long Term Jitter is equivalent to 2 times the Tracking Jitter i e co Oms _ 2 Lamar Long Term Jitter Equation 48 O gt t 2 AT OTJE Q Cycle to Cycle Jitter is based on the difference between adjacent Period Jitter measurements It is given by Equation 49 Ocrc Pore D 2T Cycle to Cycle Jitter In general each of the above calculations must be performed carefully over limits of integration to accurately calculate jitter expressions based on the finite HSPICE RF User Guide Z 2007 03 Chapter 9 Oscillator and Phase Noise Ana
42. or undesignated units and the header in the output file is Z Y or GAIN Output from the PROBE statement is written to a snxf file Reported performance log statistics are written to a lis file SNXF CPU time SNXF peak memory usage Example Based on the SN analysis the following example computes the transimpedance from isrc to v 1 SN tones 1e9 nharms 4 SNXF v 1 lin 10 le8 1 2e8 print SNXF TFV isrc TFI n3 HSPICE RF User Guide 293 Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Shooting Newton Transfer Function Analysis SNXF 294 SNXF Test Listing Test SNXF nonlinear order 2 poly equation OPTIONS PROBE OPTIONS POST 2 vlo lo 0 cos 0 1 0 1g 0 0 rlo lo 0 50 vrf1l rfl 0 0 rrfl rfl 0 50 El out 0 POLY 2 lo 0 rfl 0 0213131101 rout out 0 50 opt delmax 01n sn tones 1g nharms 5 snxf v out lin 2 100meg 200meg print sn v out v rf1 v lo print snxf tfv vrf1 tfv vlo end HSPICE RF User Guide Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses References References 1 2 3 5 6 7 8 9 S Maas Nonlinear Microwave Circuits Chapter 3 IEEE Press 1997 R Gilmore and M B Steer Nonlinear Circuit Analysis Using the Method of Harmonic Balance A Review of the Art Part I Introductory Concepts International Journal of Microwave and Millimeter
43. use the AC command to specify any other parameter sweeps of interest Use the LIN command with the AC command to activate small signal linear transfer analysis The AC command specifies the base frequency sweep for the LIN analysis The LIN analysis automatically performs multiple AC and NOISE analyses as needed to compute all complex signal transfer parameters The necessary number of port P elements numbered sequentially beginning with one to define the terminals of the multi port network For example a two port circuit must contain two port elements with one listed as port 1 and the other as port 2 The port elements define the ordering for the output quantities from the LIN command for example the terminals for port 1 are used for S11 Y11 and Z11 measurements Much of the LIN analysis is automated so the HSPICE input netlist often does not require the following AC signal sources The LIN command computes transfer parameters between the ports with no additional AC sources needed DC sources You can analyze a purely passive circuit without adding sources of any kind The following tutorial example shows how to set up a LIN analysis for an NUOS low noise amplifier circuit This netlist is shipped with the HSPICE RF distribution as gsmina sp and is available in the directory lt installdir gt demo hspicerf examples 16 HSPICE RF User Guide Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 1 Using L
44. 0 probe SNNOISE ONOISE print SNNOISE ONOISE end Simulation Status Output During the simulation the simulation status is displayed on the screen In addition to the screen display more detailed status cpu time and memory usage information is also be written to the phasefreqdet lis file Parsing Loading Netlist Data Checking Compressing Simulation Time Domain RF Analysis Steady State Time Domain Analysis will use the matrix implicit gmres solver starting Shooting Newton Sweep Shooting Newton Sweep 0 0 PHASE 0 000000 etc 0 0 sec ett 0 0 sec Input period 2e 09 begin circuit stabilization time 20 40 60 80 100 circuit stabilization time done kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk kkkkkkkkkk Shooting Newton iteration 1 kkkkkkkkkkkkkxk kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk 20 40 60 80 100 number of transient points 276 start GMRES number of GMRES iterations 3 current residual 6 79374e 09 voltage residual 0 000183225 46 HSPICE RF User Guide Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 7 Using Shooting Newton Analysis on a Driven Phase Frequency Circuit and a Ring Oscillator kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk KKKKKKKKKK Shooting Newton iteration 4 KRKEKKKKKKKKKKKK kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk 20 40 60 80 100 number of transient points 1471 start GMRES
45. 0 le 5 sqrt HERTZ CONVOLUTION 1 Capacitors The following general input syntax is for a capacitor Cxxx nodel node2 lt modelname gt lt C gt capacitance lt TCI val gt lt TC2 val gt lt W val gt lt L val gt lt DTEMP val gt lt M val gt lt SCALE val gt lt IC val gt Cxxx nl n2 4 C equation CTYPE 0 1 2 Polynomial form HSPICE RF User Guide 119 Z 2007 03 Cxxx nil n2 POLY c0 cl Chapter 6 Testbench Elements Passive Elements lt above_options gt Parameter Description Cxxx Capacitor element name Must begin with C followed by up to 1023 alphanumeric characters POLY Keyword to specify capacitance as a non linear polynomial c0 c1 Coefficients of a polynomial described as a function of the voltage node1 andnode2 capacitance modelname C TC1 TC2 W L M DTEMP SCALE IC equation across the capacitor cO represents the magnitude of the Oth order term c1 represents the magnitude of the 1st order term and so on You cannot use parameters as coefficient values Names or numbers of connecting nodes Nominal capacitance value in Farads Capacitance model name Capacitance at room temperature in Farads First order and second order temperature coefficient Capacitor width in meters Capacitor length in meters Multiplier to simulate multiple parallel capacitors Temperature difference between element and circuit
46. 03 359 Chapter 14 Statistical and Monte Carlo Analysis Monte Carlo Analysis 360 The value of the MONTE analysis keyword determines how many times to perform operating point DC sweep AC sweep or transient analysis Figure 32 Monte Carlo Distribution Gaussian Distribution Uniform Distribution Population Population Abs 3 Sigma Abs variation variation gt lt gt Nom_value Nom_value Rel_variation Abs_variation Nom_value Monte Carlo Setup To set up a Monte Carlo analysis use the following HSPICE statements PARAM statement sets a model or element parameter to a Gaussian Uniform or Limit function distribution DC AC or TRAN analysis enables MONTE MEASURE statement calculates the output mean variance sigma and standard deviation MODEL statement sets model parameters to a Gaussian Uniform or Limit function distribution Select the type of analysis to run such as operating point DC sweep AC sweep or TRAN sweep Operating Point DC MONTE lt firstrun numl1 gt or DC MONTE list lt gt lt numl num2 gt lt num3 gt lt num5 num6 gt lt num7 gt lt gt DC Sweep DC vin 1 5 0 25 sweep MONTE val lt firstrun num1 gt HSPICE RF User Guide Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Monte Carlo Analysis O r DC vin 1 5 0 25 sweep MONTE list lt gt lt numl num2 gt lt num3 gt lt nu
47. 0c eee HBOSC Output Syntax Oscillator Analysis Using Shooting Newton SNOSC Phase Noise Analysis PHASENOISE Input Syntax Phase Noise Algorithms Measuring PHASENOISE Analyses with MEASURE PHASENOISE Output Syntax 200 201 203 205 206 208 211 212 212 214 216 218 219 219 220 222 224 225 229 229 229 230 233 234 235 236 236 238 240 242 243 245 vii Contents viii 10 Phase Noise Analysis Options 000000 cee eee eee eae Jitter Analysis Jitter Input Syntax MEASURE Statements to Support Jitter Analysis RMS JITTER Measurement References 00 Large Signal Periodic AC Transfer Function and Noise Analyses Multitone Harmonic Balance AC Analysis HBAC 20 00 Prerequisites and Limitations Input Syntax Output Syntax HBAC Output Data Files Errors and Warnings Shooting Newton AC Analysis SNAC 006000 0c eee eee Prerequisites and Limitations Input Syntax Output Syntax SNAC Output Data Files Errors and Warnings SNAC Example Multitone Harmonic Balance Noise HBNOISE 20005 Supported Features Input Syntax Output Syntax Output Data Files Measuring HBNOISE Analyses wit
48. 1 sl nl n2 n3 n ref mname smodel model smodel s n 3 fqmodel sfqmodel zo 50 fbase 25e6 fmax 1e9 Example 2 sl ni n2 n3 n ref fqmodel sfqmodel zo 50 fbase 25e6 fmax 1e9 Examples 1 and 2 return the same result Example 3 sl nl n2 n3 n ref mname smodel zo 100 model smodel s n 3 fqmodel sfqmodel zo 50 fbase 25e6 fmax 1e9 In this example the characteristic impedance of each port is 100 ohms instead of 50 ohms as defined in smode1 because parameters defined in the S element statement have higher priority than those defined in the S model statement Example 4 sl nl n2 n3 n ref mname smodel model smodel s n 3 fqmodel sfqmodel zo 50 50 100 In this example the characteristic impedance of port1 and port2 are 50 ohms and the characteristic impedance of port3 is 100 ohms Example 5 sl nl n2 n3 n ref mname smodel model smodel s tstonefile expl s3p In this example the name of the tstone file expl s3p reveals that the network has three ports Example 6 sl nl n2 n3 n ref mname smodel model smodel s fqmodel sfqmodel tstonefile expl s3p citifile expl citio0 HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Multi Terminal Linear Elements In this example fqmodel tstonefile and citifile are all declared HSPICE uses only the fqmodel ignores tstonefile and citifile and reports warning messages Example 7 sl nl n2 n3 n ref mname smodel fqmodel sfqmodel 1 model smodel s n 3 fqmodel sfqmodel 2
49. 1 1 v1 f L f G f C f Vol4 nza 21 as ia Vi o Signal Conductors Vo o i o T i P l l m HIN ivil voly M r 2 Reference conductor E o 0 x For additional information about the W element see the W element Modeling of Coupled Transmission Lines chapter in the HSPICE Signal Integrity User Guide T element Ideal Transmission Lines General form Txxx in refin out refout ZO0 val TD val lt L val gt lt IC v1 i1 v2 12 gt Txxx in refin out refout ZO0 val F val lt NL val gt lt IC v1 il v2 i2 gt U Model form HSPICE RF User Guide 143 Z 2007 03 Chapter 6 Testbench Elements Multi Terminal Linear Elements 144 Txxx in refin out refout mname L val Parameter Description TXXX Lossless transmission line element name Must begin with T followed by up to 1023 alphanumeric characters in Signal input node refin Ground reference for the input signal out Signal output node refout Ground reference for the output signal ZO Characteristic impedance of the transmission line TD Signal delay from a transmission line in seconds per meter L Physical length of the transmission line in units of meters Default 1 IC v1 i1 v2 i2 Initial conditions of the transmission line Specify the voltage on the input port v1 current into the input port i1 voltage on the output port v2 and the current into the output port i2 F Frequency at which the transmission line has the e
50. 1 621684E 18 UC 3 422111E 11 VSAT 1 145012E5 AO 1 119634 AGS 0 1918651 BO 1 800933E 6 Bl 5E 6 KETA 3 313177E 3 Al 0 A2 1 RDSW 984 149934 PRWG 1 133763E 3 PRWB 7 19717E 3 WR 1 WINT 9 590106E 8 LINT 1 719803E 8 XL 5E 8 XW 0 DWG 2 019736E 9 DWB 6 217095E 9 VOFF 0 1076921 NFACTOR 0 CIT 0 CDSC 2 4E 4 CDSCD 0 CDSCB 0 ETAO 0 0147171 ETAB 7 256296E 3 DSUB 0 3377074 PCLM 1 1535622 PDIBLC1L 2 946624E 4 PDIBLC2 4 171891E 3 PDIBLCB 0 0497942 DROUT 0 0799917 PSCBE1 3 380501E9 PSCBE2 1 69587E 9 PVAG 0 4105571 DELTA 0 01 MOBMOD 1 PRT 0 UTE 1 5 KT1 0 11 KT1L 0 KT2 0 022 UAL 4 31E 9 UB1 7 61E 18 UC1 5 6E 11 AT 3 34 WL 0 WLN 1 WW 1 22182E 15 WWN 1 1657 WWL 0 LL 0 LLN 1 LW 0 LWN 1 LWL 0 CAPMOD 2 XPART 0 4 CGDO 3 73E 10 CGSO 3 73E 10 CGBO 1E 11 CI 8 988141E 4 PB 0 8616985 MJ 0 3906381 CUSW 2 463277E 10 PBSW 0 5072799 MJSW 0 1331717 PVTHO 0 0143809 PRDSW 81 683425 WRDSW 107 8071189 PK2 1 210197E 3 WKETA 1 00008E 3 LKETA 6 1699E 3 PAGS 0 24968 The following is the BUT model file bjt inc used in oscillator example It is available in directory lt installdir gt demo hspicerf examples HSPICE RF User Guide 27 Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 4 Using HBOSC Analysis for a Colpitts Oscillator RF Wideband NPN Transistor die SPICE MODEL
51. 15u M2 IN IP cs gnd NMOS 1 0 35u w 15u Bias fet bias at Vdd too high Mb cs vdd gnd gnd NMOS 1 0 35u w 15u fets used as varactors Mtl vc IP vc gnd NMOS 1 0 35u w 2u M 50 Mt2 vc IN vc gnd NMOS 1 0 35u w 2u M 50 Second oscillator section Low Q resonator with Vdd at center tap of inductors Rla_b QP ri_b 100k These R s set the Q Rlb b ri_b QN 100k L1 b QP vdd 16 5nH L2 b vdd QN 16 5nH Ccl_b QP gnd Cval Q to I Cc2_b QN gnd Cval Q to I Differential fets M1_b QP QN cs_b gnd NMOS 1 0 35u w 15u M2_b QN QP cs_b gnd NMOS 1 0 35u w 15u Bias fet bias at Vdd too high 2nd in parallel Mb_b cs_b vdd gnd gnd NMOS 1 0 35u w 15u xx fets used as varactors Mtl_b vc QP vc gnd NMOS 1 0 35u w 2u M 50 Mt2_b vc QN vc gnd NMOS 1 0 35u w 2u M 50 Differentiators Coupling transistors for quadrature param Cdiff 0 14p difMsize 50u vidiff dbias gnd 1 25 viqdiff vdcdif gnd 1 75 Midiff1 dQP dbias gnd gnd NMOS Midiff2 dQN dbias gnd gnd NMOS Midiff 3 dIN dbias gnd gnd NMOS Midiff4 dIP dbias gnd gnd NMOS Cdiffl dQP QP Cdiff Cdiff2 dQN QN Cdiff Cdiff3 dIN IN Cdiff Cdiff4 dIP IP Cdiff Mc_QP1 IP vdcdif dQP gnd NMOS 1 0 35u w difMsize Mc_QN2 IN vdcdif dQN gnd NMOS 1 0 35u w difMsize Mc_QN3 QP vdcdif dIN gnd NMOS 1 0 35u w difMsize Mc_QP4 QN vdcdif dIP gnd NMOS 1 0 35u w difMsize 35u w difMsize 35u w difMsize 35u w difMsize 35u w difMsize HSPICE RF User Guide 33 Z 2007 03 Chapter 3
52. 1720829 UA 1 300598E 9 UB 2 308197E 18 UC 2 841618E 11 VSAT 1 482651E5 AO 1 6856991 AGS 0 2874763 BO 1 833193E 8 Bl 1E 7 KETA 2 395348E 3 Al 0 A2 0 4177975 RDSW 178 7751373 PRWG 0 3774172 PRWB 0 2 WR 1 WINT 0 LINT 1 888394E 8 XL 3E 8 XW 4E 8 DWG 1 213938E 8 DWB 4 613042E 9 VOFF 0 0981658 NFACTOR 1 2032376 CIT 0 CDSC 2 4E 4 CDSCD 0 CDSCB 0 ETAO 5 128492E 3 ETAB 6 18609E 4 DSUB 0 0463218 PCLM 1 91946 PDIBLC1 1 PDIBLC2 4 422611E 3 PDIBLCB 0 1 DROUT 0 9817908 PSCBE1 7 982649E10 PSCBE2 5 200359E 10 PVAG 9 314435E 3 DELTA 0 01 RSH 3 7 MOBMOD 1 PRT 0 UTE 1 5 KT1 0 11 KT1L 0 KT2 0 022 UA1 4 31E 9 UB1 7 61E 18 UC1 5 6E 11 AT 3 3E4 WL 0 WLN 1 WW 0 WWN 1 WWL 0 LL 0 LLN 1 LW 0 LWN 1 LWL 0 CAPMOD 2 XPART 0 5 CGDO 5 62E 10 CGSO 5 62E 10 CGBO 1E 12 CJ 1 641005E 3 PB 0 99 MJ 0 4453094 CJSW 4 179682E 10 PBSW 0 99 MJSW 0 3413857 CJSWG 3 29E 10 PBSWG 0 99 MJSWG 0 3413857 CF 0 PVTHO 8 385037E 3 PRDSW 10 PK2 2 650965E 3 WKETA 7 293869E 3 LKETA 6 070221E 3 END First notice that we have defined variables that allow power to be swept in dBm units param Pin dBm 30 0 param Pin Pin dBm _param Pin W 1 0e 3 pwr 10 0 Pin 10 0 24 HSPICE RF User Guide Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 3 Using HB Anal
53. 1e15 ohms HSPICE RF User Guide 347 Z 2007 03 Chapter 13 Post Layout Analysis Linear Acceleration Table 27 PACT Options Continued Syntax Description OPTION SIM_LA_MINC lt value gt Minimum capacitance for linear matrix reduction value is the minimum capacitance preserved in the reduction After reduction completes SIM_LA lumps any capacitor smaller than value to ground The default is 1e 16 farads OPTION SIM_LA_MINMODE Reduces the number of nodes instead of the ON OFF number of elements OPTION SIM_LA_TIME lt value gt Minimum time for which accuracy must be preserved value is the minimum switching time for which the PACT algorithm preserves accuracy HSPICE RF does not accurately represent waveforms that occur more rapidly than this time SIM_LA_TIME is simply the dual of SIM_LA_FREQ The default is equivalent to setting LA_FREQ 1 GHz The default is 1ns OPTION SIM_LA_TOL lt value gt Error tolerance for the PACT algorithm value is the error tolerance for the PACT algorithm is between 0 0 and 1 0 The default is 0 05 Example In this example the circuit has a typical risetime of 1ns Set the maximum frequency to 1 GHz or set the minimum switching time to 1ns OPTION SIM LA FREQ 1GHz OPTION SIM LA TIME Ins However if spikes occur in 0 1ns HSPICE will not accurately simulate them To capture the behavior of the spikes use OPTION SIM LA FREQ 10GHz OPTION SIM LA TIME O 1ns N
54. 2 3 5 6 7 8 9 S Maas Nonlinear Microwave Circuits Chapter 3 IEEE Press 1997 R Gilmore and M B Steer Nonlinear Circuit Analysis Using the Method of Harmonic Balance A Review of the Art Part I Introductory Concepts International Journal of Microwave and Millimeter wave Computer Aided Engineering Volume 1 No 1 pages 22 37 1991 R Gilmore and M B Steer Nonlinear Circuit Analysis Using the Method of Harmonic Balance A Review of the Art Part Il Advanced Concepts International Journal of Microwave and Millimeter wave Computer Aided Engineering Volume 1 No 2 pages 159 180 1991 V Rizzoli F Mastri F Sgallari G Spaletta Harmonic Balance Simulation of Strongly Nonlinear Very Large Size Microwave Circuits by Inexact Newton Methods MTT S Digest pages 1357 1360 1996 S Skaggs Efficient Harmonic Balance Modeling of Large Microwave Circuits Ph D thesis North Carolina State University 1999 R S Carson High Frequency Amplifiers 2nd Edition John Wiley amp Sons 1982 S Y Liao Microwave Circuit Analysis and Amplifier Design Prentice Hall 1987 J Roychowdhury D Long P Feldmann Cyclostationary Noise Analysis of Large RF Circuits with Multitone Excitations IEEE JSCC volume 33 number 3 March 1998 Y Saad Iterative Methods for Sparse Linear Systems PWS Publishing Company 1995 10 J Roychowdhury D Long and P Feldmann Cyclo
55. 2 m delay 1 7929E 10 targ 3 4539E 10 trig 1 6610E 10 m power 6 6384E 03 from 0 0000E 00 to 1 0000E 09 In the preceding listing the m_delay value of 1 79e 10 seconds is the fastest pair delay You can also examine the Monte Carlo parameters that produced this result HSPICE RF User Guide Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Worst Case and Monte Carlo Sweep Example The information on shortest delay and so forth is also available from the statistics section at the end of the output listing While this information is useful to determine whether the circuit meets specification it is often desirable to understand the relationship of the parameters to circuit performance Plotting the results against the Monte Carlo index number does not help for this purpose You need to generate plots that display a Monte Carlo result as a function of a parameter For example Figure 43 shows the inverter pair delay to channel as a function of poly width which relates directly to device length Figure 43 Delay as a function of Poly width XL Monte Carla Results xI polyed 3005 m_delayxl polyed o 250p 200p Xl polyol Figure 44 shows the pair delay against the TOX parameter The scatter plot shows no obvious dependence which means that the effect of TOX is much smaller than XL To explore this in more detail set the XL skew parameter to a constant and run a simulation HSPICE RF User Guide
56. 230 XL 0 18u DELVTON 15V DELVTOP 0 15V INC usr meta lib cmos1 mod dat model include file ENDL FF The usr meta lib cemos1_mod dat include file contains the model MODEL NCH NMOS LEVEL 2 XL XL TOX TOX DELVTO DELVTON MODEL PCH PMOS LEVEL 2 XL XL TOX TOX DELVTO DELVTOP Note The model keyname left equals the skew parameter right Model keys and skew parameters can use the same names HSPICE RF User Guide 357 Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Worst Case Analysis Skew File Interface to Device Models Skew parameters are model parameters for transistor models or passive components A typical device model set includes MOSFET models for all device sizes by using an automatic model selector RC wire models for polysilicon metal1 and metal2 layers in the drawn dimension Models include temperature coefficients and fringe capacitance Single diode and distributed diode models for N P and well includes temperature leakage and capacitance based on the drawn dimension BJT models for parasitic bipolar transistors You can also use these for any special BUTs such as a BiCMOS for ECL BUT process includes current and capacitance as a function of temperature Metali and metal2 transmission line models for long metal lines Models must accept elements Sizes are based on a drawn dimension If you draw a cell at 2u dimension and shrink it to 1u the physical size is 0 9u The ef
57. 262 e GMRES residual e Actual Krylov iterations taken e Frequency swept input frequency values Errors and Warnings The following error and warning messages are used when HSPICE encounters a problem with a HBAC analysis Error Messages HBAC frequency sweep includes negative frequencies HBAC allows only frequencies that are greater than or equal to zero No HB statement is specified error at parser HBAC requires an HB statement to generate the steady state solution Warning Messages More than one HBAC statement warning at parser HSPICE RF uses only the last HBAC statement in the netlist No HBAC sources are specified error at parser HBAC requires at least one HBAC source GMRES Convergence Failure When GMRES Generalized Minimum Residual reaches the maximum number of iterations and the residual is greater than the specified tolerance The HBAC analysis generates a warning and then continue as if the data were valid This warning reports the following information Final GMRES Residual Target GMRES Residual m Maximum Krylov Iterations Actual Krylov Iterations taken HBAC Example The following example is shipped with the HSPICE RF distribution as mix_hbac sp and is available in directory lt installdir gt demo hspicerf examples HSPICE RF User Guide Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Shooting Newton AC Analysis SNAC Test HBAC ideal
58. 300 dBc Hz Dumps the element phase noise value to the lis file When the element has multiple noise sources such as a level 54 MOSFET which contains the thermal shot and 1 f noise sources When dumping the element phase noise value you can decide if you need to dump the contribution from each noise source You can specify either ON or OFF ON dumps the contribution from each noise source and OFF does not The default value is OFF HSPICE RF User Guide Z 2007 03 241 Chapter 9 Oscillator and Phase Noise Analysis Phase Noise Analysis 242 Phase Noise Algorithms HSPICE RF provides three algorithms for oscillator phasenoise nonlinear perturbation periodic AC and broadband calculations These algorithms are selected by setting the METHOD parameter to 0 1 or 2 respectively Each algorithm has their regions of validity and computational efficiency so some thought is necessary to obtain meaningful results from a PHASENOISE simulation For each algorithm the region of validity depends on the particular circuit being simulated However there are some general rules that can be applied to oscillator types that is ring or harmonic so that a valid region can be identified And there are techniques that can be used to check validity of your simulation results Nonlinear Perturbation Algorithm The nonlinear perturbation NLP algorithm which is the default selection is typically the fastest computation but is valid onl
59. 6 HBAC takes care of the second tone Adda HBAC command HBAC lin 1 0 8g 0 8g This command runs an analysis at a single frequency point 0 8 GHz In general HBAC analysis can sweep the RF frequency over a range of values The following is the complete mix_hbac sp netlist for HBAC analysis of this simple mixer This netlist also contains commands for performing periodic noise analysis It is available in directory lt installdir gt demo hspicerf examples HSPICE RF User Guide Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 6 Using Multi Tone HB and HBAC Analyses for a Mixer Ideal mixer example HBAC analysis OPTIONS POST vlo lo 0 1 0 sin 1 0 0 5 1 0g 0 0 90 HB 0 5 O 1 1 fref TEL LE 1 0 gl 0 if cur 1 0 v lo v rf mixer element cl 0 if q 1 0e 9 v lo v rf mixer element rout if ifg 1 0 vetrl ifg 0 0 0 hl out 0 vetrl 1 0 convert I to V rhi out 0 1 0 vrf rfl 0 sin 0 0 001 0 8GHz 0 0 114 HBAC 0 001 24 opt sim_accuracy 100 hb tones 1g nharms 6 hbac lin 1 0 8g 0 8g Noise analysis hbnoise v out rrfl lin 40 0 1g 4g print hbnoise onoise nf probe hbnoise onoise nf end Comparing Results After running all three netlists above you will have generated 3 output files mix_tran trO mix_hb hbO m mix_hbac hbO You can compare the results of the 3 analyses in CosmosScope 1 Torun the netlists and start CosmosScope type hspicerf mix _tran sp hspicerf mix_hb sp hspic
60. ALTER statement into a library Use the LIB statement in the main input file Puta DEL LIB statement inthe ALTER section to delete that library for the ALTER simulation run Altering Design Variables and Subcircuits The following rules apply when you use an ALTER block to alter design variables and subcircuits in HSPICE RF If the name of a new element MODEL statement or subcircuit definition is identical to the name of an original statement of the same type then the new statement replaces the old Add new statements in the input netlist file You can alter element and MODEL statements within a subcircuit definition You can also add a new element or MODEL statement to a subcircuit definition To modify the topology in subcircuit definitions put the element into libraries To add a library use LIB to delete use DEL LIB If a parameter name in a new PARAM statement in the ALTER module is identical to a previous parameter name then the new assigned value replaces the old value If you used parameter variable values for elements or model parameter values when you used ALTER use the PARAM statement to change these parameter values Do not use numerical values to redescribe elements or model parameters If you used an OPTION statement in an original input file ora ALTER block to turn on an option you can turn that option off Each ALTER simulation run prints only the actual altered input A spec
61. About This Guide Customer Support Contacting the Synopsys Technical Support Center If you have problems questions or suggestions you can contact the Synopsys Technical Support Center in the following ways Open a call to your local support center from the Web by going to http solvnet synopsys com EnterACall Synopsys user name and password required Send an e mail message to your local support center e E mail support_center synopsys com from within North America e Find other local support center e mail addresses at http www synopsys com support support_ctr Telephone your local support center e Call 800 245 8005 from within the continental United States e Call 650 584 4200 from Canada e Find other local support center telephone numbers at http www synopsys com support support_ctr HSPICE RF User Guide xix Z 2007 03 About This Guide Customer Support XX HSPICE RF User Guide Z 2007 03 P HSPICE RF Features and Functionality Introduces HSPICE RF features and functionality HSPICE RF is a special set of analysis and design capabilities that support the design of RF and high speed circuits This functionality built on top of the standard HSPICE feature set is also useful for analog and signal integrity applications Although the HSPICE and HSPICE RF simulators share a common set of device models and simulation capabilities HSPICE RF includes several modeling simulation and measurement ad
62. Analysis Port element FIBLIN Analysis 2 04si0 eGo ange ueade aie ae E EE tans a dine Output Syntax Output Data Files 1 2 2 0 0 00 ccc eee Large Signal S parameter HBLSP Analysis 0 00000 Limitations Input Syntax Output Syntax Output Data FileS seori ee eihan ie kee eee a a eds ates Envelope Analysis Envelope Simulation Envelope Analysis Commands 0000 eee eee eee eee Output Syntax 281 282 284 287 287 288 289 289 290 290 291 291 291 292 293 293 294 295 297 298 300 301 301 304 305 305 306 307 309 309 311 311 312 315 Contents 13 14 Envelope Output Data File Format Post Layout Analysis 0 0 00 eee eee Post Layout Back Annotation 000000 eee eee Standard Post Layout Flow 00000 cues Selective Post Layout Flow 0000 ease Additional Post Layout Options 04 Selective Extraction Flow 00 cece eee Overview of DSPF FileS 00 0c cee uee Overview of SPEF FileS 0 00000 e ce eeuas Linear Acceleration 0020 nananana PACT Algorithm nesas ieee 0c eee ee PI AIQO rit ee ras eine Red eee a eee we ct Re ee ee Linear Acceleration Control Options Summary Statistical and Monte Carlo Analysis OVErNIOW ie ee a Pe Le a ea Application of Statistical Analysis
63. CP X1 XA 1 0 0 1P In this example Cbypass is a straightforward 10 pF capacitor C1 calls the CBX model and its capacitance is not constant cBis a 10 pF capacitor with an initial voltage of 4v across it CPisa 0 1 pF capacitor Example 2 V1 1 0 pwl 0n Ov 100n 10v V2 2 0 pwl 0n Ov 100n 10v C1 1 0 C V 1 V 2 1le 12 CTYPE 2 Example 3 HSPICE RF Only C2 1 0 C 1 TIME Time varying capacitor Charge Based Capacitors You can also specify capacitors using behavioral equations for charge HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Passive Elements Syntax Cxxx nl n2 Q equation dQ f C qv V V n1 n2 is equivalent to Cxxx a b Q f V a b In the preceding equations d x we X Example 1 Cl a b Q sin V a b V c d V a b This example is equivalent to C1 a b C eos V a b Vie d Example 2 C3 3 0 Q TIME TIME supported in HPICE RF only Linear Capacitors Cxxx nodel node2 lt modelname gt lt C gt value lt TCl val gt lt TC2 val gt lt W val gt lt L val gt lt DTEMP val gt lt M val gt lt SCALE val gt lt IC val gt Parameter Description Cxxx Name of a capacitor Must begin with C followed by up to 1023 alphanumeric characters node1 and node2 Names or numbers of connecting nodes value Nominal capacitance value in Farads modelname Name of the capacitor model C Capacitance in Farad
64. Circuit and a Ring Oscillator 44 initialization time The initial transient analysis is used for circuit stabilization before the steady state solution is found Larger initialization values typically result in convergence that is more robust m For oscillator circuits a SNOSC command is used to activate the analysis The SNOSC command specifies e The approximate frequency of oscillation specified either as a frequency in Hertz or as the time domain period e The number of high frequency harmonics Alternatively a time resolution in seconds can be specified e A transient initialization time that is used by HSPICE RF to run a basic transient simulation of this length before attempting Newton Raphson iterations to converge on a steady state solution This parameter is optional If it is not specified the period of the specified frequency of oscillation is used as the initialization time For oscillators we recommend specifying a transient initialization time since the default initialization time is usually too short to effectively stabilize the circuit e Anode at which to probe for oscillation conditions e If the tuning curve of a VCO is to be analyzed the optional parameter MAXTRINITCYCLES can be specified One or more signal sources for driving the circuit in SN analysis if the circuit is driven In the case of autonomous oscillator analysis no signal source is required The sources are required to be time domain so
65. DC isolation Input and output have a common pin and both inductors have the same value Note that Rload 4 Rin kk all K s ideal kin O o out in L1 1 RR O ENE o 0 L2 1 RR eC o out2 kk xx With all K s ideal the actual L s values are not important only their ratio to each other subckt BALUN2 in out2 L1 in gnd L 1 L2 gnd out2 L 1 K12 L1 L2 IDEAL ends Example 3 This example is a 3 pin ideal balun transformer with shared DC path no DC isolation All inductors have the same value here set to unity HSPICE RF User Guide 133 Z 2007 03 Chapter 6 Testbench Elements Passive Elements kk all K s ideal o outl eX Lo2 1 BOR T o 0 E Lol 1 EA a S RGE o out2 x in Lin 1 a o in kk subckt BALUN3 in outi out2 Lo2 gnd outi L 1 Lol out2 gnd L 1 Lin in out2 L 1 K12 Lin Lol IDEAL K13 Lin Lo2 IDEAL K23 Lol Lo2 IDEAL ends Coupled Inductor Element This section describes the multiport syntax for coupled inductor elements This syntax extends the existing linear Lxxx and mutual Kxxx inductor elements Two syntax configurations are available areluctance format that is used by Star RCXT for inductance extraction an ideal transformer format that can be used to create balanced converter that is balun models in HSPICE RF Reluctance Format Syntax Reluctance Inline Form Lxxx nlp nin nNp nNn RELUCTANCE r1 cl vall r2 c2 val2
66. Dependent Noise Analysis PTDNOISE Ideal mixer noise source prints total noise PSD at the output 2 47e 20 V 2 when q 0 Single sideband noise figure 3 01 dB double sideband noise figure 0 dB OPTION PROBE OPTION POST 2 vlo lo 0 0 0 cos 0 1 0 1 0g 0 0 0 Ilo lo 0 0 rsrc rfin rfl 1 0S Noise source gl 0 if cur 1 0 v lo v rfin mixer element cl 0 if g 1 0e 9 v lo v rfin mixer element rout if 0 1 0 vrf rfl 0 hbac 2 0 0 0 option delmax 0 002n SN tones 1G nharms 4 trstab 10n SNNOISE rout rsrc lin 11 0 90g 0 92g probe SNNOISE onoise ssnf dsnf print SNNOISE onoise ssnf dsnf end Periodic Time Dependent Noise Analysis PTDNOISE While HBNOISE and SNNOISE calculate a time averaged power spectral density there are applications where a characterization of the time dependence of the noise is required These applications include computation of jitter associated with a noisy signal crossing a threshold and computation of the noise associated with discretizing an analog signal which computes the noise in a periodically driven circuit at a point in time Periodic Time Dependent noise analysis PTDNOISE calculates the noise spectrum and the total noise at a point in time Jitter in a digital threshold circuit can then be determined from the total noise and the digital signal slew rate Circuits driven by large periodic signals produce cyclostationary noise that is the noise characteristi
67. Elements Steady State Voltage and Current Sources Parameter Description lt rac val gt AC analysis Series resistance overrides z0 lt RHBAC val gt HSPICE RF HBAC analysis Series resistance overrides z0 lt rhb val gt HSPICE RF HB analysis Series resistance overrides z0 lt rtran val gt Transient analysis Series resistance overrides z0 lt TRANFORHB 0 1 gt Example 1 O0 default The transient description is ignored if an HB value is given or a DC value is given If no DC or HB value is given and TRANFORHB 0 then HB treats the source as a DC source and the DC source value is the time 0 value 1 HB analysis uses the transient description if its value is VMRF SIN PULSE PWL or LFSR If the type is a non repeating PWL source then the time infinity value is used as a DC source value For example the following statement is treated as a DC source with value 1 for HB vi 10 PWL 00 1n 1 1u 1 TRANFORHB 1 In contrast the following statement is a OV DC source vi 10 PWL 00 1n 1 1u 1 TRANFORHB 0 The following statement is treated as a periodic source with a 1us period that uses PWL values vi 10 PWL 00 1n 1 0 999u 1 1u 0 R TRANFORHB 1 To override the global TRANFORHB option explicitly set TRANFORHB for a V I source This example shows an HB source for a single tone analysis hb tones 100MHz harms 7 11 1 2 dc 1mA hb 3mA 0 1 1 I1 is a current source with a the followin
68. GRAPH simulation output uses PLOT model type WIDTH and OPTION CO OPTION ACCT Element template output Group time delay parameters in AC analysis output DISTO distortion analysis and associated output commands SAVE and LOAD HSPICE RF User Guide Z 2007 03 HSPICE Simulation and Analysis User Guide HSPICE Simulation and Analysis User Guide HSPICE and RF Command Reference HSPICE and RF Command Reference HSPICE and RF Command Reference HSPICE Simulation and Analysis User Guide HSPICE and RF Command Reference HSPICE Simulation and Analysis User Guide HSPICE and RF Command Reference HSPICE Simulation and Analysis User Guide HSPICE and RF Command Reference HSPICE Simulation and Analysis User Guide HSPICE Simulation and Analysis User Guide HSPICE Simulation and Analysis User Guide HSPICE and RF Command Reference Chapter 1 HSPICE RF Features and Functionality HSPICE and HSPICE RF Differences Table 1 HSPICE Features Not in HSPICE RF Continued Feature See Options that activate unsupported features in HSPICE and HSPICE RF Command Reference HSPICE RF FAST GSHDC GSHUNT LIMPTS OFF RESMIN TIMERES All version options Options ignored by HSPICE RF because they are not needed since they are replaced by automated algorithms ABSH ABSV ABSVAR BELV BKPSIZ CHGTOL CONVERGE CSHDC CVTOL DCFOR DCHOLD DCON DCSTEP DI DV DVDT FAST FS FT GMAX GRAMP GSHDC GSHUNT ICSWEEP IMAX IMIN
69. Guide Z 2007 03 Chapter 9 Oscillator and Phase Noise Analysis Phase Noise Analysis cyclo or cyclostationary means anything bias dependent 2 The following syntax is frequency dependent and bias independent print phasenoise phnoise stationary 3 The following syntax is frequency dependent and bias dependent print phasenoise phnoise flicker 4 The following syntax is frequency dependent and bias dependent print phasenoise phnoise cycloflicker Also acceptable is print phasenoise phnoise cyclostationaryflicker The phnoise_fdep is a combination of phnoise_ flicker and phnoise_ cycloflicker The phnoise_findep is a combination of phnoise stationary and phnoise_ cycloflicker Table 21 Summary of Noise Type Dependences Noise type phnoise Stationary phnoise Cyclostationary phnoise Flicker phnoise CycloFlicker phnoise Fdep phnoise Findep frrequency dependent bias dependent No No No Yes Yes No Yes Yes is the union of phnoise Flicker and phnoise CycloFlicker noise types is the union of phnoise Stationary and phnoise CycloFlicker noise types HSPICE RF User Guide Z 2007 03 247 Chapter 9 Oscillator and Phase Noise Analysis Jitter Analysis Phase Noise Analysis Options Table 22 lists the control options specific to PHASENOISE applications Table 22 PHASENOISE Analysis Options Parameter Description BPNMATCHTOL val PHASENOISEKRYLOVDIM PHASENOISEKRYLOVITER
70. Guide which is available in PDF format only Other Related Publications xvi For additional information about lt Product Name gt see The HSPICE release notes available on SolvNet See Known Limitations and Resolved STARs below Documentation on the Web which provides PDF documents and is available through SolvNet at http solvnet synopsys com DocsOnWeb You might also want to refer to the documentation for the following related Synopsys products m CosmosScope Aurora HSPICE RF User Guide Z 2007 03 About This Guide Conventions Raphael VCS Known Limitations and Resolved STARs You can find information about known problems and limitations and resolved Synopsys Technical Action Requests STARs in the lt Product Name gt Release Notes in SolvNet To see the latest lt Product Name gt Release Notes 1 Go to https solvnet synopsys com ReleaseNotes If prompted enter your user name and password If you do not have a Synopsys user name and password follow the instructions to register with SolvNet 2 Click lt Product Name gt then click the release you want in the list that appears at the bottom Conventions The following conventions are used in Synopsys documentation Convention Description Courier Indicates command syntax Italic Indicates a user defined value such as object_name Bold Indicates user input text you type verbatim in syntax and examples Denotes optiona
71. I Q mixer OPTIONS PROBE OPTIONS POST 2 vlo lo 0 1 0 cos 1 0 0 5 1g TRANFORHB 1 Periodic Large Signal Input rlo lo 0 1 0 rrf rf 0 1 0 Noise source rrfi rfi rf 1 0 Noise source gl 0 if cur 1 0 v lo v rf mixer element cl 0 if g 1 0e 9 v lo v rf mixer element rout if ifg 1 0 vetrl ifg 0 0 0 hl out 0 vcetrl 1 0 rhi out 0 1 0 vrf rfl 0 hbac 001 0 Small signal source hb tones 1 0g nharms 3 hbac lin 1 0 8g 0 8g print hb v rf1 v lo v out probe hb v rf1 v lo v out measure hb voutl find v out 1 1 at 0 8g end Shooting Newton AC Analysis SNAC You use the Shooting Newton AC SNAC statement for analyzing linear behavior in large signal periodic systems The SNAC statement uses a periodic AC PAC and Shooting Newton algorithm to perform linear analysis of nonautonomous driven circuits where the linear coefficients are modulated by a periodic steady state signal The following section describes the periodic AC analysis based on a Shooting Newton algorithm This functionality is similar to the Harmonic Balance HBAC for periodic AC analysis Prerequisites and Limitations The following prerequisites and limitations apply to SNAC Requires one and only one SNAC statement If you use multiple SNAC statements HSPICE RF uses only the last SNAC statement Requires one and only one SN statement Requires placing the parameter sweep in the SN statement Requires at leas
72. IEEE daha al me eee a wate 64 Hierarchy Paths ee ie aaae a Ea E a ea A a E 66 NUMDEIS a ood Saeed EE ee ARAE A E E dees 66 Parameters and Expressions 000 eee eee eee eee 67 Input Netlist File Structure 0 0 0 0 0000 c cee eee ee 68 Schematic Netlists 0 0000 c eee 68 Input Netlist File Composition 0 0 0 eee eee 70 Title oF SiMUlAtION ees desea del dees eek Tele ae Soha ala aid 71 Comments and Line Continuation 0 0000 cee eee eee 71 Element and Source Statements 0000 0c eee eee 72 Defining Subcircuits ri a i a ae eee 75 Node Naming Conventions 0000 0c cece eee eee 75 Element Instance and Subcircuit Naming Conventions 78 Subcircuit Node Names 0020 cece eae 78 Path Names of Subcircuit Nodes 20000 0c eee eee eee 79 Automatic Node Name Generation 0000e ceca eee 80 Global Node Names 00 0c cece eee eee 80 Circuit Temperature 0000 c eee 80 Data Driven Analysis orse rerea ie neee nmana a e a eee 81 Library Calls and Definitions aaaea 81 6 Defining Parameters Deleting a Library Ending aNetlist Contents Condition Controlled Netlists IF ELSE 2 00 00000 Using Subcircuits Hierarchical Parameters DDL Library Access Vendor Libraries 00005 Sub
73. MAX are frequency values in units of Hz If the FSPTS analysis finds an approximate oscillation frequency the TONE parameter will be ignored 231 Chapter 9 Oscillator and Phase Noise Analysis Input Syntax for Harmonic Balance Oscillator Analysis Parameter Description SWEEP Specifies the type of sweep You can sweep up to three variables You can specify either LIN DEC OCT POI SWEEPBLOCK DATA OPTIMIZE or MONTE Specify the nsteps start and stop frequencies using the following syntax for each type of sweep LIN nsteps start stop DEC nsteps start stop OCT nsteps start stop POI nsteps freq_values SWEEPBLOCK nsteps freq1 freq2 freqn DATA dataname OPTIMIZE OPT xxx MONTE val SUBHARMS Allows subharmonics in the analysis spectrum The minimum non DC frequency in the analysis spectrum is f subharms where f is the frequency of oscillation Use this option if your oscillator circuit includes a divider or prescaler that will result in frequency terms that are subharmonics of the fundamental oscillation frequency Example 1 HBOSC tone 900MEG nharms 9 probenode gate gnd 0 65 Performs an oscillator analysis searching for frequencies in the vicinity of 900 MHz This example uses nine harmonics with the probe inserted between the gate and gnd nodes The probe voltage estimate is 0 65 V Example 2 HBOSC tone 2400MEG nharms 11 probenode drainP drainN 1 0 fspts 20 2100MEG 2700MEG Performs an oscillator analy
74. Newton with Fourier Transform SNFT on page 224 HSPICE RF User Guide 219 Z 2007 03 Chapter 8 Steady State Shooting Newton Analysis SN Analysis Syntax Shooting Newton based AC analysis SNAC See Shooting Newton AC Analysis SNAC on page 263 Shooting Newton based noise analysis SGNNOISE See Oscillator Analysis Using Shooting Newton SNOSC on page 236 SN Analysis Syntax 220 Shooting Newton provides two syntaxes Syntax 1 is recommended when you are using making Time Domain sources and measurements for example going from TRAN to SN Syntax 2 effectively supports Frequency Domain sources and measurements and should be used for example when going from HB to SN Syntax 1 SN TRES lt Tr gt PERIOD lt T gt TRINIT lt Ti gt SWEEP parameter sweep MAXTRINITCYCLES lt integer gt or Syntax 2 SN TONE lt F1 gt NHARMS lt N gt TRINIT lt Ti gt SWEEP parameter sweep MAXTRINITCYCLES lt integer gt where Parameter Description TRES The time resolution to be computed for the steady state waveforms in seconds PERIOD The expected period T seconds of the steady state waveforms Enter an approximate value when using for oscillator analysis The period of the steady state waveform may be entered either as PERIOD or its reciprocal TONE TONE The fundamental frequency in Hz NHARMS Specifies the number of high frequency harmonic components to include in the analys
75. Option HBCONTINUE HBJREUSE HBJREUSETOL HBKRYLOVDIM HBKRYLOVTOL HBLINESEARCHFAG HSPICE RF User Guide Z 2007 03 Description Specifies whether to use the sweep solution from the previous simulation as the initial guess for the present simulation HBCONTINUE 1 default Use solution from previous simulation as the initial guess HBCONTINUE 0 Start each simulation in a sweep from the DC solution Controls when to recalculate the Jacobian matrix HBJREUSE 0 recalculates the Jacobian matrix at each iteration HBJREUSE 1 reuses the Jacobian matrix for several iterations if the error is sufficiently reduced The default is 0 if HBSOLVER 1 or 2 or 1 if HBSOLVER 0 Determines when to recalculate Jacobian matrix if HBJREUSE 1 The percentage by which HSPICE RF must reduce the error from the last iteration so you can use the Jacobian matrix for the next iteration Must be a real number between 0 and 1 The default is 0 05 Dimension of the Krylov subspace that the Krylov solver uses Must be an integer greater than zero Default is 40 The error tolerance for the Krylov solver Must be a real number greater than zero The default is 0 01 The line search factor If Newton iteration produces a new vector of HB unknowns with a higher error than the last iteration then scale the update step by HBLINESEARCHFAC and try again Must be a real number between 0 and 1 The default is 0 35 203 Chapte
76. PHASENOISETOL PHNOISELORENTZ val Determines the minimum required match between the NLP and PAC phase noise algorithms An acceptable range is 0 05dB to 5dB The default is 0 5dB Specifies the dimension of the Krylov subspace that the Krylov solver uses This must be an integer greater than zero The default is 500 Specifies the maximum number of Krylov iterations that the phase noise Krylov solver takes Analysis stops when the number of iterations reaches this value The default is 1000 Specifies the error tolerance for the phase noise solver This must be a real number greater than zero The default is 1e 8 Turns on a Lorentzian model for the phase noise analysis val 0 uses a linear approximation to a Lorentzian model val 1 default applies a lorentzian model to all noise sources val 2 applies a Lorentzian model to all non frequency dependent noise sources val 3 Lorentzian model applied to white noise source Gaussian model applied to flicker noise sources Jitter Analysis HSPICE RF provides several jitter measurements This section defines describes and compares the various jitter measurements in HSPICE RF Jitter measurements are derived from the results of an HSPICE RF phase noise 248 HSPICE RF User Guide Z 2007 03 Chapter 9 Oscillator and Phase Noise Analysis Jitter Analysis analysis The relationships between phase noise and the latest jitter measurements are presented here
77. PROBE Statements PRINT HBLIN Smn Smn TYPE S m n S m n TYPE PROBE HBLIN Smn Smn TYPE S m n S m n TYPE PRINT HBLIN SXYmn SXYmn TYPE SXY m n SXY m n TYPE PROBE HBLIN SXYmn SXYmn TYPE SXY m n SXY m n TYPE PRINT HBLIN lt NF gt lt SSNF gt lt DSNF gt PROBE HBLIN lt NF gt lt SSNF gt lt DSNF gt Parameter Description Smn Smn TYPE Complex 2 port parameters Where S m n S m n TYPE m 1or2 SXYmn SXYmn TYPE a n 10r2 SXY m n SXY m n TYPE X and Y are used for mixed mode S parameter output The values for X and Y can be D differential C common or S single end TYPE R M P PD D DB or DBM R real imaginary M magnitude P PD phase in degrees D DB decibels DBM decibels per 1 0e 3 304 HSPICE RF User Guide Z 2007 03 Chapter 11 S parameter Extraction Large Signal S parameter HBLSP Analysis Parameter Description NF NF and SSNF both output a single side band noise SSNF figure as a function of the IFB points NF SSNF 10 Log SSF Single side band noise factor SSF Total Noise at output at OFB originating from all frequencies Load Noise originating from OFB Input Source Noise originating from IFB DSNF DSNF outputs a double side band noise figure as a function of the IFB points DSNF 10 Log DSF Double side band noise factor DSF Total Noise at output at the OFB originatin
78. Passing Table 15 PARHIER LOCAL vs PARHIER GLOBAL Results Continued Element PARHIER Local PARHIER Global r2 2 0 1 0 r3 3 0 1 0 Parameter Passing Solutions The checklist below determines whether you will see simulation differences when you use the default scoping rules These checks are especially important if your netlists contain devices from multiple vendor libraries m Check your sub circuits for parameter defaults on the SUBCKT or MACRO line m Check your sub circuits fora PARAM statement within a SUBCKT definition To check your circuits for global parameter definitions use the PARAM statement If any of the names from the first three checks are identical set up two HSPICE simulation jobs one with OPTION PARHIER GLOBAL and one with OPTION PARHIER LOCAL Then look for differences in the output HSPICE RF User Guide 111 Z 2007 03 Chapter 5 Parameters and Functions Parameter Scoping and Passing 112 HSPICE RF User Guide Z 2007 03 6 Testbench Elements Describes the syntax for the basic and specialized elements supported by HSPICE RF for high frequency analysis and characterization Elements are local and sometimes customized instances of a device model specified in your design netlist For descriptions of the standard device models on which elements instances are based see the HSPICE Elements and Device Models Manual and the HSPICE MOSFET Models Manual For signal integr
79. RF User Guide Z 2007 03 Chapter 7 Steady State Harmonic Balance Analysis Harmonic Balance Analysis Parameter Description NHARMS INTMODMAX SS_TONE SWEEP Number of harmonics to use for each tone Must have the same number of entries as TONES You must specify NHARMS INTMODMAX or both INTMODMAxX is the maximum intermodulation product order that you can specify in the analysis spectrum You must specify NHARMS INTMODMAxX or both Small signal tone number for HBLIN analysis The value must be an integer number The default value is 0 indicating that no small signal tone is specified For additional information see Frequency Translation S Parameter HBLIN Extraction on page 298 Type of sweep You can sweep up to three variables You can specify either LIN DEC OCT POI SWEEPBLOCK DATA OPTIMIZE or MONTE Specify the nsteps start and stop frequencies using the following syntax for each type of sweep LIN nsteps start stop DEC nsteps start stop OCT nsteps start stop POI nsteps freq_values SWEEPBLOCK nsteps freq1 freq2 freqn DATA dataname OPTIMIZE OPT xxx MONTE val HB Analysis Spectrum The NHARMS and INTMODMAX input parameters define the spectrum If INTMODMAX N the spectrum consists of all f a f b fo n fn frequencies so that f gt 0 and a b n lt N The a b n coefficients are integers with absolute value lt N f you do not specify INTMODMAX it defaul
80. Time values are in col1 and voltage or current values are in col2 By default coli 1 and col2 2 Repeat function When an argument is not specified the source repeats from the beginning of the function The argument repeated is the time in seconds which specifies the start point of the waveform being repeat The repeat time must be less than the greatest time point in the file Time delay in seconds of the PWL function Any standard V or source options 188 HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements SWEEPBLOCK in Sweep Analyses Example vit It plus It neg PWL PWLFILE Imod dat SWEEPBLOCK in Sweep Analyses You can use the SWEEPBLOCK statement to specify complicated sweeps Sweeps affect DC sweep analysis Parameter sweeps around TRAN AC or HB analyses Frequency values used in AC or HBAC analyses Currently HSPICE supports the following types of sweeps Linear sweeps sweeps a variable over an interval with a constant increment The syntax is one of the following e variable start stop increment e variable lin npoints start stop Logarithmic sweeps sweeps a variable over an interval To obtain each point this sweep multiplies the previous point by a constant factor You can specify the factor as a number of points per decade or octave as in e variable dec npoints start stop e variable oct npoints start stop Point sweeps a variable takes on specific values t
81. Tools gt Calculator to open the Waveform Calculator tool This tool can be used to generate new waveforms from existing ones It is described in detail in the CosmosScope User Guide The waveform calculator has no RF specific features Tools gt Measurement opens the Measurement Tool Three RF measurements have been added under the RF submenu of the measurement selection menu e 1db compression point 1DB CP e IP3 OIP3 HSPICE RF User Guide 13 Z 2007 03 Chapter 2 Getting Started Using the CosmosScope Waveform Display e Spurious free dynamic range SFDR Tools gt RF Tool opens the RF Tool which generates contour plots on Smith or Polar charts In HSPICE RF the plotfile must be a file with a sc extension thata LIN command generates HSPICE RF automatically finds the S parameter and noise parameter data in the sc file and uses it to generate noise gain and stability circles 14 HSPICE RF User Guide Z 2007 03 3 HSPICE RF Tutorial Provides a quick start tutorial for users new to HSPICE RF This tutorial assumes you are familiar with HSPICE and general HSPICE syntax but new to RF analysis features The most basic RF analysis features are presented here using simple examples The end of this chapter contains a listing of HSPICE RF demonstration files available for your use when you have access to the HSPICE RF installation directory This tutorial covers the following examples Example 1 Usi
82. Wiley amp Sons 1986 A Hajimiri S Limotyrakis and T H Lee Jitter and phase noise in ring oscillators IEEE J Solid State Circuits vol 34 no 6 pp 790 804 June 1999 Jitter Analysis Techniques for High Data Rates Application Note 1432 Agilent Technologies Feb 2003 Characterization of Clocks and Oscillators NIST Technical Note 1337 National Institute of Standards and Technology 10 G V Klimovitch Near carrier oscillator spectrum due to flicker and white 256 noise Proc ISCAS 2000 Geneva May 2000 HSPICE RF User Guide Z 2007 03 10 Large Signal Periodic AC Transfer Function and Noise Analyses Describes how to use both harmonic balance based and Shooting Newton based AC and transfer function analyses as well as nonlinear steady state noise analysis The following topics are presented in this section m Multitone Harmonic Balance AC Analysis HBAC Shooting Newton AC Analysis SNAC Multitone Harmonic Balance Noise HBNOISE Shooting Newton Noise Analysis GSGNNOISE Periodic Time Dependent Noise Analysis PTDNOISE m Multitone Harmonic Balance Transfer Function Analysis HBXF Shooting Newton Transfer Function Analysis SNXF Multitone Harmonic Balance AC Analysis HBAC You use the HBAC Harmonic Balance AC statement for analyzing linear behavior in large signal periodic systems The HBAC statement uses a periodic AC PAC algorithm to pe
83. With HB Measurements The required statements are HSPICE RF User Guide 409 Z 2007 03 Chapter 16 Advanced Features Optimization 410 m Analysis statement HB TONES lt fi gt lt f2 gt lt fn gt lt NHARMS lt hi1 gt lt h2 gt lt hn gt gt SWEEP parameter sweep OPTIMIZE OPTxxx RESULT measname MODEL mname Measure statement MEASURE HB measname FIND out_vari AT val GOAL val Optimization With HBNOISE PHASENOISE or HBTRAN Measurements The required statements are Analysis statement HB TONES lt fi gt lt f2 gt lt fn gt lt NHARMS lt hi1 gt lt h2 gt lt hn gt gt SWEEP OPTIMIZE OPTxXxx RESULT measname MODEL mname For example HBOSC tones 1lg nharms 5 optimize optl result yl y2 model ml model ml opt level 0 PHASENOISE dec 1 1k 1g meas phasenoise yl find phnoise at 10k goal 150dbc meas phasenoise y2 RMSJITTER phnoise units sec goal 1 0e 12 Measure statement MEASURE HBNOISE measname FIND out_vari AT val GOAL val MEASURE PHASENOISE measname FIND out_varl AT val GOAL val MEASURE HBTRAN measname FIND out_vari AT val GOAL val Optimizing HBOSC Analysis There are two types of optimizations with HBOSC analyses Optimization with only HB measurements m Optimization with HBNOISE PHASENOISE or HBTRAN measurements Optimization With HB Measurements The required statements are HSPICE RF User Guide Z 2007 03 Chapter 16 Advan
84. and a model name If you specify other parameters the nodes and model name must be first and the other parameters can appear in any order Example 1 The D1 diode with anode and cathode connects to nodes 1 and 2 Diode1 specifies the diode model D1 1 2 diodel Example 2 The Dprot diode with anode and cathode connects to both the output node and ground references the firstd diode model and specifies an area of 10 unitless for LEVEL 1 model The initial condition has the diode OFF Dprot output gnd firstd 10 OFF Example 3 The Ddrive diode with anode and cathode connects to the driver and output nodes The width and length are 500 microns This diode references the model_d diode model Ddrive driver output model _d W 5e 4 L 5e 4 IC 0 2 HSPICE RF User Guide 161 Z 2007 03 Chapter 6 Testbench Elements Active Elements 162 Bipolar Junction Transistor BJT Element Oxxx nc nb ne lt ns gt mname lt area gt lt OFF gt lt IC vbeval vceval gt lt M val gt lt DTEMP val gt Oxxx nc nb ne lt ns gt mname lt AREA area gt lt AREAB val gt lt AREAC val gt lt OFF gt lt VBE vbeval gt lt VCE vceval gt lt M val gt lt DTEMP val gt Parameter Description Qxxx BJT element name Must begin with Q then up to 1023 alphanumeric characters nc Collector terminal node name nb Base terminal node name ne Emitter terminal node name ns Substrate terminal node name which is optional You can
85. arbitrary offset The noise figure measurement is also dependent on this index term See Measuring SNNOISE Analyses with MEASURE below Prints the element noise value to the lis file You can specify at which frequencies the element noise value is printed The frequencies must match the sweep_frequency values defined in the parameter_sweep otherwise they are ignored In the element noise output the elements that contribute the largest noise are printed first The frequency values can be specified with the NONE or ALL keyword which either prints no frequencies or every frequency defined in parameter_sweep Frequency values must be enclosed in parentheses For example list freq none listfreq all listfreq 1 0G listfreq 1 0G 2 0G The default value is NONE Prints the element noise value to the lis file which is sorted from the largest to smallest value You do not need to print every noise element instead you can define listcount to print the number of element noise frequencies For example 1istcount 5 means that only the top 5 noise contributors are printed The default value is 1 277 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Shooting Newton Noise Analysis SNNOISE 278 Parameter Description listfloor Prints the element noise value to the lis file and defines a minimum meaningful noise value in V Hz units Only those elements with noise values larger than lis
86. current source CCVS Keyword for current controlled current source n n Positive and negative controlled source connecting nodes VMRF Keyword that identifies and activates the vector modulated RF signal source lin lin Node names for input I t signal Qin Qin Node names for input Q t signal VI VQ FREQ Carrier frequency in Hertz Set fc 0 0 to generate baseband I Q signals PHASE Carrier phase in degrees If fc 0 0 ph 0 and baseband I t is generated ph 90 and baseband Q t is generated SCALE Unit less amplitude scaling parameter HSTIXEPYcepl ue 187 Chapter 6 Testbench Elements Complex Signal Sources and Stimuli Example Emod1l inpl innl VMRF It plus It_neg Qt plus Qt neg freq lg phase 0 scale 1 5 File Driven PWL Source Vxxx nl n2 PWL PWLFILE filename lt coll lt col2 gt gt lt R repeat gt lt TD delay gt lt options gt Ixxx nl n2 PWL PWLFILE filename lt coll lt col2 gt gt lt R repeat gt lt TD delay gt lt options gt Parameter VXXX Description Independent voltage source Ixxx ni n2 PWL PWLFILE coll lt col2 gt TD options Independent current source Positive and negative terminal node names Keyword for piecewise linear Text file containing the PWL data consisting of time and voltage or current pairs This file should not contain a header row unless it is a comment The PWL source data is obtained by extracting col1 and col2 from the file
87. example RF 1K GAIN SHOULD BE 100 S MAY THE FORCE BE WITH MY CIRCUIT VIN 10 PL O 0 5V 5NS 10v 50ns R12 1 0 1MEG FEED BACK PARAM a lwS comment a 1 w treated as a space and ignored PARAM a 1kScomment a 1e3 k is a scale factor A dollar sign is the preferred way to indicate comments because of the flexibility of its placement within the code Line continuations require a plus sign as the first character in the line that follows Here is an example of comments and line continuation in a netlist file ABC Title Line HSPICE or HSPICE RF ignores the netlist keyword on this line because the first line is always a comment This is a comment line MODEL nl NMOS this is an example of an inline comment This is a comment line and the following line is a continuation LEVEL 3 Element and Source Statements Element statements describe the netlists of devices and sources Use nodes to connect elements to one another Nodes can be either numbers or names Element statements specify HSPICE RF User Guide Z 2007 03 Chapter 4 Input Netlist and Data Entry Input Netlist File Composition Type of device Nodes to which the device is connected Operating electrical characteristics of the device Element statements can also reference model statements that define the electrical parameters of the element Table 9 lists the parameters of an element statements Table 9 Element Parameters Parameter Descrip
88. follow names See Delimiters below Names can be up to 1024 characters long and are not case sensitive Do not use any of the time keywords as a parameter name or node name in your netlist a k The following symbols are reserved operator keywords Do not use these symbols as part of any parameter or node name that you define Using any of these reserved operator keywords as names causes a syntax error and HSPICE RF stops immediately Special Characters The following table lists the special characters that can be used as part of node names element parameter names and element instance names For detailed discussion see the appropriate sections in this chapter Note To avoid unexpected results or error messages do not use these A mathematical characters and in a parameter name in HSPICE Table 4 HSPICE HSPICE RF Netlists Net Name Special Characters Special Character Note character legal anywhere in the string first or included Node Name Instance Name Parameter Name Delimiters cannot be the cannot be the first first character character element element key key letter only is letter only tilde HSPICE y Included only Included only n a Included only for HSPICE RF exclamation V Included only Included only n a point at sign Included only J Included only n a pound sign V Included only Included only n a HSPICE RF User Guide 59 Z 2007 03 Chapter 4 In
89. frequency dependent and correlated noise effects Swept parameter analysis Results are independent of the number of SNAC sources in the netlist Prerequisites and Limitations The following prerequisites and limitations apply to SNNOISE Requires one SN statement which determines the steady state solution Requires at least one Periodic source Does not recognize HB sources Requires placing the parameter sweep in the SN statement HSPICE RF User Guide 275 Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Shooting Newton Noise Analysis SNNOISE Input Syntax SNNOISE output insrc parameter sweep lt n1 1 gt lt listfreq frequencies none all gt lt listcount val gt lt listfloor val gt lt listsources on off gt Parameter Description output insrc parameter_sweep 276 Output node pair of nodes or 2 terminal element HSPICE RF references equivalent noise output to this node or pair of nodes Specify a pair of nodes as V n n If you specify only one node V n then HSPICE RF assumes that the second node is ground You can also specify a 2 terminal element name that refers to an existing element in the netlist An input source If this is a resistor HSPICE RF uses it as a reference noise source to determine the noise figure If the resistance value is 0 the result is an infinite noise figure Frequency sweep range
90. get bed en x1 OUE 9 Aaaa x2 TT O oseph x2 OU Lemeki Total Power gt Subckt Name inv Instance Name Port Max A Min A Avg A mn mn mn mn mp mp mp mp Total Power TNO ATHY DA POWER Analysis The POWER statement in HSPICE RF creates a table which by default contains the measurements for AVG RMS MAX and MIN for every signal specified For example POWER lt signals gt lt REF vname FROM start_time TO end_time gt By default the scope of these measurements are set from 0 to the maximum timepoint specified in the TRAN statement For syntax and description of POWER statement see POWER in the HSPICE and HSPICE RF Command Reference Example 1 In this example no simulation start and stop time is specified for the x1 in signal so the simulation scope for this signal runs from the start Ops to the last tran time 100ps power xl in tran 4ps 100ps Example 2 You can use the FROM and TO times to specify a separate measurement start and stop time for each signal In this example HSPICE RF User Guide 417 Z 2007 03 Chapter 16 Advanced Features Detecting and Reporting Surge Currents The scope for simulating the x2 in signal is from 20ps to 80ps The scope for simulating the x0 in signal is from 30ps to 70ps param myendtime 80ps power x2 in REF al23 from 20ps to 80ps power x0 in REF abc from 30ps to myendtime 10ps Setting Default St
91. in radians VDB dB units VDBM dB relative to 1 mV Current type current magnitude and phase in degrees IR real component Il imaginary component IM magnitude IP Phase in degrees IPD Phase in degrees IPR Phase in radians IDB dB units IDBM GB relative to 1 mV Power type P Frequency type HERTZ i for single tone analysis HERTZ i j for two tone analysis HERTZ i j k for 3 tone analysis etc You must specify the harmonic index integer for the HERTZ keyword The frequency of the specified harmonics is dumped NODES or ELEM can be one of the following Voltage type a single node name n1 or a pair of node names n1 n2 Current type an element name elemname Power type a resistor resistorname or port portname element name 207 Chapter 7 Steady State Harmonic Balance Analysis Harmonic Balance Analysis 208 Parameter Description INDICES Index to tones in the form n1 n2 nN where nj is the index of the HB tone and the HB statement contains N tones If INDICES is used then wildcards are not supported HB data can be transformed into the time domain and output using the following syntax PRINT hbtran ovl lt ov2 gt PROBE hbtran ovl lt ov2 gt Where ov1 are the output variables to print or probe Calculating Power Measurements After HB Analyses Two types of power measurements are availabl
92. in this case can be written as V n2 02 b P rms Equation 39 L ef a S y Un A 4 2 This model for oscillator noise shows that sidebands about the fundamental due to noise are directly related to the spectrum of the phase fluctuations 6 t The more general definition of phase noise relates it to the spectral density of phase fluctuations i e 02 P Equation 40 SoA a 2L f HSPICE RF uses several sophisticated analysis algorithms to predict the power spectrum for the phase variations from which to compute phase noise HSPICE RF uses several sophisticated analysis techniques for computing the HSPICE RF User Guide 239 Z 2007 03 Chapter 9 Oscillator and Phase Noise Analysis Phase Noise Analysis 240 power spectrum of the phase variations to yield the phase noise response This information can be used to predict the spectrum of the oscillator about the fundamental frequency and also used to predict its random jitter characteristics PHASENOISE Input Syntax PHASENOISE lt output gt lt frequency_sweep gt lt method 0 1 2 gt lt carrierindex int gt lt listfreq frequencies none al1 gt lt listcount val gt lt listfloor val gt lt listsources on off gt Parameter Description output frequency_sweep method An output node pair of nodes or 2 terminal element HSPICE RF references phase noise calculations to this node or pair of nodes Specify a pair of nodes as V n
93. into a subcircuit branch including all lower subcircuit hierarchies X0 returns only current flowing into a subcircuit branch minus any current flowing into lower subcircuit hierarchies Figure 48 on page 406 illustrates the difference between the X and XO variables The dotted line boxes represent subcircuits and the black circles are the external nodes The X X1 vc1 path returns the current of the X1subcircuit through the vc7 node including the current to the X1 X1 and X1 X2 subcircuits as represented by the white black outlined arrows In contrast XO X1 vc2 returns only the current flowing through vc2 to the top level of the X1 subcircuit as shown by the black arrows HSPICE RF User Guide 405 Z 2007 03 Chapter 16 Advanced Features Probing Subcircuit Currents 406 Figure 48 Probing Subcircuit Currents VDD1 VDD2 X X1 ve1 X0 X1 vc2 X X2 vd2 vci vc S vdi vd2 sti x1 gt N x2 Sas iP X1 X1 X1 X2 Example 1 In this example the first five lines constitute the definition of the sb1 subcircuit with external nodes named node1 node2 and clr The line beginning with X1 is an instance of sb1 with nodes named 11 references node1 12 references node2 0 references clr subckt sb1 nodel node2 clr subckt elements R1 nodel node2 1K C1 clr nodel 1U ends subcircuit instance X1 11120 sbi PRINT X X1 nodel X X1 clr I X1 R1 To find the current flowing into n
94. is the index of the j th HB tone and the HB statement contains K tones 1 is the index of the HBAC tone Wildcards are not supported if this parameter is used You can transform HB data into the time domain and output by using the following syntax PRINT HBTRAN ov1 ov2 ovN PROBE HBTRAN ov1 ov2 ovN See TYPE above for voltage and current type definitions HBAC Output Data Files An HBAC analysis produces these output data files Output from the PRINT statement is written to a printhb file This data is against the IFB points e The header contains the large signal fundamental and the range of small signal frequencies e The columns of data are labeled as F Hz followed by the output variable names Each variable name has the associated mixing pair value appended All N variable names and all M mixing pair values are printed for each swept small signal frequency value a total of N M for each frequency value m Output from the PROBE statement is written to a hb file This data is against the IFB points Reported performance log statistics are written to a lis file e Number of nodes e Number of FFT points e Number of equations e Memory in use e CPU time e Maximum Krylov iterations e Maximum Krylov dimension e Target GMRES residual HSPICE RF User Guide 261 Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Multitone Harmonic Balance AC Analysis HBAC
95. logarithm of the absolute value of logarithm x with the sign of x sign of x log x log10 x base 10 math Returns the base 10 logarithm of the absolute value logarithm of x with the sign of x sign of x log x exp x exponential math Returns e raised to the power x e db x decibels math Returns the base 10 logarithm of the absolute value HSPICE RF User Guide Z 2007 03 of x multiplied by 20 with the sign of x sign of x 2010g4 x 101 Chapter 5 Parameters and Functions Built In Functions and Variables Table 12 Synopsys HSPICE Built in Functions Continued HSPICE Form Function Class Description int x integer math Returns the integer portion of x The fractional portion of the number is lost nint x integer math Rounds x up or down to the nearest integer sgn x return sign math Returns 1 if x is less than 0 Returns 0 if x is equal to 0 Returns 1 if x is greater than 0 sign x y transfer sign math Returns the absolute value of x with the sign of y sign of y x def x parameter control Returns 1 if parameter x is defined defined Returns 0 if parameter x is not defined min x y smaller of control Returns the numeric minimum of x and y two args max x y largeroftwo control Returns the numeric maximum of x and y args val element get value various Returns a parameter value for a specified element For example val r1 returns the resistance value of the r7 resistor val element
96. lt FMIN param_expr6 gt lt FMAX param_expr7 gt HSPICE RF User Guide 225 Z 2007 03 Chapter 8 Steady State Shooting Newton Analysis Shooting Newton with Fourier Transform SNFT Arguments Argument Description output_var START FROM STOP TO NP FORMAT WINDOW ALFA FREQ 226 Can be any valid output variable such as voltage current or power Start of the output variable waveform to analyze Defaults to the START value in the SN statement which defaults to 0 An alias for START in SNFT statements End of the output variable waveform to analyze Defaults to the TSTOP value in the SN statement An alias for STOP in SNFT statements Number of points to use in the SNFT analysis NP must be a power of 2 If NP is not a power of 2 HSPICE automatically adjusts it to the closest higher number that is a power of 2 The default is 1024 Specifies the output format NORM normalized magnitude default UNORM unnormalized magnitude Specifies the window type to use RECT simple rectangular truncation window default BART Bartlett triangular window HANN Hanning window HAMM Hamming window BLACK Blackman window HARRIS Blackman Harris window GAUSS Gaussian window KAISER Kaiser Bessel window Parameter to use in GAUSS and KAISER windows to control the highest side lobe level bandwidth and so on 1 0 lt ALFA lt 20 0 The default is 3 0 Frequency to ana
97. n2 nk 1 gt HSPICE RF User Guide 269 Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Multitone Harmonic Balance Noise HBNOISE lt listfreq frequencies none all gt lt listcount val gt lt listfloor val gt lt listsources on off gt Parameter Description output insrc parameter_sweep n1 n2 nk 1 270 Output node pair of nodes or 2 terminal element HSPICE RF references equivalent noise output to this node or pair of nodes Specify a pair of nodes as V n n If you specify only one node V n then HSPICE RF assumes that the second node is ground You can also specify a 2 terminal element name that refers to an existing element in the netlist An input source If this is a resistor HSPICE RF uses it as a reference noise source to determine the noise figure If the resistance value is 0 the result is an infinite noise figure Frequency sweep range for the input signal Also referred to as the input frequency band IFB or fin You can specify LIN DEC OCT POI SWEEPBLOCK DATA MONTE or OPTIMIZE sweeps Specify the nsteps start and stop frequencies using the following syntax for each type of sweep LIN nsteps start stop DEC nsteps start stop OCT nsteps start stop POI nsteps freq_values SWEEPBLOCK nsteps freq freq2 freqn Index term defining the output frequency band OFB or fout at which the noise is evaluated General
98. netlist file and store results in either an output listing file or a graph data file An input file with the name lt design gt sp contains the following Design netlist subcircuits macros power supplies and so on Statement naming the library to use optional Specifies the type of analysis to run optional Specifies the type of output desired optional An input filename can be up to 1024 characters long The input netlist file cannot be in a packed or compressed format To generate input netlist and library input files HSPICE or HSPICE RF uses either a schematic netlister or a text editor Statements in the input netlist file can be in any order except that the first line is a title line and the last ALTER submodule must appear at the end of the file and before the END statement Note If you do not place an END statement at the end of the input netlist file HSPICE RF issues an error message HSPICE RF User Guide 57 Z 2007 03 Chapter 4 Input Netlist and Data Entry Input Netlist File Guidelines 58 Netlist input processing is case insensitive except for file names and their paths HSPICE RF does not limit the identifier length line length or file size Input Line Format The input reader can accept an input token such as e astatement name e anode name e aparameter name or value Any valid string of characters between two token delimiters is a token You can not use a char
99. node2 are the positive and negative nodes that connect to the noise source The noise expression can contain the bias frequency or other parameters Data form Gxxx nodel node2 noise data dataname Exxx nodel node2 noise data dataname data dataname pnamel pname2 freql noisel freq2 noise2 F eats enddata The data form defines a basic frequency noise table The DATA statement contains two parameters frequency and noise to specify the noise value at each frequency point The unit for frequency is hertz and the unit for noise is A Hz for G current noise source or V2 Hz for E voltage noise source Example The following netlist shows a 1000 ohm resistor g1 using a G element The ginoise element placed in parallel with the g1 resistor delivers the thermal noise expected from a resistor The r1 resistor is included for comparison The noise due to r1 should be the same as the noise due to glnoise Resistor implemented using g element vl 101 ri 1 2 1k gl 1 2 cur v 1 2 0 001 glnoise 1 2 noise 4 1 3806266e 23 TEMPER 273 15 0 001 rout 2 0 l1lmeg ac lin 1 100 100 noise v 2 v1 1 end HSPICE RF User Guide 177 Z 2007 03 Chapter 6 Testbench Elements Function Approximations for Distributed Devices Function Approximations for Distributed Devices 178 High order rational function approximations constructed for distributed devices used at RF frequencies are obtained in the pole residu
100. number of GMRES iterations 3 current residual 1 37423e 17 voltage residual 2 12608e 13 20 40 60 80 100 Starting SNNOISE analysis TD PAC FFT Noise Algorithm Using GMRES Solver 4 completed Estimated 4 6 Min Remaining 10 completed Estimated 3 8 Min Remaining 21 completed Estimated 3 2 Min Remaining 30 completed Estimated 2 8 Min Remaining 41 completed Estimated 2 4 Min Remaining 51 completed Estimated 119 Sec Remaining 60 completed Estimated 97 Sec Remaining 71 completed Estimated 72 Sec Remaining 80 completed Estimated 50 Sec Remaining 91 completed Estimated 24 Sec Remaining Finished SNNOISE TDFD analysis 265 16 sec Steady State Time Domain has converged successfully in 3 iterations HSPICE RF User Guide Z 2007 03 47 Chapter 3 HSPICE RF Tutorial Example 7 Using Shooting Newton Analysis on a Driven Phase Frequency Circuit and a Ring Oscillator 48 Finished SNNOISE TDFD analysis 405 86 sec Steady State Time Domain has converged successfully in 4 iterations finished Shooting Newton Sweep 2200 5 sec Shooting Newton time 2200 53 cpu sec Total Shooting Newton memory 228 86 Mbytes Viewing Results in CosmosScope To view the time domain phasefreqdet sn0 file the frequency domain phasefreqdet snf0 file and the noise results phasefreqdet snpn0d file in CosmosScope 1 Enter cscope at the prompt to start CosmosScope 2 The time domain resu
101. of oscillation frequency found 3 351176768144e 08 20 40 60 80 100 Finished Shooting Newton Analysis Steady State Time Domain has converged successfully in 5 iterations DC operating point time 0 cpu seconds Shooting Newton time 26 04 cpu sec Total Shooting Newton memory 11 16 Mbytes Viewing Results in CosmosScope You can view the time domain ringoscSN sn0 file the frequency domain ringoscSN snf0 file and the phase noise ringoscSN snpnoO file in CosmosScope 1 Enter cscope at the prompt to start CosmosScope HSPICE RF User Guide Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 7 Using Shooting Newton Analysis on a Driven Phase Frequency Circuit and a Ring Oscillator 2 Use the File gt Open gt Plotfiles dialog to open the ringoscSN sn0 file Remember to set the file type filter to HSPICE RF 3 From the signal manager double click on the signal v 7 This is the time domain trace shown at the top of Figure 8 Use the File gt Open gt Plotfiles dialog to open the ringoscSN snfo file From the signal manager double click on the signals v 7 This is the frequency domain spectrum shown at the bottom of Figure 8 Figure 8 Ring Oscillator Output 6 To view the phase noise of the ring oscillator use the File gt Open gt Plotfiles dialog to open the ringoscSN snpn0 file 7 Open anew XY graph by clicking the waveform icon on the left side of the icon bar 8 From the signal manage
102. ook eee 147 155 155 159 159 159 162 164 166 169 169 172 174 175 176 178 178 178 179 180 180 182 189 190 190 191 191 191 192 195 197 198 199 200 Input Syntax HB Analysis Spectrum HB Analysis Options Contents Harmonic Balance Output Measurements 2 20020000 Output Syntax Calculating Power Measurements After HB Analyses Calculations for Time Domain Output 00 0c eee Output Examples Using MEASURE with HB Analyses 00 000 e eee euee HB Output Data Files Errors and Warnings References 000 Steady State Shooting Newton Analysis 04 SN Steady State Time Domain Analysis 00000 0 eee eee SN Analysis Syntax SN Analysis Output Shooting Newton with Fourier Transform SNFT 00000000 eae SNFT Input Syntax Oscillator and Phase Noise Analysis 0020000 Harmonic Balance or Shooting Newton for Oscillator Analysis Harmonic Balance Analysis for Frequency of Oscillation Input Syntax for Harmonic Balance Oscillator Analysis HB Simulation of Ring Oscillators HBOSC Analysis Using Transient Initialization 0 0000 Additional HBOSC Analysis Options 00000
103. operating point LIN analysis System impedance used when converting to a power source inserted in series with the voltage source Currently this only supports real impedance When power 0 zO defaults to 0 When power 1 zO defaults to 50 ohms You can also enter zo val 157 Chapter 6 Testbench Elements Port Element 158 Parameter Description lt RDC val gt DC analysis Series resistance overrides z0 lt RAC val gt AC analysis Series resistance overrides z0 lt RHBAC val gt HSPICE RF HBAC analysis Series resistance overrides z0 lt RHB val gt HSPICE RF HB analysis Series resistance overrides zo lt RTRAN val gt Transient analysis Series resistance overrides z0 lt power 0 1 2 W dom gt HSPICE RF power switch When 0 default element treated as a voltage or current source When 1 or W element treated as a power source realized as a voltage source with a series impedance In this case the source value is interpreted as RMS available power in units of Watts When 2 or dbm element treated as a power source in series with the port impedance Values are in dbms You can use this parameter for Transient analysis if the power source is either DC or SIN Example For example the following port element specifications identify a 2 port network with 50 ohm reference impedances between the in and out nodes Pl in gnd port 1 z0 50 P2 out gnd port
104. parameter When you use this feature Monte Carlo analysis can use a parameterized schematic netlist without additional modifications Syntax PARAM xx UNIF nominal _ val rel_variation lt multiplier gt PARAM xx AUNIF nominal _val abs_variation lt multiplier gt PARAM xx GAUSS nominal_val rel variation sigma lt multiplier gt PARAM xx AGAUSS nominal _val abs_variation sigma lt multiplier gt HSPICE RF User Guide Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Monte Carlo Analysis PARAM xx LIMIT nominal_val abs_variation Argument Description XX Distribution function calculates the value of this parameter UNIF Uniform distribution function by using relative variation AUNIF Uniform distribution function by using absolute variation GAUSS Gaussian distribution function by using relative variation AGAUSS Gaussian distribution function by using absolute variation LIMIT Random limit distribution function by using absolute variation Adds abs_variation to nominal_val based on whether the random outcome of a 1 to 1 distribution is greater than or less than 0 nominal_val Nominal value in Monte Carlo analysis and default value in all abs_variation rel_ variation sigma multiplier other analyses AUNIF and AGAUSS vary the nominal_val by abs_variation UNIF and GAUSS vary the nominal_val by nominal_val rel_variation Specifies abs_variat
105. represent the characteristics of the W element transmission line Internal field solver model name References the PETL internal field solver as the source of the transmission line characteristics for syntax see Using the Field Solver Model section in the HSPICE Signal Integrity Guide String that assigns each index of the S parameter matrix to one of the W element terminals This string must be an array of pairs that consists of a letter and a number for example Xn where X 1 i N orn to indicate near end input side terminal of the W element X O i F orfto indicate far end output side terminal of the W element The default value for NODEMAP is 1112I3 InNO10203 On S Model name reference which contains the S parameters of the transmission lines for the S Model syntax see the HSP CE Signal Integrity Guide Name of the frequency dependent tabular model Example 1 The W1 lossy transmission line connects the in node to the out node HSPICE RF User Guide Z 2007 03 141 Chapter 6 Testbench Elements Multi Terminal Linear Elements 142 W1 in gnd out gnd RLGCfile cable rlgc N 1 L 5 Where Both signal references are grounded The RLGC file is named cable rlgc The transmission line is 5 meters long Example 2 The Wcable element is a two conductor lossy transmission line Weable inl in2 gnd outl out2 gnd Umodel umod_1 N 2 L 10 Where int and in2 input nodes connect
106. signal also referred to as the input frequency band IFB or fin You can specify LIN DEC OCT POI or SWEEPBLOCK Specify the nsteps start and stop frequencies using the following syntax for each type of sweep LIN nsteps start stop DEC nsteps start stop OCT nsteps start stop POI nsteps freq_values SWEEPBLOCK nsteps freq freq2 freqn DATA dataname HBAC Analysis Options The following options directly relate to a HBAC analysis and override the corresponding PAC options if specified in the netlist 258 HSPICE RF User Guide Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Multitone Harmonic Balance AC Analysis HBAC OPTION HBACTOL default 1x10 8 Range 1x10 14 to Infinity m OPTION HBACKRYLOVDIM default 300 Range 1 to Infinity OPTION HBACKRYLOVITR default 1000 Range 1 to Infinity If these parameters are not specified then the following conditions apply If HBACTOL gt HBTOL then HBACTOL HBTOL f HBACKRYLOVDIM lt HBKRYLOVDIM then HBACKRYLOVDIM HBKRYLOVDIM Output Syntax This section describes the syntax for the HBAC PRINT and PROBE statements These statements are similar to those used for HB analysis PRINT and PROBE Statements PRINT HB TYPE NODES ELEM INDICES PROBE HB TYPE NODES ELEM INDICES HSPICE RF User Guide 259 Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyse
107. sin 2nf3 t Piri a N cos 2nAN b N sin 2nfN f Where m v t is the resulting time domain waveform m N 1 is the total number of harmonics including DC in the frequency domain spectrum of the hb0 file the zero th data point represents DC afi is the real component at the ith frequency b i is the imaginary component at the ith frequency f i is the ith frequency value with i 0 representing the zero frequency DC term These frequencies need not be harmonically related This frequency domain Fourier coefficient representation can be converted into a steady state time domain waveform output representation by using the PRINT or PROBE HBTRAN output option or by invoking the To Time Domain function on complex spectra within CosmosScope Output Syntax This section describes the syntax for the HB PRINT and PROBE statements PRINT and PROBE Statements PRINT HB TYPE NODES or ELEM INDICES 206 HSPICE RF User Guide Z 2007 03 Chapter 7 Steady State Harmonic Balance Analysis Harmonic Balance Analysis PROBE HB TYPE NODES or ELEM INDICES Parameter Description TYPE NODES or ELEM Specifies a harmonic type node or element TYPE can be one of the following HSPICE RF User Guide Z 2007 03 Voltage type V voltage magnitude and phase in degrees VR real component VI imaginary component VM magnitude VP Phase in degrees VPD Phase in degrees VPR Phase
108. so on You can also use them in sweep or statistical analysis For descriptions of RF commands referenced in this chapter see Chapter 3 RF Netlist Commands in the HSPICE and RF Command Reference Using Parameters in Simulation PARAM Defining Parameters Parameters in HSPICE are names that you associate with numeric values See Assigning Parameters on page 97 You can use any of the methods described in Table 10 to define parameters Table 10 PARAM Statement Syntax Parameter Description Simple PARAM lt SimpleParam gt 1e 12 assignment HSPICE RF User Guide 95 Z 2007 03 Chapter 5 Parameters and Functions Using Parameters in Simulation PARAM Table 10 PARAM Statement Syntax Continued Parameter Algebraic definition Description PARAM lt AlgebraicParam gt SimpleParam 8 2 SimpleParam excludes the output variable You can also use algebraic parameters in PRINT and PROBE statements For example PRINT AlgebraicParam par algebraic expression You can use the same syntax for PROBE statements See Using Algebraic Expressions on page 99 User defined function PARAM lt MyFunc x y gt Sqrt x x y y Character string PARAM lt paramname gt str string definition Subcircuit SUBCKT lt SubName gt lt ParamDefName gt lt Value gt str string default MACRO lt SubName gt lt ParamDefName gt lt Value gt str string Pre
109. statement on the first line of the netlist However HSPICE or HSPICE RF does not require the TITLE syntax The first line of the input file is always the implicit title If any statement appears as the first line in a file simulation interprets it as a title and does not execute it An ALTER statement does not support use the TITLE statement To change a title fora ALTER statement place the title content in the ALTER statement itself Comments and Line Continuation The first line of a netlist is always a comment regardless of its first character comments that are not the first line of the netlist require an asterisk as the HSPICE RF User Guide 71 Z 2007 03 Chapter 4 Input Netlist and Data Entry Input Netlist File Composition 72 first character in a line or a dollar sign directly in front of the comment anywhere on the line For example lt comment on a line by itself gt lt HSPICE statement gt lt comment_ following HSPICE_input gt You can place comment statements anywhere in the circuit description The dollar sign must be used for comments that do not begin in the first character position on a line for example for comments that follow simulator input on the same line If it is not the first nonblank character then the dollar sign must be preceded by either Whitespace Comma Valid numeric expression You can also place the dollar sign within node or element names For
110. the Port element s impedance is matched with the network input impedance The actual value of Vp will still depend on the port and network impedances Defines the hierarchy delimiter in the active nodes file in RCXT format Directs HSPICE RF to consider transistors with matching geometries except for NRD and NRS as identical for pre characterization purposes Activates detection of the atto unit Otherwise HSPICE RF assumes that a represents amperes Changes the default voltage supply range for characterization Begins range expression negative td a negative time delay defaults to zero port element _ voltage _ matchload rext divider skip _nrd_nrs unit _atto v_supply 3 wildcard left_ range 401 Chapter 16 Advanced Features Using Wildcards in HSPICE RF Table 29 Configuration File Options Continued Keyword Description Example wildcard_match_all Matches any group of characters wildcard match_ all wildcard_match_one Matches any single character wildcard _match_ one wildcard_right_range Ends range expression wildcard right _ range Note For more information about wildcards see Using Wildcards in HSPICE RF on page 402 Inserting Comments in a hspice File To insert comments into your hspicerf file include a number sign character as the first character in a line For example this configuration file shows how to use comments in a hspicerf file
111. the integration range using the FROM and TO parameters The measurements for PERJITTER and CTCJITTER use the full offset frequency sweep range given for the phase noise analysis to compute values the FROM and TO parameters are ignored if entered As given currently in HSPICE RF the frequency intervals can be modified for these jitter calculations if desired although not recommended and UNITS can be selected between seconds radians and Unit Intervals The following table specifies the calculation used for units seconds for each jitter measurement MEASURE name Calculation used Units sec RMSJITTER Op ms PHJITTER Tp ns TRJITTER Gop Orms PERJITTER OpER LTJITTER OAT gt 2O D HSPICE RF User Guide Z 2007 03 Chapter 9 Oscillator and Phase Noise Analysis Jitter Analysis MEASURE name Calculation used Units sec CTCJITTER Sore Peak to Peak As noted in MEASURE Statements to Support Jitter Analysis an additional BER Bit Error Rate parameter is supported This parameter allows you to convert any jitter value from an RMS value into a Peak to Peak value The RMS jitter values correspond to a 1 sigma standard deviation value for the Gaussian distribution of the jitter Peak to peak values represent the full span of the Gaussian distribution Since this span is theoretically unbounded for truly random distributions the conversion to peak to peak values has to be i
112. the parameter value within the condition expression HSPICE or HSPICE RF updates the parameter value only after it selects the IF ELSE block You can nest IF ELSE blocks Youcaninclude SUBCKT and MACRO statements within an IF ELSE block You can include an unlimited number of ELSEIF statements within an IF ELSE block You cannot use an IF ELSE block within another statement In the following example HSPICE or HSPICE RF does not recognize the IF ELSE block as part of the resistor definition ri1o0 if r_val gt 10k 10k else r val endif Using Subcircuits Reusable cells are the key to saving labor in any CAD system This also applies to circuit simulation in HSPICE or HSPICE RF T create and simulate a reusable circuit construct it as a subcircuit Use parameters to expand the utility of a subcircuit Traditional SPICE includes the basic subcircuit but does not provide a way to consistently name nodes However HSPICE or HSPICE RF provides a simple method for naming subcircuit nodes and elements use the subcircuit call name as a prefix to the node or element name In HSPICE RF you cannot replicate output commands within subcircuit subckt definitions HSPICE RF User Guide 87 Z 2007 03 Chapter 4 Input Netlist and Data Entry Using Subcircuits 88 Figure 12 Subcircuit Representation J o oO gt MN MP INV The following
113. the screen display more detailed status cpu time and memory usage information is also written to the ringoscSN lis file Parsing Loading Netlist Data Checking Compressing Simulation Time Domain RF Analysis Steady State Time Domain Analysis will use the matrix implicit gmres solver Input period 2 98507e 09 Starting Shooting Newton Analysis begin circuit stabilization time 20 40 60 80 100 Stabilization time done initial frequency 3 33854e 08 computed period 2 99532e 09 kkkkkkkkkkkxkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk KKKKKKKKKK Shooting Newton iteration L RRR KKKEKKKKKKEKE kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkxk 20 40 60 80 HSPICE RF User Guide 51 Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 7 Using Shooting Newton Analysis on a Driven Phase Frequency Circuit and a Ring Oscillator 52 100 number of transient points 162 start GMRES number of GMRES iterations 4 updated period 2 98848e 09 current residual 6 79374e 09 voltage residual 0 000183225 kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk kkkkkkkkkk Shooting Newton iteration 5 kkkkkkkkkkkkkxk kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk 20 40 60 80 100 number of transient points 2911 start GMRES number of GMRES iterations 4 updated period 2 98403e 09 current residual 3 20824e 11 voltage residual 3 42679e 10 period relative error 6 7151e 12 value
114. to delete that library for the ALTER simulation run Connecting Nodes Use a CONNECT statement to connect two nodes in your HSPICE netlist so that simulation evaluates two nodes as only one node Both nodes must be at the same level in the circuit design that you are simulating you cannot connect nodes that belong to different subcircuits You also cannot use this statement in HSPICE RF Deleting a Library Use a DEL LIB statement to remove library data from memory The next time you run a simulation the DEL LIB statement removes the LIB call statement with the same library number and entry name from memory You can then use a LIB statement to replace the deleted library You can use a DEL LIB statement with a ALTER statement HSPICE RF does not support the ALTER statement HSPICE RF User Guide 85 Z 2007 03 Chapter 4 Input Netlist and Data Entry Input Netlist File Composition 86 Ending a Netlist An END statement must be the last statement in the input netlist file Text that follows the END statement is a comment and has no effect on the simulation An input file that contains more than one simulation run must include an END statement for each simulation run You can concatenate several simulations into a single file Condition Controlled Netlists IF ELSE You can use the IF ELSE structure to change the circuit topology expand the circuit set parameter values for each device instance select di
115. to the out and out2 output node Both signal references are grounded umod_1 references the U model The transmission line is 10 meters long Example 3 The Wnet1 element is a five conductor lossy transmission line Wnet1 il 12 i3 i4 i5 gnd ol gnd o3 gnd o5 gnd FSmodel boardl1 N 5 L 1m Where The i1 i2 i3 i4 and i5 input nodes connect to the 01 03 and 05 output nodes The i5 input and three outputs 01 03 and 05 are all grounded board1 references the Field Solver model The transmission line is 1 millimeter long Example 4 S Model Example Wnet1 il i2 gnd ol o2 gnd Smodel smod_1 nodemap ili20102 N 2 L 10m Where ini and in2 input nodes connect to the out1 and out2 output node Both signal references are grounded HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Multi Terminal Linear Elements smod 1 references the S Model The transmission line is 10 meters long You can specify parameters in the W element card in any order You can specify the number of signal conductors N after the node list You can also mix nodes and parameters in the W element card You can specify only one of the RLGCfile FSmodel Umodel or Smodel models in a single W element card Figure 17 shows node numbering for the element syntax Figure 17 Terminal Node Numbering for the W element N 1 conductor line lah Rif L f G f C f liz
116. two points This option overrides EXTRAPOLATION in MODEL SP HIGHPASS Specifies high frequency extrapolation 0 Use zero in Y dimension open circuit 1 Use highest frequency 2 Use linear extrapolation with the highest two points 3 Apply window function default This option overrides EXTRAPOLATION in MODEL SP PRECFAC Preconditioning factor to avoid a singularity in the form of an infinite 150 admittance matrix See Pre Conditioning S parameters on page 154 for more information The default 0 75 HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Multi Terminal Linear Elements Parameter Specifies DELAYHANDLE DELAYFREQ MIXEDMODE DATATYPE XLINELENGTH Delay handler for transmission line type parameters 1 or ON activates the delay handler See Group Delay Handler in Time Domain Analysis on page 153 0 or OFF default deactivates the delay handler You must set the delay handler if the delay of the model is longer than the base period specified in the FBASE parameter If you set DELAYHANDLE OFF but DELAYFQ is not zero HSPICE simulates the S element in delay mode Delay frequency for transmission line type parameters which is the frequency point when HSPICE RF extracts the matrix delay The default is the FMAX value which is the maximum frequency used in the transient analysis If you set DELAYHANDLE to OFF but DELAYFREQ is not zero HSPICE still simulates the S e
117. two specific nodes Unique identifier for resistance between two specific nodes Unique identifier for inductance between two specific nodes First of two nodes between which you are specifying a capacitance resistance or inductance value Second of two nodes between which you are specifying a capacitance resistance or inductance value For a capacitance value if you do not specify a second node name HSPICE RF assumes that the second node is ground 339 Chapter 13 Post Layout Analysis Post Layout Back Annotation Table 26 SPEF Parameters Continued Parameter Definition capacitance Specifies the capacitance value assigned to a cap_id identifier capacitance_unit defines the units of capacitance For example if you set capacitance to 5 and capacitance_unit to 10 PF then the actual capacitance value is 50 picoFarads resistance Specifies the resistance value assigned to a res_id identifier resistance_unit defines the units of resistance For example if you set resistance to 5 and resistance_unit to 5 KOHM then the actual resistance value is 25 kilo ohms inductance Specifies the resistance value assigned to an induc_id identifier 340 inductance_unit defines the units of inductance For example if you set inductance to 6 and inductance_unit to 2 UH then the actual inductance value is 12 microhenries SPEF File Example SPEF IEEE 1481 1998 DESIGN My design DATE 11 26 34 Friday June 28 2
118. value from the list The power value should be as large as possible but still well within the linear range of the amplifier Try 25dbm Click the Apply button Result CosmoScope will show the linear gain of the amplifier and the 1dBcompression point The 3rd order intercept point is also measured by using the measurement tool Use the down arrow at the end of the Measurement field and select RF and IP3 SFDR The PowerOut1 field should contain the Pr rload 1 0 trace and the PowerOuts field should contain the Pr rload 2 1 trace Select a Powerln value from the list The power value should be a value that is as large as possible but still well within the linear range of the amplifier Try 25dbm Click Apply Result CosmosScope will show the 3rd order intercept point of the amplifier HSPICE RF User Guide Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 3 Using HB Analysis for an Amplifier Device Model Cards The following is an NMOS model in cmos49_model inc file used in the power amplifier example It is available in directory lt installdir gt demo hspicerf examples MODEL CMOSN NMOS LEVEL 49 VERSION 3 1 TNOM 27 TOX 7 9E 9 XJ 1 5E 7 NCH 1 7E17 VTHO 0 5047781 K1 0 5719698 K2 0 0197928 K3 33 4446099 K3B 3 1667861 wo 1E 5 NLX 2 455237E 7 DVTOW 0 DVT1W 0 DVT2W 0 DVTO 2 8937881 DVT1 0 6610934 DVT2 0 0446083 U0 421 8714618 UA 1 18967E 10 UB
119. way You can specify the wafer distribution in the MOSFET model to set the speed and power dissipation characteristics Monte Carlo Examples Gaussian Uniform and Limit Functions You can find the sample netlist for this example in the following directory installdir demo hspice apps mont1 sp Figure 33 Uniform Functions MONT1 SP TEST OF MONTE CARLO GAUSSIAN UNIFORM AND LIMIT FUNCTIONS May 15 2003 11 41 23 119 182 DE _ be ee MONTI SVO a r a A A a S a Z RUNIF_1 MOOREA ARS e OSes oe CK Siae BOS a B A a ae a A j a wy 8 a a oo lle ae Se a ee A 100 05 ae ae K et r a aii 3 ee gt on aay cS Bhs yore oe ee amas TE wht E E Ge E z 80 1384 ea bikes NAR a Bag 8k ek ko Tages ere MONT1_SVO O Oan n BAS nak i i rA Aa Ae RUNIF_10 I a 110 07 5 Pe s 100 0 gt 90 0 E a 80 0402 a A 44 2 1 i gt 1 ind a Sj LAI A Am S 1 0 10 0 20 0 30 0 40 0 50 0 60 0 MONTE CARLO LIN HSPICE RF User Guide 365 Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Monte Carlo Analysis Figure 34 Gaussian Functions MONT1 SP TEST OF MONTE CARLO GAUSSIAN UNIFORM AND LIMIT FUNCTIONS May 15 2003 11 41 23 MONT1_SV ee a ER A Std a RGA 1 115 0 A RGAUSS I L USS_ J o a v Sa A 110 0 ee ae Pr a o 4 100 0 a ue ee TTTS C SAA A z A zZ wt A A A i ee I 95 04 a A eS F 90 0 A
120. which HSPICE RF computes the phase noise The phase noise output is normalized to this carrier harmonic Default 1 Dumps the element phase noise value to the lis file You can specify which frequencies the element phase noise value dumps The frequencies must match the sweep_frequency values defined in the parameter_sweep otherwise they are ignored In the element phase noise output the elements that contribute the largest phase noise are dumped first The frequency values can be specified with the NONE or ALL keyword which either dumps no frequencies or every frequency defined in the parameter_sweep Frequency values must be enclosed in parentheses For example listfreq none listfreq all listfreq 1 0G listfreq 1 0G 2 0G The default value is the first frequency value Dumps the element phase noise value to the lis file which is sorted from the largest to smallest value You do not need to dump every noise element instead you can define listcount to dump the number of element phase noise frequencies For example listcount 5 means that only the top 5 noise contributors are dumped The default value is 20 Dumps the element phase noise value to the lis file and defines a minimum meaningful noise value in dBc Hz units Only those elements with phase noise values larger than the listfloor value are dumped For example listfloor 200 means that all noise values below 200 dBc Hz are not dumped The default value is
121. 0 cece Example 1 Using LIN Analysis for a NMOS Low Noise Amplifier Example 2 Using HB Analysis for a Power Amplifier Example 3 Using HB Analysis for an Amplifier Device Model Cards 0 000 cece eee Example 4 Using HBOSC Analysis for a Colpitts Oscillator Example 5 Using HBOSC Analysis fora CMOS GPS VCO Example 6 Using Multi Tone HB and HBAC Analyses for a Mixer Contents xiii XV xvi xvi xvii xviii 10 10 11 12 15 15 19 22 27 28 31 38 Contents Two tone HB Approach 0 ee eee 39 HBAC Approach aoi oen a e AA tees 40 Comparing Results anana uaaa aaa 41 Example 7 Using Shooting Newton Analysis on a Driven Phase Frequency Circuit and a Ring Oscillator 0 2 0 0 000 cee 43 INt FOCUCHON is tiie face E e aie ae eS ae Makes Gor eee ees 43 Shooting Newton Analysis Setup 0000 02 eee eee eee 43 Driven Phase Frequency Example 000 eee eee eee 44 Ring Oscillator Example 0 00 000 cece eee ae 50 Other Shooting Newton Analyses 0000 0c cece eee eee 54 Demonstration Input Files 0 0 0 0 0 cece ee eee 55 Input Netlist and Data Entry 20 0 0 0 eee eee eee 57 Input Netlist File Guidelines 2 0 0 0 000000 cee eee 57 Input Line Format sro oiia ae EDRED eee 58 De limiters sce is te e a a Be ce E a E E A 64 Node Identifiers 0 0 cee 64 Instance Names zover aenea a
122. 0 file A time domain waveform appears 4 View the v out signal from the pa hbo file This should be a histogram with lines at 950MHz and multiples thereof up to 9 5GHz 5 Right click on the waveform label for v out from the pa hb0 file and choose To Time Domain Change the X End sec value to 10n Click OK to accept the default interval value You should now see a new waveform called timedomain v out 8 Left click on the timedomain v out label hold and drag the signal to the plot containing v out This should overlay the v out and timedomain v out signals on the same panel Zoom into the transitions to see the slight differences between the waveforms Example 3 Using HB Analysis for an Amplifier This example takes the LNA circuit of Example 1 and performs a simulation using two closely spaced steady state tones to study the compression and third order distortion properties of the amplifier The example file gsmina P3 sp is located at lt install_dir gt demo hspicerf examples 22 HSPICE RF User Guide Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 3 Using HB Analysis for an Amplifier kk NMOS 0 25um Cascode LNA for GSM applications Test bench setup for two tone power sweep in dBm to extract IP3 kk temp 27 options post 2 param Vdd 2 3 global gnd param Pin dBm 30 0 param Pin Pin dBm param Pin W 1 0e 3 pwr 10 0 Pin 10 0 Change to Watts for sources kk Casc
123. 002 VENDOR Synopsys Inc PROGRAM Star RCXT VERSION 2002 2 DESIGN FLOW EXTERNAL LOADS EXTERNAL SLEWS MISSING NETS DIVIDER DELIMITER BUS DELIMITER T UNIT 1 NS C UNIT 1 PF R UNIT 1 OHM L UNIT 1 HENRY POWER_ NETS VDD GND NETS VSS PORTS CONTROL O L 30 S 0 0 FARLOAD O L 30 S 0 0 INVX1IFNTC_IN I L 30 S 5 5 NEARLOAD O L 30 S 0 0 TREE O L 30 S 0 0 HSPICE RF User Guide Z 2007 03 HSPICE Z 2007 03 Chapter 13 Post Layout Analysis Post Layout Back Annotation If you use triplet format the above section would look like this PORTS CONTROL O L 30 30 30 S 0 0 FARLOAD O L 30 30 30 S 0 0 INVX1IFNTC_IN I L 30 30 30 NEARLOAD O L 30 30 30 S 0 TREE O L 30 30 30 S 0 0 0 This triplet formatting principle applies to the rest of this example RF User Guide 341 Chapter 13 Post Layout Analysis Post Layout Back Annotation 342 D NET INVX1FNTC_IN 0 033 CONN XP INVX1IFNTC_IN I I FL 1281 A L 0 033 END D NET INVX1FNTC 2 033341 CONN xT sn x I Ie FL 1281 X O L I1184 A I L 0 FL 1000 A I L NL 1000 A I L TR_1000 A I L CAP 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 FL 1000 A 0 3 I1184 A 0 344 INVX1FNTC_IN INVX1FNTC_IN INVX1FNTC_IN INVX1FNTC_IN INVX1FNTC_IN INVX1FNTC_IN INVX1FNTC_IN INVX1FNTC_IN INVX1FNTC_IN INVX1FNTC_
124. 03 Chapter 6 Testbench Elements Phase Differences Between HB and SIN Sources Isrc in gnd HB 1 0e 3 0 1 1 power 1 z0 50 Example 3 Five series voltage sources sum to produce a stimulus of five equally spaced frequencies at and above 2 44 GHz using modharm and modtone parameters These are commensurate tones an integer relation exists therefore you only need to specify two tones when invoking the HB analysis param Vin 1 0 param f0 2440MEG param deltaf 312 5K param fcenter f0 2 0 deltaf Vrfa in ina HB Vin 0 1 1 2 440625 GHz Vrfb ina inb HB Vin 0 1 1 1 2 S 2 4403125 GHz Vrfc inb inc HB Vin 0 1 1 2 2 S 2 440 GHz Vrfd ine ind HB Vin 0 1 1 1 2 S 2 4409375 GHz Vrfe ind gnd HB Vin 0 1 1 2 2 2 44125 GHz HB tones fcenter deltaf intmodmax 5 Phase Differences Between HB and SIN Sources 174 The HB steady state cosine source has a phase variation compared to the TRAN time domain SIN source The SIN source with no offset delay or damping follows the equation Equation 11 Asin ar while the HB sources follow Equation 12 Acos In order for the two sources to yield identical results it is necessary to align them by setting their phase values accordingly using Equation 13 Acos a Asin at 0 90 Equation 14 Asin a Acos art o 90 To specify sources with matching phase for HB and TRAN analysis use a convention similar to HSPICE RF User Guide Z 2007 0
125. 0meg print hb v out v rf1 v lo print hbxf tfv vrf1 tfv vlo end Shooting Newton Transfer Function Analysis SNXF The SNXF command calculates transfer functions from an arbitrary number of small signal sources to a designated output in a circuit under periodic steady state conditions Frequency conversion is calculated from multiple input frequencies to a single output at a single frequency that is specified on the command line Prerequisites and Limitations The following prerequisites and limitations apply to the SNXF command Only one SNXF statement is required If you use multiple SNXF statements HSPICE RF only uses the last one issued Atleast one SN statement is required which determines the steady state solution Parameter sweeps must be placed in SN statements Input Syntax SNXF out_var lt freq sweep gt HSPICE RF User Guide 291 Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Shooting Newton Transfer Function Analysis SNXF 292 Parameter Description Parameter Description out_var Specify i 2 _port_elem or V n1 lt n2 gt freq_sweep Frequency sweep range for the input signal also referred to as the input frequency band IFB or fin A sweep of type LIN DEC OCT POI or SWEEPBLOCK Specify the nsteps start and stop frequencies using the following syntax for each type of sweep LIN nsteps start stop DEC nsteps start stop OCT n
126. 1W 0 DVT2W 0 HSPICE RF User Guide 17 Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 1 Using LIN Analysis for a NMOS Low Noise Amplifier 0 7186877 1 300598E 9 1 482651E5 1 833193E 8 0 0 3774172 0 4E 8 0 0981658 2 4E 4 5 128492E 3 1 91946 0 L 5 200359E 10 31 1 5 0 022 5 6E 11 62E 10 s99 79 9 lt 99 OOOWUNOOF 8 385037E 3 7 293869E 3 DVT2 UB AO B1 A2 PRWB LINT DWG NFACTOR CDSCD ETAB PDIBLC1 DROUT PVAG MOBMOD KT1 UA1 AT WW LL LWN XPART CGBO MJ MJSW MJSWG PRDSW LKETA 0 5 2 3082E 18 1 6856991 1E 7 0 4177975 0 2 1 88839E 8 1 2139E 8 2032376 18609E 4 1 0 6 1 0 9817908 9 31443E 3 1 0 11 4 31E 9 3 3E4 0 0 1 0 5 1E 12 0 4453094 0 3413857 0 3413857 10 6 070E 3 This invokes a LIN analysis and activates noise calculations and S Specifies that an input port is assumed between terminals rfin and ground that it is has a 50 ohm termination and it has a built in DC bias of DVTO 0 5334651 DVT1 U0 289 1720829 UA UC 2 841618E 11 VSAT AGS 0 2874763 BO KETA 2 395348E 3 Al RDSW 178 7751373 PRWG WR 1 WINT XL 3E 8 XW DWB 4 613042E 9 VOFF CIT 0 CDSC CDSCB 0 ETAO DSUB 0 0463218 PCLM PDIBLC2 4 422611E 3 PDIBLCB PSCBE1 7 982649E10 PSCBE2 DELTA 0 01 RSH PRT 0 UTE KT1L 0 KT2 UB1 7 61E 18 UC1 WL 0 WLN WWN 1 WWL LLN 1 LW LWL 0 CAPMOD CGDO 5 6
127. 2 z0 50 Computing scattering parameters requires z0 reference impedance values The order of the port parameters in the P element determines the order of the S Y and Z parameters Unlike the NET command the LIN command does not require you to insert additional sources into the circuit To calculate the requested transfer parameters HSPICE automatically inserts these sources as needed at the port terminals You can define an unlimited number of ports HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Active Elements Using the Port Element for Mixed Mode Measurement You can use a port element with three terminals as the port element for measuring the mixed mode S parameters Except for the number of external terminals the syntax of the port element remains the same The LIN analysis function internally sets the necessary drive mode common differential of these mixed mode port elements For analyses other than the LIN analysis such as DC AC TRAN and so on the mixed mode P element acts as a differential driver that drives positive nodes with half of their specified voltage and the negative nodes with a negated half of the specified voltage Figure 20 shows the block diagram of the mixed mode port element Figure 20 Mixed Mode Port Element P1 Port element A O n1 ZO V o ZO V n1_ref PI nl nl nl_ref Zo 50 Active Elements This section describes t
128. 2002 14 12 43 VENDOR Synopsys PROGRAM Star RC VERSION Star RCXT 2002 2 DIVIDER DELIMITER SUBCKT BUFFER OUT IN Description of Nets GROUND_NET VSS NET IN 1 221451PF P IN 1 0 0 0 10 I DF1 A DF1 A I 0 0PF 10 0 10 0 I DF1 B DF1 B I 0 0PF 10 0 20 0 S IN 1 5 0 10 0 IN 2 5 0 20 0 C1 IN VSS 0 117763PF C2 IN 1 VSS 0 276325PF C3 IN 2 VSS 0 286325PF C4 DF1 A VSS 0 270519PF C5 DF1 B VSS 0 270519PF R20 IN N 1 1 70333E00 1 1 2 R21 IN 1 DF1 A 1 29167E 01 R22 IN 1 IN 2 1 29167E 01 R23 IN 2 DF1 B 1 70333E 01 NET BF 0 287069PF I DF1 C DF1 C O 0 0PF 12 0 15 0 I INV1 IN INV1 IN I 0 0PF 30 0 15 0 C6 DF1 C VSS 0 208719PF C7 INV1 IN VSS 0 783500PF R24 DF1 C INV1 IN 1 80833E 01 NET OUT 0 148478PF S OUT 1 45 0 15 0 P OUT O 0 0PF 50 0 5 0 I INV1 OUT INV1 OUT O 0 0PF 40 0 15 0 C8 INV1 OUT VSS 0 147069PF C9 OUT 1 VSS 0 632813PF C10 OUT VSS 0 776250PF R25 INV1 OUT OUT 1 3 11000E00 R26 OUT 1 OUT 3 03333E00 Description of Instances XDF1 DF1 A DF1 B DF1 C DFF XINV1 INV1 IN INV1 OUT INV ENDS END 332 HSPICE RF User Guide Z 2007 03 Chapter 13 Post Layout Analysis Post Layout Back Annotation Overview of SPEF Files The Standard Parasitics Exchange Format SPEF file structure is described in IEEE standard EEE 1487 For information about how to obtain the complete SPEC EEE 1481 specification or any other documents from IEEE see http www ieee org produc
129. 298 301 limitations 300 output syntax 304 HBLSP 305 example 308 input syntax 307 limitations 306 output data files 305 309 output syntax 309 HBOSC options 235 HBOSC analysis 424 Colpitts oscillator 28 VCO 31 HBOSC statement 229 HBXF command 287 hertz variable 104 hier_delimiter configuration option 400 hierarchical designs flattened 68 hl file 305 hold time verification 414 hspice ini file 91 hspicerf command 9 hspicerf file 399 hspicerf test 400 html configuration option 400 IBIS buffers 155 ideal transformer 137 INCLUDE statement 68 83 84 91 93 individual element temperature 353 inductor coupled 134 frequency dependent 130 inductors element 125 node names 125 138 input data for data driven analysis 81 files character case 58 compression 57 netlist 57 structure 68 table of components 69 netlist 70 netlist file 70 86 input data 85 input files demonstration 55 input files demo examples 55 int x function 102 integer function 102 integer_node configuration option 400 internal nodes referencing 79 invoking HSPICE RF 9 IR drop checking 415 J JFETs elements 164 length 164 width 164 jitter random with clock source 191 jitter random clock source 191 K keywords DTEMP 352 MONTE 360 PAR 100 L L Element inductor 130 large signal S parameter extraction 305 LENGTH model parameter 368 LIB call statement 81 statement 68 93 in ALTER blocks 81 83 84 with DEL LIB 85 wi
130. 2E 10 CGSO CJ 1 641005E 3 PB CUSW 4 179682E 10 PBSW CUSWG 3 29E 10 PBSWG CF 0 PVTHO PK2 2 650965E 3 WKETA END A LIN analysis also includes the following m LIN command LIN noisecalc 1 sparcalc 1 parameter output files Two port elements Pl rfin gnd port 1 z0 50 dc 0 595 0 595 V The output Second port is 18 HSPICE RF User Guide Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 2 Using HB Analysis for a Power Amplifier P2 rfo _vdd port 2 z0 255 This syntax specifies that the output port is between terminals rfo and _vdd and is being used as a pull up resistor with impedance of 255 ohms A PRINT command for plotting the output S parameters in dB and the noise figure minimum To run this netlist type the following command hspicerf gsmlna sp This produces two output files named gsmina sc0O and gsmina printaco containing the S parameter and noise parameter results and the requested PRINT data To view the output 1 Type cscope to invoke CosmosScope 2 Open gsmina sc0 in the File gt Open gt Plotfiles dialog Be sure to change the Files of Type filter to find the scO file 3 To open a blank Smith chart click the Smith chart icon on the left side of the upper toolbar 4 Using the signal manager select the S 1 1 and S 2 2 signals under the S Par heading from the gsmina scO file You should see them plotted on the Smith chart 5 To open a blank Polar chart clic
131. 3 Chapter 6 Testbench Elements Behavioral Noise Sources Example 1 with equivalent HB and SIN sources SIN source is given 90 phase shift param freql 2400MEG Vin 1 0 Vsre in gnd DC 0 HB Vin 0 1 1 SIN O Vin freql 0 0 90 HB tones freql intmodmax 7 Example 2 with equivalent HB and SIN sources HB source is given 90 phase shift to align with SIN param freql 2400MEG Vin 1 0 Vsre in gnd DC 0 HB Vin 90 1 1 SIN 0 Vin freql 0 HB tones freql intmodmax 7 Example 3 with equivalent HB and TRAN sources SIN source is activated for HB using TRANFORHB param freql 2400MEG Vin 1 0 Vsre in gnd DC 0 SIN 0 Vin fregql 0 TRANFORHB 1 HB tones freql intmodmax 7 Behavioral Noise Sources In HSPICE RF you can use the G element to specify noise sources Frequency domain noise analyses NOISE HBNOISE and PHASENOISE take these noise sources into account You can attach noise sources to behavioral models For example you can use a G element with the VCCAP parameter to model a varactor which includes a noise model You can also simulate effects such as substrate noise including its effect on oscillator phase noise You can also use this G element syntax to simulate behavioral descriptions of substrate noise during any frequency domain noise analysis which includes phase noise analysis For example gname nodel node2 noise noise equation gname nodel node2 node3 node4 noise noise equation
132. 3 receives the same random resistance value during each Monte Carlo run param r_random agauss param r_global r_ random rl 1 2 r r_global r2 3 4 r r_global r3 5 6 r r_global Monte Carlo Parameter Distribution Each time you use a parameter Monte Carlo calculates a new random variable If you do not specify a Monte Carlo distribution then HSPICE RF assumes the nominal value f you specify a Monte Carlo distribution for only one analysis HSPICE RF uses the nominal value for all other analyses You can assign a Monte Carlo distribution to all elements that share a common model The actual element value varies according to the element distribution If you assign a Monte Carlo distribution to a model keyword then all elements 364 HSPICE RF User Guide Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Monte Carlo Analysis that share the model use the same keyword value You can use this feature to create double element and model distributions For example the MOSFET channel length varies from transistor to transistor by a small amount that corresponds to the die distribution The die distribution is responsible for offset voltages in operational amplifiers and for the tendency of flip flops to settle into random states However all transistors on a die site vary according to the wafer or fabrication run distribution This value is much larger than the die distribution but affects all transistors the same
133. 377 Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Worst Case and Monte Carlo Sweep Example Figure 44 Sensitivity of Delay with TOX Monte Carla Results ii tox toxed o00p m_delayttax toxcd 170 0 160 0 180 0 200 0 210 0 220 0 230 0 tox toxcd The plot in Figure 45 overlays the skew result with the ones from Monte Carlo The skew simulation traverses the design space with all parameters changing in parallel and then produces a relationship between power and delay which shows as a Single line Monte Carlo exercises a variety of independent parameter combinations and shows that there is no simple relationship between the two results Since the distributions were defined as Gaussian in the netlist parameter values close to the nominal are more often exercised than the ones far away With the relatively small number of samples the chance of hitting a combination at the extremes is very small In other words designing for 3 sigma extreme for every parameter is probably not a good solution from the point of view of economy 378 HSPICE RF User Guide Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Worst Case and Monte Carlo Sweep Example Figure 45 Superimposing Sigma Sweep Over Monte Carlo Monte Carla Results 2 m_delay 3 0m m_power m_delay o 8 0m im f s_delay T S_power s_delay 6 0m z 5 0m 4 0m 100p 200p 300p 40 0p m_delay 4 a a 100p 200p 300p ARIP s_delay
134. 5 SIM_DELTAI option 396 SIM_DELTAV option 396 SIM_DSPF option 320 389 390 396 SIM_DSPF_ACTIVE option 320 323 SIM_DSPF_INSERROR option 325 SIM_DSPF_LUMPCAPS option 325 SIM_DSPF_MAX_ITER option 324 SIM_DSPF_RAIL option 324 SIM_DSPF_SCALEC option 324 SIM_DSPF_SCALER option 324 SIM_DSPF_VTOL option 323 SIM_LA option 320 321 344 347 SIM_LA_FREQ option 347 SIM_LA_MAXR option 347 SIM_LA_MINC option 348 SIM_LA_MINMODE option 348 SIM_LA_TIME option 348 SIM_LA_TOL option 348 SIM_ORDER option 389 SIM_POSTAT option 404 SIM_POSTDOWN option 404 SIM_POSTSCOPE option 405 SIM_POSTSKIP option 403 404 SIM_POWERDC_ACCURACY option 416 SIM_POWERED_HSPICE option 416 SIM_POWERPOST option 418 SIM_POWERSTART option 418 SIM_SPEF option 320 SIM_SPEF_ACTIVE option 323 SIM_SPEF_INSERROR option 325 SIM_SPEF_LUMPCAPS option 325 SIM_SPEF_MAX_ITER option 324 SIM_SPEF_PARVALUE option 325 SIM_SPEF_RAIL option 324 SIM_SPEF_SCALEC option 324 SIM_SPEF_SCALER option 324 SIM_SPEF_VTOL option 323 simulation multiple runs 86 title 71 simulation engine 1 sin x function 100 sinh x function 101 skew file 358 parameters 354 skip_nrd_nrs configuration option configuration options skip_nrd_nrs 401 slew rate example 412 verification 412 small signal noise parameter extraction 305 small signal S parameter extraction 305 SN steady state time domain analysis 219 SNAC input syntax 263 output data files 266 SNFT 224 SNNOISE 275 input syntax 276 output data
135. 6 initial conditions 167 node names 166 perimeter 167 source 167 168 squares 167 temperature differential 167 zero bias voltage threshold shift 168 multiple ALTER statements 83 84 multiply parameter 88 115 multi tone HB analysis mixer 38 mutual inductor 128 N natural log function 101 negative_td configuration option 401 netlist 68 file example 70 flat 68 input files 57 schematic 68 structure 70 netlist file example 70 nodes connection requirements 79 floating supply 79 internal 79 MOSFET s substrate 79 names 75 80 automatic generation 80 ground node 79 period in 76 subcircuits 78 79 numbers 75 terminators 79 noise HBNOISE 268 275 noise parameter extraction 426 small signal 305 nonlinear perturbation algorithm 242 numerical integration 389 390 NW output format 394 O operators 100 optimization 407 syntax 407 OPTION ALTER blocks 83 84 MAXORD 391 PURETP 392 SIM_ACCURACY 390 SIM_DELTAI 396 SIM_DELTAV 396 SIM_DSPF 320 SIM_DSPF_ACTIVE 320 323 SIM_DSPF_INSERROR 325 SIM_DSPF_LUMPCAPS 325 SIM_DSPF_MAX_ITER 324 SIM_DSPF_RAIL 324 SIM_DSPF_SCALEC 324 SIM_DSPF_SCALER 324 SIM_DSPF_VTOL 323 SIM_LA 320 321 344 347 SIM_LA_FREQ 347 SIM_LA_MAXR 347 SIM_LA_MINC 348 SIM_LA_MINMODE 348 SIM_LA_TIME 348 SIM_LA_TOL 348 SIM_ORDER 389 391 SIM_OSC_DETECT_TOL 392 SIM_POSTAT 404 SIM_POSTDOWN 404 SIM_POSTSCOPE 405 SIM_POSTSKIP 403 404 SIM_POWERDC_ACCURACY 416 SIM_POWERDC_HSPICE 416 SIM_POW
136. 8 DC statement 353 DDL 91 DDLPATH environment variable 91 decibel function 101 DEFW option 108 DEL LIB statement 68 in ALTER blocks 83 84 with ALTER 85 with LIB 85 with multiple ALTER statements 84 85 DELVTO model parameter 355 demo files 55 55 demonstration files RF 55 demonstration input files 55 Detailed Standard Parasitic Format See DSPF deviation average 351 device model cards 27 diodes junction 160 models 160 polysilicon capacitor length 160 DSPF expansion 327 file structure 321 DTEMP parameter 352 353 E edge condition 413 element active BJTs 162 diodes 159 JFETs 164 MESFETs 164 MOSFETs 166 C capacitor 123 identifiers 65 L inductor 130 markers mutual inductors 128 names 78 passive 113 capacitors 119 inductor 125 mutual inductor 128 R resistor 116 statements 72 91 temperature 353 transmission line 140 143 element parameters ALTER blocks 83 84 BJTs 162 capacitors 120 DTEMP 352 inductors 125 127 JFETs and MESFETs 164 165 linear inductors 125 137 MOSFETs 166 168 mutual inductors Kxxx 128 resistors 114 115 transmission lines T Element 144 W Element 140 140 141 elements coupled inductor 134 END statement for multiple HSPICE runs 86 in libraries 82 location 86 missing 57 with ALTER 88 85 ENDL statement 81 ENV statement 312 Envelope Analysis ENV 311 envelope simulation 311 ENVFFT 314 ENVFFT statement 314 environment variables 91 ENVOSC 313 ENVOSC
137. 94 NE 794 NE 794 NE 794 NE 794 NE 794 NE 794 NE 794 NE 794 NE 794 NE 794 NE 794 NE 794 NE 794 NE 794 NE 794 NE 794 NE 794 NE 794 NE 794 NE 794 NE 794 NE 794 3 4 5 5 6 7 7 8 9 26 0 00184653 0789158 oOoOmaArNIoauw ss Ww 07968 07968 07899 07899 07939 2 O OCO O GGO G 26 26 91 91 92 0789158 NL 1039 X 0 00871972 NL 1040 A 0 344453 NL 2039 A 0 343427 1 NE_794 13 66 1953 311289 1 NE_794 2 0 794 794 794 794 794 794 794 11 13 14 15 15 17 19 2 20 21 25 26 3 4 e3 22224242424 ma eo 12 14 19 16 20 18 21 0 353289 365644 227289 239644 14 0511746 a O G OO OG 311289 NE 794 9 65 9153 N N N N NE 794 NE 794 NE 794 NE 794 NE 794 NE 794 NE 794 NE 794 NE 794 NE 794 Free e 794 23 1 15117 1039 X 3 01917 794 26 0 166349 1040 A 0 651175 10 65 9153 4 0 311289 17 66 18 66 6 0 311289 11 65 12 65 8 0 311289 16 66 3213 10 0 311289 NL 1039 X NE 794 25 1 00317 NL 2039 A NE 794 23 0 171175 5437 5437 98853 9853 Linear acceleration by using the SIM_LA option accelerates the simulation of circuits that include large linear RC networks To achieve this acceleration 344 HSPICE RF User Guide Z 2007 03 Chapter 13 Post Layout Analysis Linear Acceleration HSPICE RF reduces all matrices that represent RC networks The result is a smaller ma
138. AC frequency sweep includes negative frequencies SNAC allows only frequencies that are greater than or equal to zero No SN statement is specified error at parser SNAC requires an SN statement to generate the steady state solution Warning Messages More than one SNAC statement warning at parser HSPICE RF uses only the last SNAC statement in the netlist No SNAC sources are specified error at parser SNAC requires at least one SNAC source GMRES Convergence Failure When GMRES Generalized Minimum Residual reaches the maximum number of iterations and the residual is greater than the specified tolerance The SNAC analysis generates a warning and then continue as if the data were valid This warning reports the following information Final GMRES Residual Target GMRES Residual m Maximum Krylov Iterations Actual Krylov Iterations taken SNAC Example The following example is shipped with the HSPICE RF distribution as mix_snac sp and is available in directory lt installdir gt demo hspicerf examples HSPICE RF User Guide 267 Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Multitone Harmonic Balance Noise HBNOISE Test SNAC ideal I Q mixer rrd OPTIONS PROBE OPTIONS POST 2 OPTIONS snmaxiter 100 OPTIONS SNACCURACY 5 vlo lo 0 1 0 cos 1 0 0 5 1g Periodic Large Signal SN Input rlo lo 0 1 0 rrf rf 0 1 0 Noise source rrfi rfl rf 1 0 Noise source
139. ALL duration RISE FALL nodel lt node2 gt lt hi lo hi_th low _th gt Fora SETUP condition this is the minimum time before the triggering event during which the specified nodes cannot rise or fall Figure 52 SETUP Example nodeA 7 v1 HI HI_ thresh te es eek LO thresh i i LO 1 t gt 2ns For syntax and description of this statement see CHECK SETUP in the HSPICE and HSPICE RF Command Reference Fora HOLD condition this is minimum time required after the triggering event before the specified nodes can rise or fall 414 HSPICE RF User Guide Z 2007 03 Chapter 16 Advanced Features POWER DC Analysis Figure 53 HOLD Example vin g nodeA HI Htl_thresh LO_thresh LO gt t gt 2ns For syntax and description of this statement see CHECK HOLD in the HSPICE and HSPICE RF Command Reference IR Drop Detection You use the CHECK IRDROP statement to verify that the IR drop does not exceed or does not fall below a specified value for a specified duration For example CHECK IRDROP volt _val time nodel lt node2 gt lt hi lo hi_th low_th gt Figure 54 IR Drop Example t lt ins For syntax and description of this statement see CHECK IRDROP in the HSPICE and HSPICE RF Command Reference POWER DC Analysis You use the POWERDC standby current statement to calc
140. Apply the following considerations when using PERJITTER T should be forced to be between lt 7 lt 2 T since period cannot go negative and the curve should be symmetrical It is reasonable to require that PERJITTER lt T Otherwise the jitter would result in very large period changes and many would be 7 lt 0 To establish a waveform reference the first period should be T i e no jitter in the first period This helps to establish good eye diagrams 194 HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements References References 1 2 3 4 5 6 7 Lael L J Greenstein and M Shafi Microwave Digital Radio IEEE Press 1988 N Sheikholeslami and P Kabal A Family of Nyquist Filters Based on Generalized Raised Cosine Spectra Proceedings of the 19th Biennial Symposium on Communications Kingston Ontario pages 131 135 June 1998 IEEE Standard Definitions of Physical Quantities for Fundamental Frequency and Time Metrology Random Instabilities IEEE Std 1139 1999 A van der Ziel Noise in Solid State Devices and Circuits John Wiley amp Sons 1986 A Demir A Mehrotra and J Roychowdhury Phase noise in oscillators A unifying theory and numerical methods for characterization IEEE Trans Circuits Syst vol 47 pp 655 674 May 2000 A Hajimiri S Limotyrakis and T H Lee Jitter and phase noise in ring oscillators
141. CE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Port Element Parameter Description lt TRANFORHB 0 1 gt DCOPEN lt z0 val gt HSPICE RF User Guide Z 2007 03 0 default The transient description is ignored if an HB value is given or a DC value is given If no DC or HB value is given and TRANFORHB 0 then HB analysis treats the source as a DC source and the DC source value is the time 0 value 1 HB analysis uses the transient description if its value is VMRF SIN PULSE PWL or LFSR If the type is a non repeating PWL source then the time infinity value is used as a DC analysis source value For example the following statement is treated as a DC source with value 1 for HB analysis vi 10 PWL 00 1n1 1u 1 TRANFORHB 1 In contrast the following statement is a OV DC source vi 10 PWL 00 1n1 1u 1 TRANFORHB 0 The following statement is treated as a periodic source with a 1us period that uses PWL values vi 10 PWL 00 1n 1 0 999u 1 1u 0 R TRANFORHB 1 To override the global TRANFORHB option explicitly set TRANFORHB for a voltage or current source Switch for open DC connection when DC magis not set 0 default P element behaves as an impedance termination 1 P element is considered an open circuit in DC operating point analysis DCOPEN 1 is mainly used in LIN analysis so the P element will not affect the self biasing device under test by opening the termination at the
142. CK command cannot refer to another SWEEPBLOCK to build its list of values You cannot include data sweeps in a SWEEPBLOCK statement Clock Source with Random Jitter In many applications involving signal integrity RF analog and mixed signal design it is desirable to have an ideal signal source such as a sine wave or square wave that also includes a non ideal random drift in phase jitter Such a source is useful for representing non ideal clock sources during time domain transient simulation Modeling jitter in this way can be used to examine eye diagram behavior or study how jitter may propagate through a circuit or system A source with jitter is useful for representing non ideal clock sources during time domain transient simulation HSPICE RF User Guide 191 Z 2007 03 Chapter 6 Testbench Elements Clock Source with Random Jitter 192 The PERJITTER option allows you to add periodic jitter to SIN COS and PULSE time domain sources Syntax of SIN COS and Pulse Sources The syntax of SIN source is Vxxx n n SIN lt gt vo va lt freq lt td lt q lt j gt gt gt gt lt gt lt PERJITTER val SEED val gt gt Ixxx n n SIN lt gt vo va lt freq lt td lt q lt j gt gt gt gt lt gt lt PERJITTER val SEED val gt gt Parameter Description VXXX Independent voltage source xxx Independent current source PERJITTER Period jitter PWL Keyword for piecewise linear PWLFILE Text file con
143. DIVIDER divider DELIMITER delimiter SUBCKT GROUND_NET path divider net_name NET path divider net_name path divider instance name pin_name net capacitance P pin name pin type pinCap resistance unit o capacitance unit F x coordinate y_coordinate I path divider instance_name delimiter pin _name path divider instance name pin_name pin type pinCap resistance unit 0 capacitance unit F x coordinate y coordinate S path divider net_name path divider instance name delimiter pin_name pin_name instance number x coordinate y_coordinate capacitor statements resistor statements subcircuit call statements ENDS 328 HSPICE RF User Guide Z 2007 03 END Chapter 13 Post Layout Analysis Post Layout Back Annotation Table 25 DSPF Parameters Parameter Definition DSPF Specifies that the file is in DSPF format version Version number of the DSPF specification optional design_name date vendor program_name program_version divider delimiter path net_name HSPICE RF User Guide Words that start with are keywords Or use the option either preceding or following For example P I means you can use either the P option or the I option Name of your circuit design optional Date and time when a parasitic extraction tool such as Star RCXT generated the DSPF file optional
144. E Output Syntax and File Format You can specify which frequencies the element noise is printed The frequencies must match the sweep_frequency values defined in the frequency_sweep otherwise they are ignored In the element noise output the elements that contribute the largest noise are printed first The frequency values can be specified with the NONE or ALL keyword which either prints no frequencies or every frequency defined in frequency_sweep Frequency values must be enclosed in parentheses For example listfreq none listfreq all listfreq 1 0G listfreq 1 0G 2 0G The default value is NONE listcount Prints the element noise value to the is file which is sorted from the largest to smallest value You do not need to print every noise element instead you can define listcount to print the number of element noise frequencies For example listcount 5 means that only the top 5 noise contributors are printed The default value is 1 listfloor Prints the element noise value to the is file and defines a minimum meaningful noise value in V Hz units Only those elements with noise values larger than list floor are printed The default value is 1 0e 14 V Hz listsources Prints the element noise value to the is file when the element has multiple noise sources such as a FET which contains the thermal shot and 1 f noise sources You can specify either ON or OFF ON prints the contribution from each noise s
145. ERPOST 418 SIM_POWERSTART 418 SIM_SPEF 320 SIM_SPEF_ACTIVE 323 SIM_SPEF_INSERROR 325 SIM_SPEF_LUMPCAPS 325 SIM_SPEF_MAX_ITER 324 SIM_SPEF_PARVALUE 325 SIM_SPEF_RAIL 324 SIM_SPEF_SCALEC 324 SIM_SPEF_SCALER 324 SIM_SPEF_VTOL 323 SIM_TG_THETA 391 SIM_TRAP 391 options configuration file 400 oscillator HB analysis 229 phase noise 238 oscillator analysis 229 oscillator example 28 output files 10 format DSPF 327 NW 394 tabulated data 393 WDB 393 generating 10 restricting 403 variables function 103 P p2d file 310 packed input files 57 PAR keyword 100 PARAM statement 82 350 in ALTER blocks 83 84 parameters algebraic 99 100 analysis 99 assignment 97 cell geometry 105 constants 97 data type 97 data driven analysis 81 defaults 108 defining 95 106 evaluation order 97 hierarchical 88 104 inheritance 107 108 input netlist file 67 libraries 106 108 M 88 measurement 99 Index modifying 81 optimization 105 overriding 106 108 PARHIER option 108 passing 104 111 order 97 problems 111 Release 95 1 and earlier 111 scope 104 105 111 simple 97 subcircuit 88 user defined 98 PARHIER option 108 passive element 113 path names 79 periodic AC algorithm 242 periodic pime dependent noise analysis 281 phase noise 238 phase noise analysis 238 PHASENOISE 238 240 PHASENOISE algorithms 242 PHOTO model parameter 368 PI linear acceleration algorithm 346 port_element_voltage_matchload configuration option 401 p
146. ERTZ keyword to form frequency dependent resistors HSPICE RF accurately analyzes these in all time domain and frequency domain simulations In this example R4 has resistance with both DC and skin effect contributions R4 in out R 100 0 sqrt HERTZ 1000 0 Linear Resistors Rxxx nodel node2 lt modelname gt lt R gt value lt TCl val gt lt TC2 val gt lt W val gt lt L val gt lt M val gt lt C val gt lt DTEMP val gt lt SCALE val gt Parameter Description RXxxx Name of a resistor node1 and node2 Names or numbers of the connecting nodes HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Passive Elements Parameter Description modelname Name of the resistor model value Nominal resistance value in ohms R Resistance in ohms at room temperature TC1 TC2 Temperature coefficient W Resistor width L Resistor length M Parallel multiplier C Parasitic capacitance between node2 and the substrate DTEMP Temperature difference between element and circuit SCALE Scaling factor Example R1 1 2 10 0 Rload 1 GND RVAL param rx 100 R3 2 3 RX TC1 0 001 TC2 0 RP X1 A X2 X5 B 5 MODEL RVAL R In the example above R1 is a simple 10Q linear resistor and Rload calls a resistor model named RVAL which is defined later in the netlist Note If a resistor calls a model then you do not need to specify a constant resistance as you do with R1 R3 ta
147. F G and H Element Statements For E F G and H elements you can use the VMRF function to modulate I t and Q t signals with a RF carrier signal The and Q signal are driven by PWL sources that might be generated by an external tool such as MATLAB The PWL source accepts a text file containing time and voltage or current pairs When the VMRF function is used with controlled sources it is anticipated that the in phase I and quadrature Q signals are not digital but continuous time analog signals The VMRF function therefore includes no filtering and merely serves to create the complex modulation on the RF carrier Exxx n n lt VCVS gt VMRF lt gt Iin Iin Qin Qin FREQ fc PHASE ph lt SCALE A gt lt gt Fxxx n n lt CCCS gt VMRF lt gt VI VQ FREQ fc PHASE ph lt SCALE A gt lt gt HSPICE RF User Guide Z 2007 03 GXXX N n Chapter 6 Testbench Elements Complex Signal Sources and Stimuli lt VCCS gt VMRF lt gt Iin Iin Qin Qin FREQ fc PHASE ph lt SCALE A gt lt gt Hxxx n n lt CCVS gt VMRF lt gt VI VQ FREQ fc PHASE ph lt SCALE A gt lt gt Parameter Description Exxx Voltage controlled voltage source FXXX Current controlled current source Gxxx Voltage controlled current source Hxxx Current controlled current source VCVS Keyword for voltage controlled voltage source CCCS Keyword for current controlled current source VCCS Keyword for voltage controlled
148. F V KV QQ V a Y aQ V I HSPICE harmonic balance analysis HBOSC adds the fundamental frequency of oscillation to the list of unknown circuit quantities To accommodate the extra unknown the phase or equivalently the imaginary part of one unknown HSPICE RF User Guide 229 Z 2007 03 Chapter 9 Oscillator and Phase Noise Analysis Input Syntax for Harmonic Balance Oscillator Analysis variable generally a node voltage is set to zero The phases of all circuit quantities are relative to the phase at this reference node Additionally HSPICE HB tries to avoid the degenerate solution where all non DC quantities are zero Although this is a valid solution of the above equation it is the correct solution if the circuit does not oscillate HB analysis might find this solution incorrectly if the algorithm starts from a bad initial solution Harmonic balance follows the technique described by Ngoya et al which uses an internally applied voltage probe to find the oscillation voltage and frequency The source resistance of this probe is a short circuit at the oscillation frequency and an open circuit otherwise HSPICE RF uses a two tier Newton approach to find a non zero probe voltage which results in zero probe current HSPICE HB approach uses the DC solution as a starting point for non autonomous HB analysis In addition to the DC solution autonomous circuits need an accurate initial value for both the osc
149. Functions HSPICE Form Function Class Description sin x sine trig Returns the sine of x radians cos x cosine trig Returns the cosine of x radians tan x tangent trig Returns the tangent of x radians asin x arc sine trig Returns the inverse sine of x radians acos x arc cosine trig Returns the inverse cosine of x radians 100 HSPICE RF User Guide Z 2007 03 Chapter 5 Parameters and Functions Built In Functions and Variables Table 12 Synopsys HSPICE Built in Functions Continued HSPICE Form Function Class Description atan x arc tangent trig Returns the inverse tangent of x radians sinh x hyperbolic trig Returns the hyperbolic sine of x radians sine cosh x hyperbolic trig Returns the hyperbolic cosine of x radians cosine tanh x hyperbolic trig Returns the hyperbolic tangent of x radians tangent abs x absolute math Returns the absolute value of x x value sqrt x square root math Returns the square root of the absolute value of x sqrt x sqrt x pow x y absolute math Returns the value of x raised to the integer part of y power x integer part of y pwr x y signed math Returns the absolute value of x raised to the y power power with the sign of x sign of x x x y power If x lt 0 returns the value of x raised to the integer part of y If x 0 returns 0 If x gt 0 returns the value of x raised to the y power log x natural math Returns the natural
150. HBLIN statement to extract frequency translation S parameters and noise figures Frequency translation S parameter describes the capability of a periodically linear time varying systems to shift signals in frequency The S parameters for a frequency translation system are similar to the S parameters of a linear time varying system it is defined as Equation62 gt S a S Akg pen C The incident waves a w and reflected waves b w are defined by using these equations oe Vi w nwo Zo1 w nwo in AR Vi w nwo Zot j w nWo a Equation 63 bi 298 HSPICE RF User Guide Z 2007 03 Chapter 11 S parameter Extraction Frequency Translation S Parameter HBLIN Extraction Where Wg is the fundamental frequency tone nis a signed integer iis the port number a w is the input wave at the frequency w nw on the i port Dew is the reflected wave at the frequency w nw on the itn port V w nw is the Fourier coefficient at the frequency w nw of the voltage at port i I w wng is the Fourier coefficient at the frequency w nw of the current at port i Zp is the reference impedance at port i V and definitions are Fourier coefficients rather than phasors For a multi tone analysis it can be expressed as Equation 64 b w PAR 1 jim My MN No Ny a Bie h w cy 1 Dee N ip Due k jVp en W 7 HSPICE RF User Guide 299 Z 2007 03
151. HSPICE RF User Guide Version Z 2007 03 March 2007 SYNOPSYS Copyright Notice and Proprietary Information Copyright 2007 Synopsys Inc All rights reserved This software and documentation contain confidential and proprietary information that is the property of Synopsys Inc The software and documentation are furnished under a license agreement and may be used or copied only in accordance with the terms of the license agreement No part of the software and documentation may be reproduced transmitted or translated in any form or by any means electronic mechanical manual optical or otherwise without prior written permission of Synopsys Inc or as expressly provided by the license agreement Right to Copy Documentation The license agreement with Synopsys permits licensee to make copies of the documentation for its internal use only Each copy shall include all copyrights trademarks service marks and proprietary rights notices if any Licensee must assign sequential numbers to all copies These copies shall contain the following legend on the cover page This document is duplicated with the permission of Synopsys Inc for the exclusive use of and its employees This is copy number Destination Control Statement All technical data contained in this publication is subject to the export control laws of the United States of America Disclosure to nationals of other countries contrary to United States law is prohibited
152. HSPICE RF Tutorial Example 5 Using HBOSC Analysis fora CMOS GPS VCO Transient Analysis Test Bench stimulate oscillation with 2mA pulse iosc IP IN PULSE 0 2m Oln Oln Oln 10n lu probe tran v IP v IN print tran v IP v IN TRAN Oln 10n Harmonic Balance Test Bench sweepblock vtune_sweep 0 50 2 2 3 0 1 HBOSC tones 1550e6 nharms 12 PROBENODE IP QN 4 sweep Vtune sweepblock vtune_sweep kk phasenoise dec 10 100 1e7 print phasenoise phnz probe phasenoise phnz print hb v IP IN v IP IN 1 v QP QN v QP QN 1 probe hb v IP IN v IP IN 1 v QP QN v QP QN 1 probe hb hertz 1 1 NMOS Device from MOSIS 0 35um Process BSIM3 VERSION 3 1 PARAMETERS DATE Mar 8 00 LOT n9co WAF 07 Temperature _parameters Default MODEL NMOS NMOS LEVEL 49 VERSION 3 1 TNOM 27 TOX 7 9E 9 XJ 1 5E 7 NCH 1 7E17 VTHO 0 5047781 K1 0 5719698 K2 0 0197928 K3 33 4446099 K3B 3 1667861 wo 1E 5 NLX 2 455237E 7 DVTOW 0 DVT1W 0 DVT2W 0 DVTO 2 8937881 DVT1 0 6610934 DVT2 0 0446083 U0 421 8714618 UA 1 18967E 10 UB 1 621684E 18 UC 3 422111E 11 VSAT 1 145012E5 AO 1 119634 AGS 0 1918651 BO 1 800933E 6 B1 5E 6 KETA 3 313177E 3 A1 0 A2 1 RDSW 984 149934 PRWG 1 133763E 3 PRWB 7 19717E 3 WR 1 WINT 9 590106E 8 LINT 1 719803E 8 XL 5E 8 XW 0 DWG 2 019736E 9 DWB 6 217095E 9 V
153. Hz To run this simulation type the following line at the command line hspicerf gsmlnaIP3 sp Viewing Results using CosmoScope For this analysis the print statement will generate a lt design_name gt printhbO file Assume you want to find out the output power through the load resistor at the first tone when the input power is 0 1mW HSPICE RF User Guide 25 Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 3 Using HB Analysis for an Amplifier 26 To view the file AN Click the 4 Analysis button and then click on the Print tab Click the 3 Simulation button Invoke CosmosScope by clicking on the Waveform button Choose File gt Open gt Plotfiles to open the lt design_name gt hb0 file Be sure to select HSPICERF hb pn hr jt from the Files of type pulldown to find the lt design_name gt hb0 file Plot the signals Pr rload 1 0 Pr rload 2 0 and Pr rload 2 1 on top of each other The X axis will be the input power and the Y axis will be the output power Result CosmosScope will display the input and output power in dBm But there will be a W or Watt after the dBm label this is incorrect To measure the 1dB compression point of the amplifier open the measurement tool by clicking on the caliper icon at the bottom tool bar Use the down arrow at the end of the Measurement field and select RF and P1dB The PowerOut field should contain the Pr rload 1 0 trace Select a Powerln
154. IDEAL SPICE FILE ideal_spice netlist sp NETLIST IDEAL SPICE TYPE layout NETLIST IDEAL SPICE HIER YES xxx for XREF COMPLETE flow NETLIST IDEAL SPICE FILE ideal spice netlist sp NETLIST IDEAL SPICE TYPE schematic NETLIST IDEAL SPICE HIER YES Note Before version 2002 2 Star RCXT used NETLIST IDEAL SPICE SKIP CELLS to generate the hierarchical ideal SPICE netlist HSPICE RF can still simulate post layout designs using the brute force flow but the post layout flow is preferable in HSPICE RF 318 HSPICE RF User Guide Z 2007 03 Chapter 13 Post Layout Analysis Post Layout Back Annotation HSPICE RF supports these post layout flows to address your post layout simulation needs Standard Post Layout Flow Selective Post Layout Flow Additional Post Layout Options Standard Post Layout Flow Use this flow mainly for analog or mixed signal design and high coverage verification runs when you need to back annotate RC parasitics into the hierarchical LVS ideal netlist In this flow HSPICE RF expands all nets from the DSPF or SPEF file To expand only selected nets use see Selective Post Layout Flow on page 322 Figure 23 Standard Post Layout Flow Extraction Tool DSPF Ideal Netlist SPEF HSPICE RF Back annotation li HSPICE RF User Guide 319 Z 2007 03 Chapter 13 Post Layout Analysis Post Layout Back Annotation
155. IN INVX1FNTC_IN INVX1FNTC_IN INVX1FNTC_IN INVX1FNTC_IN INVX1FNTC_IN INVX1FNTC_IN INVX1FNTC_IN INVX1FNTC_IN INVX1LFNTC_IN INVX1FNTC_IN INVX1FNTC_IN INVX1FNTC_IN NL 1000 A 0 3 TR_000 A 0 34 RES 152 153 154 15 5 156 INVX1FNTC_IN INVX1FNTC_IN INVX1FNTC_IN INVX1FNTC_IN INVX1FNTC_IN 0 0 343 0 343 0 343 0 343 46393 053 0 10 0 154198 11 0 117827 12 0 463063 13 0 0384381 14 0 00246845 15 0 00350198 16 0 00226712 17 0 0426184 18 0 0209701 2 0 0699292 20 0 019987 21 0 0110279 24 0 0192603 25 0 0141824 3 0 0520437 0527105 1184749 0468458 0391578 0113856 9 0 0142528 44804 506 aNA us D O OOGG INVX1FNTC IN 18 8 39117 INVX1FNTC_IN 5 25 1397 11 INVXIFNTC_IN 20 4 59517 12 INVXIFNTC_IN 13 3 688 13 INVXIFNTC_IN 17 25 102 HSPICE RF User Guide Z 2007 03 157 158 LS 9 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 END INVX1FNTC_IN INVX1FNTC_IN INVX1FNTC_IN INVX1FNTC_IN INVX1FNTC_IN INVX1FNTC_IN 14 14 L5 15 17 18 INVX1FNTC_IN 16 0 NL_1000 A 0 804 INVX1FNTC_IN 16 INVX1FNTC_IN 24 INVX1FNTC_IN 25 5 FL 1000 A 1 36317 Chapter 13 Post Layout Analysis Post Layout Back Annotation 0856444 1 73764 0 307175 65517 INVXIFNTC_IN 2 INVX1FNTC_IN 4 6 95371 INVXIFNTC_IN 2 INVX1FNTC_IN 5 50 9942 INVX1FNTC_IN NVX1FNTC_ IN NVX1FNTC_ IN NVX1FNTC_ IN NVX1FNTC_ IN NVX1FNTC
156. IN Analysis for a NMOS Low Noise Amplifier NMOS 0 25um Cascode LNA for GSM applications setup for s parameter and noise parameter measurements Revision 2 0 change to HB analysis and add measurements 2 tone HB analysis 1 tone as input 1 tone as interfere Port element as power source sweep input power kk temp 27 options post 2 param Vdd 2 3 global gnd kk Cascode LNA tuned for operation near 1 GHz kk M1 n4 n3 n5 n5 CMOSN l 0 25u w 7 5u as 15p ad 15p ps 19u pd 19u m 80 M2 n6 nli n4 n4 CMOSN 1 0 25u w 7 5u as 15p ad 15p ps 19u pd 19u m 80 M3 rfo _n6 gnd gnd CMOSN 1 0 25u w 7 5u as 15p ad 15p ps 19u pd 19u m 40 ri vdd _n6 400 11 _n5 gnd 1 0 9nH 12 rfin _n3 1 13nH vvb _n1 gnd dc 1 19 bias for common base device vvdd _vdd gnd dc vdd rfb rfo _n6 120 feedback kk 50 Ohm input port incl bias 255 Ohm output port kk P1 rfin gnd port 1 z0 50 dc 0 595 input port includes DC bias P2 rfo _vdd port 2 z0 255 port doubles as pull up resistor kk Measure s parameters and noise parameters k AC DEC 50 100MEG 5G LIN noisecalc 1 sparcalc 1 PRINT S11 DB S21 DB S12 DB S22 DB NFMIN kk Approximate parameters for TSMC 0 25 Process MOSIS run T17B kk MODEL CMOSN NMOS LEVEL 49 3 1 TNOM 27 TOX 5 8E 9 XJ 1E 7 NCH 2 3549E17 VTHO 0 3819327 K1 0 477867 K2 2 422759E 3 K3 1E 3 K3B 2 1606637 WO 1E 7 NLX 1 57986E 7 DVTOW 0 DVT
157. ISE Measuring HBNOISE Analyses with MEASURE Note A MEASURE HBNOTSE statement cannot contain an expression that uses a HBNOISE variable as an argument Also you cannot use a MEASURE HBNOISE statement for error measurement and expression evaluation of HBNOISE The MEASURE HBNOISE syntax supports four types of measurements m Find when MEASURE HBNOISE result FIND out_varl At Input Frequency Band value The previous measurement yields the result of a variable value at a specific IFB point MEASURE HBNOISE result FIND out _varl WHEN out _var2 out _var3 The previous measurement yields the result at the input frequency point when out_var2 out_var3 MEASURE HBNOISE result WHEN out_var2 out_var3 The previous measurement yields the input frequency point when out_var2 out_vars3 Average RMS min max and peak to peak MEASURE HBNOISE result lt RMS gt out_var lt FROM IFB1 gt lt TO IFB2 gt Integral evaluation MEASURE HBNOISE result INTEGRAL out_var lt FROM IFB1 gt lt TO IFB2 gt This measurement integrates the out_var value from the IFB1 frequency to the IFB2 frequency Derivative evaluation MEASURE HBNOISE result DERIVATIVE out_var AT IFB1 This measurement finds the derivative of out_var at the IFB1 frequency point HSPICE RF User Guide 273 Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Multitone Harmonic Balan
158. ITL3 ITL5 ITLPZ LIMPTS LVLTIM MAXAMP MBYPASS NEWTOL RELH RELI RELQ RELV RELVAR TRTOL All matrix options All error options HSPICE and HSPICE RF Command Reference Some Transient AC input output I O options HSPICE RF does support POST and PROBE options Sub circuit cross listing in a pa file HSPICE Simulation and Analysis User Guide Chapter 3 r command line argument for a remote host HSPICE Simulation and Analysis User Guide OP supports node voltage for any time but HSPICE and HSPICE RF Command Reference supports element values only for t 0 Sensitivity analysis SENS HSPICE Simulation and Analysis User Guide HSPICE and HSPICE RF Command Reference 6 HSPICE RF User Guide Z 2007 03 Chapter 1 HSPICE RF Features and Functionality HSPICE and HSPICE RF Differences Table 1 HSPICE Features Not in HSPICE RF Continued Feature See DC mismatch analysis DCMATCH HSPICE Simulation and Analysis User Guide HSPICE and HSPICE RF Command Reference Table2 Device Models Not in HSPICE RF Model See B element IBIS buffer HSPICE Signal Integrity Guide Bname n1 n2 parameters data driven element current source HSPICE Elements and Device Models Manual data driven V element voltage source HSPICE Elements and Device Models Manual BJT LEVEL 10 MODELLA HSPICE Elements and Device Models Manual Chapter 5 MOSFET Levels 4 8 HSPICE MOSFET Models Manual Common Model Interface CMI HSPICE MOSFET M
159. It is the reader s responsibility to determine the applicable regulations and to comply with them Disclaimer SYNOPSYS INC AND ITS LICENSORS MAKE NO WARRANTY OF ANY KIND EXPRESS OR IMPLIED WITH REGARD TO THIS MATERIAL INCLUDING BUT NOT LIMITED TO THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE Registered Trademarks Synopsys AMPS Cadabra CATS CRITIC CSim Design Compiler DesignPower DesignWare EPIC Formality HSIM HSPICE iN Phase in Sync Leda MAST ModelTools NanoSim OpenVera PathMill Photolynx Physical Compiler PrimeTime SiVL SNUG SolvNet System Compiler TetraMAX VCS Vera and YIELDirector are registered trademarks of Synopsys Inc Trademarks AFGen Apollo Astro Astro Rail Astro Xtalk Aurora AvanWaves Columbia Columbia CE Cosmos CosmosEnterprise CosmosLE CosmosScope CosmosSE DC Expert DC Professional DC Ultra Design Analyzer Design Vision DesignerHDL Direct Silicon Access Discovery Encore Galaxy HANEX HDL Compiler Hercules Hierarchical Optimization Technology HSIMP YS HSPICE Link iN Tandem i Virtual Stepper Jupiter Jupiter DP JupiterXT JupiterXT ASIC Liberty Libra Passport Library Compiler Magellan Mars Mars Xtalk Milkyway ModelSource Module Compiler Planet Planet PL Polaris Power Compiler Raphael Raphael NES Saturn Scirocco Scirocco i Star RCXT Star SimXT Taurus TSUPREM 4 VCS Express VCSi VHDL Compiler VirS
160. LIST_TYPE_VHDL87 NETLIST_TYPE_VHDL93 or NETLIST_TYPE_EDIF Specifies the type of naming conventions used in the SPEF file If you specify more than one format in one SPEF file HSPICE RF reports an error ROUTING_CONFIDENCE positive_integer Specifies a default routing confidence value for all nets in the SPEF file ROUTING_CONFIDENCE_ENTRY positive_integer character_string Specifies one or more characters that represent additional routing confidence values which you can assign to nets in the SPEF file HSPICE RF User Guide Z 2007 03 Chapter 13 Post Layout Analysis Post Layout Back Annotation Table 26 SPEF Parameters Continued Parameter Definition flow_type continued divider delimiter bus_prefix bus_suffix time_unit capacitance_unit resistance_unit inductance_unit name_index HSPICE RF User Guide Z 2007 03 NAME_SCOPE LOCAL FLAT Specifies whether paths in the SPEF file are LOCAL relative to the current SPEF file or FLAT relative to the top level of your circuit design SLEW_THRESHOLDS low high Specifies low and high default input slew thresholds for your circuit design as a percentage of the voltage level for the input pin PIN_CAP NONE INPUT_OUTPUT INPUT_ONLY Specifies the type of pin capacitance to include when calculating the total capacitance for all nets in the SPEF file either no capacitance all input and output capacitances or only input capacitances Charac
161. M 1 91le9 TO 2 0e9 Finds max output power MEASURE HB MinPwr MIN P Rload 1 FROM 1 91le9 TO 2 0e9 Finds min output power Example 2 In the following example the independent variable is the power variable and the MEASURE values return results based on the power sweep Units are in Watts HARMONIC BALANCE power sweep for amplifier param freql 1 91e9 power 1le 3 HB tones freql nharms 10 sweep power DEC 10 1e 6 le 3 MEASURE HB PatluW FIND P Rload 1 AT le 6 Pout at luW MEASURE HB PinlW WHEN P Rload 1 1 Pin 1 Watt Pout MEASURE HB PrangelW TRIG AT 1 92e9 TARG P Rload 1 VAL 1 CROSS 2 IW oper range MEASURE HB ssGain DERIV P Rload 1 AT 1e 5 relative power gain at 10uW input MEASURE HB Gain3rd DERIV P Rload 3 AT 1le 5 3rd harmonic gain at 10uW input MEASURE HB PAE1W FIND P Rload 1 power P Vdc 0 WHEN P Rload 1 1 PAE at 1 Watt output Example 3 In this example the independent variable is again the power variable and the MEASURE values return results based on the power sweep This is a two tone sweep where both input frequency sources are at the same power level in Watts HSPICE RF User Guide 213 Z 2007 03 Chapter 7 Steady State Harmonic Balance Analysis HB Output Data Files HARMONIC BALANCE two tone sweep for amplifier An IP3 calculation is made at 10uW in the sweep param fregql 1 91le9 freq2 1 91e9 power le 3 HB tones f
162. NVOLUTION Indicates which method is used 0 Acts the same as the conventional method This is the default 1 Applies recursive convolution and if the rational function is not accurate enough it switches to linear convolution 2 Applies linear convolution 118 HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Passive Elements Parameter Description FBASE Specifies the lower bound of the transient analysis frequency For CONVOLUTION 1 mode HSPICE starts sampling at this frequency For CONVOLUTION 2 mode HSPICE uses this value as the base frequency point for Inverse Fourier Transformation For recursive convolution the default value is OHz and for linear convolution HSPICE uses the reciprocal of the transient period FMAX Specifies the possible maximum frequency of interest The default value is the frequency point where the function reaches close enough to infinity value assuming that the monotonous function is approaching the infinity value and that it is taken at 10THz The equation should be a function of HERTZ If CONVOLUTION is turned on when a HERTZ keyword is not used in the equation it is automatically be turned off to let the resistor behave as conventional The equation can be a function of temperature but it cannot be node voltage or branch current and time The equation can only be a function of time independent variables such as hertz and temperature Example R1 1 2 r 1
163. OFF 0 1076921 NFACTOR 0 CIT 0 CDSC 2 4E 4 CDSCD 0 34 HSPICE RF User Guide Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 5 Using HBOSC Analysis fora CMOS GPS VCO CDSCB 0 ETAO 0 0147171 ETAB 7 256296E 3 DSUB 0 3377074 PCIM 1 1535622 PDIBLC1 2 946624E 4 PDIBLC2 4 171891E 3 PDIBLCB 0 0497942 DROUT 0 0799917 PSCBE1 3 380501E9 PSCBE2 1 69587E 9 PVAG 0 4105571 DELTA 0 01 MOBMOD 1 PRT 0 UTE 1 5 KT1 0 11 KT1L 0 KT2 0 022 UA1 4 31E 9 UB1 7 61E 18 UC1 5 6E 11 AT 3 3E4 WL 0 WLN 1 WW 1 22182E 15 WWN 1 1657 WWL 0 LL 0 LLN 1 LW 0 LWN 1 LWL 0 CAPMOD 2 XPART 0 4 CGDO 3 73E 10 CGSO 3 73E 10 CGBO 1E 11 Co 8 988141E 4 PB 0 8616985 MJ 0 3906381 CJSW 2 463277E 10 PBSW 0 5072799 MJSW 0 1331717 PVTHO 0 0143809 PRDSW 81 683425 WRDSW 107 8071189 PK2 1 210197E 3 WKETA 1 00008E 3 LKETA 6 1699E 3 PAGS 0 24968 AF 1 0 KF 1 0E 30 END Figure 1 VCO Schematic HSPICE RF User Guide Z 2007 03 Quadrature LC VCO 35 Chapter 3 HSPICE RF Tutorial Example 5 Using HBOSC Analysis fora CMOS GPS VCO The results of the analysis are displayed in Figure 2 on page 36 Figure 3 on page 37 and Figure 4 on page 38 using CosmosScope for VCO waveforms tuning curves and phase noise response Figure 2 VCO Waveforms Output in CosmosScope VCO IQ Waveformsand Spectra V t s timedomain v m1d_b m2d_b
164. ON SIM LA PACT VEC dspf adder vec TRAN 1n 5u vdd vdd 0 3 3 OPTION POST END HSPICE RF User Guide Z 2007 03 Chapter 13 Post Layout Analysis Post Layout Back Annotation To expand only active nodes such as those that move include the SIM _DSPF_ ACTIVE option in your netlist For example OPTION SIM DSPF_ACTIVE active net filename This option is most effective when used with a large design for example over 5K transistors Smaller designs lose some of the performance gain due to internal overhead processing For syntax and description of SIM _DSPF ACTIVE option see OPTION SIM_DSPF_ACTIVE in the HSPICE and RF Command Reference When you have included the appropriate control option run HSPICE RF using the ideal netlist The structure of a DSPF file is DSPF 1 0 DESIGN demo Date October 6 1998 SUBCKT lt name gt lt pins gt Net Section C1 R1 Instance Section ENDS HSPICE RF User Guide 321 Z 2007 03 Chapter 13 Post Layout Analysis Post Layout Back Annotation 322 Selective Post Layout Flow Figure 24 Selective Post Layout Flow Extraction Tool i Ideal Netlist rests vy HSPICE RF Active Nodes lt Back annotation HSPICE RF pe Get You can use the selective post layout flow to simulate a post layout design for a memory or digital circuit and
165. OPTION POST 1 saves the results in binary format OPTION POST 2 saves the results in ASCII format XP Output Format HSPICE RF outputs XP binary format to a file with the xp extension This format is compatible with the HSPICE TR binary format For example to output to a xp file enter OPTION POST xp NW Output Format HSPICE RF outputs the NW format to a file with the nw extension Synopsys developed this format you need a Synopsys waveform display tool to process a file in NW format For example to output to a nwi file enter OPTION POST nw You can compress NW files For additional information see Compressing Analog Files on page 396 VCD Output Format To output your waveforms from HSPICE RF in VCD Value Change Dump format set the VCD option in conjunction with the LPRINT statement For example HSPICE RF User Guide Z 2007 03 Chapter 15 Using HSPICE with HSPICE RF RF Transient Analysis Output File Formats OPTION VCD LPRINT 0 5 4 5 v 0 v 2 v 6 LPRINT Statement You use the LPRINT statement to produce output in VCD file format from transient analysis For example LPRINT v1 v2 output varable list For additional information see LPRINT in the HSPICE and HSPICE RF Command Reference turboWave Output Format To use turboWave output format TW enter OPTION POST tw This format supports analog compression as described in Compressing Analog Files on page 396 Undertow Outpu
166. PICE RF User Guide Z 2007 03 Chapter 4 Input Netlist and Data Entry Input Netlist File Guidelines Table 7 Scale Factors Continued Scale Factor Prefix Symbol Multiplying Factor P pico p 1e 12 F femto f 1e 15 A atto a 1e 18 Note Scale factor A is not a scale factor in a character string that contains amps For example HSPICE interprets the 20amps string as 20e 18mps 20 Samps but it correctly interprets 20amps as 20 amperes of current not as 20e 18mps 20 8amps Numbers can use exponential format or engineering key letter format but not both 1e 12 or 1p but not 1e 6u To designate exponents use D or E The OPTION EXPMAX limits the exponent size Trailing alphabetic characters are interpreted as units comments Units comments are not checked The OPTION INGOLD controls the format of numbers in printouts The OPTION NUMDGT x controls the listing printout accuracy The OPTION MEASDGT x controls the measure file printout accuracy The OPTION VFLOOR x specifies the smallest voltage for which HSPICE or HSPICE RF prints the value Smaller voltages print as 0 Parameters and Expressions Parameter names in HSPICE RF use HSPICE name syntax rules except that names must begin with an alphabetic character The other characters must be either a number or one of these characters S 6 To define parameter hierarchy overrides and defaults use the OPTION PARHIER global local statement
167. RCXT y DSPF SPEF y Post Layout Flow Note HSPICE RF generates an active node file in both Star RC and Star RCXT format It then expands the active node file to the Star RCXT command file to extract only active parasitics 326 HSPICE RF User Guide Z 2007 03 Chapter 13 Post Layout Analysis Post Layout Back Annotation Overview of DSPF Files In general an SPF Standard Parasitic Format file describes interconnect delay and loading due to parasitic resistance and capacitance DSPF Detailed Standard Parasitic Format is a specific type of SPF file that describes the actual parasitic resistance and capacitance components of a net DSPF is a standard output format commonly used in many parasitic extraction tools including Star RCXT The HSPICE RF circuit simulator can read DSPF files DSPF File Structure The DSPF standard is published by Open Verilog International OVI For information about how to obtain the complete DSPF specification or any other documents from OVI see http www ovi org document html The OVI DSPF specification requires the following file structure in a DSPF file Parameters in braces are optional HSPICE RF User Guide 327 Z 2007 03 Chapter 13 Post Layout Analysis Post Layout Back Annotation DSPF file DSPF version DESIGN design name DATE date VENDOR vendor PROGRAM program_name VERSION program version
168. S A a A 3 Q I I l A l 4 3 i er ar ae roa MONT1_SV 118 375 A Aa A a AI RGAUSS_1 SR es A Agp a 4 110 0 A 4 A Z 100 0 4 d a asa A i A X amp B 90 0 aA Aa EO cs 4 80 9998 m te a a ke ae ae ae oe ee TE a a A LAI Loin at 1 0 10 0 20 0 30 0 40 0 50 0 60 0 MONTE CARLO LIN 366 HSPICE RF User Guide Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Monte Carlo Analysis Figure 35 Limit Functions MONT1 SP TEST OF MONTE CARLO GAUSSIAN UNIFORM AND LIMIT FUNCTIONS May 15 2003 11 41 23 MONT1 SVO LIMIT so 120 0 AAAAA AKAN At A AAAAA WA A AA o 7 A ASN AA 115 0 110 0 105 0 100 0 VOLT LIN Ke gi e A A ANE AOE EN E ae A jo ee 1 AAA l LAA AAA A I A AAALAAAT MAANA a k ATA 10 0 20 0 30 0 40 0 50 0 60 0 MONTE CARLO LIN 1 Major and Minor Distribution In MOS IC processes manufacturing tolerance parameters have both a major and a minor statistical distribution The major distribution is the wafer to wafer and run to run variation It determines electrical yield The minor distribution is the transistor to transistor process variation It is responsible for critical second order effects such as amplifier offset voltage and flip flop preference HSPICE RF User Guide 367 Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Monte Carlo A
169. Simulation 312 for HB simulations at each point in time All other sources are associated with the transient timescale Also the input waveforms can be represented in the frequency domain as RF carriers modulated by an envelope by identifying a VMRF signal source in a V or element statement The amplitude and phase values of the sampled envelope are used as the input signal for HB analysis Some typical applications for envelope simulation are amplifier spectral regrowth adjacent channel power ration ACPR and oscillator startup and shutdown analyses Envelope Analysis Commands This section describes those commands specific to envelope analysis These commands are Standard envelope simulation ENV m Oscillator simulation both startup and shutdown ENVOSC Envelope Fast Fourier Transform ENVFFT Nonautonomous Form ENV TONES f1 lt f2 fn gt NHARMS hi lt h2 hn gt ENV STEP tstep ENV_STOP tstop Parameter Description TONES Carrier frequencies in hertz NHARMS Number of harmonics ENV_STEP Envelope step size in seconds ENV_STOP Envelope stop time in seconds Description You use the ENV command to do standard envelope simulation The simulation proceeds just as it does in standard transient simulation starting at time 0 and continuing until time env_stop An HB analysis is performed at each step in time You can use Backward Euler BE trapezoidal TRAP or level 2 Gear GEAR integration
170. The mulvsat instance parameter is supplied to facilitate efficient modeling of mismatch local variation and mechanical stress and proximity effects Variations Specified on Model Parameters In this section we investigate the method of specifying distributions on parameters and using these parameters to define values of model parameters With this approach the netlist does not have to be parameterized The modmonte option can be used to distinguish between global variations all devices of a particular model have the same parameter set or local variations every device has a unique random value for the specified parameters _test10 sp shows a simple case where the model parameter for sheet resistivity is assigned a distribution defined on the parameter rsheet The results show that all resistors have the same value for each Monte Carlo sample but a different one for different samples This setup is useful for studying global variations m test11 sp has option modmonte 1 added Now every resistor has a different value Note that option modmonte has no effect on any other approach presented here In summary assigning parameters with specified distributions to model parameters allows for investigating the effects of global or local variations but not both The possibility of selecting one or the other with a simple option is misleading in the sense that the underlying definitions for global and local variations are not the same
171. URE SNNOISE result FIND out _varl WHEN out _var2 out _var3 The previous measurement yields the result at the input frequency point when out_var2 out_var3 MEASURE SNNOISE result WHEN out_var2 out_var3 The previous measurement yields the input frequency point when out_var2 out_vars Average RMS min max and peak to peak MEASURE SNNOISE result lt RMS gt out_var lt FROM IFB1 gt lt TO IFB2 gt Integral evaluation MEASURE SNNOISE result INTEGRAL out var lt FROM IFB1 gt lt TO IFB2 gt This measurement integrates the out_var value from the IFB1 frequency to the IFB2 frequency Derivative evaluation MEASURE SNNOISE result DERIVATIVE out_var AT IFB1 This measurement finds the derivative of out_var at the IFB1 frequency point Note MEASURE SNNOISE cannot contain an expression that uses an hbnoise variable as an argument You also cannot use MEASURE SNNOTSE for error measurement and expression evaluation of SNNOISE SNNOISE Analysis Example This example performs an SN analysis then runs an SNNOISE analysis over a range of frequencies from 9 0e8 to 9 2e8 Hz Simulation outputs the output noise at V out and the single side band noise figure versus IFB from 9 0e8 to 9 2e8 Hz to the pn0 file The netlist for this example is shown immediately following HSPICE RF User Guide Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Periodic Time
172. Y 05u 1 sigma 5angstrom oxide thickness AT 1SIGMA PARAM ELPOLY AGAUSS 10U 0 02U 1 MODPOLY AGAUSS 0 05U 1 POLYCAP AGAUSS 200e 10 5e 10 1 MODEL CMOD C THICK POLYCAP DEL MODPOLY Electrical Approach The electrical approach assumes no physical interpretation but requires a local element distribution and a global model distribution In this example You can match the capacitors to 1 for the 2 sigma population The process can maintain a 10 variation from run to run for a 2 sigma distribution Cla 1 0 CMOD SCALE ELCAP Clb 1 0 CMOD SCALE ELCAP C1C 1 0 CMOD SCALE ELCAP C1D 1 0 CMOD SCALE ELCAP PARAM ELCAP Gauss 1 01 2 1 at 2 sigma MODCAP Gauss 25p 1 2 10 at 2 sigma MODEL CMOD C CAP MODCAP Worst Case and Monte Carlo Sweep Example The following example measures the delay and the power consumption of two inverters Additional inverters buffer the input and load the output This netlist contains commands for two sets of transient analysis parameter sweep from 3 to 3 sigma and a Monte Carlo analysis It creates one set of output files mtO and trO for the sigma sweep and one set mt1 and tr1 for Monte Carlo HSPICE RF User Guide 371 Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Worst Case and Monte Carlo Sweep Example inv sp sweep mosfet 3 sigma to 3 sigma use measure output param vref 2 5 sigma 0 global 1 vec 10 5 0 vin in 0 pwl 0 0 0 2n 5 xl in 2 in
173. _ IN NVX1FNTC_IN NVX1FNTC_IN NVX1FNTC_IN INVX1FNTC_IN INVX1FNTC_IN INVX1FNTC_IN T I ab I a L ak I TR_1000 A 0 D NET NE 794 1 98538 CONN I NL 1039 X O L 0 D INVX I NL 2039 A I L 0 343 I NL 1040 A I L 0 343 CAP 3387 3388 3389 3390 3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 HSPICE RF User Guide Z 2007 03 NE 794 0 NE 794 1 0 0792492 NE 794 10 0 0789158 NE 794 11 0 0789991 NE 794 12 0 0789991 NE 794 13 0 0792992 NE 794 14 0 00093352 NE 794 15 0 00063346 NE 794 16 0 0792992 NE 794 17 0 80116 NE 794 18 0 80116 NE 794 19 0 00125452 NE 794 2 0 0789158 NE 794 20 0 00336991 NE 794 21 0 00668512 NE 794 23 0 00294932 NE 794 25 0 00259882 INVX1FNTC_IN I1184 A 0 403175 923175 INVX1FNTC_IN INVX1FNTC_IN INVX1FNTC_IN INVX1FNTC_IN INVX1FNTC_IN NVX1FNTC_IN NVX1FNTC_IN NVX1FNTC_IN INVX1FNTC_IN 21 4 71035 12 31 7256 4 11 9254 7 25 3618 6 23 3057 24 8 64717 8 7 46529 10 2 04729 10 10 8533 11 1 05164 343 Chapter 13 Post Layout Analysis Linear Acceleration 3404 3405 3406 3407 3408 3409 3410 3411 3412 3413 3414 RES 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900 2901 2902 2903 2904 END Linear Acceleration NE 794 NE 794 NE 794 NE 794 NE 794 NE 794 NE 794 NE 794 NE 794 NE 7
174. a function of tuning voltage as well as its phase noise characteristics as a function of tuning voltage As in previous examples the oscillator analysis is activated using the HBOSC command The TONE parameter sets an approximate oscillation frequency near 1550 MHz The NHARMS parameter sets the harmonic content to 11th order The PROBENODE parameters identify the drain pins across the first oscillator section as the pair of oscillating nodes This is a differential oscillator and the approximate value for this differential amplitude is 6 1 V The FSPTS parameters set the search frequency range between 1500 and 1600 MHz The SWEEP parameters set a tuning voltage sweep from 2 0 to 3 2 V The following example is based on demonstration netlist gosvco sp which is available in directory lt installdir gt demo hspicerf examples This netlist simulates the oscillator schematic shown in Figure 1 and performs phase noise analysis HSPICE RF User Guide 31 Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 5 Using HBOSC Analysis fora CMOS GPS VCO 32 kk NMOS IC Quadrature VCO circuit for GPS local oscillator kk Twin differential negative resistance VCOs using NMOS transistors for varactors coupled to produce quadrature resonances Design based on 0 35um CMOS process kk References gt P Vancorenland and M S J Steyaert A 1 57 GHz fully integrated very low phase noise quadra
175. a means to provide spectrum analysis Spectrum analysis represents a time domain signal within the frequency domain SNFT uses the Fourier transform a Discrete Fourier Transform DFT uses sequences of time values to determine the frequency content of analog signals in circuit simulation The SNFT statement uses the internal time point values By default the SNFT statement uses a second order interpolation to obtain waveform samples based on the number of points that you specify HSPICE RF User Guide Z 2007 03 Chapter 8 Steady State Shooting Newton Analysis Shooting Newton with Fourier Transform SNFT You can use windowing functions to reduce the effects of waveform truncation on the spectral content You can also use the SNFT command to specify output format frequency number of harmonics total harmonic distortion THD SNFT Input Syntax The SNFT command an take arguments with either alphanumeric or numerics and expressions Syntax 1 Alphanumeric input SNFT lt output_var gt lt START value gt lt STOP value gt lt NP value gt lt FORMAT keyword gt lt WINDOW keyword gt lt ALFA value gt lt FREQ value gt lt FMIN value gt lt FMAX value gt Syntax 2 Numerics and expressions SNFT lt output_var gt lt START param_exprl gt lt STOP param_expr2 gt lt NP param_expr3 gt lt FORMAT keyword gt lt WINDOW keyword gt lt ALFA param_expr4 gt lt FREQ param_expr5 gt
176. acter string as a parameter value in HSPICE RF See Delimiters on page 64 An input statement or equation can be up to 1024 characters long HSPICE RF ignores differences between upper and lower case in input lines except in quoted filenames To continue a statement on the next line enter a plus sign as the first non numeric non blank character in the next line To indicate to the power of in your netlist use two asterisks For example 2 5 represents two to the fifth power 2 To continue all HSPICE or HSPICE RF statements including quoted strings such as paths and algebraics use a single whitespace followed by a backslash or a double backslash at the end of the line that you want to continue e A single backslash preserves white space Names must begin with an alphabetic character but thereafter can contain numbers and the following characters ite amp ee fees P74 bee e When you use an asterisk or a question mark with a PRINT PROBE LPRINT HSPICE RF or CHECK HSPICE RF statement HSPICE or HSPICE RF uses the character as a wildcard For additional information see Using Wildcards on Node Names on page 76 e When you use curly brackets HSPICE converts them to square brackets automatically HSPICE RF User Guide Z 2007 03 Chapter 4 Input Netlist and Data Entry Input Netlist File Guidelines Names are input tokens Token delimiters must precede and
177. ali n V i n Wo Zoli I i n Wo 2 sqrt Z fil bfi n V i n Wo Zoli I i n Wo 2 sqrt Zofi Where Wo is the fundamental frequency tone nis a signed integer iis the port number ali n is the input wave at the frequency n Wo on the ith port bfi n is the reflected wave at the frequency n Wo on the ith port V i n Wo is the Fourier coefficient at the frequency n Wo of the voltage at port i I i n Wo is the Fourier coefficient at the frequency n Wo of the current at port i m Zoli is the reference impedance at port i An HBLSP analysis only extracts the S parameters on the first harmonic that is n 1 Limitations The HBLSP analysis has these known limitations Power dependent S parameter extraction is a 2 port analysis only Multiport power dependent S parameters are not currently supported The intermodulation data block IMTDATA in the p2d file is not supported The internal impedance of the P port Element can only be a real value Complex impedance values are not supported HSPICE RF User Guide Z 2007 03 Chapter 11 S parameter Extraction Large Signal S parameter HBLSP Analysis Input Syntax HBLSP NHARMS nh lt POWERUNIT dbm watt gt lt SSPCALC 1 0 YES NO gt lt NOISECALC 1 0 YES NO gt lt FILENAME file name gt lt DATAFORMAT ri ma db gt FREQSWEEP freq sweep POWERSWEEP power sweep Parameter Desc
178. all other parameters are not supported in HBLSP PRINT and PROBE statements PRINT and PROBE Statements PRINT HBLSP Smn Smn TYPE S m n S m n TYPE small signal 2 port noise params PROBE HBLSP Smn Smn TYPE S m n S m n TYPE small signal 2 port noise params Parameter Description Smn Smn TYPE Complex 2 port parameters Where S m n S m n TYPE m 1or2 n 1or2 TYPE R M P PD D DB or DBM R real imaginary M magnitude P PD phase in degrees D DB decibels DBM decibels per 1 0e 3 small signal 2 port noise G_AS NF RN YOPT GAMMA_OPT NFMIN parameters VN2 ZCOR GN RHON YCOR ZOPT IN2 For a description of these parameters see Linear Network Parameter Analysis in the HSPICE Simulation and Analysis User Guide Output Data Files An HBLSP analysis produces these output data files The large signal S parameters from the PRINT statement are written to a printls file The small signal S parameters from the PRINT statement are written to a printss file HSPICE RF User Guide 309 Z 2007 03 Chapter 11 S parameter Extraction Large Signal S parameter HBLSP Analysis The large signal S parameters from the PROBE statement are written to a ls file The small signal S parameters from the PROBE statement are written to a ss file The extracted large and small signal S and noise parameters are writt
179. all signal you can use a faster linear analysis to analyze its effect For example if a mixer s LO is a large signal but RF is a small signal a single tone HB analysis using the LO frequency can be combined with HBAC in place of a 2 tone HB analysis HSPICE RF User Guide Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 6 Using Multi Tone HB and HBAC Analyses for a Mixer To demonstrate both techniques this example analyzes an ideal mixer built using behavioral elements It is based on demonstration netlist mix_tran sp which is available in directory lt installdir gt demo hspicerf examples Ideal mixer example transient analysis OPTIONS POST vlo lo 01 0 sin 1 0 0 5 1 0g 0 0 90 rrtl TEl FELO gl 0 if cur 1 0 v lo v rf mixer element cl 0 if gq 1 0e 9 v lo v rf mixer element rout if ifg 1 0 vetrl ifg 0 0 0 hl out 0 vetrl 1 0 convert I to V rhi out 0 1 0 vrf rfl 0 sin 0 0 001 0 8GHz 0O 0 114 tran 10p 10n opt sim_accuracy 100 end This example uses behavioral controlled current and charge sources to simulate a mixer The LO signal is driven by a 0 5 Volt sinusoid at 1 GHz and RF is driven by a 10mV signal at 800 MHz The mixer output is the voltage at node out v out Two tone HB Approach To analyze this circuit using 2 tone HB add HB source for LO add HB 0 5 0 1 1 to the LO voltage source this sets the amplitude to 0 5 no phase shift for the first harmonic of the first tone wh
180. alled globw A parameter called globwidth is assigned the value of globw The parameter globwidth is assigned a different random value for each Monte Carlo sample The parameter globwidth is used to define the width of the physical resistors r1 r2 r3 and r4 with model resistor Since parameter globwidth does not have its own distribution defined but rather gets its value from the parameter globw the value for globwidth is the same wherever it is used thus the HSPICE RF User Guide 381 Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Simulating the Effects of Global and Local Variations with Monte Carlo 382 resistors have the same width for each Monte Carlo sample and therefore the same resistance When plotting the simulation results v1 v2 v3 and v4 from the meas file the waveforms overlay perfectly This type of setup is typically used to model global variations which means variations that affect all devices the same way test2 sp has a distribution parameter defined called locwidth This parameter is used to define the width of the physical resistors r1 r2 r3 and r4 with model resistor Since the parameter has its own distribution defined its value will be different for each reference and of course for each Monte Carlo sample Therefore the resistors will always have different values and the voltages will be different This type of setup is typically used to model local variations which means var
181. ally starts from the DC solution and looks for potential resonances in the linear portion of the circuit to determine the initial guess for the oscillation frequency However these resonances generally do not exist in ring oscillators which do not contain many linear elements HB analysis provides a second method of obtaining a good initial guess for the oscillation frequency which is specifically intended for ring oscillators Instead of starting from the results of a DC analysis this method starts from the result of a transient analysis This method is called Transient Initialization and also provides a good initial guess for all the voltages and currents in the circuit HSPICE RF User Guide 233 Z 2007 03 Chapter 9 Oscillator and Phase Noise Analysis HBOSC Analysis Using Transient Initialization HBOSC Analysis Using Transient Initialization To perform an HBOSC analysis use the following options in your HSPICE RF netlist Table 19 HBOSC Analysis Options for Transient Initialization Option Description HBTRANINIT lt time gt Tells HB to use transient analysis to initialize all state variables lt time gt is when the circuit has reached or is near steady state Default 0 HBTRANPTS lt npts gt lt npts gt specifies the number of points per period for converting the time domain data results from transient analysis into the frequency domain lt npts gt must be an integer greater than 0 The units are in n
182. also use the BULK parameter to set this name in the BUT model mname BJT model name reference area Emitter area multiplying factor which affects currents resistances and AREA area_ capacitances Default 1 0 OFF Sets initial condition for this element to OFF in DC analysis Default ON IC vbeval Initial internal base emitter voltage vbeval and collector emitter vceval VBE voltage vceval HSPICE or HSPICE RF uses this value when VCE the TRAN statement includes UIC The IC statement overrides it M Multiplier to simulate multiple BUTs in parallel The M setting affects all currents capacitances and resistances Default 1 DTEMP The difference between the element temperature and the circuit temperature in degrees Celsius Default 0 0 AREAB Base area multiplying factor which affects currents resistances and capacitances Default AREA HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Active Elements Parameter Description AREAC Collector area multiplying factor which affects currents resistances and capacitances Default AREA The only required fields are the collector base and emitter nodes and the model name The nodes and model name must precede other fields in the netlist Example 1 In the Q1 BJT element below Q1 1 2 3 model_1 The collector connects to node 1 The base connects to node 2 The emitter connects to node 3 model_1 references the BUT model
183. amed in the tones list does not have TRANFORHB specified Source named in the tones list has no transient description Source named in the tones list must be HB SIN PULSE PWL or VMRF Tone specification for the source is inconsistent with its frequency HB oscillator analysis has reached the NULL solution HSPICE RF User Guide Z 2007 03 Chapter 7 Steady State Harmonic Balance Analysis HB Output Data Files Table 17 HB Analysis Error Messages Continued File Description HB_ERR 18 Bad subharms format HB_ERR 19 Modtone may not be set to the same value as tone Table 18 HB Analysis Warning Messages File Description HB_WARN 1 hb multiply defined Last one will be used HB_WARN 2 Tone specified for V I source not specified in HB command HB_WARN 3 HB convergence not achieved HB_WARN 4 Source specifies both HB and transient description HB description will be used HB_WARN 5 Source specifies exponential decay HB will ignore it HB_WARN 6 Source specifies a non positive frequency HB_WARN 7 Source does not fit the HB spectrum HB_WARN 8 amp Source cannot be used with the TRANFORHB option HB_WARN 9 Frequency not found from transient analysis HB_WARN 10 hb hbosc will be ignored due to env envosc HB_WARN 11 HBTRANINIT does not support more than one input tone HSPICE RF User Guide Z 2007 03 217 Chapter 7 Steady State Harmonic Balance Analysis References References 1
184. amed r1 and a voltage source named vin then PRINT i prints the current for both of these elements i r1 and i vin HSPICE RF User Guide Z 2007 03 Chapter 4 Input Netlist and Data Entry Input Netlist File Composition e And PRINT v o prints the voltages for all nodes whose names start with o if your netlist contains nodes named in and out this example prints only the v out voltage _ matches any character tht appears within the brackets For example 123 matches 1 2 or 3 A hyphen inside the brackets indicates a character range For example 0 9 is the same as 0123456789 and matches any digit For example the following prints the results of a transient analysis for the voltage at the matched node name PRINT TRAN V 9 t u Wildcards must begin with a letter or a number for example PROBE v correct format PROBE incorrect format PROBE x correct format Here are some practical applications for these wildcards m If your netlist includes a resistor named r1 and a voltage source named vin then PRINT i prints the current for both elements i r1 and i vin The statement PRINT v o prints the voltages for all nodes whose names start with o if your netlist contains nodes named in and out this example prints only the v out voltage f your netlist contains nodes named 0 1 2 and 3 then PRINT v 0 or PRINT v 0 prints the voltage between node 0 and each of th
185. ameter DTEMP and a model reference parameter TREF If you specify DTEMP in an element statement the element temperature for the simulation is element temperature circuit temperature DTEMP Specify the DTEMP value in the element statement resistor capacitor inductor diode BUT JFET or MOSFET statement or in a subcircuit element Assign a parameter to DTEMP then use the Dc statement to sweep the parameter The DTEMP value defaults to zero If you specify TREF in the model statement the model reference temperature changes TREF overrides TNOM Derating the model parameters is based on the difference between circuit simulator temperature and TREF instead of TNOM HSPICE RF User Guide 353 Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Worst Case Analysis TEMP Statement To specify the temperature of a circuit fora HSPICE RF simulation use the TEMP statement Worst Case Analysis 354 Circuit designers often use worst case analysis when designing and analyzing MOS and BJT IC circuits To simulate the worst case set all variables to their 2 or 3 sigma worst case values Because several independent variables rarely attain their worst case values simultaneously this technique tends to be overly pessimistic and can lead to over designing the circuit However this analysis is useful as a fast check Model Skew Parameters The HSPICE RF device models include physically measurable model paramete
186. ance AC HBAC for periodic AC analysis see Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Multitone Harmonic Balance AC Analysis HBAC on page 257 Shooting Newton AC analysis see Shooting Newton AC Analysis SNAC on page 263 Harmonic Balance Noise HBNOISE for periodic time varying AC noise analysis see Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses HSPICE RF User Guide 197 Z 2007 03 Chapter 7 Steady State Harmonic Balance Analysis Harmonic Balance Analysis Shooting Newton noise analysis see Shooting Newton Noise Analysis SNNOISE on page 275 Harmonic balance transfer functions see Multitone Harmonic Balance Transfer Function Analysis HBXF on page 287 Shooting Newton transfer functions see Shooting Newton Transfer Function Analysis SNXF on page 291 Frequency translation S parameter extraction for describing N port circuits that exhibit frequency translation effects see Frequency Translation S Parameter HBLIN Extraction on page 298 Envelope Analysis ENV see Chapter 12 Envelope Analysis You can use steady state analysis on a circuit if it contains only DC and periodic sources These analyses assume that all start up transients have completely died out with only the steady state response remaining Sources that are not periodic or DC are treated as zero valued in these analyses Harmonic Balance Analysis 198
187. ance delimiter logical _pin physical_ node I B O HSPICE RF User Guide Z 2007 03 XC L S D Chapter 13 Post Layout Analysis Post Layout Back Annotation coordinate par_value rising_slew falling_slew low_threshold high_threshold cell type N net_name delimiter net_number coordinate CAP cap_id nodel node2 capacitance RES res_id nodel node2 resistance INDUC induc_id nodel node2 inductance END Table 26 SPEF Parameters Parameter SPEF version design_name date vendor program_name program_version HSPICE RF User Guide Z 2007 03 Definition Specifies that the file is in SPEF format Version number of the SPEF specification such as IEEE 1481 1998 Words that start with an asterisk are keywords Or For example NS PS means choose either nanoseconds or picoseconds as the time units Name of your circuit design Date and time when a parasitic extraction tool such as Star RCXT generated the SPEF file Name of the vendor such as Synopsys whose tools you used to generate the SPEF file optional Name of the program such as Star RCXT that generated the SPEF file Version number of the program that generated the SPEF file 335 Chapter 13 Post Layout Analysis Post Layout Back Annotation Table 26 SPEF Parameters Continued Parameter Definition flow_type 336 One or more of the following flow types EXTERNAL_LOADS T
188. and HSPICE RF RF Numerical Integration Algorithm Control In HSPICE RF you can select either the Backward Euler or Trapezoidal integration algorithm Each of these algorithms has its own advantages and disadvantages for specific circuit types For pre charging simulation or timing critical simulation the Trapezoidal algorithm usually improves accuracy You use the SIM_ORDER option to control the amount of Backward Euler BE to mix with the Trapezoidal TRAP method for hybrid integration For example OPTION SIM ORDER x Setting SIM_ORDER to its lowest value selects Backward Euler integration algorithm and setting it to its highest value selects Trapezoidal integration For the syntax and description of this control option see OPTION SIM_ORDER in the HSPICE and HSPICE RF Command Reference RF Transient Analysis Accuracy Control The default time step method in HSPICE RF mixes timestep algorithms Trapezoidal and second order Gear Gear 2 This yields a more accurate scheme than Trapezoidal or Backward Euler Also detection of numerical HSPICE RF User Guide 389 Z 2007 03 Chapter 15 Using HSPICE with HSPICE RF RF Transient Analysis Accuracy Control 390 oscillations inserts fewer Backward Euler steps than in previous HSPICE versions OPTION SIM_ACCURACY You use the SIM_ ACCURACY option to modify the size of timesteps in HSPICE RF For example OPTION SIM ACCURACY lt value gt A timestep is a time interval a
189. and their means for calculation The types of jitter measurements include Timing Jitter RMS Phase jitter Timing jitter is a measurement of oscillator uncertainty in the time domain For clock applications time domain measurements are preferable since most specifications of concern involve time domain values Timing jitter is the standard deviation of the timing uncertainty which is a function of the auto correlation function in the power spectrum of the phase variations Timing Jitter is the square root of the variance standard deviation squared of the timing uncertainty between two clock edges separated by an interval given byt N T where T is the ideal clock period It can be written as a function of the auto correlation function of the power spectrum of phase variations as Equation 41 or t 1R 0 R t O where TIE is the Time Interval error The Weiner Khintchine Theorem 1 relates the auto correlation function to the power spectrum of phase variations as in the following equation 1 JOT Equation 42 Rt F Soei d 2 L f cos 2nft df where SD is the double sided power spectrum of phase variations and L f is the single sideband phase noise The auto correlation for t 0 is given by Equation 43 RQ ms 2f LO sin nfi 0 which defines in HSPICE RF known as RMS Phase Jitter Using the identity 2sin a 1 cos2o we can then write 2 8 S Equation 44 Srp 1 gt f L f sin nfi df 0
190. ands of HSPICE RF simulation runs Use yield analyses to modify DC operating points DC sweeps AC sweeps Transient analysis HSPICE RF User Guide Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Simulating Circuit and Model Temperatures CosmosScope can generate scatter plots from the operating point analysis or a family of curve plots for DC AC and transient analysis Use MEASURE statements to save results for delay times power or any other characteristic extracted ina MEASURE statement HSPICE RF generates a table of results in an mt file in ASCII format You can analyze the numbers directly or read this file into CosmosScope to view the distributions Also if you use MEASURE statements in a Monte Carlo or data driven analysis then the HSPICE RF output file includes the following statistical results in the listing Kah ae eS Mean ea N x Mean x Mean Variance No Sigma J Variance x Mean o Ix Mean Average Deviation NLI Simulating Circuit and Model Temperatures Temperature affects all electrical circuits Figure 28 shows the key temperature parameters associated with circuit simulation m Model reference temperature you can model different models at different temperatures Each model has a TREF temperature reference parameter Element junction temperature each resistor transistor or other element generates heat so an element is hotter than the ambi
191. aracter element legal anywhere in the element key key letter only string first or included letter only backslash HSPICE Included only Illegal in HSPICE Continuation requires a included only Included only in character for whitespace HSPICE RF HSPICE RF quoted strings before to use preserves asa whitespace continuation A double HSPICE Illegal Continuation Illegal backslash included only character for requires a HSPICE RF y quoted strings whitespace before to use asa continuation pipe HSPICE v Included only Included only n a Include only for HSPICE RF 7 comma Illegal Illegal Illegal Token delimiter period Illegal Included only Included only Netlist keyword i e TRAN DC etc Hierarchy delimiter when used in node names colon Included only Included only Included only Delimiter for element attributes semi colon Included only Included only Included only n a double quotes _ Illegal Illegal Illegal Expression and filename delimiter 62 HSPICE RF User Guide Z 2007 03 Chapter 4 Input Netlist and Data Entry Input Netlist File Guidelines Table 4 HSPICE HSPICE RF Netlists Net Name Special Characters Special Character Node Name Instance Name Parameter Name Delimiters Note character is cannot be the cannot be the first first character character element legal anywhere in the element key key letter only string first or included letter only single quotes _ Illegal I
192. arameter Extraction Describes how to do frequency translation and large signal S parameter extraction as well as noise parameter calculation This chapter focuses on Frequency Translation S Parameter HBLIN Extraction on page 298 Large Signal S parameter HBLSP Analysis on page 305 This chapter discusses various techniques supported in HSPICE RF for extracting circuit scattering parameters Since RF circuits can operate under large signal and small signal conditions there are several types of scattering parameters that are useful to measure Linear small signal scattering parameters represent the RF frequency domain transfer characteristics for a circuit that is operating at its DC bias condition but the stimulus and response signals are sufficiently small that they do not influence the operating point This type of analysis is performed using the LIN analysis which is supported in both HSPICE and HSPICE RF For information on doing small signal S parameter analysis LIN please see Chapter 11 Linear Network Parameter Analysis in the HSPICE Simulation and Analysis User Guide In the case of RF mixers and receiver front ends some of the input and output frequencies of interest involve a frequency translation This translation is intentional and caused by nonlinear mixing in the circuit due to devices being driven by large signal periodic waveforms This type of scattering parameter analysis therefore must begin by solving the l
193. arge signal periodic response and then finding the small signal behavior about this large signal operating point This capability is provided by the HBLIN analysis which has setup and analysis control options similar to LIN but is capable of extracting S parameters about a large signal periodic steady state operating point HSPICE RF User Guide 297 Z 2007 03 Chapter 11 S parameter Extraction Frequency Translation S Parameter HBLIN Extraction In the case of circuits such as power amplifiers the extraction of scattering parameters is also important but the circuit stimulus and response signals may themselves be large signal periodic waveforms And it can be important to analyze how these S parameter vary as a function of input power levels This capability is provided by the HBLSP Large Signal S parameter analysis which uses large signal stimulus signals for the S parameter extractions Frequency Translation S Parameter HBLIN Extraction Frequency translation scattering parameter S parameter extraction is used to describe N port circuits that exhibit frequency translation effects such as mixers The analysis is similar to the existing LIN analysis except that the circuit is first linearized about a periodically varying operating point instead of a simple DC operating point After the linearization the S parameters between circuit ports that convert signals from one frequency band to another are calculated You use the
194. armonic Balance Noise HBNOISE The modulated noise source thermal shot or flicker is modeled as a cyclostationary noise source A PAC algorithm solves the modulated transfer function You can also use the HBNOISE PAC method with correlated noise sources including the MOSFET level 9 and level 11 models and the behavioral noise source in the G Element Voltage Dependent Current Source You use the HBNOISE statement to perform a Periodic Noise Analysis Supported Features HBNOISE supports the following features All existing HSPICE RF noise models Uses more than one single tone harmonic balance to generate the steady state solution Unlimited number of HB sources using the same tone possibly multiple harmonics Includes stationary cyclostationary frequency dependent and correlated noise effects Swept parameter analysis Results are independent of the number of HBAC sources in the netlist Prerequisites and Limitations The following prerequisites and limitations apply to HBNOISE Requires one HB statement which determines the steady state solution Requires at least one HB source or one TRANFORHEB source Requires placing the parameter sweep in the HB statement The requested maximum harmonic in HBNOISE must be less than or equal to half the number of harmonics used in harmonic balance that is max_harm lt num_hb_harms 2 Input Syntax HBNOISE output insrc parameter sweep lt nl1
195. art and Stop Times In addition to using FROM and TO times in a POWER statement you can also use the SIM_POWERSTART and SIM _POWERSTOP options with POWER statements to specify default start and stop times for measuring signals during simulation These times apply to all signals that do not have their own defined FROM and TO measurement times For example OPTION SIM POWERSTART lt time gt OPTION SIM POWERSTOP lt time gt These options control the power measurement scope the default is for the entire run For syntax and description of these options see OPTION SIM_POWERSTART or OPTION SIM_POWERSTOP in the HSPICE and HSPICE RF Command Reference Controlling Power Analysis Waveform Dumps You use the SIM_POWERPOST option to control power analysis waveform dumping For example OPTION SIM POWERPOST ON OFF Considering the potentially enormous number of signals there is no waveform dumping by default for the signals in the POWER statement Setting SIM_POWERPOST ON turns on power analysis waveform dumping Detecting and Reporting Surge Currents The SURGE statement in HSPICE RF automatically detects and reports a current surge that exceeds the specified surge tolerance For example 418 HSPICE RF User Guide Z 2007 03 HSPICE RF User Guide Z 2007 03 Chapter 16 Advanced Features Detecting and Reporting Surge Currents SURGE surge threshold surge width nodel lt node2 noden gt This statement repor
196. at option off Each ALTER simulation run prints only the actual altered input A special ALTER title identifies the run ALTER processing cannot revise LIB statements within a file that an INCLUDE statement calls However ALTER processing can accept INCLUDE statements within a file thata LIB statement calls Using Multiple ALTER Blocks The following is the process for using multiple ALTER blocks For the first simulation run HSPICE RF reads the input file up to the first ALTER statement and performs the analyses up to that ALTER statement After it completes the first simulation HSPICE RF reads the input between the first ALTER statement and either the next ALTER statement or the END statement HSPICE RF then uses these statements to modify the input netlist file HSPICE RF then resimulates the circuit For each additional ALTER statement HSPICE RF performs the simulation that precedes the first ALTER statement HSPICE RF then performs another simulation using the input between the current ALTER statement and either the next ALTER statement or the END statement HSPICE RF User Guide 83 Z 2007 03 Chapter 4 Input Netlist and Data Entry Input Netlist File Composition 84 If you do not want to rerun the simulation that precedes the first ALTER statement every time you run an ALTER simulation then do the following 1 2 3 Put the statements that precede the first
197. ate gnd 0 65 PHASENOISE V gate gnd DEC 10 100 1 0e7 METHOD 0 CARRIERINDEX 1 Suse NLP algorithm This example performs an oscillator analysis searching for frequencies in the vicinity of 900 MHz followed by a phase noise analysis at frequency offsets from 100 Hz to 10 MHz Example 2 HBOSC TONE 2400MEG NHARMS 11 PROBENODE drainP drainN 1 0 FSPTS 20 2100MEG 2700MEG SWEEP Vtune 0 0 5 0 0 2 PHASENOISE V drainP drainN DEC 10 100 1 0e7 METHOD 0 CARRIERINDEX 1 Suse NLP algorithm This example performs a VCO analysis searching for frequencies in the vicinity of 2 4 GHz This example uses eleven harmonics and sweeps the VCO tuning voltage from 0 to 5 V HSPICE RF uses the nonlinear perturbation NLP algorithm to perform a phase noise analysis about the fundamental frequency for each tuning voltage value Frequency Dependent and Frequency Independent Sources The phnoise_fdep keyword variable will collect all frequency dependent noise sources contributions to the phase noise The phnoise findep keyword variable will collect all frequency independent noise sources contributions print phasenoise phnoise fdep print phasenoise phnoise findep Frequency and Bias Dependencies The four keywords below are obvious in their names 1 The following syntax is frequency independent and bias dependent _print phasenoise phnoise cyclo Also acceptable is print phasenoise phnoise cyclostationary Where HSPICE RF User
198. between zero and one Because the reflection coefficient is easy to measure in real microwave circuits the S parameter can be a very useful tool for microwave engineers HSPICE RF User Guide 147 Z 2007 03 Chapter 6 Testbench Elements Multi Terminal Linear Elements You can use the S element with a MODEL SP or with data files that describe the frequency response of a network and provide discrete frequency dependent data Touchstone and CITIfile You can measure this data directly using network analyzers such as Hewlett Packard s MDS Microwave Design System or HFSS High Frequency Structure Simulator HSPICE can also extract the S element from a real circuit system For a description of the S parameter and SP analyses see S parameter Model in the HSPICE Signal Integrity Guide S element Syntax The nodes of the S element must come first If MNAME is not declared you must specify the FOMODEL You can specify all the optional parameters in both the S element and S model statements except for MNAME argument You can enter the optional arguments in any order and the parameters specified in the element statement have a higher priority If the number of nodes in the element card is smaller than the number specified in the model card or external file by 1 then the reference node is the default The default reference node is 0 gnd Figure 18 Terminal Node Notation N 1 terminal system vinc N vine
199. butions on parameters and applying them to model parameters can be used on some models and the DEV LOT approach on others in the same simulation m test13 sp has DEV LOT specified for model res1 and the parameter width for model res2 The values for the resistors with model res1 are different and the values for resistors with model res2 are the same m test14 sp is similar to test7 sp and has modmonte 1 specified All four resistors have different values However note that in reality the sigma for width would be different when simulating local or global variations test15 sp has instance parameter variations specified on two resistors and DEV LOT on two others From the waveforms v3 and v4 form a first pair and v1 and v2 a second pair HSPICE RF User Guide Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Simulating the Effects of Global and Local Variations with Monte Carlo It is also possible to mix variations on instance parameters and model parameters in the same setup test16 sp has small instance parameter variations specified on width and relatively large model parameter variations on the sheet resistivity rsh The results show four different waveforms with a common behavior test17 sp shows instance and model parameter variations as in the previous test case but option modmonte is set to 1 thus the model variations affect every device in a different way The results show completely independent b
200. ce Noise HBNOISE 274 Note MEASURE HBNOISE cannot contain an expression that uses an hbnoise variable as an argument You also cannot use MEASURE HBNOISE for error measurement and expression evaluation of HBNOISE The HSPICE RF optimization flow can read the measured data from a MEASURE HBNOISE analysis This flow can be combined in the HSPICE RF optimization routine with a MEASURE HBTR analysis see Using MEASURE with HB Analyses on page 212 and a MEASURE PHASENOISE analysis see Measuring PHASENOISE Analyses with MEASURE on page 243 Errors and Warnings HBNOISE Errors See the list of HBAC Errors and Warnings on page 262 HBNOISE Example This example performs an HB analysis then runs an HBNOISE analysis over a range of frequencies from 9 0e8 to 9 2e8 Hz Simulation outputs the output noise at V out and the single side band noise figure versus IFB from 1e8 to 1 2e8 Hz to the pn0 file The netlist for this example is shown immediately following Ideal mixer noise source prints total noise at the output 1 57156p V sqrt Hz Single sideband noise figure 3 01 dB double sideband noise figure 0 dB OPTION PROBE OPTION POST 2 vlo lo 0 0 0 hb 1 0 01 1S Periodic HB Input Ilo lo 0 0 rsrc rfin rfl 1 0 Noise source cl 0 if g 1 0e 9 v lo v rfin mixer element gl 0 if cur 1 0 v lo v rfin S mixer element rout if 0 1 0 vrf rfl 0 hbac 2 0 0 0 hb tones 1 0g nharms 4 swe
201. ced Features Using CHECK Statements Analysis statement HBOSC TONES lt fi gt lt f2 gt lt fn gt lt NHARMS lt hi1 gt lt h2 gt lt hn gt gt SWEEP parameter sweep OPTIMIZE OPTxxx RESULT measname MODEL mname Measure statement MEASURE HB measname FIND out_varl AT val GOAL val Optimization With HBNOISE PHASENOISE or HBTRAN Measurements The required statements are Analysis statement HBOSC TONES lt fil gt lt f2 gt lt fn gt lt NHARMS lt hi1 gt lt h2 gt lt hn gt gt SWEEP OPTIMIZE OPTxxx RESULT measname MODEL mname For example HBOSC tones 1lg nharms 5 sweep x 1 5 1 optimize optl result yl y2 model ml model ml opt level 0 PHASENOISE dec 1 1k 1g meas phasenoise yl find phnoise at 10k goal 150dbc meas phasenoise y2 RMSJITTER phnoise units sec goal 1 0e 12 Measure statement MEASURE HBNOISE measname FIND out_vari AT val GOAL val MEASURE PHASENOISE measname FIND out_varl AT val GOAL val MEASURE HBTRAN measname FIND out_vari1 AT val GOAL val Optimization with HBNOISE PHASENOISE or HBTRAN measurements must not be used in combination with HB measurement optimization as shown in Optimization With HB Measurements Using CHECK Statements The CHECK statements in HSPICE RF offer the following instrumentation Setting Global Hi Lo Levels Slew Rise and Fall Conditions Edge Timing Verification HSPICE RF User Guide 411 Z 2007 03
202. ch Elements Steady State HB Sources HB TONES 1900MEG 1910MEG INTMODMAX 5 This specifies two fundamental frequencies f rone 1 1 9GHz and fltone 2 1 91GHz Their mixing product at 10 MHz can then be referenced using indices as f 2 11 while their 3rd order intermodulation product at 1 89 GHz can be referenced as 2 11 121l Steady state voltage and current sources are identified with the HB keyword according to lt HB lt mag lt phase lt harm lt tone lt modharm lt modtone gt gt gt gt gt gt gt The source is mathematically equivalent to a cosine signal source that follows the equation Acos a 0 where A mag 2n harm f tone modharm f modtone 730 phase Values for tone and modtone an optional modulating tone must be non negative integers that specify index values for the frequencies specified with the HB TONES command Values for harm harmonic and modharm modulating tone harmonic must be integers negative values are OK that specify harmonic indices Example 1 The following example is a 1 0 Volt peak steady state cosine voltage source which is at the fundamental HB frequency with zero phase and with a zero volt DC value Vsre in gnd DC 0 HB 1 0 0 1 1 Example 2 The following example is a steady state cosine power source with 1 0mW available power which is implemented with a Norton equivalent circuit and a 50 ohm input impedance HSPICE RF User Guide 173 Z 2007
203. ch output file is the same as the input netlist file s base name The at the end of each file extension represents the ALTER run from which the file came Chapter 2 Getting Started Generating Output Files In general text output from PRINT commands is intended to be read by humans while binary output from PROBE or OPTION POST is intended to be read by the CosmosScope waveform display tool Table 3 HSPICE RF Output File Types Command Text Output Output for CosmosScope AC analysis AC printac ac AC noise analysis NOISE printac ac DC sweep DC printsw SW Envelope analysis ENV printev ev Envelope FFT ENVFFT none fe Harmonic Balance HB printhb hb Harmonic Balance AC printhb hb HBAC HBLIN analysis HBLSP large signal HBLSP small signal HBAC noise HBNOISE Harmonic Balance OSC HBOSC HSPICE RF User Guide PRINT output printhl S param output SnP PRINT output printls S param output p2d PRINT output printss S noise output S2P printsnpn printhb PROBE output hl S paramr output SnP PROBE output ls S param output p2d PROBE output ss S noise output S2P pn hb Chapter 2 Getting Started Using the CosmosScope Waveform Display Table 3 HSPICE RF Output File Types Command Text Output Output for CosmosScope Harmonic Balance TRAN printhr hr HBTRAN Transfer Function
204. circuit Therefore it is possible to calculate the value of a model parameter within the subcircuit for example as a function of geometry information When specifying variations on these parameters the effects of local variations between subcircuits are created If this method is used at the extreme with one device per subcircuit then each device has its own model This approach leads to a substantial overhead in the simulator and is therefore not recommended Indirect Variations on a Model Parameter In sections Variations Specified on Geometrical Instance Parameters and Variations Specified in the Context of Subcircuits variations on geometrical parameters were presented If we want to specify variations on a model parameter for example the threshold of a MOS device then the approach explained in the previous section with one model per device in a subcircuit could be used However this is impractical because the netlist needs to be created to call each device as a subcircuit and because of the overhead Since variations are of interest only on a few model parameters an indirect method of varying model parameters can be used Some special instance parameters are available for this purpose For example for MOS devices the parameter delvtO defines a shift in threshold Referencing a parameter with a distribution as value for delvto creates the effect of local threshold variations A significant number of parameters of this type are ava
205. circuit Library Structure Parameters and Functions Using Parameters in Simulation PARAM 00000 0 eee euee Defining Parameters Assigning Parameters User Defined Function Parameters Predefined Analysis Function Measurement Parameters PRINT and PROBE Parameters Using Algebraic Expressions Built In Functions and Variables Parameter Scoping and Passing Library Integrity Reusing Cells Creating Parameters in a Library Parameter Defaults and Inheritance Parameter Passing Solutions Testbench Elements Passive Elements 5 ReSISIONS ona oe eee aad d enous Capacitors 000 e eee Inductors 0 0 00 e eee Multi Terminal Linear Elements W element Distributed Transmission Lines 006 T element Ideal Transmission Lines 82 85 86 86 87 88 91 92 93 95 95 95 97 98 99 99 99 99 100 104 105 106 106 108 111 113 113 114 119 125 139 139 143 Contents vi Scattering Paramete Port Element Port Element Syntax r Data Element s raora annn Rev eek et Using the Port Element for Mixed Mode Measurement Active Elements Diode Element Bipolar Junction Transistor BUT Element 200005 JFETS and MESFEMS nice
206. cosine Nyquist filter RECT rectangular filtering FILCOEF Filter parameter for the COS filter O lt filpar lt 1 RATE Bit rate for modulation bits per second For BPSK modulation the data rate and the symbol rate are the same For QPSK modulation the symbol rate is half the data rate The Rb value must be greater than zero HSPICE RF User Guide 183 Z 2007 03 Chapter 6 Testbench Elements Complex Signal Sources and Stimuli Parameter Description BITSTREAM A binary b or hexadecimal h string that represents an input data stream Valid data string characters are 0 or 1 for binary b mode 0 1 2 3 4 5 6 7 8 9 A B C D E F a b c d e or f for hexadecimal h mode For example 0101001 1b binary OF647A30EQ9h hexadecimal You can also use the standard V source and source options for non transient simulations Such aS DC val and AC mag ph a with the VMRF source Example BITSTREAM 01010010011100b data 1 dr BPSK I and Q Signals 707 1 dr 184 HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Complex Signal Sources and Stimuli QPSK Signal 707 QPSK Q Signal 707 1 dr The Rb parameter represents the data rate The associated symbol rate represents how fast the and Q data streams change The period for each bit of data is Equation 18 T E b The symbol rate depends on whether you select BPSK or QPSK modulation
207. cs are periodic in time Cyclostationary noise can be characterized in several ways with the particular application determining which is appropriate 9 The time average power spectral density PSD ignores frequency correlations in the noise but is adequate when the fundamental frequency of the cyclostationary noise is much larger than the bandwidth of interest The time average PSD is calculated in the HBNOISE SNNOISE analyses 10 The harmonic power spectral density HPSD or equivalently the auto correlation function R t1 t2 contains the correlation information between HSPICE RF User Guide 281 Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Periodic Time Dependent Noise Analysis PTDNOISE noise sidebands that is necessary to build behavioral cyclostationary noise sources and to separate the amplitude modulation AM and phase modulation PM noise components The time dependent power spectral density TDPSD can be integrated over frequency to yield the time dependent noise TDN TDN can then be used to determine jitter associated with a noisy signal crossing a threshold PTDNOISE analysis allows the calculation of TDPSD TDN and jitter By measuring the jitter associated with a noisy signal crossing a threshold jitter is modeled by displacing the time in a noise free signal v t with a stochastic process j t Equation 56 Vji t vt j We can also determine the voltage at thi
208. culates the DC voltage across the circuit s nodes In all other non DC analyses a DC voltage source of this value represents the DC block that is HSPICE RF does not then allow dv dt variations The following input syntax is for the Choke ideal infinite inductor Syntax Lxxx nodel node2 lt L gt INFINITY lt IC val gt HSPICE RF does not support any other inductor parameters because HSPICE RF assumes that the infinite inductance value is independent of temperature and scaling factors The choke acts as a short circuit for all DC analyses HSPICE RF calculates the DC current through the inductor In all other non HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Multi Terminal Linear Elements DC analyses a DC current source of this value represents the choke that is HSPICE RF does not then allow di dt variations Multi Terminal Linear Elements A multi terminal linear element such as a transmission line is a passive element that connects any two conductors at any distance apart One conductor sends the input signal through the transmission line and the other conductor receives the output signal from the transmission line The signal is voltage between the conductors that is transmitted from one end of the pair to the other end Examples of transmission lines include Power transmission lines Telephone lines Waveguides Traces on printed circuit boards and multi chip modules MCMs Bon
209. culation Errors Warnings Error handling and recovery is exercised to capture obvious errors in input specifications The following error checking is performed Calculations are be performed if oscillator or phase noise analysis fails ERROR if L f gt 1 over any part of the frequency sweep non dB form ERROR if L f lt 0 over any part of the frequency sweep non dB form Error if any time or frequency samples are negative values ERROR if BER lt 0 for any Jitter measurement WARNING if BER gt 1 for any Jitter measurement WARNING if fO lt 10 Hz Message Jitter calculations may be ineffective for offset frequencies under 10 Hz HSPICE RF User Guide Z 2007 03 Chapter 9 Oscillator and Phase Noise Analysis Jitter Analysis RMS JITTER Measurement Based on the phase noise data the syntax of the RMS JITTER measurement is provided where word is in units of sec Seconds rad radiens or UI Unit Interval The default is sec MEASURE phasenoise integralOutMag RMSJITTER phnoise lt FROM start _frequency gt lt TO end _frequency gt lt UNITS word gt Example meas phasenoise rj RMSJITTER phnoise from 1K to 100K units rad The RMSJITTER is calculated as endfrequency 1l ph j Equation 52 msl 2 0 10 0 phasenoise k startfrequency With sec units the RMSUITTER is calculated as rms 1 20 m fO in which PI 3 1415926 and f0 is the tone frequency of the oscillator E
210. curate results is to run both algorithms over a broad frequency range and check that the curves have some range in frequency where they overlap Typically you will see the NLP curve rolling off at 20 to 30 dB decade as frequency increases characteristic of white noise or 1 f noise behavior Also the PAC curve will at first be flat or even noisy close to the carrier At some point though you will see this curve match the NLP roll off The lowest frequency at which the curves overlap defines the point fpac above which the PAC algorithm is valid Sometimes by increasing the number of HB harmonics it is possible to move fpac to lower frequencies The highest frequency at which the curves overlap defines the point fy p below which the NLP algorithm is valid A rough rule of thumb is that fpac f p Q where f is the carrier frequency and Q is the oscillator Q value This implies that for high Q oscillators such as crystal and some harmonic oscillators that PAC will be accurate to values quite close to the carrier Broadband Phasenoise Algorithm The broadband phasenoise BPN algorithm has been added to HSPICE RF to allow phasenoise simulation over a broad frequency range The BPN algorithm actually runs both the NLP and PAC algorithms and then connects them in the overlap region to generate a single phasenoise curve This algorithm is ideal for verifying the NLP and PAC accuracy regions and when you require a phasenoise curve over a broad fre
211. current Because the DDL library devices are based on HSPICE circuit level models simulation automatically compensates for the effects of supply voltage loading and temperature HSPICE or HSPICE RF accesses DDL models in several ways The installation script creates an hspice ini initialization file HSPICE or HSPICE RF writes the search path for the DDL and vendor libraries into a OPTION SEARCH lt lib path gt statement This provides immediate access to all libraries for all users It also automatically includes the models in the input netlist If the input netlist references a model or subcircuit HSPICE or HSPICE RF searches the directory to which the DDLPATH environment variable points for a file with the same name as the reference name This file is an include file so its filename suffix is inc HSPICE installation sets the DDLPATH variable in the meta cfg configuration file Set OPTION SEARCH lt lib path gt in the input netlist Use this method to list the personal libraries to search HSPICE first searches the default libraries referenced in the hspice ini file then searches libraries in the order listed in the input file Directly include a specific model using the INCLUDE statement For example to use a model named T2N2211 store the model in a file named T2N2211 inc and put the following statement in the input file HSPICE RF User Guide 91 Z 2007 03 Chapter 4 Input Netlist and Data Entry
212. d instruct HSPICE RF to output phase noise results to the osc pnO and osc printpno files k Uses emitter resistor limiting to keep output sinusoidal Output can be taken at the emitter eml node kk Options for Oscillator Harmonic Balance Analysis k OPTIONS post sim_accuracy 100 hbsolver 0 Bias NPN transistor for 5V Vce 10mA Ic Emitter follower Colpitts design Vec collector 0 9V Q1 collector base emitter emitter RF_WB_NPN Rel emitter eml 100 RLoad eml 0 300 Rb1 collector base 4300 Rb2 base 0 5600 Capacitive feedback network Ce 0 eml 100pF Cfb base eml 100pF Cbb base bb 470pF Lb bb 0 6uH SN a E OEN Pe re E E EE AE E E E EE EN a EA HSPICE RF User Guide 29 Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 4 Using HBOSC Analysis for a Colpitts Oscillator Simulation control for automated oscillator analysis k HBOSC tones 1 0e7 nharms 15 PROBENODE emitter 0 4 27 FSPTS 40 9 e6 1 1e7 PHASENOISE V emitter DEC 10 10K 1MEG METHOD 0 CARRIERINDEX 1 print hbosc vm eml vp eml vr emitter vi emitter print hbosc vm emitter vp emitter P Rload print phasenoise phnoise probe phasenoise phnoise probe hbosc v emitter v eml1 include bjt ine END After you run this netlist examine the osc printhbo file At the top is the oscillator frequency about 10 14 MHz and the PRINT HBOSC output The first 2 lines show that the eml node oscillates aroun
213. d 3V with an amplitude of about 2 85V The emitter node oscillates around 4V with an amplitude of about 4 27V Also examine the osc printpn0 file which contains the phase noise results in text form You can view the osc hbO and osc pn0 files in CosmosScope 1 To start CosmosScope type cscope 2 Use the File gt Open gt Plotfiles dialog to open osc hb0 Remember to set the file type filter to HSPICE RF HB hb 3 From the signal manager double click on v emitter to see that node s spectrum 4 Right click on the v emitter label in the chart and choose To Time Domain to create a time domain waveform 5 To accept the defaults for range and interval click OK You should see an oscillating time domain waveform To run a transient simulation for comparison 1 Usethe TRAN 1n 10u command 30 HSPICE RF User Guide Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 5 Using HBOSC Analysis fora CMOS GPS VCO 2 Add ic 10n to the Lb inductor The resulting waveforms should be the same as those from HB oscillator analysis Example 5 Using HBOSC Analysis for a CMOS GPS VCO This second oscillator analysis example involves two negative resistance oscillators coupled at 90 degrees MOS capacitors are used as varactors This VCO topology is common for GPS applications and produces quadrature LO outputs near 1550 MHz The purpose of this example is to generate the VCO tuning curve output level and frequency as
214. d Passing 110 Each of the three resistors obtains its simulation time resistance from the Val parameter The netlist defines the Va parameter in four places with three different values Figure 16 Hierarchical Parameter Passing Problem TEST OF PARHIER OPTION list node post 2 ingold 2 Sub1 Sub2 Sub3 parhier lt Local Global gt PARAM Val 1 xl n0 0 Subl SubCkt Sub1 nl n2 Val 1 ri ni n2 Val Pi ri r2 r3 x2 nl n2 Sub2 Ends Sub1 SubCkt Sub2 nl n2 Val 2 1V r2 nl n2 Val x3 nl n2 Sub3 Ends Sub2 SubCkt Sub3 nl n2 Val 3 r3 nl n2 Val Ends Sub3 OP END The total resistance of the chain has two possible solutions 0 3333Q and 0 54550 You can use OPTION PARHIER to specify which parameter value prevails when you define parameters with the same name at different levels of the design hierarchy Under global scoping rules if names conflict the top level assignment PARAM Val 1 overrides the subcircuit defaults and the total is 0 33339 Under local scoping rules the lower level assignments prevail and the total is 0 54550 one two and three ohms in parallel The example in Figure 16 produces the results in Table 15 based on how you set OPTION PARHIER to local global Table 15 PARHIER LOCAL vs PARHIER GLOBAL Results Element PARHIER Local PARHIER Global ri 1 0 1 0 HSPICE RF User Guide Z 2007 03 Chapter 5 Parameters and Functions Parameter Scoping and
215. d continue the evaluation of the next instance SIM_SPEF_INSERROR The default is OFF ON skips unmatched instances and continues the evaluation SIM_SPEF_PARVALUE This option affects only values in a SPEF file that have triplet format float float float which this option interprets as best average worst In such cases f SIM_SPEF_PARVALUE 1 HSPICE RF uses best If SIM_SPEF_PARVALUE 2 default HSPICE RF uses average f SIM_SPEF_PARVALUE 3 HSPICE RF uses worst Unsupported SPEF Options HSPICE RF does not yet support the following IEEE 481 SPEF options Hierarchical SPEF definition multiple SPEF files connected with a hierarchical definition DEFINE and PDEFINE R_NET and R_PNET definition D_ PNET definition HSPICE RF User Guide 325 Z 2007 03 Chapter 13 Post Layout Analysis Post Layout Back Annotation Selective Extraction Flow Use the selective extraction flow if disk space is limited Especially use this option when simulating a full chip post layout design where block latency is high HSPICE RF feedbacks the active net information to Star RCXT to extract only the active parasitic The major advantage of this flow is a smaller DSPF or SPEF file which saves disk space Figure 25 Selective Extraction Flow Star RCXT vy y DSPF SPEF Ideal Netlist v OR Post Layout Flow HSPICE RF y Active Nodes Star
216. d to the given node The transfer function is computed on the output variables and input current or voltage NODES ELEM NODES or ELEM can be one of the following Voltage type a single node name n1 ora pair of node names n1 n2 Current type an element name elemname Power type a resistor resistorname or port portname element name Output Data Files An HBXF calculation produces these output data files Output from the PRINT statement is written to a printxf file e The output is in ohms siemens or undesignated units and the header in the output file is Z Y or GAIN Output from the PROBE statement is written to a xf file Reported performance log statistics are written to a lis file e HBXF CPU time e HBXF peak memory usage Example Based on the HB analysis the following example computes the trans impedance from isrc to v 1 hb tones 1le9 nharms 4 hbxf v 1 lin 10 1e8 1 2e8 print hbxf tfv isrce tfi n3 HSPICE RF User Guide Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Shooting Newton Transfer Function Analysis SNXF HBXF Test Listing Test HBXF nonlinear order 2 poly equation OPTIONS PROBE OPTIONS POST 2 vlo lo 0 cos 0 1 0 1g 0 0 tranforhb 1 rlo lo 0 50 vrf1l rfl 0 0 rrfl rfl 0 50 El out 0 POLY 2 lo O rfl O 0 1 1 1 10 1 rout out 0 50 hb tones 1g nharms 5 hbxf v out lin 2 100meg 20
217. data X A points that are ON the line plotting 3 segments But NW and WDB both eliminate these data points WDB eliminates these data which are within DELTAV or DELTAI of the previous points plotting only ONE data point and are not ON the plotted waveform line segment for this line For a additional information see OPTION SIM_DELTAV in the HSPICE and HSPICE RF Command Reference Eliminating Current Datapoints You use the SIM_DELTAI option to determine the selection criteria for HSPICE RF current waveforms in WDB or NW format For example HSPICE RF User Guide Z 2007 03 Chapter 15 Using HSPICE with HSPICE RF Compressing Analog Files OPTION SIM DELTAI lt value gt For a additional information see OPTION SIM_DELTAI in the HSPICE and RF Command Reference HSPICE RF User Guide 397 Z 2007 03 Chapter 15 Using HSPICE with HSPICE RF Compressing Analog Files 398 HSPICE RF User Guide Z 2007 03 16 Advanced Features Describes how to invoke HSPICE RF and how to perform advanced tasks including redirecting input and output HSPICE RF accepts a netlist file from stdin and delivers the ASCII text simulation results to an HTML file or to stdout Error and warning messages are forwarded to standard error output This chapter describes how to do this as well as how to invoke HSPICE RF and redirect input and output Creating a Configuration File HSPICE Z 2007 03 You can create a configuration file called hs
218. defined PARAM lt mcVar gt Agauss 1 0 0 1 analysis function MEASURE MEASURE lt DC AC TRAN gt result TRIG statement TARG lt GOAL val gt lt MINVAL val gt lt WEIGHT val gt lt MeasType gt lt MeasParam gt PRINT PRINT PROBE PROBE outParam Par_Expression A parameter definition in HSPICE always uses the last value found in the input netlist Subject to local versus global parameter rules The definitions below assign a value of 3 to the DupParam parameter PARAM DupParam 1 PARAM DupParam 3 HSPICE assigns 3 as the value for all instances of DupParan including instances that are earlier in the input than the PARAM DupParam 3 statement 96 HSPICE RF User Guide Z 2007 03 Chapter 5 Parameters and Functions Using Parameters in Simulation PARAM All parameter values in HSPICE are IEEE double floating point numbers The parameter resolution order is 1 Resolve all literal assignments 2 Resolve all expressions 3 Resolve all function calls Table 11 shows the parameter passing order Table 11 Parameter Passing Order OPTION PARHIER GLOBAL OPTION PARHIER LOCAL Analysis sweep parameters Analysis sweep parameters PARAM statement library SUBCKT call instance SUBCKT call instance SUBCKT definition symbol SUBCKT definition symbol PARAM statement library Assigning Parameters You can assign the following types of values to parameters Con
219. del optmod1 e Selecting an optimization model model optmodl1 opt level 1 Bisection method itropt 40 relin le 4 relout le 6 accuracy settings e Measurement statements to tune the optimization parameters measure HB vif find vdb if 1 1 at 10e 6 measure HB vrf find vdb rf 0 1 at 10e 6 measure HB gain param vif vrf goal 2 HSPICE RF User Guide Z 2007 03 Chapter 16 Advanced Features Optimization e Measurement statement to find the fundamental frequency from HB analysis measure HB frequency max FIND HERTZ 1 at 0 Optimizing AC DC and TRAN Analyses The HSPICE syntax is followed for optimizing AC DC and TRAN analyses The required statements are m Optimization PARAM statement PARAM lt ParamName gt OPTXxx lt Init gt lt LoLim gt lt HiLim gt Optimizing TRAN statement TRAN tincrl tstop1 lt tincr2 tstop2 tincrN tstopN gt SWEEP OPTIMIZE OPTxxx RESTULTS measname MODEL optmod Optimizing MODEL statement MODEL mname OPT LEVEL 0 1 Where e Ospecifies the Modified Levenberg Marquardt method You would use this setting with multiple optimization parameters and goals e 1 specifies the Bisection method You would use this setting with one optimization parameter Optimizing HB Analysis There are two types of optimizations with HB analyses Optimization with only HB measurements m Optimization with HBNOISE PHASENOISE or HBTRAN measurements Optimization
220. dependence given according to the following equation v x 1 Re Ad Mypa Bx va 1 ref B phere Zo Zo The A represents the incident voltage B represents the reflected voltage Zo is the characteristic impedance and is the propagation constant The latter are related to the transmission line inductance L and capacitance C by the following equation B a LC The L and C terms are in per unit length units Henries meter Farads meter The following equation gives the phase velocity HSPICE RF User Guide 145 Z 2007 03 Chapter 6 Testbench Elements Multi Terminal Linear Elements 146 1 D p B JLC At the end of the transmission line x the propagation term B becomes the following equation Bre aC ie om p This is equivalent to an ideal delay with the following value Pe Ses Vp Where T absolute time delay sec 1 physical length L meters Vp phase velocity meters sec Using standard distance velocity time relationships the HSPICE T element parameter values are related to these terms according to Vp f 7 Where f frequency gt wavelength 4 relative time delay TD sec meter Where 1 physical length L meters a normalized length NL f frequency at NL F Hz TD L A2 JEC HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Multi Terminal Linear Elements HSPICE therefore allows you to specify a tra
221. dependent variable between the MEASURE HB and MEASURE HBTR statements these two types of measurements cannot be mixed in a HSPICE RF optimization Buta MEASURE HBTR statement can be combined with a MEASURE PHASENOISE statement see Measuring PHASENOISE Analyses with MEASURE on page 243 and a MEASURE HBNOISE statement see Measuring HBNOISE Analyses with MEASURE on page 273 in a HSPICE RF optimization flow HB Output Data Files The results of an HB analysis are complex spectral components at each frequency point The ali is the real part and b i is the imaginary part of the 214 HSPICE RF User Guide Z 2007 03 Chapter 7 Steady State Harmonic Balance Analysis HB Output Data Files complex voltage at frequency index i The conversion to a steady state time domain is then given by the Fourier series expansion An HB analysis produces these output data files Output from the PRINT HB statement is written to a printhb file e The header contains the large signal fundamental frequencies e The columns of data are labeled as HERTZ followed by frequency indices and then the output variable names e The sum of the frequency indices multiplied by the corresponding fundamental frequencies add up to the frequency in the first column Output from the PROBE HB statement is written to a hb file It is in the same format as the HSPICE transient analysis tr file Besides the output waveform it contains the information o
222. descriptions of V I sources and to use only HB descriptions To override this option specify TRANFORHEB in the source description 204 HSPICE RF User Guide Z 2007 03 Chapter 7 Steady State Harmonic Balance Analysis Harmonic Balance Analysis Harmonic Balance Output Measurements This section explains the harmonic balance output measurements you receive after HSPICE runs an HB simulation Harmonic Balance Signal Representation The HB cosine sources can be interpreted in real imaginary and polar formats according to v t Acos ort o Repet Rede Rede cos of jsin or Re Vp jV cos at jsin or Equation 26 Vpcos at V sin at Acos cos at A sin sin a Note that real imaginary and polar formats are related with the standard convention Equation 27 VatiV Ae Vp Acos o V Asin 0 A Vp V Vi tand Vr The result of HB analysis is a complex voltage current spectrum at each circuit node or specified branch Let ali be the real part and b i be the imaginary part of the complex voltage at the ith frequency index Conversion to a steady state time domain waveform is given by the Fourier series expansion HSPICE RF User Guide 205 Z 2007 03 Chapter 7 Steady State Harmonic Balance Analysis Harmonic Balance Analysis Equation 28 v t a O a 1 cos 2nf1 t b 1 sin 2nf1 8 a 2 cos 2nf2 f b 2 sin 2nf2 t a 3 cos 2rf3 b 3
223. ding wires in semiconductor IC packages m On chip interconnections The following sections discuss W element Distributed Transmission Lines T element Ideal Transmission Lines Scattering Parameter Data Element W element Distributed Transmission Lines The W element supports five different formats to specify the transmission line properties Model 1 RLGC Model specification e Internally specified in a model statement e Externally specified in a different file Model 2 U Model specification e RLGC input for up to five coupled conductors e Geometric input planer coax twin lead HSPICE RF User Guide 139 Z 2007 03 Chapter 6 Testbench Elements Multi Terminal Linear Elements 140 e Measured parameter input e Skin effect Model 3 Built in field solver model Model 4 Frequency dependent tabular model Model 5 S parameter Model W element Statement The general syntax for a lossy W element transmission line element is RLGC file form Wxxx inl lt in2 lt inx gt gt refin outi lt out2 lt outx gt gt refout lt RLGCfile filename gt N val L val U Model form Wxxx inl lt in2 lt inx gt gt refin outi lt out2 lt outx gt gt refout lt Umodel modelname gt N val L val Field solver form Wxxx inl lt in2 lt inx gt gt refin outi lt out2 lt outx gt gt refout lt FSmodel modelname gt N val L val The number of ports on a single transmission line are not li
224. ditions that augment the ultimate accuracy analog circuit simulation capabilities of HSPICE Note This manual describes the additional features and capabilities of HSPICE RF Where necessary the manual describes differences between HSPICE RF and HSPICE For information about standard HSPICE device models syntax and simulation control you can refer to one of the other HSPICE manuals in the HSPICE documentation set listed in The HSPICE Documentation Set on page xv HSPICE RF Overview HSPICE RF User Guide Z 2007 03 HSPICE RF consists of The hspicerf simulation engine The CosmosScope cscope waveform display tool The hspicerf simulation engine contains extensions to HSPICE for RF design These extensions are in the form of new analysis commands and new elements The hspicerf simulation engine processes command and element syntax for new RF simulation features but also accepts standard HSPICE netlist files as input Chapter 1 HSPICE RF Features and Functionality HSPICE RF Overview The CosmosScope waveform display tool has been enhanced with special features for reading and analyzing data created by the HSPICE RF simulation engine For a basic overview on how to use CosmosScope to view HSPICE RF output see Using the CosmosScope Waveform Display on page 12 HSPICE RF Features This section briefly introduces the features of both the simulation engine and the waveform display tool HSPICE RF supports most HSPICE ca
225. dy state Oscillator phase noise analysis including both a nonlinear perturbation method and a PAC method and includes stationary cyclostationary frequency dependent and correlated noise effects Frequency translation S parameter and noise figure extraction with the HBLIN command Envelope analysis The ENV command invokes standard envelope simulation The ENVOSC command invokes envelope startup simulation The ENVFFT command invokes envelope Fast Fourier Transform simulation OPTION HBTRANINIT HBTRANPTS and HBTRANSTEP for transient analysis of ring oscillators Convolution for transient analysis of S parameter data models S element Calculation of the transfer function from an arbitrary source and harmonic in the circuit to a designated output with the HBXF command Reading encrypted netlists HSPICE RF User Guide 3 Z 2007 03 Chapter 1 HSPICE RF Features and Functionality HSPICE RF Overview OPTION SIM ACCURACY provides simplified accuracy control for all simulations while OPTION SIM ORDER and SIM_TRAP improve transient analysis simulation controls DSPF Flow for fast analysis using parasitic data from layout OPTION SIM_LAprovides linear acceleration for RC network reduction for faster simulation Saving PRINT simulation output to a separate file HERTZ variable for frequency dependent equations IC OFF in element statements IC parameter initial conditions Shooting Newton steady state t
226. e dissipated power in resistors and delivered power to port elements The following subtle differences between these two measurements are described in this section Power Dissipated in a Resistor All power calculations make use of the fundamental phasor power relationship given as the following equation where voltage V and current are complex phasors given in peak values not rms nor peak to peak Equation 29 Pms sRe Vi In the case of a simple resistor its current and voltage are related according to V l R The power dissipated in a resistor of real value R at frequency index nis then given by Equation 30 P resistor n Power Delivered to a Port Element The port element can be either a source or sink for power You can use a special calculation that computes the power flowing into a port element even if the port element itself is the source of that power In the following figure is a HSPICE RF User Guide Z 2007 03 Chapter 7 Steady State Harmonic Balance Analysis Harmonic Balance Analysis port element connected to a circuit the port element may or may not include a voltage source Figure 21 Port Element Zo In AAA Port 7 Beas Vs Element n Circuit Let V be the peak voltage across the terminals of the port element at frequency index n Let be the peak current into the 1st terminal of the port element at frequency index n Le
227. e during transient analysis time temper current circuit control Uses parameters to define the current simulation temperature temperature during transient temperature analysis hertz current control Uses parameters to define the frequency during simulation AC analysis frequency Parameter Scoping and Passing If you use parameters to define values in sub circuits you need to create fewer similar cells to provide enough functionality in your library You can pass Circuit parameters into hierarchical designs and assign different values to the same parameter within individual cells when you run simulation 104 HSPICE RF User Guide Z 2007 03 Chapter 5 Parameters and Functions Parameter Scoping and Passing For example if you use parameters to set the initial state of a latch in its subcircuit definition then you can override this initial default in the instance call You need to create only one cell to handle both initial state versions of the latch You can also use parameters to define the cell layout For example you can use parameters in a MOS inverter to simulate a range of inverter sizes with only one cell definition Local instances of the cell can assign different values to the size parameter for the inverter In HSPICE you can also perform Monte Carlo analysis or optimization on a cell that uses parameters How you handle hierarchical parameters depends on how you construct and analyze your cells You can c
228. e Formats 392 The default output format for transient analysis in HSPICE RF is the same as in HSPICE the trO file format See Transient Analysis in the HSPICE Simulation and Analysis User Guide HSPICE RF supports these output formats which are described in this section Tabulated Data Output WDB Output Format XP Output Format NW Output Format m VCD Output Format turboWave Output Format tw Undertow Output Format ut CSDF Output Format If your netlist includes an unsupported output format HSPICE RF prints a warning message indicating that the selected format is unsupported HSPICE RF then automatically defaults the output to TRO format HSPICE RF User Guide Z 2007 03 Chapter 15 Using HSPICE with HSPICE RF RF Transient Analysis Output File Formats You can use the waveform viewer to view certain output formats wdb XP CosmosScope Recommended nw XP AvanWaves xp XP AvanWaves CosmosScope Note If your waveform file is larger than 2GB use split waveforms Tabulated Data Output HSPICE RF outputs all analog waveforms specified ina PRINT statement HSPICE RF saves these waveforms as ASCII tabulated data into a file with the PRINT extension To display waveforms graphically CosmosScope can directly read the tabulated data For more information about CosmosScope see the CosmosScope User s Guide and Reference Note Tabulated data excludes waveforms specified in PROBE statements
229. e HSPICE RF and how to perform advanced tasks including redirecting input and output HSPICE RF User Guide Z 2007 03 About This Guide The HSPICE Documentation Set The HSPICE Documentation Set This manual is a part of the HSPICE documentation set which includes the following manuals Manual Description HSPICE Simulation and Analysis User Guide HSPICE Signal Integrity Guide HSPICE Applications Manual HSPICE RF User Guide HSPICE Elements and Device Models Manual HSPICE MOSFET Models Manual HSPICE RF User Guide AMS Discovery Simulation Interface Guide for HSPICE AvanWaves User Guide Describes how to use HSPICE to simulate and analyze your circuit designs This is the main HSPICE user guide Describes how to use HSPICE to maintain signal integrity in your chip design Provides application examples and additional HSPICE user information Provides reference information for HSPICE and HSPICE RF commands and options Describes standard models you can use when simulating your circuit designs in HSPICE including passive devices diodes JFET and MESFET devices and BJT devices Describes standard MOSFET models you can use when simulating your circuit designs in HSPICE Describes a special set of analysis and design capabilities added to HSPICE to support RF and high speed circuit design Describes use of the Simulation Interface with other EDA tools for HSPICE Describes t
230. e form also known as Foster canonical form The popular method of recursive convolution also uses this form HSPICE supports the pole residue form for its frequency dependent controlled sources G and E elements You can enter the pole residue form directly without first converting to another form Foster Pole Residue Form for Transconductance or Gain The Foster pole residue form for transconductance G s or gain E s has the form Nt A As Equation 15 G s k k s gt L S Pi s p i l Where ko ky are real constants residues Aj and poles p are complex numbers or real as a special case of complex asterisk denotes the expression s complex conjugate Advantages of Foster Form Modeling The advantages of Foster canonical form modeling are m models high order systems It can theoretically model systems having infinite poles without numerical problems equivalent to Laplace and Pole zero models popular method of recursive convolution uses this form HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Function Approximations for Distributed Devices G and E Element Syntax Transconductance G s form GXXX n n Gain E s form EXXX N n FOSTER in in k0 kl Re A1 Im A1 Re p1 Im p1 Re A2 Im A2 Re p2 Im p2 Im A3 Re p3 Im p3 Re A3 FOSTER in in k0 kl Re A1 Im A1 Re p1 Im p1 Re A2
231. e inv mt7 2 Open the Calculator select TOX left mouse button transfer to calculator middle mouse button and then select and transfer XL 3 On the WAVE pulldown in the calculator select f x and then click the plot icon 4 Using the right mouse button on the plotted waveform select Attributes to change from the line plot to symbols HSPICE RF User Guide Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Worst Case and Monte Carlo Sweep Example Figure 41 Scatter Plot XL and TOX Monte Carla Results G xl polyca 20 0 tox toxcdix kE polyed o 200 0 150 0 500n 0 0 500n X1I polyed The next graph see Figure 42 is a standard scatter plot showing the measured delay for the inverter pair against the Monte Carlo index number HSPICE RF User Guide 375 Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Worst Case and Monte Carlo Sweep Example 376 Figure 42 Scatter Plot of Inverter Pair Delay Monte Carla Results If a particular result looks interesting for example if the simulation 68 monte carlo index 68 produces the smallest delay then you can obtain the Monte Carlo parameters for that simulation xxx monte carlo index 68 MONTE CARLO PARAMETER DEFINITIONS polycd xl 1 6245E 07 nactcd xwn 3 4997E 08 pactcd xwp 3 6255E 08 toxcd tox 191 0 vtoncd delvton 2 2821E 02 delvtop 4 1776E 02 vtopcd rshned rshn 45 16 rshpcd rshp 166
232. e node or element TYPE can be one of the following Voltage type V voltage magnitude and phase in degrees VR real component VI imaginary component VM magnitude VP Phase in degrees VPD Phase in degrees VPR Phase in radians VDB GB units VDBM dB relative to 1 mV Current type current magnitude and phase in degrees IR real component Il imaginary component IM magnitude IP Phase in degrees IPD Phase in degrees IPR Phase in radians IDB dB units IDBM dB relative to 1 mV Power type P Frequency type hertz index hertz index1 index2 You must specify the harmonic index for the hertz variable The frequency of the specified harmonics is dumped NODES NODES or ELEM can be one of the following ELEM Voltage type a single node name n1 or a pair of node names n1 n2 Current type an element name elemname Power type a resistor resistorname or port portname element name Frequency type the harmonic index for the hertz variable The frequency of the specified harmonics is dumped HSPICE RF User Guide 265 Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Shooting Newton AC Analysis SNAC Parameter Description INDICES Index to tones in the form n1 1 1 is the index of the SN tone 4 1 is the index of the SNAC tone Wildcards are not supported if this parameter is used You can transfo
233. e other nodes v 0 1 v 0 2 andv 0 3 Examples The following examples use wildcards with PRINT PROBE and LPRINT statements Probe node voltages for nodes at all levels PROBE v Probe all nodes whose names start with a For example a1 a2 a3 a00 ayz HSPICE RF User Guide 77 Z 2007 03 Chapter 4 Input Netlist and Data Entry Input Netlist File Composition 78 PROBE v a Print node voltages for nodes at the first level and all levels below the first level where zero level are top level nodes For example X1 A X4 554 Xab abc123 PRINT v Probe node voltages for all nodes whose name start with x at the first level and all levels below the first level where zero level are top level nodes For example x1 A x4 554 xab abc123 PROBE v x Print node voltages for nodes whose names start with x at the second level and all levels below the second level For example x1 x2 a xab xdff in PRINT v x m Match all first level nodes with names that are exactly two characters long For example x1 in x4 12 PRINT v x In HSPICE RF print the logic state of all top level nodes whose names start with b For example b1 b2 b3 b56 bac LPRINT 1 4 b Element Instance and Subcircuit Naming Conventions Instances and subcircuits are elements and as such follow the naming conventions for elements Element names in HSPICE or HSPICE RF be
234. e9 34 7e9 9 95e9 114e9 34 7e9 103e9 103e9 34 7e9 9 95e9 114e9 34 7e9 103e9 103e9 34 7e9 9 95e9 114e9 34 7e9 9 103e9 HORTALL no IGNORE COUPLING no ONOHDWOAWAUAUNTAITAN SEA al 1 1 4 4 7 2 2 2 5 5 8 3 3 3 6 6 9 S ttt ttt t tt t t tt t t t t t t Alternatively the same element could be specified by using L ThreeNetsal1l22a1b445 5 b1c 7177 8 8 c_1 RELUCTANCE FILE reluctance dat SHORTALL no IGNORE COUPLING no Where reluctance dat contains HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Passive Elements 103e9 34 7e9 9 95e9 114e9 34 7e9 103e9 103e9 34 7e9 9 95e9 114e9 34 7e9 103e9 103e9 34 7e9 9 95e9 114e9 34 7e9 103e9 ONNDWWWOAUUNNNNAA KBR HR BR t etetetetetetetestest CONDOWOHDWOOUHDUNIYYAYAPH The following shows the mapping between the port numbers and node pairs Ports 1 3 4 5 6 7 8 9 hese ee a 1 15 2 PEEN b 4 4 5 ee a c 7 7 8 Mate a5 Ideal Transformer Format in HSPICE RF The ideal transformer format simplifies modeling of baluns Previously baluns were modeled using mutual inductors K elements with the IDEAL keyword Multiple L and K elements were needed for a given balun model The ideal transformer model allows modeling of a balun using a single L element In the ideal transformer format no absolute inductance or reluctance values are specified Instead the transformers coupling charact
235. eature set that support the design of RF and high speed circuits Where necessary the manual describes differences that might exist between HSPICE RF and HSPICE Note This manual discusses only HSPICE RF features For information on other HSPICE applications see the other HSPICE manuals listed in The HSPICE Documentation Set on page xv Inside This Manual This manual contains the chapters described below For access to the other manuals in the HSPICE documentation set see the next section Searching Across the HSPICE Documentation Set on page xvi Chapter Description Chapter 1 HSPICE RF Features and Functionality Chapter 2 Getting Started Chapter 3 HSPICE RF Tutorial Chapter 4 Input Netlist and Data Entry Chapter 5 Parameters and Functions HSPICE RF User Guide Z 2007 03 Introduces HSPICE RF features and functionality Describes how to set up your environment invoke HSPICE RF customize your simulation and redirect input and output Provides a quick start tutorial for users new to HSPICE RF Describes the input netlist file and methods of entering data in HSPICE or HSPICE RF Describes how to use parameters within HSPICE RF netlists xiii About This Guide Inside This Manual xiv Chapter Description Chapter 6 Testbench Elements Chapter 7 Steady State Harmonic Balance Analysis Chapter 8 Steady State Shooting Newton Analysis Chapter 9 Oscillator and Pha
236. ectory lt installdir gt hspicerf examples This example performs analysis on a D flipflop phase frequency divider with charge pump implemented in 50nm technology The example is configured to measure the gain volts per degree of the DFF PFD and tri state output combination 228 HSPICE RF User Guide Z 2007 03 9 Oscillator and Phase Noise Analysis Describes how to use HSPICE RF to perform oscillator and phase noise analysis on autonomous oscillator circuits Harmonic Balance or Shooting Newton for Oscillator Analysis HSPICE RF includes special analysis algorithms for finding the steady state solution for oscillator circuits In oscillators there are no driving sources that set the frequencies of operation but rather the fundamental oscillation frequency is one of the unknowns that is being solved by the simulator HSPICE RF provides two approaches either harmonic balance or analysis based on the Shooting Newton algorithm The following sections are presented in this chapter Harmonic Balance Analysis for Frequency of Oscillation Oscillator Analysis Using Shooting Newton SNOSC Phase Noise Analysis Harmonic Balance Analysis for Frequency of Oscillation Because the frequency of oscillation is not determined by the frequencies of driving sources oscillator circuits are called autonomous Autonomous simulation solves a slightly different set of nonlinear equations as shown in the following equation Equation 34
237. ee the Passive Device Models chapter in the HSPICE Elements and Device Models Manual for temperature dependent relations 125 Chapter 6 Testbench Elements Passive Elements 126 Parameter Description TC2 Second order temperature coefficient for the inductor SCALE Element scale parameter scales inductance by its value Default 1 0 IC Initial current through the inductor in amperes HSPICE or L inductance DTEMP R L equation LTYPE POLY HSPICE RF uses this value as the DC operating point voltage when you specify UIC in the TRAN statement The IC statement overrides it Inductance value This can be a numeric value in henries a parameter in henries a function of any node voltages a function of branch currents any independent variables such as time hertz and temper Multiplier used to simulate parallel inductors Default 1 0 Temperature difference between the element and the circuit in degrees Celsius Default 0 0 Resistance of the inductor in ohms Default 0 0 Inductance at room temperature specified as a function of any node voltages a function of branch currents any independent variables such as time hertz and temper Calculates inductance flux for elements using inductance equations If the L inductance is a function of I Lxxx then set LTYPE 0 Otherwise set LTYPE 1 Use this setting correctly to ensure proper inductance calculations and correct sim
238. efault 0 0 Sets initial condition for this element to OFF in DC analysis Default ON This command does not work for depletion devices Initial voltage across external drain and source vds gate and source vgs and bulk and source terminals vbs Use these arguments with TRAN UIC IC statements override these values Multiplier to simulate multiple MOSFETs in parallel Affects all channel widths diode leakages capacitances and resistances Default 1 The difference between the element temperature and the circuit temperature in degrees Celsius Default 0 0 167 Chapter 6 Testbench Elements Active Elements 168 Parameter Description GEO Source drain sharing selector fora MOSFET model parameter value of ACM 3 Default 0 0 DELVTO Zero bias threshold voltage shift Default 0 0 The only required fields are the drain gate and source nodes and the model name The nodes and model name must precede other fields in the netlist If you did not specify a label use the second syntax with the OPTION WL statement to exchange the width and length options Example In the following M1 MOSFET element M1 1 2 3 model 1 The drain connects to node 1 The gate connects to node 2 The source connects to node 3 model_1 references the MOSFET model In the following Mopamp1 MOSFET element Mopampl dl g3 s2 b 1lstage L 2u W 10u The drain connects to the d1 node The gate connects to the g3 node
239. egals by default mp1 lt NodeList gt lt Model gt L 1lu W Wid 2 mnl lt NodeList gt lt Model gt L lu W Wid Ends Invoke symbols in a design xl A Y1 Inv Incorrect width x2 A Y2 Inv Wid lu Incorrect Both x1 and x2 simulate with mpl 10u and mnl 5u instead of 2u and lu Under global parameter scoping rules simulation succeeds but incorrectly HSPICE does not warn you that the x1 inverter has no assigned width because the global parameter definition for Wid overrides the subcircuit default Note Similarly sweeping with different values of Wid dynamically changes both the Wid library internal parameter value and the pulse width value to the Wid value of the current sweep In global scoping the highest level name prevails when resolving name conflicts Local scoping uses the lowest level name When you use the parameter inheritance method you can specify to use local scoping rules When you use local scoping rules the Example 2 netlist correctly aborts in x1 for W 0 default Wid 0 in the SUBCKT definition has higher precedence than the PARAM statement This results in the correct device sizes for x2 This change can affect your simulation results if you intentionally or accidentally create a circuit such as the second one shown above As an alternative to width testing in the Example 2 netlist you can use OPTION DEFW to achieve a limited version of library integrity This option HSPICE RF U
240. ehavior of all four resistors If an instance parameter or instance parameter variations and model parameter variations are specified on the same parameter then the instance parameter always overrides the model parameter Because only few parameters can be used in both domains this case is rather seldom but it needs to be considered to avoid unexpected results test18 sp has model variation specified on width with a parameter Two resistors have width also defined on instance The resistors with instance parameter do not vary at all The other two resistors vary independently as expected because option modmonte is set to 1 test19 sp is similar to test18 sp with option modmonte set to 0 The two resistors that do not have width defined on the instance line vary together test20 sp has DEV LOT specified Instance parameters override variations on selected resistors The DEV LOT approach has no mechanism to describe variation as a function of an element parameter HSPICE RF User Guide 387 Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Simulating the Effects of Global and Local Variations with Monte Carlo 388 HSPICE RF User Guide Z 2007 03 15 Using HSPICE with HSPICE RF Describes how various analysis features differ in HSPICE RF as compared to standard HSPICE This first section of this chapter describes topics related to transient analysis and the other section describe other differences between HSPICE
241. eit ohatue ah bleedin weed peda ted 4 MOSFETs Steady State Voltage and and V Element Syn Current SourceS 0 0 000000 cee eee MAX eth a a oth dan deacon nae laed Pie Sahay eel a G Steady State HB Sources 20 0 0 0c ccc eens Phase Differences Between HB and SIN Sources 00 Behavioral Noise Sources 0 eee eee ee Power Supply Current and Voltage Noise Sources Function Approximations for Distributed Devices 4 Foster Pole Residue Advantages of Foste Form for Transconductance or Gain r Form Modeling 00 e eee eee eee G and E Element Syntax 0 0 00 eee Complex Signal Sources and Stimuli 2 0 2 0 0 0 00 cece ee eee Vector Modulated RF Source 0 00 eens Voltage and Current Source Elements 000000000 ee eee SWEEPBLOCK in Sweep Analyses 000 0c cece eee eens Input Syntax Using SWEEPBLOC K ina DC Parameter Sweep Using in Parameter Sweeps in TRAN AC and HB Analyses Limitations Clock Source with Random Jitter 0 00000000 c eee eee Syntax of SIN COS References Steady State Harmonic and Pulse SourcesS 0 0 0000 eae Balance Analysis 2 055 Harmonic Balance Analysis 00000 c eee eee Harmonic Balance E Features Supported QualiONS sides ei ee ee
242. elements supports the syntax for specifying power sources In this case the source value is interpreted as a power value in Watts or dBm units and the Port element is HSPICE RF User Guide Z 2007 03 Chapter 1 HSPICE RF Features and Functionality HSPICE RF Overview implemented as a voltage source with a series impedance The HBLSP command invokes periodically driven nonlinear circuit analyses for power dependent S parameters Harmonic Balance HB analysis using Direct and Krylov solvers The HB command invokes the single and multitone Harmonic Balance algorithm for periodic steady state analysis TRANFORHB element parameter to recognize V I sources that include SIN and PULSE transient descriptions as well as PWL and VMRF sources Harmonic balance based periodic AC analysis The HBAC command invokes periodic AC analysis for analyzing small signal perturbations on circuits operating in a large signal periodic steady state Harmonic Balance based Periodic Noise analysis HBNOISE for noise analysis of periodically modulated circuits includes stationary cyclostationary and frequency dependent noise effects Autonomous Harmonic Balance analysis The HBOSC command invokes the multitone oscillator capable Harmonic Balance algorithm for periodic steady state analysis Perturbation analysis for Oscillator Phase Noise The HBAC command invokes phase periodic AC noise for oscillators circuits operating in a large signal stea
243. en present has precedence over the SIM_POSTSKIP option For additional information see OPTION SIM_POSTAT in the HSPICE and HSPICE RF Command Reference SIM_POSTDOWN Option You use the SIM_POSTDOWN option to include an instance and all children of that instance in the output For example OPTION SIM POSTDOWN lt instance gt It can be used in conjunction with the SIM_POSTTOP option and when present has precedence over the SIM_POSTSKIP option HSPICE RF User Guide Z 2007 03 Chapter 16 Advanced Features Probing Subcircuit Currents For additional information see OPTION SIM_POSTDOWN in the HSPICE and HSPICE RF Command Reference SIM_POSTSCOPE Option You use the SIM_POSTSCOPE option to specify the signal types to probe from within a scope For example OPTION SIM _POSTSCOPE net port all For additional information see OPTION SIM_POSTSCOPE in the HSPICE and HSPICE RF Command Reference Probing Subcircuit Currents To provide subcircuit power probing utilities HSPICE RF uses the X and X0 extended output variables You can use these X variables in PROBE PRINT or MEASURE statements The following syntax is for the output variable X X subcircuit_node_path XO subcircuit_node_ path subcircuit_node_path specifies the subcircuit path and the subcircuit node name definition The node must be either an external node in a subcircuit definition or a global node X returns the total current flowing
244. en to a p2d file The large and small signal S parameters from the PROBE statement are viewable in CosmosScope 310 HSPICE RF User Guide Z 2007 03 12 Envelope Analysis Describes how to use envelope simulation Envelope Simulation Envelope simulation combines features of time and frequency domain analysis Harmonic Balance HB solves for a static set of phasors for all the circuit state variables as shown in this equation N Equation 66 v t ag X a cos cat b sin at i l In contrast envelope analysis finds a dynamic time dependent set of phasors as this equation shows N Equation 67 v t ag t a f cos ayr b t sinar i l Thus in envelope simulation each signal is described by the evolving spectrum Envelope analysis is generally used on circuits excited by signals with significantly different timescales An HB simulation is performed at each point in time of the slower moving 7 timescale In this way for example a 2 tone HB simulation can be converted into a series of related 1 tone simulations where the transient analysis proceeds on the 7 timescale and 1 tone HB simulations are performed with the higher frequency tone as the fundamental frequency In HSPICE RF any voltage or current source identified as a HB source either in a V orl element statement or by an OPTION TRANFORHB command is used HSPICE RF User Guide 311 Z 2007 03 Chapter 12 Envelope Analysis Envelope
245. ent by the square root of the product of the self inductances When using the mutual inductor element to calculate the coupling between more than two inductors HSPICE or HSPICE RF can automatically calculate an approximate second order coupling See the third example below for a specific situation Note The automatic inductance calculation is an estimation and is accurate for a subset of geometries The second order coupling coefficient is the product of the two first order coefficients which is not correct for many geometries Example 1 The Lin and Lout inductors are coupled with a coefficient of 0 9 K1 Lin Lout 0 9 Example 2 The Lhigh and Llow inductors are coupled with a coefficient equal to the value of the COUPLE parameter Kxfmr Lhigh Llow K COUPLE The K1 mutual inductor couples L1 and L2 The K2 mutual inductor couples L2 and L3 RF User Guide 129 Chapter 6 Testbench Elements Passive Elements 130 Example 3 The coupling coefficients are 0 98 and 0 87 HSPICE or HSPICE RF automatically calculates the mutual inductance between L1 and L3 witha coefficient of 0 98 0 87 0 853 K1 L1 L2 0 98 K2 L2 L3 0 87 Linear Inductors Lxxx nodel node2 lt L gt inductance lt TCl val gt lt TC2 val gt lt M val gt lt DTEMP vals gt lt IC vals gt Parameter Description Lxxx Name of an inductor node1 and node2 Names or numbers of the connecting nodes inductance Nominal inductance value i
246. ent temperature Part temperature at the system level each part has its own temperature System temperature a collection of parts form a system which has a local temperature Ambient temperature the ambient temperature is the air temperature of the system HSPICE RF User Guide 351 Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Simulating Circuit and Model Temperatures 352 Figure 28 Part Junction Temperature Sets System Performance _ gt _ gt Ambient Temperature System Temperature Part Temperature source drain source drain gate gate aN A gt Model Junction Temperature Part Junction Temperature HSPICE RF calculates temperatures as differences from the ambient temperature Equation 68 Tambient Asystem Apart Ajunction Tjunction Equation 69 Ids f Tjunction Tmodel Every element includes a DTEMP keyword which defines the difference between junction and ambient temperature Example The following example uses DTEMP in a MOSFET element statement M1 drain gate source bulk Model _name W 10u L 1lu DTEMP 20 Temperature Analysis You can specify three temperatures HSPICE RF User Guide Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Simulating Circuit and Model Temperatures Model reference temperature specified ina MODEL statement The temperature parameter is usually TREF but can be TEMP or TNOM in some models T
247. ep mval 1 2 1 HBNOISE rout rsrc lin 11 0 90g 0 92g print HBNOISE onoise ssnf dsnf probe HBNOISE onoise ssnf dsnf end HSPICE RF User Guide Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Shooting Newton Noise Analysis SNNOISE Shooting Newton Noise Analysis SNNOISE A SNNOISE Shooting Newton noise analysis simulates the noise behavior in periodic systems It uses a Periodic AC PAC algorithm to perform noise analysis of nonautonomous driven circuits under periodic steady state tone conditions SNNOISE is similar to the HBNOISE analysis The PAC method simulates noise assuming that the stationary noise sources and or the transfer function from the noise source to a specific output are periodically modulated The modulated noise source thermal shot or flicker is modeled as a cyclostationary noise source A PAC algorithm solves the modulated transfer function You can also use the SNNOISE PAC method with correlated noise sources including the MOSFET Level 9 and Level 11 models and the behavioral noise source in the G Element Voltage Dependent Current Source You use the SNNOISE statement to perform a Periodic Noise Analysis Supported Features SNNOISE supports the following features All existing HSPICE RF noise models Uses Shooting Newton to generate the steady state solution Unlimited number of sources Includes stationary cyclostationary
248. eqg 1 0G listfreq 1 0G 2 0G The default value is NONE Prints the element noise value to the lis file which is sorted from the largest to smallest value You do not need to print every noise element instead you can define listcount to print the number of element noise frequencies For example 1istcount 5 means that only the top 5 noise contributors are printed The default value is 1 Prints the element noise value to the lis file and defines a minimum meaningful noise value in V Hz units Only those elements with noise values larger than listfloor are printed The default value is 1 0e 14 V Hz 2 Prints the element noise value to the lis file when the element has multiple noise sources such as a FET which contains the thermal shot and 1 f noise sources You can specify either ON or OFF ON Prints the contribution from each noise source and OFF does not The default value is OFF Output Syntax This section describes the syntax for the HBNOISE PRINT and PROBE statements PRINT and PROBE Statements PRINT HBNOISE lt ONOISE gt lt NFs lt SSNF gt lt DSNF gt HSPICE RF User Guide Z 2007 03 271 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Multitone Harmonic Balance Noise HBNOISE 272 PROBE HBNOISE lt ONOISE gt lt NF gt lt SSNF gt lt DSNF gt Parameter Description ONOISE Outputs the voltage noise at the output frequency band OFB acros
249. er than the right operator operand Otherwise returns 0 greater para x y gt z y greater than z than gt relational Returns 1 if the left operand is greater than or equal operator to the right operand Otherwise returns 0 greater para x y gt z y greater than or equal to z than or equal equality Returns 1 if the operands are equal Otherwise returns 0 para x y z y equal to z l inequality Returns 1 if the operands are not equal Otherwise returns 0 para x y z y not equal to z amp amp Logical Returns 1 if neither operand is zero Otherwise AND returns 0 para x y amp amp z y AND z HSPICE RF User Guide Z 2007 03 103 Chapter 5 Parameters and Functions Parameter Scoping and Passing Table 12 Synopsys HSPICE Built in Functions Continued HSPICE Form Function Class Description Logical OR Returns 1 if either or both operands are not zero Returns 0 only if both operands are zero para x y z y OR 2 Example parameters pl 4 p2 5 p3 6 rl 1 0 value pl p2 1 p3 HSPICE reserves the variable names listed in Table 13 on page 104 for use in elements such as E G R C and L You can use them in expressions but you cannot redefine them for example this statement would be illegal param temper 100 Table 13 Synopsys HSPICE Special Variables HSPICE Form Function Class Description time current control Uses parameters to define the current simulation simulation tim
250. erf mix_hbac sp cscope amp 2 Open the mix_tran trO file choose File gt Open gt Plotfiles and select mix_tran trO 3 To plot v out double click v out in the signal manager Open the mix_hb hb0 file choose File gt Open gt Plotfiles and select mix_hb hbo HSPICE RF User Guide 41 Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 6 Using Multi Tone HB and HBAC Analyses for a Mixer 42 You might need to change the Files of Type filter to HSPICERF HB hb Plot v out by double clicking v out in the signal manager A histogram displays Open the mix_hbac hb0 file choose File gt Open gt Plotfiles and select mix_hbac hb0 You might need to change the Files of Type filter to HSPICERF HBAC hb Plot v out by double clicking v out in the signal manager You should see a histogram similar to the one from mix_hb hb0 Convert the HB and HBAC histograms to time domain For each of the two v out histogram signals right click on the v out label and choose To Time Domain Accept the default range and interval settings Two new time domain waveforms should appear Overlay the three time domain plots Right click on each timedomain v out label and choose Stack Region Analog 0 The bottom panel should now display all three time domain signals All three are almost indistinguishable from each other You can also use HBAC to perform noise analysis on RF circuits by using the
251. eristics are specified using inductor number of turns values The behavior of the ideal transformer depends on ratios of the inductors number of turns Syntax LXXX nip nin nNp nNn TRANSFORMER NT ntl1 ntN Parameter Description Lxxx Inductor element name Must begin with L followed by up to 1023 alphanumeric characters HSPICE RF User Guide 137 Z 2007 03 Chapter 6 Testbench Elements Passive Elements 138 Parameter Description nipnin nNpnNn Positive and negative terminal node names The number of terminals must be even Each pair of reports represents the location of an inductor TRANSFORMER_NT Number of turns values These parameters must match the number of inductors The ideal transformer element obeys the standard ideal transformer equations Y Y2 LYN nt nt nty inti int iynty 0 Example L1 1 0 0 2 3 0 transformer_nt 1 2 2 DC Block and Choke Elements In HSPICE RF you can specify an INFINITY value for capacitors and inductors to model ideal DC block and choke elements The following input syntax is for the DC block ideal infinite capacitor Syntax Cxxx nodel node2 lt C gt INFINITY lt IC val gt HSPICE RF does not support any other capacitor parameters for DC block elements because HSPICE RF assumes that the infinite capacitor value is independent of temperature and scaling factors The DC block acts as an open circuit for all DC analyses HSPICE RF cal
252. es These statements can call netlists model parameters test vectors analysis and option macros into a file from library files or other files The input netlist file also can call an external data file which contains parameterized data for element sources and models You must enclose the names of included or internally specified files in single or double quotation when they begin with a number 0 9 Schematic Netlists HSPICE RF typically use netlisters to generate circuits from schematics and accept either hierarchical or flat netlists HSPICE RF User Guide Z 2007 03 Chapter 4 Input Netlist and Data Entry The process of creating a schematic involves Symbol creation with a symbol editor Circuit encapsulation Property creation Symbol placement Symbol property definition Wire routing and definition Table 8 Input Netlist File Sections Input Netlist File Guidelines Sections Examples Definition Title TITLE The first line in the netlist is the title of the input netlist file optional in HSPICE RF Set up OPTION IC or NODESET Sets conditions for simulation PARAM GLOBAL Initial values in circuit and subcircuit Set parameter values in the netlist Set node name globally in netlist Sources Sources and digital inputs Sets input stimuli I or V element Netlist Circuit elements Circuit for simulation SUBKCT ENDS or Subcircuit definitions MACRO EOM Analysis DC TRAN AC and so on Stateme
253. esented here with multiple coupled inductors Ij is the current into the first terminal of Ly Msg VES ee VO Vs Equation 1 Ss JLI JL2 vNJI3 a EA Equation2 il JL1 i2 JL2 i3 JL3 i4 JLA HSPICE RF can solve any i or v in terms of L ratios For two inductors non DC values Equation 3 Wi WA JLI sf ED Equation4 0 il JL1 i2 JL2 Equation5 v2 vl L2 L1 Equation 6 i2 il El L2 DC is treated as usual inductors are treated as short circuits DC ignores mutual coupling You can couple inductors that use the INFINITY keyword to IDEAL K elements All inductors involved must have the INFINITY value and for K IDEAL the ratios of all L values is unity Then for two L values v1 il v2 i2 HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Passive Elements Example 1 This example is a standard 5 pin ideal balun transformer subcircuit Two pins are grounded for standard operation With all K values being IDEAL the absolute L values are not crucial only their ratios are important kk all K s ideal o outl x Lol 25 KA Ormen Topora o 0 Lin 1 Lo2 25 BR VUOMOso ase te ees o out2 kk subckt BALUN1 in outi out2 Lin in gnd L 1 Lol outl gnd L 0 25 Lo2 gnd out2 L 0 25 K12 Lin Lol IDEAL K13 Lin Lo2 IDEAL K23 Lol lLo2 IDEAL ends Example 2 This example is a 2 pin ideal 4 1 step up balun transformer subcircuit with shared DC path no
254. esults of an SN analysis are complex spectral components at each frequency point The ali is the real part and b i is the imaginary part of the HSPICE RF User Guide 223 Z 2007 03 Chapter 8 Steady State Shooting Newton Analysis Shooting Newton with Fourier Transform SNFT complex voltage at frequency index i The conversion to a steady state time domain is then given by the Fourier series expansion An SN analysis produces these output data files Output from the PRINT SN statement is written to a printsn file e The header contains the large signal fundamental frequencies e The columns of data are labeled as HERTZ followed by frequency indices and then the output variable names e The sum of the frequency indices multiplied by the corresponding fundamental frequencies add up to the frequency in the first column Output from the PROBE SN statement is written to a sn file in the same format as the HSPICE transient analysis tr file It contains the information of harmonic indices and basic tone frequencies plus the output waveform Reported performance log statistics are written to a lis file e Name of SN data file e Simulation time DC operating point op time SN time Total simulation time e Memory used e Size of matrix nodes harmonics e Final SN residual error Shooting Newton with Fourier Transform SNFT 224 The SNFT command is to the SN analysis what FFT is to the TRAN analysis
255. etlist File Composition In HSPICE RF you can use multiple TEMP statements to specify multiple temperatures for different portions of the circuit HSPICE permits only one temperature for the entire circuit Multiple TEMP statements in a circuit behave as a sweep function Data Driven Analysis In data driven analysis you can modify any number of parameters then use the new parameter values to perform an operating point DC AC or transient analysis An array of parameter values can be either inline in the simulation input file or stored as an external ASCII file The DATA statement associates a list of parameter names with corresponding values in the array HSPICE RF supports DATA only for Data driven analysis Inline or external data files Library Calls and Definitions To create and read from libraries of commonly used commands device models subcircuit analysis and statements in library files use the LIB call statement As HSPICE RF encounters each LIB call name in the main data file it reads the corresponding entry from the designated library file until it finds an ENDL statement You can also place a LIB call statement in an ALTER block Library Building Rules A library cannot contain ALTER statements A library can contain nested LIB calls to itself or to other libraries If you use a relative path in a nested LIB call the path starts from the directory of the parent library not from t
256. f logical_power_net physical_power_net ground_net logical_port coordinate par_value rising_slew falling_slew low_threshold high_threshold cell_type physical_port logical_instance 338 A name identifier bit path name or physical reference to map to the name_index Logical path or logical path index to a power net Physical path or physical path index to a power net You can specify multiple jogical_power_net physical_power_net pairs Name of a net to use as a ground net You can specify multiple ground net names Logical name of an input output or bidirectional port Geometric location of a logical or physical port Either a single float value or a triplet in float float float form Rising slew of the waveform for the port T_UNIT defines the time unit for the waveform Rising slew of the waveform for the port T_UNIT defines the time unit for the waveform Low voltage threshold as a percentage of the port s input voltage Can bed one float value or a triplet in float float float form High voltage threshold as a percentage of the input voltage for the port Either a single float value or a triplet in float float float form Type of cell that drives the port If you do not know the cell type use the reserved word UNKNOWN_DRIVER as the cell type Physical name of an input output or bidirectional port Logical name of a subcircuit in your design_name circuit design You can spec
257. f harmonic indices and basic tone frequencies Output from the PRINT HBTRAN statement is written to a printhr file The format is identical to a print file Output from the PROBE HBTRAN statement is written to a hr file The format is identical to a tr file Reported performance log statistics are written to a lis file e Name of HB data file e Simulation time DC operating point op time HB time Total simulation time e Memory used e Size of matrix nodes harmonics e Final HB residual error HSPICE RF User Guide 215 Z 2007 03 Chapter 7 Steady State Harmonic Balance Analysis HB Output Data Files 216 Errors and Warnings Table 17 lists the errors messages and Table 18 on page 217 lists the warning messages Table 17 HB Analysis Error Messages File HB_ERR 1 HB_ERR 2 HB_ERR 3 HB_ERR 4 HB_ERR 5 HB_ERR 6 HB_ERR 7 HB_ERR 8 HB_ERR 9 HB_ERR 10 HB_ERR 11 HB_ERR 12 HB_ERR 13 HB_ERR 14 HB_ERR 15 HB_ERR 16 HB_ERR 17 Description Harmonic numbers must be positive non zero No hb frequencies given Negative frequency given Number of harmonics should be greater than zero Different number of tones nharms Bad probe node format for oscillator analysis Bad format for FSPTS Bad hb keyword Tones must be specified for hb analysis Nharms or intmodmax must be specified for hb analysis Source harmonic out of range Source named in the tones list is not defined Source n
258. facturing variations 369 capacitor charge based 122 element 119 123 frequency dependent 124 linear 123 cell characterization 350 charge based capacitor 122 CHECK EDGE statement 413 CHECK FALL statement 413 CHECK GLOBAL_LEVEL statement 412 CHECK HOLD statement 414 CHECK IRDROP statement 415 CHECK RISE statement 412 CHECK SETUP statement 414 CHECK SLEW statement 412 421 Index choke elements 138 circuit description syntax 9 circuits description syntax 70 reusable 87 subcircuit numbers 79 temperature 353 See also subcircuits clock source random jitter 191 CMOS GPS VCO 31 Colpitts oscillator 28 command PRINT ENV 315 command PROBE ENV 315 commands hspicerf 9 PTDNOISE 281 comment line netlist 71 comparing results 41 compression of input files 57 config file hspicerf 399 configuration file 399 example 402 configuration options flush_waveform 400 ground_floating_ node 400 hier_delimiter 400 htm 400 integer_node 400 max_waveform_size 400 negative_td 401 port_element_ voltage_ matchload 401 rext_divider 401 unit_atto 401 v_supply 401 wildcard_left_range 401 wildcard_match_all 402 wildcard_match_one 402 wildcard_right_range 402 continuation character parameter strings 100 continuation of line netlist 72 cos x function 100 cosh x function 101 Cosmos Scope 12 422 coupled inductor element 134 D DATA statement 81 data driven analysis 81 350 351 db x function 101 DC block elements 13
259. fective electrical size is 0 84 Account for the four dimension levels e drawn size e shrunken size e physical size e electrical size Most simulator models scale directly from drawn to electrical size HSPICE MOS models support all four size levels as in Figure 31 358 HSPICE RF User Guide Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Monte Carlo Analysis Figure 31 Device Model from Drawn to Electrical Size Drawn Size Electrical Size source drain gate LMLT WMLT LD WD Shrunken Size LELEI A 1m XL XW Physical Size source drain gt 0 9 m sF lt Monte Carlo Analysis Monte Carlo analysis uses a random number generator to create the following types of functions Gaussian parameter distribution e Relative variation variation is a ratio of the average e Absolute variation adds variation to the average e Bimodal multiplies distribution to statistically reduce nominal parameters Uniform parameter distribution e Relative variation variation is a ratio of the average e Absolute variation adds variation to the average e Bimodal multiplies distribution to statistically reduce nominal parameters Random limit parameter distribution e Absolute variation adds variation to the average e Monte Carlo analysis randomly selects the min or max variation HSPICE RF User Guide Z 2007
260. fferent model cards reference subcircuits or define subcircuits in each IF ELSE block if conditioni lt statement_block1 gt The following statement block in braces is optional and you can repeat it multiple times elseif condition2 lt statement_block2 gt we tH The following statement block in brackets is optional and you cannot repeat it else lt statement_block3 gt endif Inan IF ELSEIF or ELSE condition statement complex Boolean expressions must not be ambiguous For example change a b amp amp c gt d to a b amp amp c gt d nan IF ELSEIF or ELSE statement block you can include most valid HSPICE or HSPICE RF analysis and output statements The exceptions are e END ALTER GLOBAL DEL LIB MALIAS ALIAS LIST NOLIST and CONNECT statements e search d_ibis d_imic d 1v56 biasfi modsrh cmiflag nxx and brief options HSPICE RF User Guide Z 2007 03 Chapter 4 Input Netlist and Data Entry Using Subcircuits You can include IF ELSEIF ELSE statements in subcircuits and subcircuits in IF ELSEIF ELSE statements You can use IF ELSEIF ELSE blocks to select different submodules to structure the netlist using INC LIB and VEC statements f two or more models in an IF ELSE block have the same model name and model type they must also be the same revision level Parameters in an IF ELSE block do not affect
261. files 279 SNOSC 236 SNXF 291 source statements 72 sqrt x function 101 square root function 101 Ss file 310 starting hspicerf 9 statement 313 314 ENV 312 HBOSC 229 statements AC 353 CHECK EDGE 413 CHECK FALL 413 CHECK GLOBAL_LEVEL 412 CHECK HOLD 414 CHECK IRDROP 415 CHECK RISE 412 CHECK SETUP 414 CHECK SLEW 412 DATA 81 DC 353 element 72 Index ENDL 81 HBXF 287 LIB 81 LPRINT 395 MODEL 353 PARAM 82 POWER 417 POWERDC 415 PTDNOISE 281 source 72 SURGE 418 TEMP 80 353 354 TRAN 353 417 statistical analysis 354 380 statistics calculations 351 steady state time domain analysis Shooting Newton 219 strobejitter 286 subcircuit probing currents 405 subcircuits calling tree 79 changing in ALTER blocks 82 84 creating reusable circuits 87 hierarchical parameters 88 library structure 93 multiplying 89 node names 78 path names 79 PRINT and PLOT statements 90 SURGE statement 418 T tabulated data output 393 Taguchi analysis 350 tan x function 100 tanh x function 101 TEMP model parameter 80 353 TEMP statement 353 354 temper variable 104 temperature circuit 351 353 coefficients 115 derating 80 353 element 353 reference 80 353 429 Index variable 104 Temperature Variation Analysis 350 time domain steady state analysis 219 time variable 104 title for simulation 71 TITLE statement 71 TNOM option 80 353 TOX model parameter 355 TRAN stateme
262. find the average value of out_var Similarly you can replace lt RMS gt with lt MIN gt lt MAX gt or lt PP gt to find the result of min max or pp m integral evaluation MEASURE PHASENOISE result INTEGRAL out _ var lt FROM IFB1 gt lt TO IFB2 gt This measurement integrates the out_var value from the IFB1 frequency to the IFB2 frequency derivative evaluation MEASURE PHASENOISE result DERIVATIVE out_var AT IFB1 This measurement finds the derivative of out_var at the IFB1 frequency point 244 HSPICE RF User Guide Z 2007 03 Chapter 9 Oscillator and Phase Noise Analysis Phase Noise Analysis Note MEASURE PHASENOISE cannot contain an expression that uses an phasenoise variable as an argument You also cannot use MEASURE PHASENOTSE for error measurement and expression evaluation of PHASENOISE The HSPICE RF optimization flow can read the measured data from a MEASURE PHASENOISE analysis This flow can be combined in the HSPICE RF optimization routine with a MEASURE HBTR analysis see Using MEASURE with HB Analyses on page 212 anda MEASURE HBNOISE analysis see Measuring HBNOISE Analyses with MEASURE on page 273 PHASENOISE Output Syntax HSPICE RF supports the output of the phase noise as well as the phase noise due to a specified element In addition you can output phase noise due to the specified noise source types In addition you can use specialized keywords to output phase no
263. for a corner point verification run Instead of back annotating all RC parasitics into the ideal netlist the selective post layout flow automatically detects and back annotates only active parasitics into the hierarchical LVS ideal netlist For a high latency design the selective post layout flow is an order of magnitude faster than the standard post layout flow HSPICE RF User Guide Z 2007 03 Chapter 13 Post Layout Analysis Post Layout Back Annotation Note The selective post layout flow applies only to RF transient analyses and cannot be used with other analyses such as DC AC or HB Selective Post Layout Flow Control Options To invoke the selective post layout flow include one of the options listed in Table 23 in your netlist Table 23 Selective Post Layout Flow Options Syntax Description SIM_DSPF_ACTIVE HSPICE RF performs a preliminary verification run to or determine the activity of the nodes and generates two ASCII SIM_SPEF_ACTIVE files active_node rc and active_node rcxt These files save all active node information in both Star RC format and Star RCXT format By default a node is considered active if the voltage varies by more than 0 1V To change this value use the SIM_DSPF_VTOL or SIM_SPEF_VTOL option For descriptions and usage examples see OPTION SIM_DSPF_ACTIVE and OPTION SIM_SPEF_ACTIVE in the HSPICE and RF Command Reference SIM_DSPF_VTOL HSPICE RF performs a second simulation run by us
264. for a realistic semiconductor technology HSPICE RF User Guide 385 Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Simulating the Effects of Global and Local Variations with Monte Carlo 386 Variations Specified Using DEV and LOT The two limitations of the approach described in section Variations Specified on Model Parameters are resolved in this method by specifying global and local variations directly on a model parameter with the syntax parameterName parameterValue LOT distribution LotDist DEV distribution DevDist Where LOT keyword for global distribution DEV keyword for local distribution distribution is as explained in section Variations Specified on Geometrical Instance Parameters LotDist DevDist characteristic number for the distribution 3 sigma value for Gaussian distributions test12 sp has large global and small local variation similar to the setup in the file test3 sp The result shows four different curves with a large common part and small separate parts The amount of variation defined in the two files is the same The curves look different from the test3 sp results because different random sequences are used However the statistical results sigma converge for a large number of samples There is no option available to select only local or only global variations This can be an obstacle if the file is read only or encrypted Combinations of Variation Specifications Specifying distri
265. for the input signal Also referred to as the input frequency band IFB or fin You can specify LIN DEC OCT POI SWEEPBLOCK DATA MONTE or OPTIMIZE sweeps Specify the nsteps start and stop frequencies using the following syntax for each type of sweep LIN nsteps start stop DEC nsteps start stop OCT nsteps start stop POI nsteps freq_values SWEEPBLOCK nsteps freq freq2 freqn HSPICE RF User Guide Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Shooting Newton Noise Analysis SNNOISE Parameter Description n1 1 listfreq listcount HSPICE RF User Guide Z 2007 03 Index term defining the output frequency band OFB or fout at which the noise is evaluated Generally fout ABS n1 f1 fin Where f1 is the fundamental harmonic tone determined in the Shooting Newton analysis ni is the associated harmonic multiplier n1 n2 nk are the associated harmonic multipliers n1 can be any non negative integer lt nharm defined in the SN statement 1 is fixed either 1 or 1 fin is the IFB defined by parameter_sweep The default index term is 1 1 For a single tone analysis the default mode is consistent with simulating a low side down conversion mixer where the RF signal is specified by the IFB and the noise is measured at a down converted frequency that the OFB specifies In general you can use the n1 1 index term to specify an
266. for this example is located in the following directory installdir demo hspicerf examples The analysis finds the oscillation frequency of the ring oscillator Since the circuit is an oscillator no input source is required The oscillator is started by setting an initial condition at the input of the ring node 1 In the SNOSCc command the node that the analysis will probe for oscillation conditions is specified as well as the approximate frequency of oscillation The number of harmonics to include in the analysis is specified as well 50 HSPICE RF User Guide Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 7 Using Shooting Newton Analysis on a Driven Phase Frequency Circuit and a Ring Oscillator The phase noise characteristics of the oscillator are analyzed by using the PHASENOISE command The PHASENOISE command requires that an output node pair of nodes or a two port element and a frequency sweep be specified The frequency sweep is used to calculate the phase noise analysis at the specified offset frequencies measured from the oscillation carrier frequency For this example phase noise analysis the default Nonlinear Perturbation NLP method is used The signals v 7 will be probed in both the frequency and time domain The measure statement is used to measure the fundamental frequency of the oscillator Simulation Status Output During the simulation the simulation status is displayed on the screen In addition to
267. frequency dependent data For details see Using Instruments with ADS by Agilent Technologies http www agilent com One of the following parameter types S scattering default Y admittance Z impedance 149 Chapter 6 Testbench Elements Multi Terminal Linear Elements Parameter Specifies Zo Characteristic impedance value of the reference line frequency independent For multi terminal lines N gt 1 HSPICE assumes that the characteristic impedance matrix of the reference lines are diagonal and their diagonal values are set to Zo You can also set a vector value for non uniform diagonal values Use Zof to specify more general types of a reference line system The default is 50 FBASE Base frequency to use for transient analysis This value becomes the base frequency point for Inverse Fast Fourier Transformation IFFT If you do not set this value the base frequency is a reciprocal value of the transient period Ifyou set a frequency that is smaller than the reciprocal value of the transient then the transient analysis performs circular convolution and uses the reciprocal value of FBASE as its base period FMAX Maximum frequency for transient analysis Used as the maximum frequency point for Inverse Fast Fourier Transform IFFT LOWPASS Specifies low frequency extrapolation 0 Use zero in Y dimension open circuit 1 Use lowest frequency default 2 Use linear extrapolation with the lowest
268. frequency detector The time domain signals v cfin v fin v pu and v pdn and v lfin are be probed The gain of the phase frequency detector can be found by probing the frequency domain value of v Ifin at DC frequency indices 0 Phase Frequency Detector global vdd gnd options wl post DC sources vsup vdd 0 DC 1 0 Reference signal sine wave vref xin gnd DC 0 sin 0 5 0 5 0 5g 0 0 0 0 0 0 Input buffers square up Ref sine wave xfinl xin finl inv xfin2 finl FIN inv3 Compare Signal sine wave vcRef cin gnd DC 0 sin 0 5 0 5 0 5g 0 0 0 0 phase phase shift Input buffers Square up compare sine wave xcfinl cin cfinl inv xcfin2 cfinl cFIN inv3 Phase frequency detector xPFD CFIN FIN pdn pu phasedet Chargepump xCP LFIN Ibias pdn pu chargepump Bias voltage vIbias Ibias gnd 0 15 Sets charge pump bias Rload LFIN 0 10Meg Harmonic Balance Test Bench param NH 127 param phase 0 0 phase shift in degrees Opt snaccuracy 30 SN tone 0 5g nharms NH SWEEP phase POI 5 0 0 22 5 45 0 67 5 90 0 SNNOISE V LFIN Vref DEC 21 100 10MEG offset frequency sweep 0 1 Take low frequency noise HSPICE RF User Guide 45 Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 7 Using Shooting Newton Analysis on a Driven Phase Frequency Circuit and a Ring Oscillator probe sn v fin v cfin v pu v pdn v lfin i vibias print snfd v lfin v lfin 0 probe snfd v lfin v lfin
269. g from all frequencies Load Noise originating from the OFB Input Source Noise originating from the IFB and from the image of IFB Output Data Files An HBLIN analysis produces these output data files The S parameters from the PRINT statement are written to a printhl file The extracted S parameters from the PROBE statement are written to a hl file Large Signal S parameter HBLSP Analysis An HBLSP analysis provides three kinds of analyses for periodically driven nonlinear circuits such as those that employ power amplifiers and filters Two port power dependant large signal S parameter extraction Two port small signal S parameter extraction Two port small signal noise parameter calculation Unlike small signal S parameters which are based on linear analysis power dependent S parameters are based on harmonic balance simulation Its solution accounts for nonlinear effects such as compression and variation in power levels HSPICE RF User Guide 305 Z 2007 03 Chapter 11 S parameter Extraction Large Signal S parameter HBLSP Analysis 306 The definition for power dependent S parameters is similar to that for small signal parameters Power dependent S parameters are defined as the ratio of reflected and incident waves by using this equation b S a Sfi jJ bli n alj zn when a k n k j 0 The incident waves ali n and reflected waves bli n are defined by using these equations
270. g time domain description Il 1mA 3mA cos 2 pi 1 e8 t Example 2 This example shows HB sources used for a two tone analysis HSPICE RF User Guide Z 2007 03 171 Chapter 6 Testbench Elements Steady State HB Sources hb tones 1 e9 1 1e9 intmodmax 5 Vin lo 0 dc 0 hb 1 5 90 11 Vrf rf O dc 0 hb 0 201 2 These sources have the following time domain descriptions Vin 1 5 cos 2 pi 1l e9 t 90 pi 180 V Vrf 0 2 cos 2 pi 1 1e9 t V Example 3 The following HB source uses a modtone and modharms hb tones 2 e9 1 9e9 harms 5 5 Vm input gnd dc 0 5 hb 0 2 0 1 1 1 2 Vm has the following time domain description Vm 0 5 cos 2 pi 1 e8 t Example 4 This example uses an HB source specified with a SIN source and HBTRANINIT hb tone 1 e8 harms 7 Vt 1 2 SIN 0 1 1 0 2 e8 0 0 90 tranforhb 1 Vt is converted to the following HB source Vt 1 2 dc 0 1 hb 1 00 02 1 Example 5 This example shows a power source the units are Watts hb tones 1 1le9 harms 9 Pt Input Gnd power 1 Z0 50 1m 0 1 1 Pt delivers 1 mW of power through a 50 ohm impedance Steady State HB Sources 172 The fundamental frequencies used with harmonic balance analysis are specified with the HB TONES command These frequencies can then be referenced by their integer indices when specifying steady state signal sources For example the HB specification given by the following line HSPICE RF User Guide Z 2007 03 Chapter 6 Testben
271. g2 expression2 off Parameter Description funcname Specifies the function name This parameter must be distinct from array names and built in functions In subsequently defined functions all embedded functions must be previously defined arg1 arg2 Specifies variables used in the expression off Voids all user defined functions Example PARAM f a b POW a 2 a b g d SQRT d h e e f 1 2 g 3 HSPICE RF User Guide Z 2007 03 Chapter 5 Parameters and Functions Using Algebraic Expressions Predefined Analysis Function HSPICE includes specialized analysis types such as Optimization and Monte Carlo that require a way to control the analysis Measurement Parameters MEASURE statements produce a measurement parameter The rules for measurement parameters are the same as for standard parameters except that measurement parameters are defined in a MEASURE statement not in a PARAM statement PRINT and PROBE Parameters PRINT and PROBE statements in HSPICE produce a print parameter The rules for print parameters are the same as the rules for standard parameters except that you define the parameter directly in a PRINT or PROBE statement not ina PARAM statement Using Algebraic Expressions HSPICE Z 2007 03 Note Synopsys HSPICE uses double precision numbers 15 digits for expressions user defined parameters and sweep variables For better precision use parameters instead of constants i
272. get value various Returns a value for a specified parameter of a parameter specified element For example val rload temp returns the value of the temp temperature parameter for the rload element val model_type get value various model_name model_param Returns a value for a specified parameter of a specified model of a specific type For example val nmos mos1 rs returns the value of the rs parameter for the mos model which is an nmos model type lv lt Element gt element various or templates Ix lt Element gt 102 Returns various element values during simulation HSPICE RF User Guide Z 2007 03 Chapter 5 Parameters and Functions Built In Functions and Variables Table 12 Synopsys HSPICE Built in Functions Continued HSPICE Form Function Class Description v lt Node gt circuit various Returns various circuit values during simulation i lt Element gt output variables cond x y ternary Returns x if cond is not zero Otherwise returns y operator param z condition x y lt relational Returns 1 if the left operand is less than the right operator operand Otherwise returns 0 less than para x y lt z y less than z lt relational Returns 1 if the left operand is less than or equal to operator the right operand Otherwise returns 0 less than or para x y lt z y less than or equal to z equal gt relational Returns 1 if the left operand is great
273. gh the Table function For example gname nodel node2 noise Table argl f1 vl1 f2 v2 yen The f1 v1 f2 V2 parameters describe the noise table When arg1 f1 the function returns v1 The arg1 can be an expression of either HERTZ bias or both For example arg1 HERTZ 1 0E 3 The noise table can be input through a DATA structure DATA d1 x Yy f1 v1 f2 v2 ENDDATA gname nodel node2 noise TABLE arg1 dl1 The x y parameters in the DATA structure are two placeholder strings that can be set to whatever you prefer even if they are in conflict with other parameters in the netlist The arg1 parameter can be an expression of HERTZ and bias When arg1 f2 the function will return v2 Power Supply Current and Voltage Noise Sources You can implement the power supply noise source with G and E elements The G element for the current noise source and the E element for the voltage noise source As noise elements they are two terminal elements that represent a noise source connected between two specified nodes HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Behavioral Noise Sources Syntax Expression form Gxxx nodel node2 noise expression EXXX nodel node2 noise expression The G noise element represents a noise current source and the E noise element represents a noise voltage source The xxx parameter can be set with a value up to 1024 characters The node7 and
274. gin with a letter designating the element type followed by up to 1023 alphanumeric characters Element type letters are R for resistor C for capacitor M fora MOSFET device and so on see Element and Source Statements on page 72 Subcircuit Node Names To assign the first name HSPICE or HSPICE RF uses the extension to concatenate the circuit path name with the node name for example X1 XBIAS M5 HSPICE RF User Guide Z 2007 03 Chapter 4 Input Netlist and Data Entry Input Netlist File Composition Node designations that start with the same number followed by any letter are the same node For example 1c and 1d are the same node To indicate the ground node use either the number 0 the name GND or GND Every node should have at least two connections except for transmission line nodes unterminated transmission lines are permitted and MOSFET substrate nodes which have two internal connections Floating power supply nodes are terminated with a 1Megohm resistor and a warning message Path Names of Subcircuit Nodes A path name consists of a sequence of subcircuit names starting at the highest level subcircuit call and ending at an element or bottom level node Periods separate the subcircuit names in the path name The maximum length of the path name including the node name is 1024 characters You can use path names in PRINT NODESET and IC statements as another way to reference internal nodes nodes not appea
275. given source in the circuit to a designated output Frequency conversion is calculated from the input frequencies to a single output frequency that is specified with the command The relationship between the HBXF command and the input output is expressed in the following equation HSPICE RF User Guide 287 Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Multitone Harmonic Balance Transfer Function Analysis HBXF 288 Equation 61 Y ja SABXF JQ j AW X j Ao W Where HBXF O AA is the transfer function from input port n to the output port m Wis the set of all possible harmonics Aq is the input frequency Aw is the offset frequency mis the output node number nis the input node number is the output frequency Y is the output voltage or current m X is the input voltage or current Supported Features The HBXF command supports the following features All existing HSPICE RF models and elements Sweep parameter analysis Unlimited number of HB sources Prerequisites and Limitations The following prerequisites and limitations apply to the HBXF command Only one HBXF statement is required If you use multiple HBXF statements HSPICE RF only uses the last HBXF statement Atleast one HB statement is required which determines the steady state solution Parameter sweeps must be placed in HB statements HSPICE RF User G
276. grees VPD Phase in degrees VPR Phase in radians VDB GB units VDBM dB relative to 1 mV HSPICE RF User Guide Z 2007 03 Chapter 8 Steady State Shooting Newton Analysis SN Analysis Output Current type current magnitude and phase in degrees IR real component imaginary component IM magnitude P Phase in degrees IPD Phase in degrees IPR Phase in radians IDB dB units IDBM dB relative to 1 mV Power type P Frequency type hertz index hertz indexl index2 You must specify the harmonic index for the hertz variable The frequency of the specified harmonics is dumped Parameter Description NODES ELEM can be any of the following Voltage type a single node name n1 or a pair of node names n1 n2 Current type an element name elemname Power type a resistor resistorname or port portname element name INDEX n1 is the harmonic index of the SNFD tone Index is limited to the single tone associated with the SN analysis Output Files The time domain data are output to printsnO and sn0 files Frequency domain data are output to printsnfO and snf0 files Output Format The format for time domain output is the same as standard transient analysis For frequency domain output the format is similar to HB The main difference is that Shooting Newton output in the frequency domain is single tone only The r
277. gt gt lt gt lt PERJITTER val SEED val gt gt Ixxx n n PU lt LSE gt lt svl v2 lt td lt tr lt tf lt pw lt per gt gt gt gt gt lt gt lt PERJITTER val SEED val gt gt The effect of jitter on the PULSE source results in random shifts of the rise and fall transitions that takes place at RISE edge td n Ty Std trt n To FALL edge td pwtn To d pw tf n To The jitter effect is equivalent to introducing random shifts in the period 7 consistent with the 1st order jitter model based on Period Jitter A Gaussian random number generator computes the time deviation x t after each leading edge of the clock sources For flexibility the SEED parameter integer is supported for generating different random number sequences when different SEED integers are used for initialization SEED does not set a fixed time deviation It only changes the sequence of random samples By HSPICE Monte Carlo convention the default value for SEED is 1 An interpretation of PERJITTER is to view it as causing each period of the PULSE SIN COS to be a random variable Tj where period T will have a Gaussian distribution about the mean given period value of T The standard deviation of this Gaussian is the PERJITTER value it is considered RMS period jitter which results in a bell curve distribution centered about period 7 HSPICE RF User Guide 193 Z 2007 03 Chapter 6 Testbench Elements Clock Source with Random Jitter
278. h MEASURE Errors and Warnings HBNOISE Example Shooting Newton Noise Analysis Supported Features Input Syntax Output Syntax Output Data Files SNNOISE te csnccceasaodvsteden cous Measuring SNNOISE Analyses with MEASURE SNNOISE Analysis Example 248 248 251 251 255 256 257 257 258 258 259 261 262 263 263 264 264 266 267 267 268 269 269 271 272 273 274 274 275 275 276 278 279 279 280 11 12 Contents Periodic Time Dependent Noise Analysis PTDNOISE PTDNOISE Input Syntax 0 00 cee PTDNOISE Output Syntax and File Format 0 4 Error Handling and Warnings 00 0 e eee eee eee Multitone Harmonic Balance Transfer Function Analysis HBXF Supported FeatureS 0 0 0 cee eens Input Syntax Output Syntax Output Data FileSi 2 3 ane whi ne ae alaa date en wiht eee ae ee Example HBXF Test Listi FIG ree cocci t A Stace cen AAEE Shooting Newton Transfer Function Analysis SNXF 05 Input Syntax Output Syntax Output Data Files 1 2 0 000 0 c cece Example SNXF Test Listi References NG incendie ar sseckd A deed Aarti ete as eae eh vance de ate deg S parameter Extraction 0 00 Frequency Translation S Parameter HBLIN Extraction HB
279. h the same names The scope of a lower level parameter is local to the subcircuit where you define the parameter but global to all subcircuits that HSPICE RF User Guide 105 Z 2007 03 Chapter 5 Parameters and Functions Parameter Scoping and Passing 106 are even lower in the design hierarchy Local scoping rules in HSPICE prevent higher level parameters from overriding lower level parameters of the same name when that is not desired Reusing Cells Parameter name problems also occur if different groups collaborate on a design Global parameters prevail over local parameters so all circuit designers must know the names of all parameters even those used in sections of the design for which they are not responsible This can lead to a large investment in standard libraries To avoid this situation use local parameter scoping to encapsulate all information about a section of a design within that section Creating Parameters in a Library To ensure that the input netlist includes critical user supplied parameters when you run simulation you can use illegal defaults that is defaults that cause the simulator to abort if you do not supply overrides for the defaults If a library cell includes illegal defaults you must provide a value for each instance of those cells If you do not the simulation aborts For example you might define a default MOSFET width of 0 0 HSPICE aborts because MOSFET models require this parame
280. harms nh Default 4 nh This option is relevant only if you set OPTION HBTRANINIT HBTRANSTEP lt stepsize gt lt stepsize gt specifies the step size for the transient analysis The default is 1 4 nh f0 where nh is the nharms value and f0 is the oscillation frequency This option is relevant only if you set OPTION HBTRANINIT HBTRANFREQSEARCH lt 1 0 gt If HBTRANFREQSEARCH 1 default then HB analysis calculates the oscillation frequency from the transient analysis Otherwise HB analysis assumes that the period is 1 f where fis the frequency specified in the tones description Note You can specify either OPTION HBTRANPTS or OPTION HBTRANSTEP but not both You must also either specify the initial conditions or add a PWL or PULSE source to start the oscillator for transient analysis This source should provide a 234 HSPICE RF User Guide Z 2007 03 Chapter 9 Oscillator and Phase Noise Analysis HBOSC Analysis Using Transient Initialization brief stimulus and then return to zero HB analysis effectively ignores this type of source treating it as zero valued This method does the following 1 If HBTRANFREQSEARCH 1 transient analysis runs for several periods attempting to determine the oscillation frequency from the probe voltage signal Transient analysis continues until the time specified in HBTRANINIT Stores the values of all state variables over the last period of the transient analysis
281. hat you specify as a list The syntax is variable poi npoints pl p2 Data sweeps a DATA statement identifies the swept variables and their values The syntax is data dataname You can use the SWEEPBLOCK feature to combine linear logarithmic and point sweeps which creates more complicated sets of values over which a variable is swept The TRAN AC DC and HB commands can specify SWEEPBLOCK blockname as a sweep instead of LIN DEC OCT and so forth Also you can use SWEEPBLOCKX for frequency sweeps with the AC HBAC PHASENOISE and HBNOISE commands HSPICE RF User Guide 189 Z 2007 03 Chapter 6 Testbench Elements SWEEPBLOCK in Sweep Analyses 190 All commands that can use SWEEPBLOCK must refer to the SWEEPBLOCK sweep type In addition you must specify SWEEPBLOCK as one of the syntax types allowed for frequency sweeps with the HBAC PHASENOISE and HBNOISE commands Input Syntax The SWEEPBLOCK feature creates a sweep whose set of values is the union of a set of linear logarithmic and point sweeps To specify the set of values in the SWEEPBLOCK use the SWEEPBLOCK command This command also assigns a name to the SWEEPBLOCK For example SWEEPBLOCK swblockname sweepspec Sweepspec sweepspec You can use SWEEPBLOCK to specify DC sweeps parameter sweeps AC and HBAC frequency sweeps or wherever HSPICE accepts sweeps You can specify an unlimited number of sweepspec paramete
282. he AvanWaves tool which you can use to display waveforms generated during HSPICE circuit design simulation HSPICE RF User Guide XV About This Guide Searching Across the HSPICE Documentation Set Searching Across the HSPICE Documentation Set You can access the PDF format documentation from your install directory for the current release by entering docs on the terminal command line when the HSPICE tool is open Synopsys includes an index with your HSPICE documentation that lets you search the entire HSPICE documentation set for a particular topic or keyword In a single operation you can instantly generate a list of hits that are hyper linked to the occurrences of your search term For information on how to perform searches across multiple PDF documents see the HSPICE release notes available on SolvNet at http solvnet synopsys com ReleaseNotes or the Adobe Reader online help Note To use this feature the HSPICE documentation files the Index directory and the index pdx file must reside in the same directory This is the default installation for Synopsys documentation Also Adobe Acrobat must be invoked as a standalone application rather than as a plug in to your web browser You can also invoke HSPICE and RF documentation in a browser based help system by entering help on your terminal command line when the HSPICE tool is open This provides access to all the HSPICE manuals with the expection of the AvanWaves User
283. he SPEF file defines all external loads if any If you do not specify this flow type then some or all external loads are not defined in this SPEF file If HSPICE RF cannot find external load data outside the SPEF file it reports an error EXTERNAL_SLEWS The SPEF file defines all external slews if any If you do not specify this flow type then some or all external slews are not defined in this SPEF file lf HSPICE RF cannot find external slew data outside the SPEF file it reports an error FULL_CONNECTIVITY A SPEF file defines all net connectivity If you do not specify this flow type then some or all net connectivity is not defined in this SPEF file If HSPICE RF cannot find connectivity data outside the SPEF file it issues an error This flow does not look for presence or absence of power and ground nets or any other nets that do not correspond to the logical netlist If a SPEC file includes FULL_CONNECTIVITY and MISSING_NETS HSPICE RF reports an error MISSING_NETS If any logical nets are not defined in the netlist HSPICE RF merges missing parasitic data from another source If it does not find another source HSPICE RF rereads the netlist and estimates the missing parasitics This flow does not look for presence or absence of power and ground nets or any other nets that do not correspond to the logical netlist If you use FULL_CONNECTIVITY and MISSING_NETS in the same SPEF file HSPICE RF reports an error NETLIST_TYPE_VERILOG NET
284. he active elements diodes and transistors Diode Element Geometric LEVEL 1 or Non Geometric LEVEL 3 form Dxxx nplus nminus mname lt lt AREA gt area gt lt lt PJ gt val gt lt WP val gt lt LP val gt lt WM val gt lt LM val gt lt OFF gt lt IC vd gt lt M val gt lt DTEMP val gt Dxxx nplus nminus mname lt W width gt lt L length gt lt WP val gt HSPICE RF User Guide 159 Z 2007 03 Chapter 6 Testbench Elements Active Elements 160 lt LP val gt lt WM val gt lt LM val gt lt OFF gt lt IC vd gt lt M vals gt lt DTEMP vals gt Fowler Nordheim LEVEL 2 form Dxxx nplus nminus mname lt W val lt L val gt gt lt WP val gt lt OFF gt lt IC vd gt lt M val gt Parameter Description Dxxx Diode element name Must begin with D followed by up to 1023 alphanumeric characters nplus Positive terminal anode node name The series resistor for the equivalent circuit is attached to this terminal nminus Negative terminal cathode node name mname Diode model name reference AREA Area of the diode unitless for LEVEL 1 diode and square meters for LEVEL 8 diode This affects saturation currents capacitances and resistances diode model parameters are IK IKR JS CUO and RS The SCALE option does not affect the area factor for the LEVEL 1 diode Default 1 0 Overrides AREA from the diode model If you do not specify the AREA HSPICE or HSPICE RF calculates it from the
285. he sample netlist for this example in the following directory installdir demo hspice apps rc_monte sp HSPICE RF User Guide Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Monte Carlo Analysis Figure 37 Monte Carlo Analysis of RC Time Constant 992 750N 900 0N 800 0N 700 0N VOLT LIN 600 0N 500 0N 400 0N 300 0N FILE NOM1 SP WITH UNIFORM DISTRIBUTION May 15 2003 12 38 49 ee Va rf fl J a FO MONT1 SVO 1 SS a k z Liari lics r Lisi clii ti 10 1 Ea l ai rose ba I 1 t d 1 200 0N 400 0N 600 0N 800 0N 1 0 TIME LIN Switched Capacitor Filter Design Capacitors used in switched capacitor filters consist of parallel connections of a basic cell Use Monte Carlo techniques in HSPICE RF to estimate the variation in total capacitance The capacitance calculation uses two distributions Minor element distribution of cell capacitance from cell to cell on a single die Major model distribution of the capacitance from wafer to wafer or from manufacturing run to run HSPICE RF User Guide 369 Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Monte Carlo Analysis 370 Figure 38 Monte Carlo Distribution cap to cap element KE Cla Cib Cla Cib Cid Cic Cid Ko S run to run model You can approach this problem from phy
286. he work directory If the path starts from the work directory HSPICE can also find the library but it prints a warning The depth of nested calls is limited only by the constraints of your system configuration A library cannot contain a call to a library of its own entry name within the same library file HSPICE RF User Guide 81 Z 2007 03 Chapter 4 Input Netlist and Data Entry Input Netlist File Composition 82 AHSPICE FF library cannot contain the END statement ALTER processing cannot change LIB statements within a file that an INCLUDE statement calls Defining Parameters The PARAM statement defines parameters Parameters in HSPICE or HSPICE RF are names that have associated numeric values You can also use either of the following specialized methods to define parameters Predefined Analysis Measurement Parameters Predefined Analysis HSPICE RF provides several specialized analysis types which require a way to control the analysis For the syntax used in these PARAM commands see the description of the PARAM command in the HSPICE and HSPICE RF Command Reference HSPICE RF supports the following predefined analysis parameters Temperature functions fn Optimization guess range m Monte Carlo functions HSPICE RF does not support frequency time Measurement Parameters A MEASURE statement produces a measurement parameter In general the rules for measurement paramete
287. his parameter specifies the temperature in C at which HSPICE RF measures and extracts the model parameters Set the value of TNOM in an OPTION statement Its default value is 25 C Circuit temperature that you specify using a TEMP statement or the TEMP parameter This is the temperature in C at which HSPICE RF simulates all elements To modify the temperature for a particular element use the DTEMP parameter The default circuit temperature is the value of TNOM Individual element temperature which is the circuit temperature plus an optional amount that you specify in the DTEMP parameter To specify the temperature of a circuit in a simulation run use either the TEMP statement or the TEMP parameter in the DC AC or TRAN statements HSPICE RF compares the circuit simulation temperature that one of these statements sets against the reference temperature that the TNOM option sets TNOM defaults to 25 C unless you use the SPICE option which defaults to 27 C To calculate the derating of component values and model parameters HSPICE RF uses the difference between the circuit simulation temperature and the TNoM reference temperature Elements and models within a circuit can operate at different temperatures For example a high speed input output buffer that switches at 50 MHz is much hotter than a low drive NAND gate that switches at 1 MHz To simulate this temperature difference specify both an element temperature par
288. iN gt i iN 4 vrei UN vreflN 4 ndi o _ o ndN v 1 VIN ndR reference node S Model Syntax MODEL Smodel_name S 148 HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Multi Terminal Linear Elements lt N dimension gt FOMODEL sp_model_name TSTONEFILE filename CITIFILE filename lt TYPE s y gt lt Zo value vector value gt lt FBASE base_frequency gt lt FMAX maximum_frequency gt lt HIGHPASS 0 1 2 gt lt LOWPASS 0 1 2 gt lt PRECFAC val gt lt DELAYHANDLE 1 0 ON OFF gt lt DELAYFREQ val gt lt MIXEDMODE 0 1 gt lt DATATYPE data_string gt lt XLINELENGTH val gt Parameter Specifies Smodel_name S N FQMODEL TSTONEFILE CITIFILE TYPE HSPICE RF User Guide Z 2007 03 Name of the S model Specifies that the model type is an S model S model dimension which is the terminal number of the S element excluding the reference node Frequency behavior of the S Y or Z parameters MODEL statement of SP type which defines the frequency dependent matrices array Name of a Touchstone file Data contains frequency dependent array of matrixes Touchstone files must follow the sp file extension rule where represents the dimension of the network For details see Touchstone File Format Specification by the EIA IBIS Open Forum http www eda org Name of the CITIfile which is a data file that contains
289. ial ALTER title identifies the run ALTER processing cannot revise LIB statements within a file that an INCLUDE statement calls However ALTER processing can accept INCLUDE statements within a file thata LIB statement calls Using Multiple ALTER Blocks This section does not apply to HSPICE RF HSPICE RF User Guide Z 2007 03 Chapter 4 Input Netlist and Data Entry Input Netlist File Composition For the first simulation run HSPICE reads the input file up to the first ALTER statement and performs the analyses up to that ALTER statement After it completes the first simulation HSPICE reads the input between the first ALTER statement and either the next ALTER statement or the END statement _HSPICE then uses these statements to modify the input netlist file HSPICE then resimulates the circuit For each additional ALTER statement HSPICE performs the simulation that precedes the first ALTER statement HSPICE then performs another simulation using the input between the current ALTER statement and either the next ALTER statement or the END statement If you do not want to rerun the simulation that precedes the first ALTER statement every time you run an ALTER simulation then do the following 1 Put the statements that precede the first ALTER statement into a library 2 Usethe LIB statement in the main input file 3 Puta DEL LIB statement inthe ALTER section
290. ialization see HB Simulation of Ring Oscillators on page 233 for more information NHARMS Number of harmonics to use for oscillator HB analysis PROBENODE Circuit nodes that are probed for oscillation conditions HBOSCVPROBE VP FSPTS HSPICE RF User Guide Z 2007 03 N1 and N2 are the positive and negative nodes for a voltage probe inserted in the circuit to search for oscillation conditions VP is the initial probe voltage value one half the supply voltage is a suggested value The phase of the probe voltage is forced to zero all other phases are relative to the probe phase HSPICE RF uses this probe to calculate small signal admittance for the initial frequency estimates It should be connected near the heart of the oscillator near resonators inside the ring of a ring oscillator etc Note The PROBENODE pins and approximate voltage value can also be set by using a zero amp current source that uses the HBOSCVPROBE keyword Sets PROBENODE with a current source If a current source with HBOSCVPROBE is used the PROBENODE syntax is not necessary Specifies the frequency search points that HSPICE RF uses in its initial small signal frequency search to find an oscillation frequency Optional but recommended for high Q and most LC oscillators If the circuit is a ring oscillator see HB Simulation of Ring Oscillators on page 233 for more information on how to use the HBTRANINIT option NUM is an integer MIN and
291. iations that affect devices in a different way test3 sp has two kinds of distributions defined globw globwidth as in the first example and locwidth as in the second example The sum of the two is used to define the width of the resistors Therefore the resistors will always have different widths a common variation due to globwidth and a separate variation due to locwidth In the example the distribution for locwidth was chosen as narrower than for globwidth When overlaying the measurement results the large common variation can easily be seen however all voltages are different In summary each reference to a parameter with a specified distribution causes a new random variable to be generated for each Monte Carlo sample When referencing the parameter on an instance the effect of a local variation is created When referencing the parameter on an expression for a second parameter and using the second parameter on an instance then the effect of a global variation is created Variations Specified in the Context of Subcircuits The concept explained in the previous section applies also to subcircuits as instances and instances within subcircuits Here we again use the example of a physical resistor with variation of its width test4 sp uses a subcircuit for each resistor instead of the top level resistors in test8 sp On each subcircuit a parameter width is assigned a value by an expression which is the same for all of them Thi
292. ich is 1 GHz HB source for RF add HB 0 001 24 1 2 to the RF voltage source this sets the amplitude to 0 001 24 degrees phase shift for the first harmonic of the second tone 0 8 GHz An HB command specifying both tones hb tones 1g 0 8g nharms 6 3 only a small number of harmonics is required to resolve the signals The complete mix_hb sp netlist for 2 tone HB analysis is then HSPICE RF User Guide 39 Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 6 Using Multi Tone HB and HBAC Analyses for a Mixer 40 Ideal mixer example 2 tone HB analysis OPTIONS POST vlo lo 0 1 0 sin 1 0 0 5 1 0g 0 O 90 HB 0 5011 FEEL rfl rf 1 0 gl 0 if cur 1 0 v lo v rf mixer element cl 0 if gq 1 0e 9 v lo v rf mixer element rout if ifg 1 0 vetrl ifg 0 0 0 hl out 0 vetrl 1 0 convert I to V rhi out 0 1 0 vrf rfl 0 sin 0 0 001 0 8GHz 0 0 114 HB 0 001 24 1 2 opt sim_accuracy 100 hb tones 1lg 0 8g nharms 6 3 end This example is available in directory lt installdir gt demo hspicerf examples HBAC Approach To analyze this circuit using HBAC start with the 2 tone HB analysis setup and modify it as follows Replace the RF HB signal with an HBAC signal change HB 0 001 24 1 2 tO yBac 0 001 24 this deactivates the source for HB and activates it for HBAC with the same magnitude and phase Specify the frequency in the HBAC command Change the HB command to single tone HB tones 1g nharms
293. ier coefficients and frequencies You can recognize this swept parameter in the ev file by the keyword env_time Each row in the tabulated data of an ev file includes values for identifying frequency information the complex data for the output variables and information on the envelope time sweep For example the header for a data file HSPICE RF User Guide 315 Z 2007 03 Chapter 12 Envelope Analysis Envelope Simulation 316 dump for output variables v in and v out that follow a 2 tone envelope analysis have entries for hertz v in v out no 0 ni 1 sweep env_time S amp H Which result in data blocks with floating point values following env_time 0 f 0 a 0 v in fO nl fi 1 al1 v in fo nl fi N aN v in fO nl f1 env_time 1 0 alo v in fo nil 1 1 alipivtiny fo nil f1 N a N v in fO nl f1 env_time M 1 0 a 0 v in fO nl f1 f 1 all v in fO nil f1 N a N v in fO nl fi b 0 b 1 b N a 1 a N a 0 a 1 a N a 0 a 1 a N v out v out v out v out v out v out v out v out v out b 0 v out no b 1 v out no bIN v out no b 0 v out no b 1 v out no bIN v out no b 0 v out no b 1 v out no b N v out no Where there are M data blocks corresponding to M envelope time poi
294. ify a harmonic index for each P element Port Element Syntax Without SS_TONE Pxxx p n lt n_ref gt lt PORT portnumber gt lt HBLIN H1 H2 HN 1 gt With SS_TONE Pxxx p n lt n_ref gt lt PORT portnumber gt lt HBLIN H1 H2 1 HN gt Parameter Description n_ref Reference node used when a mixed mode port is specified PORT The port number Numbered sequentially beginning with 1 with no shared port numbers HBLIN Integer vector that specifies the harmonic index corresponding to the tones defined in the HB command The 1 term corresponds to the small signal tone specified by SS_TONE in the HB command If there is no SS_TONE in the HB command the 1 term must be at the last entry of HBLIN vector HBLIN Analysis You use the HBLIN statement to extract frequency translation S parameters and noise figures Input Syntax Without SS_TONE HBLIN lt frequency_sweep gt lt NOISECALC 1 0 yes no gt lt FILENAME file_name gt lt DATAFORMAT ri ma db gt lt MIXEDMODE2PORT dd cc cd dc sd sc cs ds gt HSPICE RF User Guide 301 Z 2007 03 Chapter 11 S parameter Extraction Frequency Translation S Parameter HBLIN Extraction With SS_TONE HBLIN lt NOISECALC 1 0 yes no gt lt FILENAME file name gt lt DATAFORMAT ri ma db gt lt MIXEDMODE2PORT dd cc cd dc sd sc cs ds gt Parameter Description frequency_sweep Frequency sweep ra
295. ify more than one logical_instance Whenever you specify a logical instance name you must set NAME_SCOPE to FLAT If you connect a logical net to a physical port HSPICE RF reports an error HSPICE RF User Guide Z 2007 03 Chapter 13 Post Layout Analysis Post Layout Back Annotation Table 26 SPEF Parameters Continued Parameter Definition physical_instance routing_confidence logical_pin physical_node net_name cap_id res_id induc_id node1 node2 HSPICE RF User Guide Z 2007 03 Physical name of a subcircuit in your design_name circuit design You can specify more than one physical_instance Whenever you specify a physical instance name you must set NAME_SCOPE to FLAT If you connect a physical net to a logical port HSPICE RF reports an error One of the following positive integers 10 Statistical wire load model 20 Physical wire load model 30 Physical partitions with locations no cell placement 40 Estimated cell placement with Steiner tree based route 50 Estimated cell placement with global route 60 Final cell placement with Steiner route 70 Final cell placement with global route 80 Final cell placement final route 2d extraction 90 Final cell placement final route 2 5d extraction 100 Final cell placement final route 3d extraction Logical name of a pin Physical name of a node Name of a net in a circuit or subcircuit Unique identifier for capacitance between
296. ilable in HSPICE RF for BSIM3 and BSIM4 models The variations can be tailored for each device depending on its size for example A disadvantage of this method is that the netlist needs to be parameterized properly to get the correct variations The process of preparing a basic netlist for Monte Carlo simulations with this approach is tedious and error prone therefore it is best handled with scripts Bsim3 supports the following instance parameters L w ad as pd ps nrd nrs rdc rsc off ic dtemp delvto geo sa sb sd nf stimod sa1 sa2 sa3 sa4 sa5 sa6 sa7 sa8 sa9 sa10 sb1 sb2 sb3 sb4 sb5 sb6 sb7 sb8 sb9 sb10 sw1 sw2 sw3 sw4 sw5 sw6 sw7 sw8 sw9 HSPICE RF User Guide Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Simulating the Effects of Global and Local Variations with Monte Carlo sw10 muluO mulua mulub tnodeout rthO cthO deltox delk1 delnfct and acnqsmod Bsim4 supports the following instance parameters L w ad as pd ps nrd nrs rdc rsc off ic dtemp delvto geo rosb rbdb rbpb rbps ropd trnqgsmod acnqsmod rbodymod rgatemod geomod rgeomod nf min muluO delk1 delnfct deltox sa sb sd stimod sa1 sa2 sa3 sa4 sa5 sa6 sa7 sa8 sa9 sa10 sb1 sb2 sb3 sb4 sb5 sb6 sb7 sb8 sb9 sb10 sw1 sw2 sw3 sw4 sw5 sw6 sw7 sw8 sw9 sw10 xgw ngcon sca scb scc sc delk2 delxj mulngate delrsh delrshg dellpeO deldvtO and mulvsat
297. ile Guidelines Table 4 HSPICE HSPICE RF Netlists Net Name Special Characters Special Character Node Name Instance Name Parameter Name __ Delimiters Note character is cannot be the cannot be the first first character character element legal anywhere in the element key key letter only string first or included letter only plus sign HSPICE Included only HSPICE included Continues included only only avoid usage previous line HSPICE RF y Illegal in HSPICE Except for quoted RF strings expressions paths algebraics equals Illegal Illegal Illegal Token delimiter lt gt less more than YSPICE y Included only Included only n a included only for HSPICE RF question mark WSPICE y Included only Included only Wildcard in Illegal for character in both HSPICE RF HSPICE and HSPICE RF forward slash y Included only HSPICE included n a only avoid usage Illegal in HSPICE RF curly braces HSPICE Included only Included only Auto converts to included only square brackets converts to No conversion for HSPICE RF square Include only Included only Included only n a brackets HSPICE RF User Guide Z 2007 03 61 Chapter 4 Input Netlist and Data Entry Input Netlist File Guidelines Table 4 HSPICE HSPICE RF Netlists Net Name Special Characters Special Character Node Name Instance Name Parameter Name Delimiters Note character is cannot be the cannot be the first first character ch
298. iles The following is a listing of shipped demonstration files for illustrating HSPICE RF functionality All of these example files are available at lt installdir gt demo hspicerf examples File Name Description acpr sp Envelope simulation example bjt inc Transistor model library used by osc sp mos49_model inc gpsvco sp gsmlina sp mix_hb sp mix_hbac sp mix_tran sp osc sp pa sp gsmlInalP3 sp ringoscSN sp phasefreqdet sp tsmc018 m cmos90nmWrlicker lib Transistor model library used by example circuits Oscillator and Phase Noise analysis example LNA Linear analysis example Mixer HB analysis example Mixer HBAC analysis example Mixer transient analysis example Oscillator tuning curve and phase noise analysis example Power amplifier HB analysis example 3rd order intercept point example Shooting Newton and Phase Noise analysis example Shooting Newton and noise analysis example Transistor model library used by ringoscSN sp Transistor model library used by phasefreqdet sp HSPICE RF User Guide 55 Chapter 3 HSPICE RF Tutorial Demonstration Input Files 56 HSPICE RF User Guide Z 2007 03 4 Input Netlist and Data Entry Describes the input netlist file and methods of entering data For descriptions of individual HSPICE commands referenced in this chapter see Chapter 3 RF Netlist Commands in the HSPICE and RF Command Reference Input Netlist File Guidelines HSPICE RF operates on an input
299. illation frequency and the probe voltage HBOSC analysis calculates the small signal admittance that the voltage probe sees over a range of frequencies in an attempt to find potential oscillation frequencies Oscillation is likely to occur where the real part of the probe current is negative and the imaginary part is zero You can use the FSPTS parameter to specify the frequency search You must also supply an initial guess for the large signal probe voltage A value of one half the supply voltage is often a good starting point Input Syntax for Harmonic Balance Oscillator Analysis The input syntax for HBOSC analysis supports two different formats depending on whether the PROBENODE location is specified using a circuit element current source or using the HROSC PROBENODE parameters Syntax 1 HBOSC TONE F1 NHARMS H1 PROBENODE N1 N2 VP lt FSPTS NUM MIN MAX gt lt SWEEP PARAMETER SWEEP gt lt SUBHARMS I gt Syntax 2 Uses current source to set PROBENODE ISRC N1 N2 HBOSCVPROBE VP HBOSC TONE F1 NHARMS H1 lt FSPTS NUM MIN MAX gt lt SWEEP PARAMETER SWEEP gt lt SUBHARMS I gt 230 HSPICE RF User Guide Z 2007 03 Chapter 9 Oscillator and Phase Noise Analysis Input Syntax for Harmonic Balance Oscillator Analysis Parameter Description TONE Approximate value for oscillation frequency Hz The search for an exact oscillation frequency begins from this value unless you specify an FSPTS range or transient init
300. im and VMC are trademarks of Synopsys Inc Service Marks S MAP in SVP Caf and TAP in are service marks of Synopsys Inc SystemC is a trademark of the Open SystemC Initiative and is used under license ARM and AMBA are registered trademarks of ARM Limited Saber is a registered trademark of SabreMark Limited Partnership and is used under license All other product or company names may be trademarks of their respective owners Printed in the U S A HSPICE RF User Guide Z 2007 03 ii HSPICE RF User Guide Z 2007 03 Inside This Manual 0 000 eee eee The HSPICE Documentation Set 0000 cease Searching Across the HSPICE Documentation Set Other Related Publications 0 0000 e eee eee CONVENTIONS oa nve os Breed a ae ventas aad E ARA Customer Support ose ac riau aii eee eens HSPICE RF Features and Functionality HSPICE RF Overview 00 0000 ccc eee eee HSPICE RF Features 0 0 000 cee eee eee HSPICE and HSPICE RF Differences Getting Started 0 0 eee Running HSPICE RF Simulations 0000005 Netlist Overview 1 2 0 0 00 cece Parametric Analysis Extensions 0 00 0c eee eee Generating Output Files 0 000 eee eee HSPICE RF Output File Types 0 000 Using the CosmosScope Waveform Display HSPICE RF Tutorial 0 0 0
301. ime domain analysis the Shooting Newton algorithm provides functionality to support the following commands SN SNAC SNFT SNNOISE SNOSC and SNXF Periodic Time Dependent Noise Analysis PTDNOISE calculates the noise spectrum and the total noise at a point in time Jitter in a digital threshold circuit can then be determined from the total noise and the digital signal slew rate HSPICE RF also adds the following measurement capabilities to HSPICE Small signal scattering parameters Small signal two port noise parameters 1 dB compression point Intercept points for example IP2 IP3 Mixer conversion gain and noise figure VCO output spectrum Oscillator phase noise Options simplify specifying levels of accuracy As a result HSPICE RF provides effective simulation solutions for RF high speed and PCB signal integrity circuit challenges HSPICE RF User Guide Z 2007 03 Chapter 1 HSPICE RF Features and Functionality HSPICE and HSPICE RF Differences HSPICE and HSPICE RF Differences The following tables give an overview of which features Table 1 and device models Table 2 on page 7 in HSPICE are not supported in HSPICE RF Table 1 HSPICE Features Not in HSPICE RF Feature See Read hspice ini file Short names for internal sub circuits such as 10 M1 MODEL types AMP and PLOT for graphs Parameter definition PARAM for Monte Carlo statistical functions PLOT simulation output
302. in gt 0 0 CMOSN L 0 35u W 50u AS 100p AD 100p PS 104u PD 104u M 80 Ls in gt Lin gate tuning Ld drain vdd Ld drain tuning Cd drain 0 Cd Cb drain out INFINITY DC block Rload out 0 Rload vdd vdd 0 DC vdd Vrf1 in 0 DC Vin 2 0 SIN Vin 2 Vin 2 0 0 0 90 HB Vin 2 0 0 1 1 hb tones f0 nharms 10 tran 10p 10n probe hb p Rload probe tran p Rload include cmos49 model inc end An HB analysis uses the following m An HB command hb tones f 0 nharms 10 This invokes a single tone HB analysis with base frequency 950 mHz and 10 harmonics The HB source in Vrf1 HB Vin 2 0 0 1 1 This creates a sinusoidal waveform matching the transient analysis one The amplitude is Vin 2 1 5 V and it applies to the first harmonic of the first tone 950 MHZ A PROBE command for plotting the output power probe hb p Rload To run this netlist type the following command hspicerf pa sp This produces two output files named pa trO and pa hb0 containing the transient and HB output respectively To view and compare the output 1 Type cscope to invoke CosmosScope HSPICE RF User Guide 21 Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 3 Using HB Analysis for an Amplifier 2 To open both files use the File gt Open gt Plotfiles dialog Be sure to change the Files of Type filter to find the hbo file 3 Using the signal manager view the v out signals from the pa tr
303. in the output listing file Global Node Names The GLOBAL statement globally assigns a node name in HSPICE or HSPICE RF This means that all references to a global node name used at any level of the hierarchy in the circuit connect to the same node The most common use of a GLOBAL statement is if your netlist file includes subcircuits This statement assigns a common node name to subcircuit nodes Another common use of GLOBAL statements is to assign power supply connections of all subcircuits For example GLOBAL VCC connects all subcircuits with the internal node name Vcc Ordinarily in a subcircuit the node name consists of the circuit number concatenated to the node name When you use a GLOBAL statement HSPICE or HSPICE RF does not concatenate the node name with the circuit number and assigns only the global name You can then exclude the power node name in the subcircuit or macro call Circuit Temperature To specify the circuit temperature for a HSPICE or HSPICE RF simulation use the TEMP statement or the TEMP parameter in the DC AC and TRAN statements HSPICE compares the circuit simulation temperature against the reference temperature in the TNOM control option HSPICE or HSPICE RF uses the difference between the circuit simulation temperature and the TNOM reference temperature to define derating factors for component values HSPICE RF User Guide Z 2007 03 Chapter 4 Input Netlist and Data Entry Input N
304. ing the or active_node file the DSPF or SPEF file and the hierarchical SIM_SPEF_VTOL LVS ideal netlist to back annotate only active portions of the circuit If a net is latent then HSPICE RF does not expand the net This saves simulation runtime and memory value is the tolerance of the voltage change scopencan be a subcircuit definition which has an prefix or a subcircuit instance By default HSPICE RF performs only one iteration of the second simulation run Use the SIM_DSPF_MAX_ITER or SIM_SPEF_MAX_ITER option to change it For descriptions and usage examples see OPTION SIM_DSPF_VTOL and OPTION SIM_SPEF_VTOL in the HSPICE and HSPICE RF Command Reference HSPICE RF User Guide 323 Z 2007 03 Chapter 13 Post Layout Analysis Post Layout Back Annotation Table 23 Selective Post Layout Flow Options Continued Syntax Description SIM_DSPF_MAX_ITER Or SIM_SPEF_MAX_ITER value is the maximum number of iterations for the second simulation run Some of the latent nets might turn active after the first iteration of the second run In this case Resimulate the netlist to ensure the accuracy of the post layout simulation Use SIM_DSPF_MAX_ITER or SIM_SPEF_MAX_ITER to set the maximum number of iterations for the second run If the active_node remains the same after the second simulation run HSPICE RF ignores these options For descriptions and usage examples see OPTION SIM_DSPF_MAX_ITER a
305. ingle p2 balanced available sd sc default sd Example 1 Single tone analysis with frequency translation In this example the 2 port S parameters from RF 1G del_f to IF del_f are extracted The LO signal is specified by normal voltage source Vlo The frequency on port 1 is in the RF band 1G del_f and the frequency on port 2 is in the IF band del_f The IF band is swept from 0 to 100 MHz The results are output to file ex1 s2p pl RFin gnd port 1 HBLIN 1 1 p2 IFout gnd port 2 HBLIN 0 1 Vlo LOin gnd DC 0 HB 2 5011 HB tones 1G harms 5 HBLIN lin 5 0 100meg noisecalc no filename ex1 dataformat ma Example 2 Another single tone analysis with frequency translation example In this example the 3 port S parameters are extracted Port 3 provides the periodic large signal The frequency on port 1 is del_f the frequency on port 2 is 1G 2 del_f and the frequency on port 3 is 1G 1 del_f The small signal frequency is swept from 0 to 100MHz HBNOISE calculation is required The results are output to file ex2 s3p HSPICE RF User Guide 303 Z 2007 03 Chapter 11 S parameter Extraction Frequency Translation S Parameter HBLIN Extraction pl 1 0 port 1 HBLIN 0 1 p2 2 0 port 2 HBLIN 2 1 p3 3 0 port 3 hb 0 5 0 1 1 HBLIN 1 1 HB tones 1G harms 5 HBLIN lin 5 0 100meg noisecalc yes filename ex2 Output Syntax This section describes the syntax for the HBLIN PRINT and PROBE statements PRINT and
306. input creates an instance named X1 of the INV cell macro which consists of two MOSFETs named MN and MP X1 IN OUT VD_LOCAL VS LOCAL inv W 20 MACRO INV IN OUT VDD VSS W 10 L 1 DJUNC 0 MP OUT IN VDD VDD PCH W W L L DTEMP DJUNC MN OUT IN VSS VSS NCH W W 2 L L DTEMP DJUNC EOM Note To access the name of the MOSFET inside of the INV subcircuit that X1 calls the names are X1 MP and X1 MN So to print the current that flows through the MOSFETs use PRINT I X1 MP Hierarchical Parameters You can use two hierarchical parameters the M multiply parameter and the s scale parameter M Multiply Parameter The most basic HSPICE RF subcircuit parameter is the M multiply parameter This keyword is common to all elements including subcircuits except for voltage sources The M parameter multiplies the internal component values which in effect creates parallel copies of the element To simulate 32 output buffers switching simultaneously you need to place only one subcircuit for example HSPICE RF User Guide Z 2007 03 HSPICE Z 2007 03 Chapter 4 Input Netlist and Data Entry Using Subcircuits X1 in out buffer M 32 Figure 13 How Hierarchical Multiply Works X1 in out inv M 2 e J l M 8 g e o lt mp out in vdd pch W 10 L 1 M 4 M 6 e k mn out in vss nch W 5 L 1 M 3 E UNEXPANDED EXPANDED Multiply works hierarchically For a subcircuit
307. ion or rel_variation at the sigma level For example if sigma 3 then the standard deviation is abs_variation divided by 3 If you do not specify a multiplier the default is 1 HSPICE RF recalculates many times and saves the largest deviation The resulting parameter value might be greater than or less than nominal_val The resulting distribution is bimodal Example 1 In this example each R has an unique variation param mc_var agauss 0 1 3 20 swing param val 1000 1 mc_var v_vin vin 0 dc 1 ac 1 rl vin 0 1000 1 mc_var r2 vin 0 1000 1 mc_var HSPICE RF User Guide 363 Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Monte Carlo Analysis Example 2 In this example each R has an identical variation param mc_var agauss 0 1 3 20 swing param val 1l mc_var v_vin vin 0 dc 1 ac 1 rl vin 0 1000 val r2 vin 0 1000 val Example 3 In this example local variations to an instance parameter are applied by assigning randomly generated variations directly to each instance parameter Each resistor r1 through r3 receives randomly different resistance values during each Monte Carlo run param r_local agauss rl 1 2 r r_local r2 3 4 r r_local r3 5 6 r r_local Example 4 In this example global variations to an instance parameter are applied by assigning the variation to an intermediate parameter before assigning it to each instance parameter Each resistor r1 through r
308. ir values at the operating point are considered in the calculation CONVOLUTION Indicates which method is used 0 default Acts the same as the conventional method 1 Applies recursive convolution and if the rational function is not accurate enough it switches to linear convolution 2 Applies linear convolution FBASE Specifies the lower bound of the transient analysis frequency For CONVOLUTION 1 mode HSPICE starts sampling at this frequency For CONVOLUTION 2 mode HSPICE uses this value as the base frequency point for Inverse Fourier Transformation For recursive convolution the default value is OHZ For linear convolution HSPICE uses the reciprocal of the transient period FMAX Specifies the possible maximum frequency of interest The default value is the frequency point where the function reaches close enough to infinity value assuming that the monotonous function is approaching the infinity value and that it is taken at 10THz HSPICE RF User Guide 131 Z 2007 03 Chapter 6 Testbench Elements Passive Elements 132 Example L1 1 2 L 0 5n 0 5n 1 HERTZ 1e8 CONVOLUTION 1 fbase 10 fmax 30meg Ideal Transformers You can use the IDEAL keyword with the K element to designate ideal transformer coupling Syntax Kxxx Ij Lj lt k IDEAL IDEAL gt The IDEAL keyword replaces the coupling factor value This keyword activates the following equation set for non DC values which is pr
309. is 1 or 2 performs Inverse Fast Fourier Transformation IFFT linear convolution HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Passive Elements Parameter Description FBASE FMAX Base frequency to use for transient analysis This value becomes the base frequency point for Inverse Fast Fourier Transformation IFFT when CONVOLUTION 1 or 2 If you do not set this value the base frequency is a reciprocal value of the transient period Maximum frequency to use for transient analysis Used as the maximum frequency point for Inverse Fourier Transformation If you do not set this value the reciprocal value of RISETIME is taken Example C1 1 2 C 1le 6 Inductors General form HERTZ 1e16 CONVOLUTION 1 fbase 10 fmax 30meg Lxxx nl n2 lt L gt inductance lt mname gt lt lt TCl gt val gt lt lt TC2 sval gt lt SCALE val gt lt IC val gt lt M vals gt lt DTEMP val gt lt R val gt Lxxx nl n2 L equation lt LTYPE val gt lt above_options gt Polynomial form Lxxx nl n2 POLY c0 cl lt above_options gt Magnetic winding form Lxxx nl n2 NT turns lt above_options gt Parameter Description LXXX Inductor element name Must begin with L followed by up to n1 n2 TC1 HSPICE RF User Guide Z 2007 03 1023 alphanumeric characters Positive terminal node name Negative terminal node name First order temperature coefficient for the inductor S
310. is NHARMS defaults to PERIOD TRES rounded to nearest integer NHARMS is required to run subsequent SNAC SNNOISE SNXF and PHASENOISE analyses When using Syntax 1 NHARMS is computed automatically as NHARMS Round PERIOD TRES HSPICE RF User Guide Z 2007 03 Chapter 8 Steady State Shooting Newton Analysis SN Analysis Syntax Parameter Description TRINIT This is the transient initialization time If not specified the transient initialization time will be equal to the period for Syntax 1 or the reciprocal of the tone for Syntax 2 SWEEP Specifies the parameter sweep As in all main analyses in MAXTRINITCYCLES Options HSPICE RF such as TRAN HB etc you can specify LIN DEC OCT POI SWEEPBLOCK DATA MONTE or OPTIMIZE Stops SN stabilization simulation and frequency detection when the simulator detects that maxtrinitcycles have been reached in the oscnode signal or when time trinit whichever comes first Minimum cycles is 1 In addition to all TRAN options SN analysis supports the following options Option Default Description OPTION SNMAXITER lt integer gt 40 Maximum number Shooting Newton iterations OPTION SNACCURACY lt integer gt 10 Similar to the sim_accuracy definition in TRAN i e larger values of snaccuracy result in a more accurate solution but may require more time points Because Shooting Newton must store derivative information at every time point the memory re
311. is control option see OPTION SIM_ORDER in the HSPICE and HSPICE RF Command Reference OPTION SIM_TG_THETA You use the SIM_TG THETA option to control the amount of Gear 2 method to mix with trapezoidal integration for the hybrid TRAPGEAR method For the syntax and description of this control option see OPTION SIM_TG_THETA in the HSPICE and HSPICE RF Command Reference OPTION SIM_TRAP You use the SIM_TRAP option to change the default SIM_TG THETA to 0 so that method trapgear acts like METHOD TRAP HSPICE RF User Guide 391 Z 2007 03 Chapter 15 Using HSPICE with HSPICE RF RF Transient Analysis Output File Formats For the syntax and description of this control option see OPTION SIM_TRAP in the HSPICE and HSPICE RF Command Reference OPTION PURETP You use the PURETP option to turn off insertion of Backward Euler BE steps due to auto detection of numerical oscillations For the syntax and description of this control option see OPTION PURETP in the HSPICE and HSPICE RF Command Reference OPTION SIM_OSC_DETECT_TOL You use the SIM_OSC_DETECT_TOL option to specify the tolerance for detecting numerical oscillations If HSPICE RF detects numerical oscillations it inserts Backward Euler BE steps Smaller values of this tolerance result in fewer BE steps For the syntax and description of this control option see OPTION SIM_OSC_DETECT_TOL in the HSPICE and RF Command Reference RF Transient Analysis Output Fil
312. ise due to the specified noise source types as described below Specified Element PRINT PHASENOISE phnoise phnoise element_name PROBE PHASENOISE phnoise phnoise element_name In this syntax phnoise is the phase noise parameter The PHASENOISE statement outputs raw data to the pn and printpn files HSPICE RF outputs the phnoise data in decibels relative to the carrier signal per hertz across the output nodes in the PHASENOISE statement The data plot is a function of the offset frequency Units are in dBc Hz If you use the NLP algorithm METHOD 0 default HSPICE RF calculates only the phase noise component If you use the PAC algorithm METHOD 1 HSPICE RF sums both the phase and amplitude noise components to show the total noise at the output If you use the BPN algorithm METHOD 2 HSPICE RF adds both the phase and amplitude noise components together to show the total noise at the output HSPICE RF outputs phnoise to the pn file if you set OPTION POST Element phase noise can also be analyzed through the PRINT and PROBE statements which the previous syntax shows A single phnoise keyword specifies the phase noise for the whole circuit and the HSPICE RF User Guide 245 Z 2007 03 Chapter 9 Oscillator and Phase Noise Analysis Phase Noise Analysis 246 phnoise element_name specifies the phase noise value of the specified element Example 1 HBOSC TONE 900MEG NHARMS 9 PROBENODE g
313. itance connected from node n2 to bulk Default 0 0 if you did user defined equation NOISE not specify C in a resistor model Can be a function of any node voltages element currents temperature frequency or time NOISE 0 do not evaluate resistor noise NOISE 1 evaluate resistor noise default Resistance can be a value in units of ohms or an equation Required parameters are the two nodes and the resistance or model name If you specify other parameters the node and model name must precede those parameters Other parameters can follow in any order If you specify a resistor HSPICE RF User Guide 115 Chapter 6 Testbench Elements Passive Elements 116 model see the Passive Device Models chapter in the HSP ICE Elements and Device Models Manual the resistance value is optional The following are some basic examples for HSPICE RF Example 1 R1 is a resistor whose resistance follows the voltage at node c R1 1 0 v c Example 2 R2 is a resistor whose resistance is the sum of the absolute values of nodes c and d R2 10 abs v c abs v d Example 3 R3 is a resistor whose resistance is the sum of the rconst parameter and 100 times tx1 for a total of 1100 ohms PARAM rconst 100 tx1 10 R3 4 5 rconst txl 100 R3 takes its value from the RX parameter and uses the TC1 and TC2 temperature coefficients which become 0 001 and 0 respectively Example 4 You can use the H
314. ity applications see the HSPICE Signal Integrity User Guide HSPICE RF also supports several specialized elements for high frequency analysis and characterization Passive Elements This section describes the passive elements resistors capacitors and inductors See Multi Terminal Linear Elements for discussion of the W T and S elements This section includes Resistors Linear Resistors Behavioral Resistors Skin Effect Resistors m Frequency Dependent Resistors Capacitors Charge Based Capacitors HSPICE RF User Guide 113 Z 2007 03 Chapter 6 Testbench Elements Passive Elements 114 Linear Capacitors Frequency Dependent Capacitors Inductors Mutual Inductors Linear Inductors Frequency Dependent Inductors Ideal Transformers m Coupled Inductor Element Reluctance Format deal Transformer Format in HSPICE RF DC Block and Choke Elements Resistors Rxxx nili n2 lt mname gt Rval lt TCl lt TC2 gt lt TCs gt lt SCALE val gt lt M val gt lt AC val gt lt DTEMP vals gt lt L val gt lt W val gt lt C vals gt lt NOISE val gt RxXxx nl n2 lt mname gt lt R gt resistance lt lt TCl gt val gt lt lt TC2 sval gt lt lt TC sval gt lt SCALE val gt lt M val gt lt AC val gt lt DTEMP val gt lt L val gt lt W val gt lt C val gt lt NOISE val gt Rxxx nil n2 R equation Parameter Description Rxxx Resistor eleme
315. k the Polar chart icon on the left side of the upper toolbar Now use the signal manager to select the S 2 1 signal under the S Par heading to plot the complex gain of the LNA 6 Opena blank X Y plot Use the signal manager to plot K the Rollett stability factor and Gas the associated gain under the Gain Par heading and NFMIN the noise figure minimum under the Noise Par heading Example 2 Using HB Analysis for a Power Amplifier The HB command computes periodic steady state solutions of circuits This analysis uses the Harmonic Balance HB technique for computing such solutions in the frequency domain The circuit can be driven by a voltage power or current source or it may be an autonomous oscillator The HB algorithm represents the circuit s voltage and current waveforms as a Fourier series that is a series of sinusoidal waveforms HSPICE RF User Guide 19 Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 2 Using HB Analysis for a Power Amplifier 20 To set up a periodic steady state analysis the HSPICE input netlist must contain A HB command to activate the analysis The HB command specifies the base frequency or frequencies also called tones for the analysis and the number of harmonics to use for each tone The HB command can specify base tones so that the circuit solution is represented as a multi dimensional Fourier series The number of terms in the series are determined by the number of harmonics mo
316. kes its value from the RX parameter and uses the TC1 and TC2 temperature coefficients which become 0 001 and 0 respectively RP spans across different circuit hierarchies and is 0 50 HSPICE RF User Guide 117 Z 2007 03 Chapter 6 Testbench Elements Passive Elements Behavioral Resistors HSPICE RF accepts equation based resistors and capacitors You can specify the value of a resistor or capacitor as an arbitrary equation that involves node voltages or variable parameters Unlike HSPICE you cannot use parameters to indirectly reference node voltages in HSPICE RF Rxxx nil n2 lt R gt equation Example R1 A B R V A I VDD Skin Effect Resistors RXxxx nl n2 R value Rs value The Rs indicates the skin effect coefficient of the resistor The complex impedance of the resistor can be expressed as the following equation R f Ro 1 j Rs sqrt f The Ro j and are DC resistance imaginably unit jA2 1 and frequency respectively Frequency Dependent Resistors You can specify frequency dependent resistors using the R expression with the HERTZ keyword The HERTZ keyword represents the operating frequency In time domain analyses an expression with the HERTZ keyword behaves differently according to the value assigned to the CONVOLUTION keyword Syntax RXXX n n R expression with HERTZ lt CONVOLUTION 0 1 2 gt lt FBASE value gt lt FMAX value gt gt Parameter Description CO
317. kew Parameters Gaussian FF Fast Corner Skew Parameters IDS The LIB library statement and the INCLUDE include file statement access the models and skew The library contains parameters that modify MODEL statements The following example of LIB features both HSPICE RF User Guide Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Worst Case Analysis worst case and statistical distribution data by using model skew parameters In statistical distribution the median value is the default for all non Monte Carlo analysis Example LIB TT STYPICAL P CHANNEL AND N CHANNEL CMOS LIBRARY DATE 3 4 91 PROCESS 1 0U CMOS FAB22 STATISTICS COLLECTED 3 90 2 91 following distributions are 3 sigma ABSOLUTE GAUSSIAN PARAM polysilicon Critical Dimensions polycd agauss 0 0 06u 1 xl polycd sigma 0 06u Active layer Critical Dimensions nactcd agauss 0 0 3u 1 xwn nactcd sigma 0 3u pactcd agauss 0 0 3u 1 xwp pactcd sigma 0 3u Gate Oxide Critical Dimensions 200 angstrom 10a at 1 sigma toxcd agauss 200 10 1 tox toxcd sigma 10 MM M UN BPs Threshold voltage variation vtoncd agauss 0 0 05v 1 delvton vtoncd sigma 0 05 vtopcd agauss 0 0 05v 1 delvtop vtopcd sigma 0 05 INC usr meta lib cmos1_ mod dat model include file ENDL TT LIB FF SHIGH GAIN P CH AND N CH CMOS LIBRARY 3SIGMA VALUES PARAM TOX
318. l parameters such as write_file f filename Indicates that parameters can be repeated as many times as necessary pini pin2 pinN Indicates a choice among alternatives such as low medium high Indicates a continuation of a command line HSPICE RF User Guide xvii Z 2007 03 About This Guide Customer Support Convention Description Indicates levels of directory structure Edit gt Copy Indicates a path to a menu command such as opening the Edit menu and choosing Copy Control c Indicates a keyboard combination such as holding down the Control key and pressing c Customer Support Customer support is available through SolvNet online customer support and through contacting the Synopsys Technical Support Center Accessing SolvNet SolvNet includes an electronic knowledge base of technical articles and answers to frequently asked questions about Synopsys tools SolvNet also gives you access to a wide range of Synopsys online services which include downloading software viewing Documentation on the Web and entering a call to the Support Center To access SolvNet 1 Go to the SolvNet Web page at http solvnet synopsys com 2 If prompted enter your user name and password If you do not have a Synopsys user name and password follow the instructions to register with SolvNet If you need help using SolvNet click Help on the SolvNet menu bar xviii HSPICE RF User Guide Z 2007 03
319. l voltages between nl and n2 node in time domain PROBE HB V n1 v2 PRINT HB VP out 1 PROBE HB P Pout 2 1 UV UN N N N N N NN WN Using MEASURE with HB Analyses For transient analysis TRAN the independent variable for calculating MEASURE Is time For AC analysis the independent variable for calculating MEASURE is frequency However as with DC analysis the use of a MEASURE command is peculiar for HB analysis because it has no obvious independent variable In HSPICE RF the independent variable for HB MEASURE analysis is the first swept variable specified in the HB simulation control statement This variable can be anything frequency power voltage current a component value and so on Example 1 For the following HB simulation control statement the independent variable is the swept tone frequency and the MEASURE command values return results based on this frequency sweep 212 HSPICE RF User Guide Z 2007 03 Chapter 7 Steady State Harmonic Balance Analysis Harmonic Balance Analysis HARMONIC BALANCE tone frequency sweep for amplifier param freql 1 91e9 power 1le 3 HB tones freql nharms 10 sweep freql LIN 10 1 91e9 2 0e9 MEASURE HB Patf0 FIND P Rload 1 AT 1 95e9 Power at 0 1 95Ghz MEASURE HB Frq W WHEN P Rload 1 1 freql 1 Watt MEASURE HB BW1W TRIG AT 1 92e9 TARG P Rload 1 VAL 1 CROSS 2 1 Watt bandwidth MEASURE HB MaxPwr MAX P Rload 1 FRO
320. lance Analysis To output the final frequency of oscillation use the HERTZ keyword For example heriz 1 identifies the fundamental frequency of oscillation Oscillator Analysis Using Shooting Newton SNOSC 236 The analysis described in Chapter 8 Steady State Shooting Newton Analysis also provides a very effective means for finding the steady state for oscillator circuits This approach is also very effective for ring oscillator circuits and oscillators that operate with piecewise linear waveforms HBOSC is superior for sinusoidal waveforms As with the Harmonic Balance approach the goal is to solve for the additional unknown oscillation frequency This is accomplished in Shooting Newton by considering the period of the waveform as an additional unknown and solving the boundary conditions at the waveform endpoints that coincide with steady state operation As with regular Shooting Newton analysis input may be specified in terms of time or frequency values SNOSC TONE F1 NHARMS H1 TRINIT Ti OSCNODE N1 MAXTRINITCYCLES N SWEEP PARAMETER SWEEP or SNOSC TRES Tr PERIOD Tp TRINIT Tr OSCNODE N1 MAXTRINITCYCLES I SWEEP PARAMETER SWEEP Parameter Description TONE Approximate value for oscillation frequency Hz The search for an exact oscillation frequency begins from this value NHARMS Number of harmonics to be used for oscillator SN analysis OSCNODE Node used to probe for oscillation conditions This node is
321. lay To display the Signal Menu right click a signal label in a chart Using this menu you can change how signals look delete signals or move signals from one chart panel to another e Use the Attributes menu item to control how the signal looks e Use the Stack Region menu to move signals You can move a signal to a new panel or an existing panel The existing panels are named Analog 0 Analog 1 and so on Analog 0 is the bottom panel on a chart e Use the To Time Domain command to convert a histogram plot for example from a hb0 file to a time domain signal Right click a horizontal or vertical axis to control an axis Using the Axis Attributes dialog you can use the Axis Menu to configure the axis precisely e Use the Range submenu to zoom in or out e Use the Scale submenu to switch between linear and logarithmic scales e Lock Out New Signals creates an independent axis when you create a new panel e Display Range Slider displays a region next to the axis Click in that region to pan the display right left up or down To zoom in and out use the Axis Attributes dialog the zoom buttons on the tool bar or the mouse directly on the chart window To attach a marker to a signal click on a signal label then click the Vertical Marker or Horizontal Marker icons in the tool bar You can use the mouse to drag the marker along the signal to see the signal s precise value at different points Choose
322. lectrical length specified in NL NL Normalized electrical length of the transmission line at the frequency specified in the F parameter in units of wavelengths per line length Default 0 25 which is a quarter wavelength mname U model reference name A lossy transmission line model representing the characteristics of the lossless transmission line Only one input and output port is allowed Example 1 The T1 transmission line connects the in node to the out node T1 in gnd out gnd Z0 50 TD 5n L 5 Both signal references are grounded Impedance is 50 ohms HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Multi Terminal Linear Elements m The transmission delay is 5 nanoseconds per meter The transmission line is 5 meters long Example 2 The Tcable transmission line connects the in1 node to the out1 node Tcable inl gnd outi gnd Z0 100 F 100k NL 1 Both signal references are grounded Impedance is 100 ohms The normalized electrical length is 1 wavelength at 100 kHz Example 3 The Tnet1 transmission line connects the driver node to the output node Tnetl driver gnd output gnd Umodell L 1m Both signal references are grounded Umodel1 references the U model The transmission line is 1 millimeter long Ideal Transmission Line For the ideal transmission line voltage and current will propagate without loss along the length of the line x direction with spatial and time
323. lement in delay mode Set to 1 if the parameters are represented in the mixed mode A string used to determine the order of the indices of the mixed signal incident or reflected vector The string must be an array of a letter and a number Xn where X D to indicate a differential term C to indicate a common term S to indicate a single grounded term n the port number The line length of the transmission line system where the S parameters are extracted This keyword is required only when the S Model is used in a W element The FQMODEL TSTONEFILE and CITIFILE parameters describe the frequency varying behavior of a network Only specify one of the parameters in an S model card If more than one method is declared only the first one is used and HSPICE issues a warning message FQMODEL can be set in S element and S model statements but both statements must refer to the same model name The S element is capable of reading in two port noise parameter data from Touchstone data files and then transform the raw data into a form used for noise and LIN 2PNOISE analysis HSPICE RF User Guide Z 2007 03 151 Chapter 6 Testbench Elements Multi Terminal Linear Elements 152 For example you can represent a two port system with an S element and then perform a noise analysis or any other analysis The S element noise model supports both normal and two port noise analysis NOISE and LIN NOISECALC 1 Example
324. listing and Figure 14 on page 90 X1 D Q Qbar CL CLBAR dlatch flip 0 macro dlatch D Q Qbar CL CLBAR flip vcec nodeset v din flip xinv1 din qbar inv xinv2 Qbar Q inv ml q CLBAR din nch w 5 1 1 m2 D CL din nch w 5 1 1 eom Figure 14 D Latch with Nodeset Q clbar cl E _ D L Q din C Nodeset HSPICE does not limit the size or complexity of subcircuits they can contain subcircuit references and any model or element statement However in HSPICE RF you cannot replicate output commands within subcircuit definitions To specify subcircuit nodes in PRINT statements specify the full subcircuit path and node name HSPICE RF User Guide Z 2007 03 Chapter 4 Input Netlist and Data Entry DDL Library Access DDL Library Access To include a DDL library component in a data file use the X subcircuit call statement with the DDL element call The DDL element statement includes the model name which the actual DDL library file uses For example the following element statement creates an instance of the 1N4004 diode model X1 2 1 D1N4004 Where D1N4004 is the model name See Element and Source Statements on page 72 and the HSPICE Elements and Device Models Manual for descriptions of element statements Optional parameter fields in the element statement can override the internal specification of the model For example for op amp devices you can override the offset voltage and the gain and offset
325. llegal Expression and Illegal Blank whitespace Use before or line continuations Tab Tab filename delimiter Token delimiter Token delimiter First Character The first character in every line specifies how HSPICE RF interprets the remaining line Table 5 lists and describes the valid characters Table 5 First Character Descriptions Line If the First Character is Indicates First line of a netlist Subsequent lines of period netlist and all lines of included files c C d D e E F g H i I j J K K 1 L q Q r R S S v V w Any character asterisk plus Title or comment line The first line of an included file is a normal line and not a comment Netlist keyword For example TRAN 0 5ns 20ns Element instantiation Comment line HSPICE Continues previous line HSPICE RF User Guide Z 2007 03 63 Chapter 4 Input Netlist and Data Entry Input Netlist File Guidelines 64 Delimiters An input token is any item in the input file that HSPICE RF recognizes Input token delimiters are tab blank comma equal sign and parentheses Single or double quotes delimit expressions and filenames Colons delimit element attributes for example M1 VGS Periods indicate hierarchy For example X1 X2 n1 is the n1 node on the X2 subcircuit of the X1 circuit Node Identifiers Node identifiers can be up
326. lts are used to show the input and output waveforms of the phase frequency detector Use the File gt Open gt Plotfiles dialog to open the phasefreqdet sn0 file Remember to set the file type filter to HSPICE RF 3 From the signal manager double click on the input signals v cfin and v fin and the output signals v pu and v pdn Figure 5 shows the waveforms for the selected input and output signals Figure 5 Phase Frequency Detector Signals 0 0 100p 200p 300p 400p 500p 600p 700p 600p 9200p in 1in 1 27 1 3n 14n 15n 16n 17n 18n 1 3n ifs HSPICE RF User Guide Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 7 Using Shooting Newton Analysis on a Driven Phase Frequency Circuit and a Ring Oscillator 4 The frequency domain results are used to show the gain of the phase frequency detector Use the File gt Open gt Plotfiles dialog to open the phasefreqdet snfo file 5 Open anew XY graph by clicking the waveform icon on the left side of the icon bar 6 From the signal manager double click on the signal v Ifin 0 The signal is the DC component of the v lfin signal spectrum Both the magnitude and phase of the load current are plotted To measure the gain of the phase frequency detector verses phase only the magnitude is required 7 The Y axis should be a real value not a dB value To change the Y axis right click on the Y axis and select Attributes from the menu In the Signal Attributes window change the view fro
327. lue In test8 sp the variation definition for locwidth has been moved from the top level into the subcircuit Each resistor has a common global variation and its own local variation test9 sp assigns the top level variation to a local parameter which in turn is applied to the width definition of the resistor This happens independently within each subcircuit thus we end up with the same values for the resistor pair in each subcircuit but different values for the different pairs This technique can be applied to long resistors when a middle terminal is required for connecting capacitance to the substrate The resulting two resistor pieces will have the same resistance but it will be different from other resistor pairs In summary each subcircuit has its own parameter space therefore it is possible to put groups of identical components into a subcircuit and within each group all devices have the same parameter values but between the groups parameters are different When specifying variations on these parameters the effects of local variations between the groups are created HSPICE RF User Guide 383 Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Simulating the Effects of Global and Local Variations with Monte Carlo 384 Variations on a Model Parameter Using a Local Model in Subcircuit If a model is specified within a subcircuit then the specified parameter values apply only to the devices in the same sub
328. ly fout ABS n1 f n2 f2 nk fk fin Where f1 f2 fk are the first through k th steady state tones determined from the harmonic balance solution fin is the IFB defined by parameter_sweep The default index term is 1 1 1 1 For a single tone analysis the default mode is consistent with simulating a low side down conversion mixer where the RF signal is specified by the IFB and the noise is measured at a down converted frequency that the OFB specifies In general you can use the n1 n2 nk 1 index term to specify an arbitrary offset The noise figure measurement is also dependent on this index term HSPICE RF User Guide Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Multitone Harmonic Balance Noise HBNOISE Parameter Description listfreq listcount listfloor listsources Prints the element noise value to the lis file You can specify at which frequencies the element noise value is printed The frequencies must match the sweep_frequency values defined in the parameter_sweep otherwise they are ignored In the element noise output the elements that contribute the largest noise are printed first The frequency values can be specified with the NONE or ALL keyword which either prints no frequencies or every frequency defined in parameter_sweep Frequency values must be enclosed in parentheses For example list freq none listfreq all listfr
329. lysis Jitter Analysis frequency limits provided for the phase noise analysis Linear interpolation is used but the phase noise generally follows more of a power law expansion Jitter Input Syntax The timing jitter calculations are derived from the results of phase noise analysis The phase noise output syntax supports the JITTER keyword as an output keyword in addition to the PHNOISE keyword PRINT PHASENOISE PHNOISE JITTER PROBE PHASENOISE PHNOISE JITTER If the JITTER keyword is present the PHASENOISE statement also outputs the raw jitter data to jtO and printjtO data files The PHNOISE data is given in units of dBc HZ i e dB relative to the carrier per Hz across the output nodes specified by the PHASENOISE statement The data plot is a function of offset frequency If the JITTER keyword is present PHASENOISE outputs the TIE jitter data to jt0 and printjtO data files These data are plotted as a function of time in units of seconds The jitter calculations make use of some of the parameters given in the PHASENOISE syntax see PHASENOISE Input Syntaxfor the syntax and examples The timing jitter calculations make use of the phase noise frequency sweep specification The resulting values for type nsteps start and stop result in an array of frequency points given by fi fa fz a zf These frequency values are used for the integration calculations necessary to compute jitter The output of timing jitter infor
330. lyze If FREQ is non zero the output lists only the harmonics of this frequency based on FMIN and FMAX HSPICE also prints the THD for these harmonics The default is 0 0 Hz HSPICE RF User Guide Z 2007 03 Chapter 8 Steady State Shooting Newton Analysis Shooting Newton with Fourier Transform SNFT Argument Description FMIN Minimum frequency for which HSPICE prints SNFT output into the listing file THD calculations also use this frequency T STOP START The default is 1 0 T Hz FMAX Maximum frequency for which HSPICE prints SNFT output into the listing file THD calculations also use this frequency The default is 0 5 NP FM IN Hz Example 1 SNFT v 1 SNFT v 1 2 np 1024 start 0 3m stop 0 5m freq 5 0k window kaiser alfa 2 5 SNFT I rload start 0m to 2 0m fmin 100k fmax 120k format unorm SNFT par v 1 v 2 from 0 2u stop 1 2u window harris Example 2 SNFT v 1 np 1024 SNFT v 2 np 1024 This example generates an snft0 file for the SNFT of v 1 and an snft1 file for the SNFT of v 2 SN Signal Sources SN analysis assumes that all stimuli are periodic with period T If the circuit is driven with more than one periodic stimulus then the frequencies must be all co periodic and T must match the common period or some integer multiple of it The SN analysis only supports tran time domain periodic signal sources Refer to the tran analysis for a detailed documentation on
331. m viewers and requires excessive space on the hard drive This section describes options that limit the number of nodes output to the waveform file to reduce the file size HSPICE RF supports the following options to control the output SIM_POSTTOP Option SIM_POSTSKIP Option m SIM_POSTAT Option SIM_POSTDOWN Option SIM_POSTSCOPE Option SIM_POSTTOP Option You use the SIM_POSTTOP option to limit the data written to your waveform file to data from only the top n level nodes This option outputs instances up to n levels deep For example HSPICE RF User Guide 403 Z 2007 03 Chapter 16 Advanced Features Limiting Output Data Size 404 OPTION SIM POSTTOP lt n gt Note To enable the waveform display interface you also need the POST option For additional information see OPTION SIM_POSTTOP in the HSPICE and HSPICE RF Command Reference SIM_POSTSKIP Option You use the SIM_POSTSKIP to have the SIM_POSTTOP option skip any instances and their children that the subckt_definition defines For example OPTION SIM _POSTSKIP lt subckt_ definitions For additional information see OPTION SIM _POSTSKIP in the HSPICE and HSPICE RF Command Reference SIM_POSTAT Option You use the SIM_POSTAT option to limit the waveform output to only the nodes in the specified subcircuit instance For example OPTION SIM POSTAT lt instance gt This option can be used in conjunction with the SIM_POSTTOP option and wh
332. m db y to real y Figure 6 shows the gain of the phase frequency detector Figure 6 Phase Frequency Detector Gain Phase Frequency Detector Gain 30 0 35 0 40 0 45 0 50 0 5 phase 8 Next plot the output noise of the phase frequency detector Use the File gt Open gt Plotfiles dialog to open the phasefreqdet snpn0 file 9 Open anew XY graph by clicking the waveform icon on the left side of the icon bar 10 From the signal manager double click on the signal inoise onoise The noise results are shown in Figure 7 on page 50 This displays the noise at the output v Ifin at each phase value swept in the SN command HSPICE RF User Guide 49 Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 7 Using Shooting Newton Analysis on a Driven Phase Frequency Circuit and a Ring Oscillator 11 Change the X axis scale to log by right clicking on the X axis and selecting scale gt log You can change the color of each trace and control the labeling by right clicking on the signal name and selecting Member Attributes from the menu To assign each trace a different color click on the rainbow colored button in the Member Attributes menu The labels are enabled by clicking Show All Labels Figure 7 Phase Detector Output Noise Phase Frequency Detector Nose pha 90 0 100 0k Ring Oscillator Example The Shooting Newton algorithm provides fast and effective analysis for ring oscillators The ringoscSN sp input file
333. m5 num6 gt lt num7 gt lt gt AC Sweep AC dec 10 100 1meg sweep MONTE val lt firstrun numil gt or AC dec 10 100 1lmeg sweep MONTE list lt gt lt numl1 num2 gt lt num3 gt lt num5 num6 gt lt num7 gt lt gt TRAN Sweep TRAN 1n 10n sweep MONTE val lt firstrun num1 gt Or TRAN 1n 10n sweep MONTE list lt gt lt numl num2 gt lt num3 gt lt num5 num6 gt lt num7 gt lt gt The val value specifies the number of Monte Carlo iterations to perform A reasonable number is 30 The statistical significance of 30 iterations is quite high If the circuit operates correctly for all 30 iterations there is a 99 probability that over 80 of all possible component values operate correctly The relative error of a quantity determined through Monte Carlo analysis is proportional to val The firstrun values specify the desired number of iterations HSPICE RF runs from num1 to num1 val 1 The number after firstrun can be a parameter You can write only one number after list The colon represents from to Specifying only one number makes HSPICE FF runs only a the one specified point Example 1 In this example HSPICE RF runs from the 90th to 99th Monte Carlo iterations tran 1n 10 sweep monte 10 firstrun 90 You can write more than one number after list The colon represents from to Specifying only one number makes HSPICE FF run only at that single point Example 1 In this example
334. mation uses a corresponding time sampling derived via 1 r 1 1 Equation 50 Tta 7 T 2 Ty gt Oty OT fas cae MEASURE Statements to Support Jitter Analysis The jitter specific MEASURE statements specify the jitter keywords as follows For discussion of the BER parameter see below MEASURE PHASENOISE lt Jname gt PERJITTER phnoise HSPICE RF User Guide 251 Z 2007 03 Chapter 9 Oscillator and Phase Noise Analysis Jitter Analysis 252 lt UNITS sec rad UI gt lt BER val gt MEASURE PHASENOISE lt Jname gt CTCJITTER phnoise lt UNITS sec rad UI gt lt BER val gt MEASURE PHASENOISE lt Jname gt RMSJITTER phnoise lt FROM start _frequency gt lt TO end frequency gt lt UNITS sec rad UI gt lt BER val gt MEASURE PHASENOISE lt Jname gt PHJITTER phnoise lt FROM start _frequency gt lt TO end _frequency gt lt UNITS sec rad UI gt lt BER val gt MEASURE PHASENOISE lt Jname gt TRJITTER phnoise lt FROM start frequency gt lt TO end frequency gt lt UNITS sec rad UI gt lt BER val gt MEASURE PHASENOISE lt Jname gt LTJITTER phnoise lt FROM start _frequency gt lt TO end _frequency gt lt UNITS sec rad UI gt lt BER val gt RMSJITTER PHJITTER and TRJITTER are synonymous measurements all based on the calculations described related to the RMS Phase Jitter value in units of seconds given by On Oms These measurements allow control of
335. me Must begin with K followed by up to 1023 alphanumeric characters Lbbb Lccc Lddd Names of the windings about the Kaaa core One winding element is required and each winding element must use the magnetic winding syntax All winding elements with the same magnetic core model should be written in one mutual inductor statement in the netlist mname Saturable core model name See the Passive Device Models chapter in the HSPICE Elements and Device Models Manual for more information HSPICE RF User Guide Z 2007 03 HSPICE Z 2007 03 Chapter 6 Testbench Elements Passive Elements Parameter Description MAG Initial magnetization of the saturable core You can set this to 1 0 or 1 where 1 refer to positive and negative values of the BS model parameter See the Passive Device Models chapter in the HSPICE Elements and Device Models Manual for more information magnetization In this syntax coupling is a unitless value from zero to one representing the coupling strength If you use parameter labels the nodes and model name must be first Other arguments can be in any order If you specify an inductor model see the Passive Device Models chapter in the HSP ICE Elements and Device Models Manual the inductance value is optional You can determine the coupling coefficient based on geometric and spatial information To determine the final coupling inductance HSPICE or HSPICE RF divides the coupling coeffici
336. ment Pload PROBE HB P Pload Example 7 Prints the RMS power delivered to port element Pload at the fundamental HB analysis frequency following a one tone analysis PRINT HB P Pload 1 Example 8 Outputs the RMS power delivered to port element Pload at the low side 3rd order intermodulation product following an HB two tone analysis PROBE HB P Pload 2 1 Calculations for Time Domain Output In addition to a frequency domain output HB analysis also supports a time domain output The equivalent time domain waveform is generated according to the Fourier series expansion given by Equation 33 V n1 time t SUMoygpm REALV n1 m cos Q m t IMAG V n1 m sinQ me t ot Where m starts from 0 to the number of frequency points in the HB simulation The output syntax is PRINT HBTRAN HBTR V n1 PROBE HBTRAN HBTR V n1 The output time ranges from 0 to twice the period of the smallest frequency in the HB spectra HSPICE RF User Guide 211 Z 2007 03 Chapter 7 Steady State Harmonic Balance Analysis Harmonic Balance Analysis Output Examples PRINT HB P rload RMS power spectrum dissipated at the rload resistor Differential voltage spectrum between the nl1 n2 nodes Phase of voltage at the out node at the fundamental frequency RMS power delivered to the Pout port at third order intermod PRINT HBTRAN V n1 Voltage at nl in time domain PROBE HBTRAN V nl lt n2 gt Differentia
337. meter name or A MEASURE name associated with a time domain MEASURE statement located in the netlist PTDNOISE uses the time point generated from the MEASURE statement to evaluate the noise characteristics This is useful if you want to evaluate noise or jitter when a signal reaches some threshold value A time value used to determine the slew rate of the time domain output signal Specified as TDELTA value The signal slew rate is then determined by the output signal at TIME TDELTA and dividing this difference by 2 x TDELTA This slew rate is then used in the calculation of the strobed jitter If this term is omitted a default value of 0 01 x the SN period is assumed Frequency sweep range for the output noise spectrum The upper and lower limits also specify the integral range in calculating the integrated noise value Specify LIN DEC OCT POI SWEEPBLOCK DATA sweeps Specify the nsteps start and stop frequencies using the following syntax for each type of sweep LIN nsteps start stop DECnsteps start stop OCT nsteps start stop POI nsteps freq_values SWEEPBLOCK nsteps freq1 freq2 freqn 283 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Periodic Time Dependent Noise Analysis PTDNOISE 284 Parameter Description listfreq Prints the element noise value to the lis file This information is only printed if a noise spectrum is requested ina PRINT or PROBE statement See PTDNOIS
338. ming behavior of the inverter To create the circuit 1 Define the MOSFET models for the PMOS and NMOS transistors of the inverter 2 Insert the power supplies for both VDD and GND power rails Insert the pulse source to the inverter input This circuit uses transient analysis and produces output graphical waveform data for the input and output ports of the inverter circuit 70 HSPICE RF User Guide Z 2007 03 Chapter 4 Input Netlist and Data Entry Input Netlist File Composition Sample inverter circuit KkKKK MOS models kkkkxk MODEL nl NMOS LEVEL 3 THETA 0 4 MODEL p1 PMOS LEVEL 3 xxxx Define power supplies and sources VDD VDD 0 5 VPULSE VIN 0 PULSE 0 5 2N 2N 2N 98N 200N VGND GND 0 0O xxxx x Actual circuit topology M1 VOUT VIN VDD VDD pl M2 VOUT VIN GND GND ni1 kk k k Analysis statement TRAN 1n 300n k k k Output control statements OPTION POST PROBE PROBE V VIN V VOUT END For descriptions of individual HSPICE commands referenced in netlists see Chapter 3 RF Netlist Commands in the HSPICE and RF Command Reference Title of Simulation You set the simulation title in the first line of the input file HSPICE or HSPICE RF always reads this line and uses it as the title of the simulation regardless of the line s contents The simulation prints the title verbatim in each section heading of the output listing file To set the title you can place a TITLE
339. mited You must provide one input and output port the ground references a model or file reference a number of conductors and a length S Model form Wxxx inl lt in2 lt inx gt gt refin outl lt out2 lt outx gt gt refout lt Smodel modelname gt lt NODEMAP XiYj gt N val L val Table Model form Wxxx inl in2 lt inx gt gt refin outil lt out2 lt outx gt gt refout N val L val TABLEMODEL name Parameter Description Wxxx Lossy W element transmission line element name Must start with W followed by up to 1023 alphanumeric characters inx Signal input node for x transmission line in1 is required HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Multi Terminal Linear Elements Parameter Description refin Ground reference for input signal outx Signal output node for the x transmission line each input port must have a corresponding output port refout Ground reference for output signal N Number of conductors excluding the reference conductor L Physical length of the transmission line in units of meters RLGCfile filename Umodel modelname FSmodel modelname NODEMAP Smodel TABLEMODEL File name reference for the file containing the RLGC information for the transmission lines for syntax see Using the W element in the HSPICE Signal Integrity Guide U model lossy transmission line model reference name A lossy transmission line model used to
340. n Henries L Inductance in Henries at room temperature TC1 TC2 Temperature coefficient M Multiplier for parallel inductors DTEMP Temperature difference between the element and the circuit IC Initial inductor current Example LX A B 1E 9 LR 1 O lu IC 10mA Xisa 1 nH inductor LRis a1 uH inductor with an initial current of 10 mA Frequency Dependent Inductors You can specify frequency dependent inductors using the L equation with the HERTZ keyword The HERTZ keyword represents the operating frequency In time domain analyses an expression with the HERTZ keyword behaves differently according to the value assigned to the CONVOLUTION keyword HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Passive Elements Syntax LXXX nl n2 L equation lt CONVOLUTION 0 1 2 lt FBASE valule gt lt FMAX value gt gt Parameter Description LXXX Inductor element name Must begin with L followed by up to 1023 alphanumeric characters ni n2 Positive and negative terminal node names equation The equation should be a function of HERTZ If CONVOLUTION is turned on when a HERTZ keyword is not used in the equation CONVOLUTION is automatically be turned off and the inductor behaves conventionally The equation can be a function of temperature but it does not support variables of node voltage branch current or time If these variables exist in the equation with CONVOLUTION turned on only the
341. n algebraic expressions because constants are only single precision numbers 7 digits In HSPICE an algebraic expression with quoted strings can replace any parameter in the netlist In HSPICE you can then use these expressions as output variables in PRINT statements Algebraic expressions can expand your options in an input netlist file Some uses of algebraic expressions are Parameters PARAM x y 3 RF User Guide 99 Chapter 5 Parameters and Functions Built In Functions and Variables m Functions PARAM rho leff weff 2 leff weff 2u Algebra in elements R1 1 0 r ABS v 1 i m1 10 Algebra in MEASURE statements MEAS vmax MAX V 1 MEAS imax MAX I q2 MEAS ivmax PARAM vmax imax Algebra in output statements PRINT conductance PAR i m1 v 22 The basic syntax for using algebraic expressions for output is PAR algebraic expression In addition to using quotations you must define the expression inside the PAR statement for output The continuation character for quoted parameter strings in HSPICE is a double backslash Outside of quoted strings the single backslash is the continuation character Built In Functions and Variables In addition to simple arithmetic operations you can use the built in functions listed in Table 12 and the variables listed in Table 11 on page 97 in HSPICE expressions Table 12 Synopsys HSPICE Built in
342. n area Overrides OPTION DEFAD Default DEFAD if you set the ACM 0 model parameter HSPICE RF User Guide Z 2007 03 HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Active Elements Parameter Description AS PD PS NRD NRS RDC RSC OFF IC vds vgs vbs M DTEMP Source diffusion area Overrides OPTION DEFAS Default DEFAS if you set the ACM 0 model parameter Perimeter of drain junction including channel edge Overrides OPTION DEFPD Default DEFAD if you set the ACM 0 or 1 model parameter Default 0 0 if you set ACM 2 or 3 Perimeter of source junction including channel edge Overrides OPTION DEFPS Default DEFAS if you set the ACM 0 or 1 model parameter Default 0 0 if you set ACM 2 or 3 Number of squares of drain diffusion for resistance calculations Overrides OPTION DEFNRD Default DEFNRD if you set ACM 0 or 1 model parameter Default 0 0 if you set ACM 2 or 3 Number of squares of source diffusion for resistance calculations Overrides OPTION DEFNRS Default DEFNRS when you set the MOSFET model parameter ACM 0 or 1 Default 0 0 when you set ACM 2 or 3 Additional drain resistance due to contact resistance in units of ohms This value overrides the RDC setting in the MOSFET model specification Default 0 0 Additional source resistance due to contact resistance in units of ohms This value overrides the RSC setting in the MOSFET model specification D
343. n the form of scalar quantities that represent power at a particular element To request a complete power spectrum use the following syntax PRINT HB P Elem PROBE HB P Elem To request a power value at a particular frequency tone use the following syntax PRINT HB P Elem lt nl lt n2 lt n3 gt gt gt PROBE HB P Elem lt nl lt n2 lt n3 gt gt gt The Elem is the name of either a Resistor R or Port P element and n7 n2 and n3 are integer indices used for selecting a particular frequency in the Harmonic Balance output spectrum Example 1 Prints a table of the RMS power spectrum dissipated by resistor R1 PRINT HB P R1 Example 2 Outputs the RMS power dissipated by resistor R1 at the fundamental HB analysis frequency following a one tone analysis PROBE HB P R1 1 Example 3 Prints the power dissipated by resistor R1 at DC following a one tone analysis PRINT HB P R1 0 Example 4 Outputs the RMS power dissipated by resistor R1 at the low side 3rd order intermodulation product following an HB two tone analysis HSPICE RF User Guide Z 2007 03 Chapter 7 Steady State Harmonic Balance Analysis Harmonic Balance Analysis PROBE HB P R1 2 1 Example 5 Prints the RMS power dissipated by resistor R1 at the high side 3rd order intermodulation product following an HB two tone analysis PRINT HB P R1 1 2 Example 6 Outputs the RMS power spectrum delivered to port ele
344. nalysis 368 Figure 36 Major and Minor Distribution of Manufacturing Variations major distribution ee minor distribution pop en XL polysilicon linewidth variation The following example is a Monte Carlo analysis of a DC sweep in HSPICE RF Monte Carlo sweeps the VDD supply voltage from 4 5 volts to 5 5 volts You can find the sample netlist for this example in the following directory installdir demo hspice apps mondc_a sp The M1 through M4 transistors form two inverters The nominal value of the LENGTH parameter sets the channel lengths for the MOSFETs which are set to 1u in this example All transistors are on the same integrated circuit die The LEFF parameter specifies the distribution for example a 5 distribution in channel length variation at the 3 sigma level Each MOSFET has an independent random Gaussian value The PHOTO parameter controls the difference between the physical gate length and the drawn gate length Because both n channel and p channel transistors use the same layer for the gates Monte Carlo analysis sets XPHOTO distribution to the PHOTO local parameter XPHOTO controls PHOTO lithography for both NMOS and PMOS devices which is consistent with the physics of manufacturing RC Time Constant This simple example shows uniform distribution for resistance and capacitance It also shows the resulting transient waveforms for 10 different random values You can find t
345. nce In conjunction with oscillator analysis HSPICE RF can perform phase noise analysis Phase noise analysis measures the effect of transistor noise on the oscillator frequency Phase noise analysis is activated using the PHASENOISE command this command sets a set of frequency points for phase noise analysis The PRINT and PROBE commands can be used to output phase noise values The following netlist osc sp simulates an oscillator and performs phase noise analysis This example is included with the HSPICE RF distribution as pa sp and is available in directory lt installdir gt demo hspicerf examples Use the HBOSC command with the PROBENODE and FSPTS parameters set HSPICE RF User Guide Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 4 Using HBOSC Analysis for a Colpitts Oscillator PROBENODE emitter 0 4 27 Identifies the emitter node as an oscillating node and provides a guess value of 4 27 volts for the oscillation amplitude at the emitter node FSPTS 40 9e6 1 1e7 Causes an initial frequency search using 40 equally spaced points between 9 and 11 MHz Inthe PHASENOISE PRINT and PROBE commands PHASENOISE V emitter dec 10 10k 1meg Runs phase noise analysis at the specified offset frequencies measured from the oscillation carrier frequency The frequency points specified here are ona logarithmic scale 10 points per decade 10 kHz to 1 MHz PROBE PHASENOISE PHNOTISE and the similar PRINT comman
346. nce Information lt Z0 val gt lt RDC val gt lt RAC val gt lt RHBAC val gt lt RHB val gt lt RTRAN val gt Power Switch lt power 0 1 2 W dbm gt Parameter Description port portnumber lt DC mag gt lt AC lt mag lt phase gt gt gt lt HBAC lt mag lt phase gt gt gt lt HB lt mag lt phase lt harm lt tone lt modharm lt modtone gt gt gt gt gt gt gt lt transient_waveform gt 156 The port number Numbered sequentially beginning with 1 with no shared port numbers DC voltage or power source value AC voltage or power source value HSPICE RF HBAC voltage or power source value HSPICE RF HB voltage current or power source value Multiple HB specifications with different harm tone modharm and modtone values are allowed phase is in degrees harm and tone are indices corresponding to the tones specified in the HB statement Indexing starts at 1 corresponding to the first harmonic of a tone modtone and modharm specify sources for multi tone simulation A source specifies a tone anda harmonic and up to 1 offset tone and harmonic modtone for tones and modharm for harmonics The signal is then described as V or mag cos 2 pi harm tone modharm modtone t phase Transient analysis Voltage or power source waveform Any one of waveforms AM EXP PULSE PWL SFFM or SIN Multiple transient descriptions are not allowed HSPI
347. nd OPTION SIM_SPEF_MAX_ITER in the HSPICE and HSPICE RF Command Reference Additional Post Layout Options Other post layout options are listed in Table 24 Table 24 Additional Post Layout Options Syntax Description SIM_DSPF_RAIL Or SIM_SPEF_RAIL SIM_DSPF_SCALER SIM_SPEF_SCALER Or SIM_DSPF_SCALEC SIM_SPEF_SCALEC 324 By default HSPICE RF does not back annotate parasitics of the power net To back annotate power net parasitics include one of these options in the netlist Default OFF ON expands nets in a power rail as it expands all nets Scales the resistance or capacitance values scaleR is the scale factor for resistance scaleC is the scale factor for capacitance HSPICE RF User Guide Z 2007 03 Chapter 13 Post Layout Analysis Post Layout Back Annotation Table 24 Additional Post Layout Options Continued Syntax Description SIM_DSPF_LUMPCAPS If HSPICE RF cannot back annotate an instance in a net Or because one or more instances are missing in the SIM_SPEF_LUMPCAPS hierarchical LVS ideal netlist then by default HSPICE RF does not evaluate the net Instead of ignoring all parasitic information for this net HSPICE RF includes these options to connect a lumped capacitor with a value equal to the net capacitance to this net Default ON adds lumped capacitance ignores other net contents SIM_DSPF_INSERROR_ HSPICE RF supports options to skip the unmatched Or instance an
348. ng LIN Analysis for a NMOS Low Noise Amplifier Example 2 Using HB Analysis for a Power Amplifier Example 3 Using HB Analysis for an Amplifier Example 4 Using HBOSC Analysis for a Colpitts Oscillator Example 5 Using HBOSC Analysis fora CMOS GPS VCO Example 6 Using Multi Tone HB and HBAC Analyses for a Mixer Example 7 Using Shooting Newton Analysis on a Driven Phase Frequency Circuit and a Ring Oscillator Example 1 Using LIN Analysis for a NMOS Low Noise Amplifier The LIN command simplifies the calculation of linear multi port transfer parameters and noise parameters In the LIN analysis Port P elements are used to specify port numbers and their characteristic impedances The analysis automatically computes the frequency dependent complex transfer coefficients between ports The result is a convenient means to get scattering parameters noise parameters stability parameters and gain coefficients The LIN command renders obsolete the NET command The output from the LIN HSPICE RF User Guide 15 Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 1 Using LIN Analysis for a NMOS Low Noise Amplifier command is saved in the scO file format that can in turn be referenced as a model file for the new S parameter element To set up a linear transfer parameter analysis the HSPICE input netlist must contain Use the AC command to activate small signal AC analysis and to specify a frequency sweep Also
349. nge for the input signal also referred to as the input frequency band IFB or fin You can specify LIN DEC OCT POI or SWEEPBLOCK Specify the nsteps start and stop frequencies using the following syntax for each type of sweep LIN nsteps start stop DEC nsteps start stop OCT nsteps start stop POI nsteps freq_values SWEEPBLOCK nsteps freq1 freq2 freqn DATA dataname NOISECALC Enables calculating the noise figure The default is no 0 FILENAME Specifies the output file name for the extracted S parameters or the object name after the o command line option The default is the netlist file name DATAFORMAT Specifies the format of the output data file dataformat RI real imaginary dataformat MA magnitude phase This is the default format for Touchstone files dataformat DB DB magnitude phase 302 HSPICE RF User Guide Z 2007 03 Chapter 11 S parameter Extraction Frequency Translation S Parameter HBLIN Extraction Parameter Description MIXEDMODE2PORT Describes the mixed mode data map of output mixed mode S parameter matrix The availability and default value for this keyword depends on the first two port P element configuration as follows case 1 p1 p2 single ended standard mode P element available ss default ss case 2 p1 p2 balanced mixed mode P element available dd cd dc cc default dd case 3 p1 balanced p2 single ended available ds cs default ds case 4 p1 s
350. nsmission line in three different ways Z TD L Zp NL F E with J and JLC values taken from a U model Scattering Parameter Data Element A transmission line is a passive element that connects any two conductors at any distance apart For more information about transmission lines see S parameter Modeling Using the S element in the HSPICE Signal Integrity Guide Frequency Dependent Multi Terminal S element When used with the generic frequency domain model MODEL SP an S element is a convenient way to describe the behavior of a multi terminal network The S element describes a linear time invariant system and provides a series of data that describe the frequency response of the system The S element is particularly useful for high frequency characterization of distributed passive structures A common use of the S element is in microwave circuits because electronic devices in this frequency domain no longer act as they do in low frequencies In this case distributed system parameters must be considered The S element uses the following parameters to define a frequency dependent multi terminal network S scattering parameter Y admittance parameter Note All HSPICE and HSPICE RF analyses can use the S element The S parameter is the reflection coefficient of the system which is measured through ratios of incident and reflected sinusoidal waves For passive systems the magnitude of an S parameter varies
351. nt 353 417 transfer sign function 102 transformer ideal 137 Trapezoidal TRAP integration algorithm 389 390 TREF model parameter 353 tutorial 15 overview 1 simulation engine 1 two tone HB 39 U UNIF keyword 363 uniform parameter distribution 359 unit_atto configuration option 401 V v_supply configuration option 401 variables changing in ALTER blocks 82 84 HSPICE specific 104 variance statistical 351 VCD format 393 VCO 31 vector modualted RF 180 vector modulated RF E element 186 F element 186 430 G element 186 H element 186 element 182 implementation 180 V element 182 vendor libraries 92 VMRF See vector modulated RF 180 Vnn node name in CSOS 80 Ww warnings floating power supply nodes 79 waveform display 12 WDB format 393 W elements 140 wildcard uses 76 402 wildcard_left_range configuration option 401 wildcard_match_all configuration option 402 wildcard_match_one configuration option 402 wildcard_right_range configuration option 402 worst case analysis 354 371 380 Worst Case Corners Analysis 350 X X variable 405 XL model parameter 355 XPHOTO model parameter 368 XW model parameter 355 Y yield analysis 350
352. nt name Must begin with R followed by up to 1023 alphanumeric characters n1 Positive terminal node name n2 Negative terminal node name mname Resistor model name Use this name in elements to reference a resistor model TC TC1 alias The current definition overrides the previous definition HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Passive Elements Parameter Description TC1 First order temperature coefficient for the resistor See the Passive Device Models chapter in the HSPICE Elements and Device Models Manual for temperature dependent relations TC2 Second order temperature coefficient for the resistor SCALE Element scale factor scales resistance and capacitance by its value Default 1 0 R Resistance value at room temperature This can be resistance a numeric value in ohms a parameter in ohms a function of any node voltages a function of branch currents any independent variables such as time hertz and temper M Multiplier to simulate parallel resistors For example for two parallel instances of a resistor set M 2 to multiply the number of resistors by 2 Default 1 0 AC Resistance for AC analysis Default Reff DTEMP Temperature difference between the element and the circuit in degrees Celsius Default 0 0 L Resistor length in meters Default 0 0 if you did not specify Lina resistor model W Resistor width Default 0 0 if you did not specify W in the model C Capac
353. nterpreted as spanning some number of sigma values You can arrive at this number i e sigma multiplier by specifying a corresponding Bit Error Rate The term BER corresponds to the unitless Bit Error Rate that allows for this conversion The following table shows conversion from various BER values into a sigma multiplier value which corresponds to the number of sigma standard deviations in converting from RMS to peak to peak values Bit Error Rate Sigma Multiplier 10 3 6 180 10 4 7 438 10 5 8 530 10 6 9 507 10 7 10 399 10 8 11 224 10 9 11 996 10 10 12 723 10 11 13 412 10 12 14 069 HSPICE RF User Guide 253 Z 2007 03 Chapter 9 Oscillator and Phase Noise Analysis Jitter Analysis Bit Error Rate Sigma Multiplier 10 13 14 698 10 14 15 301 10 15 15 883 10 16 16 444 254 These conversions are done in accordance with the relationship Equation 51 foe BER T where erfc is the complementary error function and is the Sigma Multiplier Support for peak to peak conversions is included for a continuous range of BER values from 10 lt x lt 10 and some values extrapolated outside this range Specification of the BER parameter results in the output of the Peak to Peak jitter value and not the RMS value Labels for the measurements show appropriate rms and p p labels A BER parameter set to BER 0 is equivalent to having no parameter and only results in the RMS cal
354. nts with each block containing N 1 rows for the frequency data The units for the env_time sweep are seconds HSPICE RF User Guide Z 2007 03 13 Post Layout Analysis Describes the post layout analysis flow including post layout back annotation DSPF and SPEF files linear acceleration check statements and power analysis Post Layout Back Annotation A traditional straightforward brute force flow runs an RC extraction tool that produces a detailed standard parasitic format DSPF file DSPF is the standard format for transferring RC parasitic information This traditional flow then feeds this DSPF file into the circuit simulation tool for post layout simulation A key problem is that the DSPF file is flat Accurately simulating a complete design such as an SRAM or an on chip cache is a waste of workstation memory disc space usage and simulation runtime Because this DSPF file is flat control and analysis are limited How do you set different options for different blocks for better trade off between speed and accuracy How do you perform a power analysis on a flat netlist to check the power consumption This traditional flow flattens all nodes after extraction so it is more difficult to compare the delay before and after extraction This traditional flow can also stress the limits of an extraction tool so reliability also becomes an issue HSPICE RF provides a flow that solves all of these problems
355. nts to perform analyses SAVE and LOAD Save and load operating point information DATA Create table for data driven analysis TEMP Set temperature analysis Output PRINT PROBE Statements to output variables MEASURE Statement to evaluate and report user defined functions of a circuit Library INCLUDE General include files Model and File Inclusion MODEL Element model descriptions HSPICE RF User Guide Z 2007 03 69 Chapter 4 Input Netlist and Data Entry Input Netlist File Composition Table 8 Input Netlist File Sections Continued Sections Examples Definition LIB Library End of END Required statement end of netlist netlist Input Netlist File Composition The HSPICE RF circuit description syntax is compatible with the SPICE input netlist format Figure 10 shows the basic structure of an input netlist Figure 10 Basic Netlist Structure Title line First line is automatically a comment Comments all lines beginning with an asterisk Input control statements Netlist body description of circuit topology ee Element and input control statements MODEL statements OPTION statements OPTION with option statements PRINT and other output statements E PE Analysis statement such as POWER TRAN control statements END The following is an example of a simple netlist file called inv_ckt in It shows a small inverter test case that measures the ti
356. ode 11 of the X1 subcircuit instance this example uses the x variable HSPICE RF maps node 11 to the node1 external node as shown in the first part of the PRINT statement The latter half of the PRINT statement illustrates that you can combine the X variable with I variables HSPICE RF User Guide Z 2007 03 Chapter 16 Advanced Features Generating Measurement Output Files Example 2 In this example the x variable finds the current through the in node of the S1 subcircuit subckt S1 in out R1 in inp 1K Cl inp O lu R2 in out 1K PROBE X in ends Generating Measurement Output Files You can make all of the same measurements with the MEASURE statement in HSPICE RF as you can in HSPICE The results of the MEASURE statements appear in a file with one of the following filename extensions mt for measurements in transient analysis ms for measurements in DC analysis ma for measurements in AC analysis mb for measurements in HB analysis mp for measurements in HBNOISE analysis For more information about MEASURE statements see the HSPICE and HSPICE RF Command Reference Optimization Like HSPICE HSPICE RF employs an incremental optimization technique This technique solves the DC parameters first then the AC parameters and finally the transient parameters To perform optimization create an input netlist file that specifies Optimization parameters with upper and lowe
357. ode LNA tuned for operation near 1 GHz kk M1 n4 n3 n5 n5 CMOSN 1 0 25u w 7 5u as 15p ad 15p ps 19u pd 19u m 80 M2 n6 nli n4 n4 CMOSN 1 0 25u w 7 5u as 15p ad 15p ps 19u pd 19u m 80 M3 rfo n6 gnd gnd CMOSN 1 0 25u w 7 5u as 15p ad 15p ps 19u pd 19u m 40 ri _vdd _n6 400 11 _n5 gnd 1 0 9nH 12 rfin n3 1 13nH 0 65n vvb _n1 gnd dc 1 19 bias for common base device vinb rfinb gnd dc 0 595 lchk rfin rfinb INFINITY Choke cblk rfin rfind INFINITY DC block vvdd _vdd gnd dc vdd rfb rfo n6 120 feedback kk kk Two tone input source DC blocked at this point kk Vin rfind gnd dc 0 power 1 z0 50 50 Ohm src HB Pin W 0141 S tone 1 HB Pin W 012 S tone 2 Rload rfo vdd R 255 kk HB test bench to measure IP3 and IP2 kk HB tones 900MEG 910MEG nharms 11 11 intmodmax 7 SWEEP Pin dBm 50 0 0 0 2 0 print HB P Rload P Rload 1 0 P Rload 2 0 P Rload 2 1 probe HB P Rload P Rload 1 0 P Rload 2 0 P Rload 2 1 kk Approximate parameters for MOSIS 0 25um process run T17B kk HSPICE RF User Guide 23 Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 3 Using HB Analysis for an Amplifier MODEL CMOSN NMOS LEVEL 49 VERSION 3 1 TNOM 27 TOX 5 8E 9 XJ 1E 7 NCH 2 3549E17 VTHO 0 3819327 K1 0 477867 K2 2 422759E 3 K3 1E 3 K3B 2 1606637 wo 1E 7 NLX 1 579864E 7 DVTOW 0 DVT1W 0 DVT2W 0 DVTO 0 5334651 DVT1 0 7186877 DVT2 0 5 U0 289
358. odels Manual HSPICE RF User Guide 7 Z 2007 03 Chapter 1 HSPICE RF Features and Functionality HSPICE and HSPICE RF Differences 8 HSPICE RF User Guide Z 2007 03 2 Getting Started Describes how to set up your environment invoke HSPICE RF customize your simulation redirect input and output and use the CosmosScope waveform display tool Before you run HSPICE RF you need to set up several environment variables You can also create a configuration file to customize your simulation run HSPICE RF accepts a netlist file from standard input and delivers the ASCII text simulation results to HTML or to standard output Error and warning messages are forwarded to standard error output Running HSPICE RF Simulations Use the following syntax to invoke HSPICE RF hspicerf a inputfile outputfile h v For a description of the hspicerf command syntax and arguments see section HSPICE RF Command Syntax in the HSPICE and HSPICE RF Command Reference Netlist Overview The circuit description syntax for HSPICE RF is compatible with the SPICE and HSPICE input netlist format For a description of an input netlist file and methods of entering data see chapter Input Netlist and Data Entry in the HSPICE Simulation and Analysis User Guide HSPICE RF User Guide 9 Z 2007 03 Chapter 2 Getting Started Parametric Analysis Extensions Parametric Analysis Extensions All major HSPICE RF analyses TRAN AC DC and HB s
359. olves for the frequency This command is intended to be applied to high Q oscillators that take a long time to reach steady state For these circuits standard transient analysis is too costly Low Q oscillators such HSPICE RF User Guide 313 Z 2007 03 Chapter 12 Envelope Analysis Envelope Simulation 314 as typical ring oscillators are more efficiently simulated with standard transient analysis Example envosc tone 250Meg nharms 10 env_step 20n env_stop 10u probenode v5 0 1 25 Fast Fourier Transform Form ENVFFT lt output_var gt lt NP value gt lt FORMAT keyword gt lt WINDOW keyword gt lt ALFA value gt Parameter Description output_var Any valid output variable NP The number of points to use in the FFT analysis NP must be a power of 2 If not a power of 2 then it is automatically adjusted to the closest higher number that is a power of 2 The default is 1024 FORMAT Specifies the output format NORM normalized magnitude UNORM unnormalized magnitude default WINDOW Specifies the window type to use RECT simple rectangular truncation window default BART Bartlett triangular window HANN Hanning window HAMM Hamming window BLACK Blackman window HARRIS Blackman Harris window GAUSS Gaussian window KAISER Kaiser Bessel window ALFA Controls the highest side lobe level and bandwidth for GAUSS and KAISER windows The default is 3 0 Description You use the ENVFFT command to pe
360. ong the representative sources available in HSPICE RF is the complex modulated RF source Also known as the Vector Modulated source it allows digital modulation of an RF carrier using in phase and quadrature components created from a binary data stream Vector Modulated RF Source Digital RF waveforms are typically constructed by modulating an RF carrier with in phase I and quadrature Q components In HSPICE RF this is accomplished using the Vector Modulated RF VMRF signal source The VMRF signal source function is supported both for independent voltage and current sources V and elements and with controlled sources E F G and H elements When used with independent sources a baseband data stream can be input in binary or hexadecimal format and the scheme used to divide the data into and Q signals can be specified m With controlled VMRF sources the modulating and Q signals can be separately specified with other signal sources such as a PWL source and then used as control inputs into the VMRF source Implementation The VMRF source is a mathematical implementation of the following block diagram HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Complex Signal Sources and Stimuli ee cos wt Serial to S t Parallel sin wt Q t Data in ee The following equation calculates the time and frequency domain stimuli from the quadrature modula
361. onstruct a design in which information flows from the top of the design down into the lowest hierarchical levels To centralize the control at the top of the design hierarchy set global parameters To construct a library of small cells that are individually controlled from within set oca parameters and build up to the block level This section describes the scope of parameter names and how HSPICE resolves naming conflicts between levels of hierarchy Library Integrity Integrity is a fundamental requirement for any symbol library Library integrity can be as simple as a consistent intuitive name scheme or as complex as libraries with built in range checking Library integrity might be poor if you use libraries from different vendors in a circuit design Because names of circuit parameters are not standardized between vendors two components can include the same parameter name for different functions For example one vendor might build a library that uses the name Tau as a parameter to control one or more subcircuits in their library Another vendor might use Tau to control a different aspect of their library If you set a global parameter named Tau to control one library you also modify the behavior of the second library which might not be the intent If the scope of a higher level parameter is global to all subcircuits at lower levels of the design hierarchy higher level definitions override lower level parameter values wit
362. ote Higher frequencies smaller times increase accuracy but only up to the minimum time step used in HSPICE 348 HSPICE RF User Guide Z 2007 03 14 Statistical and Monte Carlo Analysis Describes the features available in HSPICE RF for statistical analysis Overview Described in this chapter are the features available in HSPICE FF for statistical analysis These features are supported for HSPICE RF and differ from the enhanced statistical analysis features available for HSPICE described in the HSPICE Simulation and Analysis User Guide The following subjects are described in this chapter Application of Statistical Analysis Analytical Model Types m Simulating Circuit and Model Temperatures Worst Case Analysis Monte Carlo Analysis Worst Case and Monte Carlo Sweep Example Simulating the Effects of Global and Local Variations with Monte Carlo Application of Statistical Analysis When you design an electrical circuit it must meet tolerances for the specific manufacturing process The electrical yield is the number of parts that meet the electrical test specifications Overall process efficiency requires maximum yield To analyze and optimize the yield HSPICE RF supports statistical techniques and observes the effects of variations in element and model parameters HSPICE RF User Guide 349 Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Analytical Model Types Analytical Model Types
363. ource and OFF does not The default value is OFF PTDNOISE Output Syntax and File Format PTDNOISE output syntax allows for the output of one parameter onoise PROBE PTDNOISE lt onoise gt HSPICE RF User Guide Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Periodic Time Dependent Noise Analysis PTDNOISE PRINT PTDNOISE lt onoise gt Parameter Units Description onoise V JHe Noise voltage spectral density at each frequency point specified by frequency_sweep at the time point specified by time_value Output File Format The following PTDNOISE output files are generated depending on the user input File Description printptn Writes output from the PRINT statement when using HB to obtain the steady state solution ptn Writes output from the PROBE statement when using HB to obtain printsnptn snptn lis HSPICE RF User Guide Z 2007 03 the steady state solution Reports output from the PRINT statement when using SN to obtain the steady state solution Writes output from the PROBE statement when using SN to obtain the steady state solution Standard output file lis contains the following information Performance Statistics Log Number of Nodes Number of FFT Points Number of Equations Memory in use Maximum Krylov iterations Maximum Krylov Dimension Target GMRES Residual Gmres Residual Actual Krylov Iterations taken Freq
364. ow x y function 101 power amplifier 19 power amplifier IP3 22 POWER statement 417 power function 101 POWERDC statement 415 PRINT ENV command 315 printhl file 305 printls file 309 printss file 309 PROBE command 315 Probing Subcircuit currents 405 PTDNOISE input syntax 282 MEASURE 286 output file format 285 output syntax 284 overview 281 syntax 282 PTDNOISE command 281 pwr x y function 101 427 Index Q quality assurance 350 R R Element resistor 116 rcells reusing 106 rext_divider configuration option 401 reference temperature 80 353 reluctors 134 resistor frequency dependent 118 length parameter 115 linear 116 model name 114 node to bulk capacitance 115 width parameter 115 restricting output 403 results 41 reusing simulation output 395 415 417 RF demo files 55 tutorial examples 15 rise time example 413 verify 412 RSH model parameter 355 S S parameter extraction large signal 305 power dependent 297 small signal 305 saturable core elements 128 models 127 128 scale factors 66 SCALE parameter 115 schematic netlists 68 schematic netlists 68 scope of parameters 105 SEARCH option 93 SETUP time verification 414 sgn x function 102 Shooting Newton driven phase frequency circuit example 43 428 overview 219 ring oscillator example 50 sign function 102 signed power function 101 silicon on sapphire devices 80 SIM_ACCURACY option 390 SIM_ACTIVE option 320 323 324 32
365. owed phase is in degrees harm and tone are indices corresponding to the tones specified in the HB statement Indexing starts at 1 corresponding to the first harmonic of a tone modtone and modharm specify sources for multi tone simulation A source specifies a tone and a harmonic and up to 1 offset tone and harmonic modtone for tones and modharm for harmonics The signal is then described as V or I mag cos 2 pi harm tone modharm modtone t phase Transient analysis Any one of waveforms AM EXP PULSE PWL SFFM or SIN Multiple transient descriptions are not allowed HSPICE RF Power Switch When 0 default element treated as a voltage or current source When 1 or W element treated as a power source realized as a voltage source with a series impedance In this case the source value is interpreted as RMS available power in units of Watts When dbm element treated as a power source in series with the port impedance Values are in doms You can use this parameter for Transient analysis if the power source is either DC or SIN LIN analysis System impedance used when converting to a power source inserted in series with the voltage source Currently this only supports real impedance When power 0 z0 defaults to 0 When power 1 zO defaults to 50 ohms You can also enter zo val DC analysis Series resistance overrides z0 HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench
366. pabilities and also includes Steady state frequency domain analyses for linear and nonlinear circuits High performance transient analysis for faster simulation of high speed digital and analog circuits Port wise automated Ac analyses for S scattering parameters The LIN command invokes extraction of noise and linear transfer parameters of a multi port linear network Extracts the S parameter and generates the N port model This command is used in conjunction with the AC command to measure multiport S Y and Z parameters noise parameters stability and gain factors and matching coefficients Additionally it is used with the Port element which identifies the network ports and their impedances You can also use mixed mode with LIN The Port P element identifies ports used in LIN analysis multiport S Y or Z parameter and noise parameter extraction A port element behaves as a noiseless impedance or a voltage source in series with an impedance depending on the simulation being performed Different impedances can be specified for DC transient AC HB and HBAC analyses The S element describes a linear network using multi port S Y or Z parameters in the form of a frequency table These parameters can come from a LIN simulation or from physical measurement The standard Touchstone and CITIfile formats are supported in addition to a proprietary HSPICE format The syntax of voltage and current sources as well as Port
367. picerf to customize your HSPICE RF simulation HSPICE RF first searches for hspicerf in your current working directory then in your home directory as defined by SHOME The configuration options listed in Table 29 are available for your use RF User Guide 399 Chapter 16 Advanced Features Creating a Configuration File Table 29 Configuration File Options Keyword Description Example flush_waveform ground_floating_ node hier_delimiter html integer_node max_waveform_size 400 Flushes a waveform If you do not specify a percentage then the default value is 20 Uses IC statements to set floating nodes in a circuit to ground You can select three options for grounding floating nodes If set to 1 grounds only floating nodes gates bulk control nodes non rail bulk that are included in the IC set If setto 2 adds unconnected terminals to this set If set to 3 uses IC statements to ground all floating nodes including dangling terminals Changes the delimiter for subcircuit hierarchies from to the specified symbol Stores all HSPICE RF output in HTML format Removes leading zeros from node names For example HSPICE RF considers 0002 and 2 to be the same node Without this keyword 0002 and 2 are two separate nodes Automatically limits the waveform file size If the number is less than 5000 HSPICE RF resets it to 2G If you do not set the number HSPICE RF use
368. put Netlist and Data Entry Input Netlist File Guidelines Table 4 HSPICE HSPICE RF Netlists Net Name Special Characters Special Character Node Name Instance Name Parameter Name Delimiters Note character is cannot be the cannot be the first first character character element legal anywhere in the element key key letter only string first or included letter only dollar sign Included only Included only Included only In line comment avoid if after a character number in node name percent HSPICE y Included only HSPICE included n a only Included only one for HSPICE RF Illegal in HSPICE RF caret Included only SPICE y HSPICE included To the power of only avoid usage i e 2 5 two Included only for HSPICE RE Illegalin HSPICE raised to the fifth RF power amp ampersand HSPICE vy Included only Included only n a Included only for HSPICE RF asterisk HSPICE Included only HSPICE included Comment in both included only only avoid using HSPICE HSPICE avoid using in in parameter RF Wildcard node names names character Double Illegal for Illegalin HSPICE asterisk is To HSPICE RF RF the power of parentheses Illegal Illegal Illegal Token delimiter minus HSPICE Included only Included only n a included only avoid usage HSPICE RF y underscore V Included only Included only n a 60 HSPICE RF User Guide Z 2007 03 Chapter 4 Input Netlist and Data Entry Input Netlist F
369. quation 53 RMSJITTER With rad units the RMSUITTER is calculated as Equation 54 RMSJITTER Jrms1 With UI Unit Intervals the RMSJITTER is calculated as Equation 55 RMSJITTER S 2 0 T HSPICE RF User Guide 255 Z 2007 03 Chapter 9 Oscillator and Phase Noise Analysis References References 1 2 3 4 5 6 7 a 8 Lael 9 E Ngoya A Suarez R Sommet R Quere Steady State Analysis of Free or Forced Oscillators by Harmonic Balance and Stability Investigation of Periodic and Quasi Periodic Regimes International Journal of Microwave and Millimeter Wave Computer Aided Engineering Volume 5 Number 3 pages 210 223 1995 C R Chang M B Steer S Martin E Reese Computer Aided Analysis of Free Running Microwave Oscillators IEEE Trans on Microwave Theory and Techniques Volume 39 No 10 pages 1735 1745 October 1991 G D Vendelin Design of Amplifiers and Oscillators by the S Parameter Method John Wiley amp Sons 1982 A Demir A Mehrotra J Roychowdhury Phase Noise in Oscillators A Unifying Theory and Numerical Methods for Characterization in Proc IEEE DAC pages 26 31 June 1998 A Demir A Mehrotra and J Roychowdhury Phase Noise in Oscillators A Unifying Theory and Numerical Methods for Characterization EEE Trans Circuits System I Volume 47 pages 655 674 May 2000 A van der Ziel Noise in Solid State Devices and Circuits John
370. quency range Measuring PHASENOISE Analyses with MEASURE The MEASURE PHASENOISE syntax supports five types of measurements trigger target MEASURE PHASENOISE result TRIG trig var VAL trig val lt TD time_delay gt lt CROSS c gt lt RISE r gt lt FALL f gt TARG This measurement yields the result of the frequency difference between the trigger event and the target event find when HSPICE RF User Guide 243 Z 2007 03 Chapter 9 Oscillator and Phase Noise Analysis Phase Noise Analysis MEASURE PHASENOISE result FIND out_varl WHEN out_var2 out_val2 lt TD time_delay gt lt RISE r gt lt FALL f gt lt CROSS c gt MEASURE PHASENOISE result FIND out_varl At Input Frequency Band value The previous measurement yields the result of a variable value at a specific input frequency band IFB point MEASURE PHASENOISE result FIND out_varl WHEN out _var2 out _var3 The previous measurement yields the result at the input frequency point when out_var2 out_var3 MEASURE PHASENOISE result WHEN out_var2 out_var3 The previous measurement yields the input frequency point when out_var2 out_vars3 average RMS min max and peak to peak MEASURE PHASENOISE result lt RMS gt out_var lt FROM IFB1 gt lt TO IFB2 gt This measurement yields the RMS of out_var from frequency IFB1 to frequency IFB2 You can replace the lt RMS gt with lt AVG gt to
371. quirements may be significant if the number of time points is very large The maximum integer value is 50 OPTION LOADSNINIT filename Loads the operating point saved at the end of SN initialization which is used as initial conditions for the Shooting Newton method OPTION SAVESNINIT filename Saves the operating point at the end of HSPICE RF User Guide Z 2007 03 SN initialization sninit 221 Chapter 8 Steady State Shooting Newton Analysis SN Analysis Output SN Analysis Output 222 The output from SN analysis is generated in both time and frequency domains The time domain output variables are the same as for standard transient analysis individual nodal voltages V n7 n2 branch currents Vxx element power dissipation n element It is also possible to output the results from Shooting Newton analysis in terms of complex frequency domain output variables This output format is activated by using the SNFD keyword in the output syntax For output in the frequency domain the syntax is identical to the Harmonic Balance output syntax PRINT SNFD TYPE NODES ELEM INDICES PROBE SNFD TYPE NODES ELEM INDICES Parameter Description TYPE Specifies a harmonic type node or element TYPE can be one of the following Voltage type V voltage magnitude and phase in degrees VR real component VI imaginary component VM magnitude VP Phase in de
372. r double click on the signal nlp_I f The noise results are shown in Figure 8 Figure 9 on page 54 shows the resulting phase noise results for the oscillator HSPICE RF User Guide 53 Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 7 Using Shooting Newton Analysis on a Driven Phase Frequency Circuit and a Ring Oscillator 54 Figure 9 Ring Oscillator Phase Noise Ring Oscillator Phase Noise dBefHz 4 Hz nipi 100 0k Other Shooting Newton Analyses The following Shooting Newton Analyses are also supported by HSPICE RF but not used in this tutorial SNFT is equivalent to the FFT command in transient TRAN analysis SNFT uses Fourier transform to represent a time domain signal in the frequency domain For more information see Shooting Newton with Fourier Transform SNFT SNAC is used to perform a linear analysis of a driven or nonautonomous circuit where the linear coefficients are modulated by a periodic steady state signal The functionality is similar to the HBAC command For more information see Shooting Newton AC Analysis SNAC SNXF is used to calculate transfer functions from an arbitrary number of small signal sources to a designated output in a circuit under periodic steady state conditions For more information see Shooting Newton Transfer Function Analysis SNXF HSPICE RF User Guide Z 2007 03 Demonstration Input Files Chapter 3 HSPICE RF Tutorial Demonstration Input F
373. r model For example CO R1 C1 R2 C2 Attaching RC pairs increases the order of the equivalent circuit from the first CO order For X S and J pin types simulation ignores this generalized capacitance value but you should insert a 0 value as a place holder for format integrity The resistance value can be a real number or an exponent optionally followed by a real number You can enter an O ohms after the value capacitance Capacitance on a pin in farads for input I output O or bidirectional B pins Use as part of a resistance capacitance RC pair Optionally enter an F farads after the value unit K kilo M milli a a a a x_coordinate Location of a pin relative to the x horizontal axis y_coordinate Location of a pin relative to the y vertical axis 330 HSPICE RF User Guide Z 2007 03 Chapter 13 Post Layout Analysis Post Layout Back Annotation Table 25 DSPF Parameters Continued Parameter Definition capacitor_ statements SPICE type statements that define capacitors in the subcircuit resistor_ statements SPICE type statements that define resistors in the subcircuit subcircuit_call_ Statements that call the subcircuit from higher level circuits statements END Marks the end of the file optional HSPICE RF User Guide 331 Z 2007 03 Chapter 13 Post Layout Analysis Post Layout Back Annotation DSPF File Example DSPF 1 0 DESIGN my circuit DATE June 15
374. r 4 Input Netlist and Data Entry Subcircuit Library Structure Subcircuit Library Structure Your library structure must adhere to the INCLUDE statement specification in the implicit subcircuit You can use this statement to specify the directory that contains the lt subname gt inc subcircuit file and then reference the lt subname gt in each subcircuit call The component naming conventions for each subcircuit is lt Subname gt inc Store the subcircuit in a directory that is accessible by a OPTION SEARCH lt lib path gt statement Create subcircuit libraries in a hierarchy Typically the top level subcircuit fully describes the input output buffer any hierarchy is buried inside The buried hierarchy can include model statements lower level components and parameter assignments Your library cannot use LIB or INCLUDE statements anywhere in the hierarchy HSPICE RF User Guide 93 Z 2007 03 Chapter 4 Input Netlist and Data Entry Subcircuit Library Structure 94 HSPICE RF User Guide Z 2007 03 5 Parameters and Functions Describes how to use parameters within HSPICE RF netlists Parameters are similar to the variables used in most programming languages Parameters hold a value that you assign when you create your circuit design or that the simulation calculates based on circuit solution values Parameters can store static values for a variety of quantities resistance source voltage rise time and
375. r 7 Steady State Harmonic Balance Analysis Harmonic Balance Analysis Table 16 HB Analysis Options Continued Option Description HBMAXITER HBSOLVER HBTOL LOADHB SAVEHB TRANFORHB Specifies the maximum number of Newton Raphson iterations that the HB engine performs Analysis stops when the number of iterations reaches this value The default is 10000 Specifies a preconditioner to solve nonlinear circuits HBSOLVER 0 invokes the direct solver HBSOLVER 1 default invokes the matrix free Krylov solver HBSOLVER 2 invokes the two level hybrid time frequency domain solver The absolute error tolerance for determining convergence Must be a real number that is greater than zero The default is 1 e 9 LOADHB filename loads the state variable information contained in the specified file These values are used to initialize the HB simulation SAVEHB filename saves the final state that is the no sweep point or the steady state of the first sweep point variable values from a HB simulation in the specified file This file can be loaded as the starting point for another simulation by using a LOADHB option TRANFORHB 1 forces HB to recognize V I sources that include SIN PULSE VMRF and PWL transient descriptions and to use them in analysis However if the source also has an HB description analysis uses the HB description instead TRANFORHB 0 forces HB to ignore transient
376. r Guide 153 Z 2007 03 Chapter 6 Testbench Elements Multi Terminal Linear Elements tn ee Sij Se toe SR Equation 7 lai j do on df fis the target frequency which you can set using DELAYFREQ The default target frequency is the maximum frequency point Os is the phase of Sij After time domain analysis obtains the group delay matrix the following equation eliminates the delay amount from the frequency domain system transfer function mn A A j The convolution process then uses the following equation to calculate the delay r T Equation9 iky O may Y kap Y nay X Carey Y20 Tey gt YNtH Tey Pre Conditioning S parameters Certain S parameters such as series inductor 2 port show a singularity when converting S to Y parameters To avoid this singularity the S element adds kR ef series resistance to pre condition S matrices Equation 10 S kl 2 k S 2 bI ks Rye iS the reference impedance vector kis the pre conditioning factor To compensate for this modification the S element adds a negative resistor kRre to the modified nodal analysis NMA matrix in actual circuit compensation To specify this pre conditioning factor use the PREFAC keyword in the S model statement The default pre conditioning factor is 0 75 154 HSPICE RF User Guide Z 2007 03 Figure 19 Pre Conditioning S parameters Chapter 6 Testbench Elements Port Element gt Preconditioning l
377. r available in HSPICE RF is the Generalized Minimum Residual GMRES Solver a Krylov technique and uses a matrix implicit algorithm HSPICE RF User Guide 199 Z 2007 03 Chapter 7 Steady State Harmonic Balance Analysis Harmonic Balance Analysis 200 Features Supported HB supports the following features All existing HSPICE RF models m Unlimited number of independent input tones Sources with multiple HB specifications SIN PULSE VMRF and PWL sources with TRANFORHB 1 Prerequisites and Limitations The following prerequisites and limitations apply to HB Requires one HB statement Treats sources without a DC HB or TRANFORHB description as a zero value for HB unless the sources have a transient description in which case the time 0 value is used as a DC value Input Syntax Without SS_TONE HB TONES lt F1 gt lt F2 gt lt gt lt FN gt SUBHARMS SH lt NHARMS lt H1 gt lt H2 gt lt gt lt HN gt gt lt INTMODMAX n gt SWEEP parameter_sweep With SS_TONE HB TONES lt F1 gt lt F2 gt lt gt lt FN gt lt NHARMS lt H1 gt lt H2 gt lt gt lt HN gt gt lt INTMODMAX n gt lt SS_ TONE n gt SWEEP parameter sweep Parameter Description TONES Fundamental frequencies SUBHARMS Allows subharmonics in the analysis spectrum The minimum non DC frequency in the analysis spectrum is f subharms where f is the frequency of oscillation HSPICE
378. r boundary values along with an initial guess An AC DC TRAN HB or HBOSC optimization statement HSPICE RF User Guide 407 Z 2007 03 Chapter 16 Advanced Features Optimization 408 An optimization model statement Optimization measurement statements for optimization parameters If you provide the input netlist file optimization specifications limits and initial guess then the optimizer reiterates the simulation until it finds an optimized solution Usage Notes and Examples Optimization works for TRAN AC DC HB HBOSC and HBAC analyses You can add the GOAL options in every meaningful MEASURE statement like FIND WHEN FIND AT and so forth A data sweep is not required to be defined in the HB statement for HB optimization to use the measured result from MEASURE HBNOISE PHASENOISE Or HBTRAN statements Therefore parameter sweep is not supported for this type of optimization Optimize multiple parameters with multiple goals by selecting MODEL OPT LEVEL 0 modified Lavenberg Marquardt method Optimize single parameters in single measurement situations by selecting MODEL OPT LEVEL 1 bisection method Examples e Setting optimization parameters param W opt1 231u 100u 800u param Rs opt1 10 8 20 e Optimization analysis statement HB tones 2 25g 2 5g nharms 6 3 sweep Pin dbm 30 0 2 sweep optimize optl results gain measure result to tune the parameters mo
379. r connects from the input node to ground Its inductance is determined by the product of the current through the inductor and 1E 6 Lcoil input gnd L 1u i input LTYPE 0 Example 4 The L99 inductor connects from the in node to the out node Its inductance is determined by the polynomial L cO c1 i c2 i i where i is the current through the inductor The inductor has a specified DC resistance of 10 ohms L99 in out POLY 4 0 0 35 0 01 R 10 HSPICE RF User Guide 127 Z 2007 03 Chapter 6 Testbench Elements Passive Elements 128 Example 5 The L inductor connects from node 1 to node as a magnetic winding element with 10 turns of wire L 1 2 NT 10 Mutual Inductors General form Kxxx Lyyy Lzzz lt K coupling coupling gt Mutual core form Kaaa Lbbb lt Lccc lt Lddd gt gt mname lt MAG magnetization gt Parameter Description KXXX Mutual inductor element name Must begin with K followed by up to 1023 alphanumeric characters Lyyy Name of the first of two coupled inductors Lzzz Name of the second of two coupled inductors K coupling Coefficient of mutual coupling K is a unitless number with magnitude gt 0 and lt 1 If K is negative the direction of coupling reverses This is equivalent to reversing the polarity of either of the coupled inductors Use the K coupling syntax when using a parameter value or an equation and the keyword k can be omitted Kaaa Saturable core element na
380. re harmonics result in higher accuracy but also longer simulation times and higher memory usage One or more signal sources for driving the circuit in HB analysis if the circuit is driven In the case of autonomous oscillator analysis no signal source is required Signal sources are specified using the HB keyword on the voltage or current source syntax Power sources are specified by setting the power switch on voltage current sources to 1 in this case the source value is treated as a power value in Watts instead of a voltage or current Optionally the netlist can also contain a set of control option for optimizing HB analysis performance The following example shows how to set up a Harmonic Balance analysis on an NMOS Class C Power Amplifier The example compares transient analysis results to Harmonic Balance results The following netlist performs both a transient and a Harmonic Balance analysis of the amplifier driven by a sinusoidal input waveform The accurate option is set to ensure sufficient number of time points for comparison with HB This example is included with the HSPICE RF distribution as pa sp and is available in directory lt installdir gt demo hspicerf examples HSPICE RF User Guide Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 2 Using HB Analysis for a Power Amplifier options POST accurate param 0 950e6 PI 3 1415926 Ld 2e 9 Rload 5 Vin 3 0 _param Lin 0 1n Vdd 2 Cd 1 0 4 PI PI f 0 0 Ld M1 dra
381. rea is 100 microns Example 3 In the Jdrive JFET element below Jdrive driver in output model jfet W 10u L 10u HSPICE RF User Guide 165 Z 2007 03 Chapter 6 Testbench Elements Active Elements 166 The drain connects to the driver node The source connects to the in node The gate connects to the output node m model_jfet references the JFET model The width is 10 microns The length is 10 microns MOSFETs Mxxx nd ng ns lt nb gt mname lt lt L gt length gt lt lt W gt width gt lt AD val gt AS val gt lt PD val gt lt PS val gt lt NRD val gt lt NRS val gt lt RDC val gt lt RSC val gt lt OFF gt lt IC vds vgs vbs gt lt M val gt lt DTEMP val gt lt GEHO val gt lt DELVTO val gt OPTION WL Mxxx nd ng ns lt nb gt mname lt width gt lt length gt lt other_options gt Parameter Description Mxxx MOSFET element name Must begin with M followed by up to 1023 nd ng ns nb mname AD alphanumeric characters Drain terminal node name Gate terminal node name Source terminal node name Bulk terminal node name which is optional To set this argument in the MOSFET model use the BULK parameter MOSFET model name reference MOSFET channel length in meters This parameter overrides OPTION DEFL with a maximum value of 0 1m Default DEFL MOSFET channel width in meters This parameter overrides OPTION DEFW Default DEFW Drain diffusio
382. regl freq2 nharms 6 6 sweep power DEC 10 le 6 le 3 MEASURE HB Pf1dBm FIND 10 LOG P Rload 1 0 1 e 3 AT le 5 P f1 at 10uW input MEASURE HB P2f1 f2dBm FIND 10 LOG P Rload 2 1 1 e 3 AT le 5 P 2f1 f2 at 10uW input MEASURE HB OIP3dBm PARAM 0 5 3 Pf 1dBm P2f1 f 2dBm MEASURE HB IIP3dBm PARAM OIP3dBm Pf 1dBm 20 0 MEASURE HB AM2PM DERIV VP outp outn 1 AT 1le 5 AM to PM Conversion in Deg Watt If you do not specify an HB sweep then MEASURE assumes a single valued independent variable sweep You can apply the measurements to current voltage and power waveforms The independent variable for measurements is the swept variable Such as power not the frequency axis corresponding to a single HB steady state point HSPICE RF also supports the MEASURE HBTRAN HBTR syntax Similar to the PROBE and PRINT HBTR statements in the section Calculations for Time Domain Output on page 211 a MEASURE HBTR statement is applied on the signals obtained in the same way Moreover like a MEASURE Statement in transient analysis the independent variable in a MEASURE HBTR Statement is time HSPICE RF optimization can read the data from MEASURE HB and MEASURE HBTR statements The optimization syntax in HSPICE RF is identical to that in the HSPICE for details see Statistical Analysis and Optimization in the HSPICE Simulation and Analysis User Guide Due to the difference in the in
383. rform Fast fourier Transform FFT on envelope output This command is similar to the FFT command The only difference is that transformation is performed on real data with the FFT command and with the ENVFFT command the data being transformed is HSPICE RF User Guide Z 2007 03 Chapter 12 Envelope Analysis Envelope Simulation complex You usually want to do this for a specific harmonic of a voltage current or power signal Example envfft v out 1 Output Syntax The results from envelope simulation can be made available through the PRINT PROBE and MEASURE commands This section describes the basic syntax you can use for this purpose PRINT or PROBE You can print or probe envelope simulation results by using the following commands PRINT ENV ovl lt ov2 gt PROBE ENV ovl lt ov2 gt Where ov1 are the output variables to print or probe MEASURE In HSPICE RF the independent variable for envelope simulation is the first tone Otherwise and except for the analysis type the MEASURE statement syntax is the same as the syntax for HB for example MEASURE ENV result Envelope Output Data File Format The results of envelope simulations are written to ev data files by the PROBE statement The format of an ev data file is equivalent to an hb data file with the addition of one fundamental parameter sweep that represents the slowly varying time envelope variation t of the Four
384. rform linear analysis of autonomous oscillator or nonautonomous driven circuits where the linear coefficients are modulated by a periodic steady state signal Multitone HBAC analysis extends single tone HBAC to quasi periodic systems with more than one periodic steady state tone One application of multitone HBAC is to more efficiently determine mixer conversion gain under the influence of a strong interfering signal than is possible by running a swept three tone HB simulation HSPICE RF User Guide 257 Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Multitone Harmonic Balance AC Analysis HBAC Prerequisites and Limitations The following prerequisites and limitations apply to HBAC Requires one and only one HBAC statement If you use multiple HBAC statements HSPICE RF uses only the last HBAC statement Requires one and only one HB statement Supports arbitrary number of tones Requires placing the parameter sweep in the HB statement Requires at least one HB source Requires at least one HBAC source Supports unlimited number of HB and HBAC sources The requested maximum harmonic in a PROBE or PRINT statement must be less than or equal to half the number of harmonics specified in harmonic balance that is max_harm lt num_hb_harms 2 Input Syntax HBAC lt frequency_sweep gt Parameter Description frequency_sweep Frequency sweep range for the input
385. ring on the parameter list You can use the path name to reference any node including any internal node Subcircuit node and element names follow the rules shown in Figure 11 on page 79 Figure 11 Subcircuit Calling Tree with Circuit Numbers and Instance Names 0 CKT 1 X1 2 X2 3 X3 4 X4 n abc is circuit number instance name sig24 sig25 sig26 In Figure 11 the path name of the sig25 node in the X4 subcircuit is X1 X4 sig25 You can use this path in HSPICE or HSPICE RF statements such as PRINT v X1 X4 sig25 HSPICE RF User Guide 79 Z 2007 03 Chapter 4 Input Netlist and Data Entry Input Netlist File Composition 80 Automatic Node Name Generation HSPICE or HSPICE RF can automatically assign internal node names To check both nodal voltages and branch currents you can use the assigned node name when you print or plot HSPICE or HSPICE RF supports several special cases for node assignment for example simulation automatically assigns node 0 as a ground node For CSOS CMOS Silicon on Sapphire if you assign a value of 1 to the bulk node the name of the bulk node is B Use this name to print the voltage at the bulk node When printing or probing current for example PROBE I R1 HSPICE inserts a zero valued voltage source This source inserts an extra node in the circuit named Vnn where nn is a number that HSPICE or HSPICE RF automatically generates this number appears
386. ription NHARMS Number of harmonics in the HB analysis triggered by the HBLSP statement POWERUNIT Power unit Default is watt SSPCALC Extract small signal S parameters Default is 0 NO NOISECALC Perform small signal 2 port noise analysis Default is 0 NO FILENAME Output data p2d filename Default is the netlist name or the object name after the o command line option DATAFORMAT Format of the output data file Default is ma magnitude angle FREQSWEEP Frequency sweep specification A sweep of type LIN DEC OCT POI or SWEEPBLOCK Specify the nsteps start and stop times using the following syntax for each type of sweep LIN nsteps start stop DEC nsteps start stop OCT nsteps start stop POI nsteps freq_values SWEEPBLOCK blockname This keyword must appear before the POWERSWEEP keyword POWERSWEEP Power sweep specification A sweep of type LIN DEC OCT POI or SWEEPBLOCK Specify the nsteps start and stop frequencies using the following syntax for each type of sweep LIN nsteps start stop DEC nsteps start stop OCT nsteps start stop POI nsteps power_values SWEEPBLOCK blockname This keyword must follow the FREQSWEEP keyword HSPICE RF User Guide 307 Z 2007 03 Chapter 11 S parameter Extraction Large Signal S parameter HBLSP Analysis 308 Note The FREQSWEEP and POWERSWEEP keywords must appear at the end of an HBLSP statement Examples Example 1does 2 por
387. rm SN data into the time domain and output by using the following syntax PRINT SNTRAN ov1 ov2 ovN PROBE SNTRAN ov1 ov2 ovN See TYPE above for voltage and current type definitions 266 SNAC Output Data Files A SNAC analysis produces these output data files Output from the PRINT statement is written to a printsnac file This data is against the IFB points The header contains the large signal fundamental and the range of small signal frequencies The columns of data are labeled as F Hz followed by the output variable names Each variable name has the associated mixing pair value appended All N variable names and all M mixing pair values are printed for each swept small signal frequency value a total of N M for each frequency value Output from the PROBE statement is written to a snac file Reported performance log statistics are written to a is file Number of nodes Number of FFT points Number of equations Memory in use CPU time Maximum Krylov iterations Maximum Krylov dimension Target GMRES residual GMRES residual HSPICE RF User Guide Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Shooting Newton AC Analysis SNAC m Actual Krylov iterations taken Frequency swept input frequency values Errors and Warnings The following error and warning messages are used when HSPICE encounters a problem with a SNAC analysis Error Messages SN
388. rs Each sweepspec Can specify a linear logarithmic or point sweep by using one of the following forms start stop increment lin npoints start stop dec npoints start stop oct npoints start stop poi npoints pl p2 Example The following example specifies a logarithmic sweep from 1 to 1e9 with more resolution from 1e6 to 1e7 sweepblock freqsweep dec 10 1 1g dec 1000 1lmeg 10meg Using SWEEPBLOCK in a DC Parameter Sweep To use the sweepblock in a DC parameter sweep use the following syntax DC sweepspec Sweepspec Sweepspec Each sweepspec can be a linear logarithmic point or data sweep or it can be in the form variable SWEEPBLOCK swblockname HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Clock Source with Random Jitter The SWEEPBLOCK syntax sweeps the specified variable over the values contained in the SWEEPBLOCK Example dc vinl 0 5 0 1 vin2 sweepblock vin2vals Using in Parameter Sweeps in TRAN AC and HB Analyses To use the sweepblock in parameter sweeps on TRAN AC and HB commands and any other commands that allow parameter sweeps use the following syntax variable sweepblock swblockname Example 1 tran 1n 100n sweep rout sweepblock rvals AC and HBAC analysis frequency sweeps can use sweepblock swblockname to specify the frequency values Example 2 ac sweepblock freqsweep Limitations You cannot use recursive SWEEPBLOCK specifications That is a SWEEPBLO
389. rs The circuit simulator uses parameter variations to predict how an actual circuit responds to extremes in the manufacturing process Physically measurable model parameters are called skew parameters because they skew from a statistical mean to obtain predicted performance variations Examples of skew parameters are the difference between the drawn and physical dimension of metal postillion or active layers on an integrated circuit Generally you specify skew parameters independently of each other so you can use combinations of skew parameters to represent worst cases Typical skew parameters for CMOS technology include XL polysilicon CD critical dimension of the poly layer representing the difference between drawn and actual size XW XW active CD critical dimension of the active layer representing the difference between drawn and actual size TOX thickness of the gate oxide RSH RSH resistivity of the active layer DELVTO DELVTO variation in threshold voltage You can use these parameters in any level of MOS model within the HSPICE RF device models The DELVTO parameter shifts the threshold value HSPICE HSPICE RF User Guide Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Worst Case Analysis RF adds this value to vTo for the Level 3 model and adds or subtracts it from VFBO for the BSIM model Table 28 shows whether HSPICE RF adds or subtracts deviations from the a
390. rs are the same as those for standard parameters However measurement parameters are not defined in a PARAM statement but directly in the MEASURE statement Altering Design Variables and Subcircuits The following rules apply when you use an ALTER block to alter design variables and subcircuits in HSPICE RF HSPICE RF User Guide Z 2007 03 Chapter 4 Input Netlist and Data Entry Input Netlist File Composition If the name of a new element MODEL statement or subcircuit definition is identical to the name of an original statement of the same type then the new statement replaces the old Add new statements in the input netlist file You can alter element and MODEL statements within a subcircuit definition You can also add a new element or MODEL statement to a subcircuit definition To modify the topology in subcircuit definitions put the element into libraries To add a library use LIB to delete use DEL LIB If a parameter name in a new PARAM statement in the ALTER module is identical to a previous parameter name then the new assigned value replaces the old value If you used parameter variable values for elements or model parameter values when you used ALTER use the PARAM statement to change these parameter values Do not use numerical values to redescribe elements or model parameters If you used an OPTION statement in an original input file ora ALTER block to turn on an option you can turn th
391. s NF SSNF DSNF the output nodes in the HBNOISE statement The data is plotted as a function of the input frequency band IFB points Units are in V Hz 2 Simulation ignores ONOISE when applied to autonomous circuits NF and SSNF both output a single side band noise figure as a function of the IFB points NF SSNF 10 Log SSF Single side band noise factor SSF Total Noise at output at OFB originating from all frequencies Load Noise originating from OFB Input Source Noise originating from IFB DSNF outputs a double side band noise figure as a function of the IFB points DSNF 10 Log DSF Double side band noise factor DSF Total Noise at output at the OFB originating from all frequencies Load Noise originating from the OFB Input Source Noise originating from the IFB and from the image of IFB Output Data Files An HBNOISE analysis produces these output data files Output from the PRINT statement is written to a printpn file Output from the PROBE statement is written to a pn file Both the printon and pn files output data against the input frequency band points Standard output information is written to a lis file simulation time HBNOISE linear solver method HBNOISE simulation time total simulation time HSPICE RF User Guide Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Multitone Harmonic Balance Noise HBNO
392. s HBXF _ printxf xf Oscillator startup printev ev ENVOSC LIN analysis PRINT output printac PROBE output ac S noise output sc SnP S noise output sc citi SnP citi Phase Noise printsnpn pn PHASENOISE SN analysis printsn Sn Transient analysis TRAN printtr tr Using the CosmosScope Waveform Display CosmosScope has been enhanced to support viewing and processing of HSPICE RF output files This section presents a basic overview of how to use CosmosScope to view HSPICE RF output Type cscope on the UNIX command line to start the CosmosScope tool Choose File gt Open gt Plotfiles or just press CTRL O to open the Open Plotfiles dialog Use the Files of Type filter to find the HSPICE RF output file that you want to open Table 3 on page 11 lists the HSPICE FF file types When you open a file its contents appear in the Signal Manager window The Signal Manager lists all open plot files If you double click a plot file name a new window appears showing the contents of that plot file To plot one of the signals listed here in the active chart double click on the signal label To create a new chart use the File gt New menu Select either XY Graph Smith Chart or Polar Chart You can also use the first three icons in the toolbar to create new chart windows 12 HSPICE RF User Guide Z 2007 03 Chapter 2 Getting Started Using the CosmosScope Waveform Disp
393. s Multitone Harmonic Balance AC Analysis HBAC Parameter Description TYPE Specifies a harmonic type node or element TYPE can be one of the following Voltage type V voltage magnitude and phase in degrees VR real component VI imaginary component VM magnitude VP Phase in degrees VPD Phase in degrees VPR Phase in radians VDB dB units VDBM dB relative to 1 mV Current type current magnitude and phase in degrees IR real component Il imaginary component IM magnitude IP Phase in degrees IPD Phase in degrees IPR Phase in radians IDB dB units IDBM dB relative to 1 mV Power type P Frequency type hertz index hertz index1 index2 You must specify the harmonic index for the hertz variable The frequency of the specified harmonics is dumped NODES NODES or ELEM can be one of the following ELEM Voltage type a single node name n1 or a pair of node names n1 n2 Current type an element name elemname Power type a resistor resistorname or port portname element name Frequency type the harmonic index for the hertz variable The frequency of the specified harmonics is dumped 260 HSPICE RF User Guide Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Multitone Harmonic Balance AC Analysis HBAC Parameter Description INDICES Index to tones in the form n1 n2 nK 1 nj
394. s Parameter Scoping and Passing OPTION parhier lt global local gt PARAM DefPwid 1u SUBCKT Inv a y DefPwid 2u DefNwid 1lu Mp1 lt MosPinList gt pMosMod L 1 2u W DefPwid Mn1 lt MosPinList gt nMosMod L 1 2u W DefNwid ENDS Setthe OPTION PARHIER parameter scoping option to GLOBAL The netlist also includes the following input statements xInvO a yO Inv S override DefPwid default xInv0O Mp1 width 1lu xInvl a yl Inv DefPwid 5u override DefPwid 5u xInv1 Mp1 width 1u measure tran WidO param lv2 xInv0O Mp1 lv2 is the template for measure tran Widl param lv2 xInv1 Mp1 the channel width lv2 xInv1 Mp1 ENDS Simulating this netlist produces the following results in the listing file wid0 1 0000E 06 widl 1 0000E 06 If you change the OPTION PARHIER parameter scoping option to LOCAL xInvO a yO Inv not override param S DefPwid 2u xInvO Mp1 width 2u xInv1l a yl Inv DefPwid 5u override param S DefPwid 2u xInv1 Mp1l width 5u measure tran WidO param 1v2 xInv0 Mp1 S override the measure tran Widl param 1lv2 xInv1 Mp1 S global PARAM Simulation produces the following results in the listing file wid0 2 0000E 06 widl 5 0000E 06 Parameter Passing Figure 16 on page 110 shows a flat representation of a hierarchical circuit which contains three resistors HSPICE RF User Guide 109 Z 2007 03 Chapter 5 Parameters and Functions Parameter Scoping an
395. s at room temperature TC1 TC2 Specifies the temperature coefficient W Capacitor width L Capacitor length HSPICE RF User Guide 123 Z 2007 03 Chapter 6 Testbench Elements Passive Elements Parameter Description M Multiplier to simulate multiple parallel capacitors DTEMP Temperature difference between element and circuit SCALE Scaling factor IC Initial capacitor voltage Frequency Dependent Capacitors You can specify frequency dependent capacitors using the C equation with the HERTZ keyword The HERTZ keyword represents the operating frequency In time domain analyses an expression with the HERTZ keyword behaves differently keyword Syntax Cxxx nil n2 C lt FBASE val gt according to the value assigned to the CONVOLUTION equation lt CONVOLUTION 0 1 2 lt FMAX val gt gt Parameter Description n1 n2 equation CONVOLUTION 124 Names or numbers of connecting nodes Expressed as a function of HERTZ If CONVOLUTION 1 or 2 and HERTZ is not used in the equation CONVOLUTION is turned off and the capacitor behaves conventionally The equation can be a function of temperature but it does not support variables of node voltage branch current or time If these variables exist in the expression and CONVOLUTION 1 or 2 then only their values at the operating point are considered in calculation Specifies the method used 0 default HERTZ 0 in time domain analys
396. s a double side band noise figure as a function of the IFB points DSNF 10 Log DSF Double side band noise factor DSF Total Noise at output at the OFB originating from all frequencies Load Noise originating from the OFB Input Source Noise originating from the IFB and from the image of IFB Output Data Files An SNNOISE analysis produces these output data files Output from the PRINT statement is written to a printsnpn file Output from the PROBE statement is written to a snpn file Both the printsnpn and pn files output data against the input frequency band points Standard output information is written to a lis file simulation time SNNOISE linear solver method SNNOISE simulation time total simulation time Measuring SNNOISE Analyses with MEASURE Note A MEASURE SNNOTSE statement cannot contain an expression that uses a SNNOISE variable as an argument Also you cannot use a MEASURE SNNOISE statement for error measurement and expression evaluation of SNNOISE The MEASURE SNNOISE syntax supports four types of measurements Find when MEASURE SNNOISE result FIND out_varl At Input Frequency Band value HSPICE RF User Guide 279 Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Shooting Newton Noise Analysis SNNOISE 280 The previous measurement yields the result of a variable value at a specific IFB point MEAS
397. s are not of interest The lower the fO frequency the greater the amount of reduction For the syntax and description of this control option see OPTION SIM_LA in the HSPICE and HSPICE RF Command Reference HSPICE RF User Guide 345 Z 2007 03 Chapter 13 Post Layout Analysis Linear Acceleration You can choose one of two algorithms explained in the following sections m PACT Algorithm m PI Algorithm PACT Algorithm The PACT Pole Analysis via Congruence Transforms algorithm reduces the RC networks in a well conditioned manner while preserving network stability The transform preserves the first two moments of admittance at DC slope and offset so that DC behavior is correct see Figure 27 The algorithm preserves enough low frequency poles from the original network to maintain the circuit behavior up to a specified maximum frequency fO within the specified tolerance This approach is the most accurate of the two algorithms and is the default Figure 27 PACT Algorithm o is Z o E frequency PI Algorithm This algorithm creates a pi model of the RC network For a two port the pi model reduced network consists of e a resistor connecting the two ports and e a capacitor connecting each port to ground 346 HSPICE RF User Guide Z 2007 03 Chapter 13 Post Layout Analysis Linear Acceleration The result resembles the Greek letter pi Fora general multiport SIM_LA preserve
398. s for the various types of supported elements see the chapters about individual types of elements in this user guide Example 1 Q1234567 4000 5000 6000 SUBSTRATE BUTMODEL AREA 1 0 The preceding example specifies a bipolar junction transistor with its collector connected to node 4000 its base connected to node 5000 its emitter connected to node 6000 and its substrate connected to the SUBSTRATE node The BJTMODEL name references the model statement which describes the transistor parameters M1 ADDR SIG1 GND SBS N1 10U 100U The preceding example specifies a MOSFET named M1 where drain node ADDR gate node SIG1 source node GND substrate nodes SBS The preceding element statement calls an associated model statement N1 The MOSFET dimensions are width 100 microns and length 10 microns Example 2 M1 ADDR SIG1 GND SBS N1 wil w 1141 The preceding example specifies a MOSFET named M1 where drain node ADDR gate node sIG1 74 HSPICE RF User Guide Z 2007 03 Chapter 4 Input Netlist and Data Entry Input Netlist File Composition source node GND substrate nodes SBS The preceding element statement calls an associated model statement N1 MOSFET dimensions are algebraic expressions width w1 w and length 11 1 Defining Subcircuits You can create a subcircuit description for a commonly used circuit and include one or more references to the subcircuit in your netlist Use SUBCKT and MACRO statements to define s
399. s node including the time dependent noise n t Equation 57 Vn t v t n t by equating these two representations expanding in a Taylor series and dropping higher order terms as follows Equation 58 V t n t v t j t v t dv t dt j t t Equation 59 I t dv t dt j t In terms of variances jitter is then defined as Equation 60 ar j t n t dv t dt PTDNOISE Input Syntax PTDNOISE output time_value lt time_delta gt frequency sweep lt listfreq frequencies none all gt lt listcount val gt lt listfloor val gt lt listsources on off gt 282 HSPICE RF User Guide Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Periodic Time Dependent Noise Analysis PTDNOISE Parameter Description output time_value time_delta frequency_sweep HSPICE RF User Guide Z 2007 03 Can be an output node pair of nodes or 2 terminal element HSPICE RF references the equivalent noise output to this node or pair of nodes Specify a pair of nodes as V n n If you specify only one node V n n If you specify only one node V n then HSPICE RF assumes the second node is ground You can also specify a 2 terminal element name that refers to an existing element in the netlist Time point at which time domain noise is evaluated Specify either a time point explicitly such as TIME value where value is either numerical or a para
400. s optional JFET or MESFET model name reference Area multiplying factor that affects the BETA RD RS IS CGS and CGD model parameters Default 1 0 in units of square meters FET gate width in meters FET gate length in meters Sets initial condition to OFF for this element in DC analysis Default ON HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Active Elements Parameter Description IC vdsval Initial internal drain source voltage vdsval and gate source voltage vgsval VDS vgsval Use this argument when the TRAN statement contains VGS UIC The IC statement overrides it M Multiplier to simulate multiple JFETs or MESFETs in parallel The M setting affects all currents capacitances and resistances Default 1 DTEMP The difference between the element temperature and the circuit temperature in degrees Celsius Default 0 0 Only drain gate and source nodes and model name fields are required Node and model names must precede other fields Example 1 In the J1 JFET element below J1 1 2 3 model 1 The drain connects to node 1 The source connects to node 2 The gate connects to node 3 m model_1 references the JFET model Example 2 In the following Jopamp1 JFET element Jopampl dl g3 s2 b 1lstage AREA 100u The drain connects to the d1 node The source connects to the g3 node The gate connects to the s2 node stage references the JFET model The a
401. s the DC admittance between the ports and the total capacitance that connects the ports to ground All floating capacitances are lumped to ground Linear Acceleration Control Options Summary In addition to OPTION SIM_LA other options are available to control the maximum resistance and minimum capacitance values to preserve and to limit the operating parameters of the PACT algorithm Table 27 contains a summary of these control options For the syntax and descriptions of these control options see the respective section in the HSPICE and HSPICE RF Command Reference Table 27 PACT Options Syntax Description OPTION SIM_LA PACT PI Activates linear matrix reduction and selects between two methods OPTION SIM_LA_FREQ lt value gt Upper frequency where you need accuracy preserved value is the upper frequency for which the PACT algorithm preserves accuracy If value is 0 PACT drops all capacitors because only DC is of interest The maximum frequency required for accurate reduction depends on both the technology of the circuit and the time scale of interest In general the faster the circuit the higher the maximum frequency The default is 1GHz OPTION SIM_LA_MAXR lt value gt Maximum resistance for linear matrix reduction value is the maximum resistance preserved in the reduction SIM_LA assumes that any resistor greater than value has an infinite resistance and drops the resistor after reduction finishes The default is
402. s the default 2G If you do not set the line the file size has no limit flush waveform percents ground_floating_ node 1 hier delimiter htmlhspicerf test This example creates a file named test html in the current directory integer node max_waveform_ size 2000000000 HSPICE RF User Guide Z 2007 03 Chapter 16 Advanced Features Creating a Configuration File Table 29 Configuration File Options Continued Keyword Description Example negative_td Allows negative time delay input in pwl If you do not set port_element_ voltage_ matchload rext_divider skip_nrd_nrs unit_atto v_supply wildcard_left_range HSPICE RF User Guide Z 2007 03 piecewise linear with repeat p1 piecewise linear exp exponential rising time delay only sin damped sinusoidal pulse trapezoidal pulse and am amplitude modulation formats Allows the alternate Port element definition A Port element consists of a voltage source in series with a resistor For the explanation that follows let the user specified DC AC or transient value of the Port element be V and let the voltage across the overall port element be Vp By default HSPICE RF will set the internal voltage source value to V The value of Vp will be lower than V depending on the internal impedance and the network s input impedance With the alternate definition the internal voltage source value is adjusted to 2 V so that Vp V when
403. s value is then passed into the subcircuit and the resistor width gets this value Because the expression is the same for all subcircuits the value of parameter width will HSPICE RF User Guide Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Simulating the Effects of Global and Local Variations with Monte Carlo be the same for all subcircuits thus it expresses a global variation Therefore all resistors have the same width and the terminal voltages are the same In test5 sp if a different width is used for the subcircuits then the expressions are treated separately get local variation assigned and different values are passed into the subcircuit In test5 sp the differences inside of the expressions are kept numerically very small thus the differences from the different values of locwidth are dominant and the results look almost identical to the ones from test3 sp In test6 sp the resistor width is assigned inside of the subcircuit The variations get picked up from the top level Because each subcircuit is a separate entity the parameter w is treated as a separate reference thus each resistor will have its own value partly defined through the common value of globwidth and partly through the separate value of locwidth test7 sp has two resistors in the subcircuit Each device in each subcircuit has a separate reference to the variation therefore each device gets its own va
404. se Noise Analysis Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Chapter 11 S parameter Extraction Chapter 12 Envelope Analysis Chapter 13 Post Layout Analysis Chapter 14 Statistical and Monte Carlo Analysis Chapter 15 Using HSPICE with HSPICE RF Chapter 16 Advanced Features Describes the specialized elements supported by HSPICE RF for high frequency analysis and characterization and the syntax for the basic elements of a circuit netlist in HSPICE or HSPICE RF Describes how to use harmonic balance analysis for frequency driven steady state analysis Describes HSPICE RF steady state time domain analysis based on Shooting Newton Describes how to use HSPICE RF to perform oscillator and phase noise analysis on autonomous oscillator circuits Describes how to use harmonic balance based and Shooting Newton AC analysis as well as nonlinear steady state noise analysis and XF analysis Describes how to use periodically driven nonlinear circuit analyses as well as noise parameter calculation Describes how to use envelope simulation Describes the post layout flow including post layout back annotation DSPF and SPEF files linear acceleration check statements and power analysis Describes the features available in HSPICE RF for statistical analysis Describes how various analysis features differ in HSPICE RF as compared to standard HSPICE Describes how to invok
405. ser Guide 107 Z 2007 03 sets the default width for all MOS devices during a simulation Part of the definition is still in the top level circuit so this method can still make unwanted changes to library values without notification from the HSPICE simulator Table 14 compares the three primary methods for configuring libraries to achieve required parameter checking for default MOS transistor widths Table 14 Methods for Configuring Libraries Parameter Method Location Pros Cons Local Ona SUBCKT Protects library from global definition line circuit parameter definitions unless you override it Single location for default values Global At the global Works with all HSPICE An indiscreet user another level and versions vendor assignment or the on SUBCKT intervening hierarchy can definition lines change the library Cannot override a global value ata lower level Special OPTION DEFW Simple to do Third party libraries or other statement sections of the design might depend on OPTION DEFW Parameter Defaults and Inheritance Use the OPTION PARHIER parameter to specify scoping rules Syntax OPTION PARHIER lt GLOBAL LOCAL gt The default setting is GLOBAL Example This example explicitly shows the difference between local and global scoping for using parameters in subcircuits The input netlist includes the following HSPICE RF User Guide 108 Z 2007 03 Chapter 5 Parameters and Function
406. shows the family of transient analysis curves for the transient sweep of the sigma parameter from 3 to 3 from the file inv trO In the sweep HSPICE RF uses the values of sigma to update the skew parameters which in turn modify the actual NMOS and PMOS models Operating Point Results in Transient Analysis If you want to get OP results after every Monte Carlo simulation in transient analysis you can add the option opfile to the netlist OP results will all output to the file dpO Figure 39 Sweep of Skew Parameters from 3 Sigma to 3 Sigma 3 Sigma Skew Results 0 0 100p 200p 300p 400p 500p tis To view the measured results plot the inv mtO output file The plot in Figure 40 shows the measured pair delay and the total dissipative power as a function of the parameter sigma HSPICE RF User Guide 373 Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Worst Case and Monte Carlo Sweep Example 374 Figure 40 Sweep MOS Inverter Pair Delay and Power 3 Sigma to 3 Sigma Power and Delay as Function of Sigma i signal s_powerisigma i sigmar s_delayisigma sigmal Monte Carlo Results This section describes the output of the Monte Carlo analysis in HSPICE RF The plot in Figure 41 shows that the relationship between TOX against XL polysilicon width transistor length is completely random as set up in the input file To generate this plot in CosmosScope 1 Read in the fil
407. sical or electrical levels The physical level relies on physical distributions such as oxide thickness and polysilicon line width control The electrical level relies on actual capacitor measurements Physical Approach 1 Since oxide thickness control is excellent for small areas on a single wafer you can use a local variation in polysilicon to control the variation in capacitance for adjacent cells Next define a local poly line width variation and a global model level poly line width variation In this example e The local polysilicon line width control for a line 10 m wide manufactured with process A is 0 02 m for a 1 sigma distribution e The global model level polysilicon line width control is much wider use 0 1 m for this example The global oxide thickness is 200 angstroms with a 5 angstrom variation at 1 sigma The cap element is square with local poly variation in both directions The cap model has two distributions e poly line width distribution e oxide thickness distribution The effective length is HSPICE RF User Guide Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Worst Case and Monte Carlo Sweep Example Leff Ldrawn 2 DEL The model poly distribution is half the physical per side values Cla 1 0 CMOD W ELPOLY L ELPOLY Clb 1 0 CMOD W ELPOLY L ELPOLY C1C 1 0 CMOD W ELPOLY L ELPOLY C1D 1 0 CMOD W ELPOLY L ELPOLY 10U POLYWIDTH 0 05U 1SIGMA CAP MODEL USES 2 MODPOL
408. sis searching for frequencies in the vicinity of 2 4 GHz This example uses 11 harmonics with the probe inserted between the drainP and drainN nodes The probe voltage estimate is 1 0 V Example 3 Another method to define the probenode information is through a zero current source The following two methods define an equivalent HBOSC command 232 HSPICE RF User Guide Z 2007 03 Chapter 9 Oscillator and Phase Noise Analysis HB Simulation of Ring Oscillators Method 1 HBOSC tone 2 4G nharms 10 probenode drainP drainN 1 0 fspts 20 2 1G 2 7G Method 2 ISRC drainP drainN 0 HBOSCVPROBE 1 0 HBOSC tone 2 4G nharms 10 fspts 20 2 1G 2 7G In method 2 the PROBENODE information is defined by a current source in the circuit Only one such current source is needed and its current must be 0 0 with the HBOSC PROBENODE voltage defined through its HROSCVPROBE property HB Simulation of Ring Oscillators Ring oscillators require a slightly different simulation approach in HB Since their oscillation is due to the inherent delay in the inverters of the ring they are best modeled in the time domain and not in the frequency domain Also ring oscillator waveforms frequently approach square waves which require a large number of harmonics to be described in the frequency domain An accurate initial guess is important if they are going to be simulated accurately with HB HSPICE RF HB oscillator analysis typic
409. slewrate slewrate v sec Output signal slewrate at the time point specified by time_value HSPICE RF User Guide Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Multitone Harmonic Balance Transfer Function Analysis HBXF Measure File Format File Description msnptn Writes output from the MEASURE statement when using SN to obtain the steady state solution Error Handling and Warnings Error messages are generated under the following circumstances PTDNOISE frequency sweep includes negative frequencies PTDNOISE allows only frequencies that are greater than or equal to zero PTDNOISE time sweep includes negative times PTDNOISE allows only time points that are greater than or equal to zero m No SN statement is specified error at parser PTDNOISE requires an SN statement to generate the steady state solution Incorrect match to MEASURE statement A warning is issued for a PTDNOISE convergence failure When the gmres solver reaches the maximum number of iterations and the residual is greater than the specified tolerance PTDNOISE generates a warning and then continue as if the data were valid The Warning reports the following information Final GMRES Residual Target GMRES Residual Maximum Krylov Iterations Actual Krylov Iterations taken Multitone Harmonic Balance Transfer Function Analysis HBXF The HBXF command calculates the transfer function from a
410. source Fsub nl n2 vin 2 0 G Voltage controlled current source G12 4 0 3 0 10 H Current controlled voltage source H3 4 5 Vout 2 0 I Current source IA 2 6 le 6 J JFET or MESFET J1 7 2 3 GAASFET K Linear mutual inductor general form K1 L1 L2 1 L Linear inductor LX a b le 9 M MOS transistor M834 1 2 3 4 N1 P Port P1 in gnd port 1 z0 50 Q Bipolar transistor Q5 3 6 7 8 pnpl R Resistor R10 21 10 1000 S S parameter element S1 ndl nd2 s_model2 V Voltage source v1 8 0 5 W Transmission Line W1 inl 0 out1 0 N 1 L 1 T uu y uu HSPICE RF User Guide Z 2007 03 65 Chapter 4 Input Netlist and Data Entry Input Netlist File Guidelines 66 Table 6 Element Identifiers Continued Letter First Element Example Line Char X Subcircuit call X1 2 4 17 31 MULTI WN 100 LN 5 Hierarchy Paths A period indicates path hierarchy m Paths can be up to 1024 characters long Path numbers compress the hierarchy for post processing and listing files The OPTION PATHNUM controls whether the list files show full path names or path numbers Numbers You can enter numbers as integer floating point floating point with an integer exponent or integer or floating point with one of the scale factors listed in Table 7 Table 7 Scale Factors Scale Factor Prefix Symbol Multiplying Factor T tera T 1e 12 G giga G 1e 9 MEG or X mega M 1e 6 K kilo k 1e 3 M milli m 1e 3 U micro u 1e 6 N nano n 1e 9 HS
411. stant real number Algebraic expression of real values Predefined function Function that you define m Circuit value Model value To invoke the algebraic processor enclose a complex expression in single quotes A simple expression consists of one parameter name The parameter keeps the assigned value unless A later definition changes its value or An algebraic expression assigns a new value during simulation HSPICE does not warn you if it reassigns a parameter HSPICE RF User Guide 97 Z 2007 03 Chapter 5 Parameters and Functions Using Parameters in Simulation PARAM 98 Inline Parameter Assignments To define circuit values using a direct algebraic evaluation rl nl 0 R 1k sqrt HERTZ Resistance for frequency Parameters in Output To use an algebraic expression as an output variable in a PRINT PROBE or MEASURE statement use the PAR keyword Example PRINT DC v 3 gain PAR v 3 v 2 PAR v 4 v 2 User Defined Function Parameters You can define a function that is similar to the parameter assignment but you cannot nest the functions more than two deep An expression can contain parameters that you did not define A function must have at least one argument and can have up to 20 and in many cases more than 20 arguments You can redefine functions The format of a function is funcnamel argi arg2 expression1 funcname2 argli ar
412. statement 313 errors missing END statement 57 example comment line 72 configuration file 402 Monte Carlo 365 371 worst case 371 Index examples RF tutorials 15 exp x function 101 exponential function 101 expressions algebraic 99 Extended output variables 405 external data files 69 F fall time verification 413 files external data 69 81 hl 305 hspice ini 91 hspicerf 399 include files 68 s 310 multiple simulation runs 86 p2d 310 printhl 305 printls 309 printss 309 ss 310 files output 10 first character descriptions 63 flags 399 flush_waveform configuration option 400 format output DSPF 321 format output NW 394 WDB 393 Foster pole residue form E element 179 G element 179 frequency variable 104 frequency table model 153 frequency dependent capacitor 124 inductor 130 resistor 118 functions built in 100 104 table 100 423 Index G GAUSS functions 366 keyword 363 parameter distribution 359 generating output 10 global parameters 105 GND node 79 ground node name 79 ground_floating_ node configuration option 400 H Harmonic Balance HB 197 analysis spectrum 201 equations 199 errors 216 options 203 oscillator analysis 229 output 206 syntax 200 warnings 216 HB for HBLIN 300 HB analysis IP3 amplifier 22 power amplifier 19 HBAC 40 257 errors 262 267 example 40 output 259 315 output data files 261 syntax 258 warnings 262 267 HBAC analysis mixer 38 HBLIN
413. stationary Noise Analysis of Large RF Circuits with Multitone Excitations IEEE Journal of Solid State Circuits volume 33 pages 324 336 March 1998 11 K Kurakawa Power waves and the Scattering Matrix IEEE Trans 218 Microwave Theory Tech vol MTT 13 pp 194 202 March 1965 HSPICE RF User Guide Z 2007 03 8 Steady State Shooting Newton Analysis Describes HSPICE RF steady state time domain analysis based on a Shooting Newton algorithm SN Steady State Time Domain Analysis An advanced Shooting Newton SN algorithm provides additional performance and functionality to HSPICE RF for time domain steady state analysis Shooting Newton adds analysis capabilities for PLL components digital circuits logic such as ring oscillators frequency dividers phase frequency detectors PFDs and for other digital logic circuits and RF components that require steady state analysis but operate with waveforms that are more square wave than sinusoidal The Shooting Newton algorithm effectively analyzes applications including Ring oscillators see Chapter 9 Oscillator and Phase Noise Analysis Frequency dividers prescalers Mixer conversion gain Phase frequency detectors PFDs Mixer noise figure Functionality includes Both driven and oscillator autonomous analyses Time Domain or Frequency analysis based on advanced Shooting Newton algorithm Spectrum analysis specific to the SN analysis see Shooting
414. steps start stop POI nsteps freq_values SWEEPBLOCK BlockName Specify the frequency sweep range for the output signal HSPICE RF determines the offset frequency in the input sidebands Fin where Fin abs n FO Fout FO is the steady state fundamental tone and Fout is the output frequency SNXF then generates the transfer functions from all of the input sidebands the Fin values to the output frequency Fout Output Syntax This section describes the syntax for the SNXF PRINT and PROBE statements PRINT and PROBE Statements PRINT SNXF TYPE NODES ELEM PROBE SNXF TYPE NODES ELEM Parameter Description TYPE can be one of the following TFV existing source TFI placeholder value for the current source attached to the given node The transfer function is computed on the output variables and input current or voltage NODES ELEM NODES or ELEM can be one of the following HSPICE RF User Guide Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Shooting Newton Transfer Function Analysis SNXF Voltage type a single node name n1 or a pair of node names n1 n2 Current type an element name elemname Power type a resistor resistorname or port portname element name Output Data Files An SNXF calculation produces these output data files Output from the PRINT statement is written to a printsnxf file The output is in ohms siemens
415. t Format To use Veritools Undertow output format UT enter OPTION POST ut This format supports analog compression as described in Compressing Analog Files on page 396 The waveform list in UT format now displays in a hierarchical structure rather than one flat level as in previous versions CSDF Output Format To use CSDF output format CSDF enter OPTION POST csdf OPTION csdf overrides OPTION POST setting HSPICE RF User Guide 395 Z 2007 03 Chapter 15 Using HSPICE with HSPICE RF Compressing Analog Files Compressing Analog Files 396 Analog compression eliminates unnecessary data points from a HSPICE RF voltage or current waveform to reduce the size of the waveform file Eliminating Voltage Datapoints You use the SIM_DELTAV option to determine the selection criteria for HSPICE RF voltage waveforms in WDB or NW format For example OPTION SIM DELTAV lt value gt During simulation HSPICE RF checks whether the value of the X signal at the n timestep changes by more than the SIM_DELTAV option from its previous value at the n 7 timestep If yes then HSPICE RF saves the new data point m Otherwise this new data point is lost Typically such an algorithm yields a reduced file size with minimal resolution loss as long as you set an appropriate SIM_DELTAV value If a value for the SIM_DELTAV option is too large the waveform degrades Figure 47 Analog Compression Formats A E NW retains these
416. t Z be the impedance value of the zO port element Then the power wave flowing into the terminals of the port element at frequency index n can be computed according to 2 Equation 31 P n Va Zol Zo This power expression remains valid whether or not the port element includes an internal voltage source at the same frequency If the port element includes a voltage source at the same frequency you can use this power calculation to compute the magnitude of the related large signal scattering parameters If you expand the preceding formula the power delivered to a port element with real impedance Z is given by 2 2 2 1 Val Zo nl 1 Equation 32 P port n fe 5 Rew Th This power value represents the power incident upon and delivered to the port element s load impedance Zo due to other power sources in the circuit and due to reflections of its own generated power HSPICE RF User Guide 209 Z 2007 03 Chapter 7 Steady State Harmonic Balance Analysis Harmonic Balance Analysis 210 If the port element is used as a load resistor no internal source the preceding equation reduces to that for the simple resistor If you used the port element as a power source with non zero available power i e a non zero V and it is terminated in a matched load Zo the port power measurement returns 0 W because no power is reflected You can request power measurements in the form of complete spectra or i
417. t one Periodic source HSPICE RF User Guide 263 Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Shooting Newton AC Analysis SNAC Limited to simulations that can be reduced to a single tone SN analysis Supports unlimited number of sources The requested maximum harmonic in a PROBE or PRINT statement must be less than or equal to half the number of harmonics specified in the SN statement that is max_harm lt nharms 2 Input Syntax SNAC lt frequency_sweep gt Parameter Description frequency_sweep Frequency sweep range for the input signal also referred to as the input frequency band IFB or fin You can specify LIN DEC OCT POI or SWEEPBLOCK Specify the nsteps start and stop frequencies using the following syntax for each type of sweep LIN nsteps start stop DEC nsteps start stop OCT nsteps start stop POI nsteps freq_values SWEEPBLOCK nsteps freq1 freq2 freqn DATA dataname Output Syntax This section describes the syntax for the SNAC PRINT and PROBE statements These statements are similar to those used for HB analysis PRINT and PROBE Statements PRINT SN TYPE NODES ELEM INDICES PROBE SN TYPE NODES ELEM INDICES 264 HSPICE RF User Guide Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Shooting Newton AC Analysis SNAC Parameter Description TYPE Specifies a harmonic typ
418. t single tone power dependent S parameter extraction without frequency translation Frequency sweep The fundamental tone is swept from 0 to 1G Power sweep The power input at port 1 is swept from 6 to 10 Watts Five harmonics are required for the HB analysis Large signal S parameters are extracted on the first harmonic Five harmonics are required in the HBLSP triggered HB analysis The DC value in p1 statement is used to set DC bias which is used to perform small signal analyses Small signal S parameters are required extracted Small signal two port noise analysis is required The data will be output to the ex1 p2d file Example 1 2 Port Single Tone pl 1 0 port 1 dc lv p2 2 0 port 2 hblsp nharms 5 powerunit watt sspcalc 1 noisecalc 1 filename ex1 freqsweep lin 5 0 1G powersweep lin 5 6 10 Example 2 generates large scale S parameters as a function of input for a differential equalizer Example 2 4 Port Network hblsp example Opt post pl nl 0 port 1 ac 1 p2 n2 0 port 2 xxx put your DUT R1 nil n2 10 hblsp nharms 5 freqsweep lin 4 1k 10k powersweep lin 2 5 10 end HSPICE RF User Guide Z 2007 03 Chapter 11 S parameter Extraction Large Signal S parameter HBLSP Analysis Output Syntax This section describes the syntax for the HBLSP PRINT and PROBE statements These statements only support S and noise parameter outputs Node voltage branch current and
419. t which you evaluate a signal HSPICE RF discretely expresses the time continuum as a series of points At each point or timestep a circuit simulator evaluates the corresponding voltage or current value of a signal Thus a resulting signal waveform is a series of individual data points connecting these points results in a smooth curve You can apply different accuracy settings to different blocks or time intervals The syntax to set accuracy on a block instance or time interval is similar to the settings used for a power supply Note An OPTION SIM ACCURACY takes precedence over an OPTION ACCURATE For the syntax and description of this control option see OPTION SIM_ACCURACY in the HSPICE and HSPICE RF Command Reference Algorithm Control In HSPICE RF you can select the Backward Euler Trapezoidal Gear or hybrid method algorithms Each of these algorithms has its own advantages and disadvantages for specific circuit types These methods have tradeoffs related to accuracy avoidance of numerical oscillations and numerical damping of circuit oscillations For pre charging simulation or timing critical simulations the Trapezoidal algorithm usually improves accuracy OPTION METHOD You use the METHOD option to select a numeric integration method for a transient analysis HSPICE RF supports three basic timestep algorithms Trapezoidal TRAP second order Gear Gear 2 and Backward Euler BE Backward Euler is the HSPICE
420. taining the PWL data consisting of time and voltage or current pairs This file should not contain a header row unless it is a comment The PWL source data is obtained by extracting col1 and col2 from the file col1 lt col2 gt Time values are in col1 and voltage or current values are in col2 By default coli 1 and col2 2 R Repeat function When an argument is not specified the source repeats from the beginning of the function The argument repeated is the time in seconds which specifies the start point of the waveform being repeat The repeat time must be less than the greatest time point in the file TD Time delay in seconds of the PWL function options Any standard V or source options The sine wave behavior following the td time delay now becomes Equation21 vit e V V sin2mfo t ta 7450 OF dal HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Clock Source with Random Jitter The Syntax of COS source is Vxxx n n COS lt gt vo va lt freq lt td lt q lt j gt gt gt gt lt gt lt PERJITTER val lt SEED val gt gt Ixxx n n COS lt gt vo va lt freq lt td lt q lt j gt gt gt gt lt gt lt PERJITTER val lt SEED val gt gt The new cosine wave becomes tq Equation 22 vit e T We4V cos 2mfo t tat xte 7559 180 The syntax for the PULSE source is Vxxx n n PU lt LSE gt lt svl v2 lt td lt tr lt tf lt pw lt per gt gt gt
421. ted signal sources Equation 17 s t I t cos 2nf t d O t sin 2nfct Oy The discrete ideal in phase and Q quadrature signal components are digital Discrete values allow uniform scaling of the overall signal HSPICE RF generates data streams for the and Q signals based on interpreting the data string breaking the data string into a binary representation and then using the bit pairs to assign values for the and Q data streams For BPSK binary phase shift keying modulation the discrete signals are scaled so that Pies Q 1 Data In Data Q Data za a 2 2 1 a 1 J2 J2 HSPICE RF User Guide 181 Z 2007 03 Chapter 6 Testbench Elements Complex Signal Sources and Stimuli 182 For QPSK quadrature phase shift keying modulation the data stream is broken into bit pairs to form the correct and Q values This function is represented as the serial to parallel converter Data In Data Q Data 00 1 1 J2 J2 01 1 1 J2 J2 10 1 1 J2 J2 11 1 1 J2 J2 To generate a continuous time waveform the VMRF source takes the resulting digital and Q data streams and passes them through ideal filters Rectangular and Nyquist raised cosine filter options are available The output waveforms are therefore band limited according to the specified data rate Voltage and Current Source Elements The V and I elements can include VMRF signal sources that you can use to generate BPSK and QPSK waveforms
422. ter Example 1 Subcircuit default definition SUBCKT Inv A Y Wid 0 Inherit illegal values by default mpl lt NodeList gt lt Model gt L 1lu W Wid 2 mnl lt NodeList gt lt Model gt L lu W Wid ENDS Invoke symbols in a design x1 A Y1 Inv Bad No widths specified x2 A Y2 Inv Wid lu Overrides illegal value for Width This simulation aborts on the x1 subcircuit instance because you never set the required Wid parameter on the subcircuit instance line The x2 subcircuit simulates correctly Additionally the instances of the Inv cell are subject to accidental interference because the Wid global parameter is exposed outside the domain of the library Anyone can specify an alternative value for the HSPICE RF User Guide Z 2007 03 Chapter 5 Parameters and Functions Parameter Scoping and Passing parameter in another section of the library or the circuit design This might prevent the simulation from catching the condition on x1 Example 2 In this example the name of a global parameter conflicts with the internal library parameter named Wid Another user might specify such a global parameter in a different library In this example the user of the library has specified a different meaning for the Wid parameter to define an independent source Param Wid 5u Default Pulse Width for source vl Pulsed 0 Pulse Ov 5v Ou 0 1u 0 1u Wid 10u Subcircuit default definition SUBCKT Inv A Y Wid 0 Inherit ill
423. ter used to divide levels of hierarchy in a circuit path name Must be one of the following characters For example X1 X2 means that X2 is a subcircuit of the X1 circuit Character used to separate the name of an instance anda pin in a concatenated instance pin name Must be one of these characters eS Delimiter characters that precede and follow a bus bit or an arrayed instance number If these characters are not matching pairs HSPICE FF reports an error Valid bus delimiter prefix and suffix character pairs are brackets braces Y parentheses y or angle brackets lt gt gt A positive number For example 10 PS means use time units of 10 picoseconds 5 NS means use time units of 5 nanoseconds A positive number For example 10 PF means capacitance units of 10 picofarads 5 FF means use capacitance units of 5 femtoseconds Positive number For example 10 OHM sets resistance units to 10 ohms 5 KOHM sets resistance units to 5 kilo ohms A positive number For example 10 HENRY means use inductance units of 10 henries 5 MH means use inductance units of 5 millihenries 2 UH means use inductance units of 2 micro henries Name used throughout a SPEF file To reduce file space you can map other names to this name 337 Chapter 13 Post Layout Analysis Post Layout Back Annotation Table 26 SPEF Parameters Continued Parameter Definition name_id bit path name physical_re
424. tfloor are printed The default value is 1 0e 14 V Hz 2 listsources Prints the element noise value to the lis file when the element has multiple noise sources such as a FET which contains the thermal shot and 1 f noise sources You can specify either ON or OFF ON Prints the contribution from each noise source and OFF does not The default value is OFF Output Syntax This section describes the syntax for the SNNOISE PRINT and PROBE statements PRINT and PROBE Statements PRINT SNNOISE lt ONOISE gt lt NF gt lt SSNF gt lt DSNF gt PROBE SNNOISE lt ONOISE gt lt NF gt lt SSNF gt lt DSNF gt Parameter ONOISE NF SSNF Description Outputs the voltage noise at the output frequency band OFB across the output nodes in the SNNOISE statement The data is plotted as a function of the input frequency band IFB points Units are in V Hz 2 Simulation ignores ONOISE when applied to autonomous circuits NF and SSNF both output a single side band noise figure as a function of the IFB points NF SSNF 10 Log SSF Single side band noise factor SSF Total Noise at output at OFB originating from all frequencies Load Noise originating from OFB Input Source Noise originating from IFB HSPICE RF User Guide Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Shooting Newton Noise Analysis SNNOISE Parameter Description DSNF DSNF output
425. th multiple ALTER statements 84 85 libraries adding with LIB 85 ASIC cells 92 building 81 configuring 108 creating parameters 106 DDL 91 duplicated parameter names 106 END statement 82 integrity 105 search 91 subcircuits 93 vendor 92 LIMIT keyword 363 LIN analysis 15 linear acceleration 344 capacitor 123 inductor 130 Index matrix reduction 344 resistor 116 linear elements elements linear 139 local parameters 105 log x function 101 log10 x function 101 logarithm function 101 low noise amplifier 15 LPRINT statement 395 Is file 310 M macros 85 manufacturing tolerances 367 max x y function 102 max_waveform_size configuration option 400 mean statistical 351 measure 315 MEASURE ENV command 315 MEASURE statement parameters 99 MESFETs 164 min x y function 102 mixer 38 model cards 27 model parameters ALTER blocks 83 84 capacitance distribution 369 DELVTO 355 DTEMP 353 LENGTH 368 manufacturing tolerances 367 PHOTO 368 RSH 355 sigma deviations worst case analysis 355 skew 354 TEMP 80 353 temperature analysis 353 TOX 355 TREF 351 353 XPHOTO 368 MODEL statement 353 models Monte Carlo analysis 359 364 371 reference temperature 353 specifying 92 425 Index typical set 358 Monte Carlo analysis 350 351 371 380 distribution options 362 363 Monte Carlo analaysis operating point results in transient analysis 373 MONTE keyword 360 MOSFETs drain diffusion area 166 elements 16
426. tion elname node mname pname1 expression Element name that cannot exceed 1023 characters and must begin with a specific letter for each element type C Capacitor D Diode E F G H Dependent current and voltage sources Current inductance source JFET or MESFET Mutual inductor Inductor model or magnetic core mutual inductor model MOSFET BJT Port Resistor S parameter model T U W Transmission line V Voltage source X Subcircuit call ODVOSOSICAC Node names identify the nodes that connect to the element The node name begins with a letter and can contain a maximum of 1023 characters You cannot use the following characters in node names lt space gt HSPICE or HSPICE RF requires a model reference name for all elements except passive devices An element parameter name identifies the parameter value that follows this name Any mathematical expression containing values or parameters such as param1 val2 HSPICE RF User Guide 73 Z 2007 03 Chapter 4 Input Netlist and Data Entry Input Netlist File Composition Table9 Element Parameters Continued Parameter Description vall Value of the pname7 parameter or of the corresponding model node The value can be a number or an algebraic expression M val Element multiplier Replicates va element times in parallel Do not assign a negative value or zero as the M value For descriptions of element statement
427. tion Specifications 0005 15 Using HSPICE with HSPICE RF 00002 c eee eee RF Numerical Integration Algorithm Control 000000 eee RF Transient Analysis Accuracy Control 0 0 000 cece eee OPTION SIM_ACCURACY 0 0 0000 0 cece eee eee Algorithm Control 0 0 00 cee tee RF Transient Analysis Output File Formats 060 0e es ee eee Tabulated Data Output n s saanane WDB Output Format 0 00 00 cee TR Output Format 0 00000 cece eee XPOutputsFormMate cic tates belek wed awe eee ee bs yet Der hanes NW Output Format 0 0 00 000 ceca VCD Output Format a nus asana aaea 16 turboWave Output Format Undertow Output Format CSDF Output Format Compressing Analog Files Eliminating Voltage Datapoints 0 ce eee eee Eliminating Current Datapoints 00 0000 eee eee Advanced Features Creating a Configuration File Inserting Comments in a hspice File 000000 eens Using Wildcards in HSPICE RF Limiting Output Data Size SIM_POSTTOP Option SIM_POSTSKIP Option SIM_POSTAT Option SIM_POSTDOWN Option SIM_POSTSCOPE Option 382 384 384 385 386 386 389 389 389 390 390 392 393 393 394 394 394 394 395 395 395 396 396 396 399 399 402 402 403 403 404 404 404 405 xi Contents xii Probing Subcircuit Currents
428. to 1024 characters long including periods and extensions Node identifiers are used for node numbers and node names Node numbers are valid in the range of 0 through 9999999999999999 1 1e16 Leading zeros in node numbers are ignored Trailing characters in node numbers are ignored For example node 1A is the same as node 1 While node names can contain any of the following special characters 1 lt gt _ amp there are restrictions on their use as a first character See Node Naming Conventions on page 75 for restrictions For additional information see Node Naming Conventions on page 75 To make node names global across all subcircuits use a GLOBAL statement The 0 GND GND and GROUND node names all refer to the global HSPICE RF ground Simulation treats nodes with any of these names as a ground node and produces v 0 into the output files Instance Names The names of element instances begin with the element key letter see Table 6 except in subcircuits where instance names begin with X Subcircuits HSPICE RF User Guide Z 2007 03 Chapter 4 Input Netlist and Data Entry Input Netlist File Guidelines are sometimes called macros or modules Instance names can be up to 1024 characters long Table 6 Element Identifiers Letter First Element Example Line Char Cc Capacitor Cbypass 1 0 10pf D Diode D7 3 9 D1 E Voltage controlled voltage source Ea 1234K F Current controlled current
429. top frequencies using the following syntax for each type of sweep LIN nsteps start stop DEC nsteps start stop OCT nsteps start stop POI nsteps freq_values SWEEPBLOCK nsteps freq freq2 freqn DATA dataname OPTIMIZE OPT xxx MONTE val Example 1 SNOSC tone 900Meg nharms 9 trinit 10n oscnode gate Performs an oscillator analysis searching for periodic behavior after an initial transient analysis of 10 ns This example uses nine harmonics while searching for an oscillation at the gate node HSPICE RF User Guide Z 2007 03 237 Chapter 9 Oscillator and Phase Noise Analysis Phase Noise Analysis Example 2 SNOSC tone 2400MEG nharms 11 trinit 20n oscnode drainP Performs an oscillator analysis searching for frequencies in the vicinity of 2 4 Ghz This example uses 11 harmonics and a search at the drainP Phase Noise Analysis 238 Phase Noise analysis requires first running either harmonic balance HBOSC or Shooting Newton SNOSC analysis and then PHASENOISE analysis The PHASENOISE analysis itself is identical whether you run SNOSC or HBOSC Figure 22 shows a simple free running oscillator which includes a port with injected current Figure 22 Oscillator with Injected Current QO UN In D v t O O A An ideal oscillator would be insensitive to perturbations with a fixed amplitude frequency and phase represented by Equation 35 v t Acos ayt oo A nois
430. transient signal sources SN Reported Performance Log Statistics The following performance statistics are displayed DC operating time Initial transient time including user s time for circuit stabilization Total simulation time HSPICE RF User Guide 227 Z 2007 03 Chapter 8 Steady State Shooting Newton Analysis Shooting Newton with Fourier Transform SNFT SN time Total simulation time Memory used Final SN convergence residual error The value of the computed frequency if the circuit is autonomous Errors Warnings Error messages are displayed with convergence recommendations in cases of non convergence within the maximum number of Shooting Newton iterations Error messages are displayed for software errors such as segmentation violations and abort conditions such as unrecognized format i e unrecognized V I source faulty input values i e wrong sign out of range value unspecified values i e unspecified tone inconsistent values i e non commensurable tones duplicate values i e same entry given more than one the last one is always taken Limitations and Assumptions True distributed components such as ideal delays or transmission lines are not supported components with hidden states are not supported Example For a demonstration of using Shooting Newton analysis you can run the pdfcpGain sp file shipped with the HSPICE RF distribution located in the dir
431. trix that maintains the original port behavior yet achieves significant savings in memory and computation Thus the SIM_LA option is ideal for circuits with large numbers of resistors and capacitors such as clock trees power lines or substrate networks In general the RC elements are separated into their own network The nodes shared by both main circuit elements including PRINT PROBE and MEASURE statements and RC elements are the port nodes of the RC network All other RC nodes are internal nodes The currents flowing into the port nodes are a frequency dependent function of the voltages at those nodes The multiport admittance of a network represents this relationship The SIM_LA option formulates matrices to represent multiport admittance Then to eliminate as many internal nodes as possible it reduces the size of these matrices while preserving the admittance otherwise known as port node behavior The amount of reduction depends on the fO upper frequency the threshold frequency where SIM_LA preserves the admittance This is shown graphically in Figure 26 Figure 26 Multiport Admittance vs Frequency A admittance Jo frequency The SIM_LA option is very effective for post layout simulation because of the volume of parasitics For frequencies below fO the approx signal matches that of the original admittance Above fo the two waveforms diverge but presumably the higher frequencie
432. try If you specify lower diagonal terms the simulator converts that entry to the appropriate upper diagonal term If multiple entries are supplied for the same r c location then only the first one is used and a warning will be issued indicating that some entries are ignored All diagonal entries of the reluctance matrix must be assigned a positive value The reluctance matrix should be positive definite For the external file format the data files should contain three columns of data Each row should contain an r c val triplet separated by white space The r c and val values may be expressions surrounded by single quotes Multiple files may be specified to allow the reluctance data to be spread over several files if necessary 135 Chapter 6 Testbench Elements Passive Elements 136 Parameter Description SHORTALL SHORTALL yes all inductors in this model are converted to short circuits and all reluctance matrix values are ignored SHORTALL no default inductors are not converted to short circuits and reluctance matrix values are not ignored IGNORE_COUPLING IGNORE_COUPLING yes all off diagonal terms are ignored that is set to zero IGNORE_COUPLING no default off diagonal terms are not ignored Example This example has 9 segments or ports with 12 nodes and can potentially generate a 9x9 reluctance matrix with 81 elements L ThreeNetsal1l22a1bD4455bi1ic7788c1 RELUCTANCE 1 103
433. ts any current surge that is greater than surge_threshold for a duration of more than surge_width For additional information see SURGE in the HSPICE and RF Command Reference 419 Chapter 16 Advanced Features Detecting and Reporting Surge Currents 420 HSPICE RF User Guide Z 2007 03 Symbols IGND node 79 A abs x function 101 absolute power function 101 value function 101 AC statement 353 accuracy control 389 acos x function 100 adding library data 85 AGAUSS keyword 363 algebraic expressions 99 algorithm linear acceleration 346 nonlinear perturbation 242 numerical integration 389 390 periodic AC 242 ALTER blocks 82 83 84 statement 83 84 85 amplifier 15 19 amplifier IP3 22 analysis data driven 350 351 Monte Carlo 351 359 359 380 oscillator 229 phase noise 238 statistical 354 380 Taguchi 350 temperature 350 352 time domain steady state 219 worst case 350 354 380 yield 350 arccos x function 100 arcsin x function 100 arctan x function 101 arithmetic operators 100 Index ASIC libraries 92 asin x function 100 atan x function 101 AUNIF keyword 363 average deviation 351 B node name in CSOS 80 backslash continuation character 100 Backward Euler algorithm 389 390 integration 389 390 Behavioral resistors 118 BJTs elements names 162 block elements 138 broadband phasenoise 243 broadband phasenoise algorithm 243 buffer 155 C C Element capacitor 123 capacitance manu
434. ts onlinepubs stand standards html SPEF File Structure The IEEE 1481 specification requires the following file structure in a SPEF file Parameters in brackets are optional HSPICE RF User Guide 333 Z 2007 03 Chapter 13 Post Layout Analysis Post Layout Back Annotation 334 SPEF file SPEF version DESIGN design _name DATE date VENDOR vendor PROGRAM program_name VERSION program version DESIGN FLOW flow type flow type DIVIDER divider DELIMITER delimiter BUS DELIMITER bus prefix bus_suffix T UNIT time_unit NS PS C_UNIT capacitance unit FF PF R_ UNIT resistance unit OHM KOHM L_ UNIT inductance_unit HENRY MH UH NAME MAP name_index name_id bit path name physical_ref POWER NETS logical power net physical power net GROUND NETS ground net PORTS logical port I B O C coordinate L par_value S rising slew falling slew low_threshold high threshold D cell type PHYSICAL PORTS physical instance delimiter physical port I B O C coordinate L par_value S rising slew falling slew low_threshold high threshold D cell type DEFINE logical instance design name PDEFINE physical instance design name D NET net path total capacitance V routing confidence CONN P logical instance delimiter logical port physical port I B O C coordinate L par_value S rising slew falling slew low _ threshold high threshold D cell type I physical inst
435. ts to the largest value in the NHARMS list If entries in the NHARMS list are gt INTMODMAX HSPICE RF adds the m f frequencies to the spectrum where f is the corresponding tone and mis a value lt the NHARMS entry HSPICE RF User Guide Z 2007 03 201 Chapter 7 Steady State Harmonic Balance Analysis Harmonic Balance Analysis Example 1 hb tones f1 2 intmodmax 1 The resulting HB analysis spectrum dc f4 fo Example 2 hb tones f1 2 intmodmax 2 The resulting HB analysis spectrum dc f4 fo f4 fo fy fo 2 f4 2 fo Example 3 hb tones f1 2 intmodmax 3 The resulting HB analysis spectrum de f4 fo fy fo fy fo 2 f4 2 fo 2 f4 f2 2 fy fo 2 fo f4 2 fo fy 3 f4 3 fo Example 4 hb tones f1 2 nharms 2 2 The resulting HB analysis spectrum dc f4 fo f4 fo f4 fo 2 f4 2 fo Example 5 hb tones f1 2 nharms 2 2 intmodmax 3 The resulting HB analysis spectrum dc f4 fo fy fo fy fo 2 f4 2 fo 2 f4 f2 2 fo 2 f f4 2 fo f1 Example 6 hb tones f1 f2 nharms 5 5 intmodmax 3 The resulting HB analysis spectrum dc f4 fo fy fo fy fo 2 f4 2 f2 2 f4 fp 2 f fo 2 fo fy 2 fo f4 3 f 3 fo 4 f4 4 fo 5 f4 5 fo 202 HSPICE RF User Guide Z 2007 03 Chapter 7 Steady State Harmonic Balance Analysis Harmonic Balance Analysis HB Analysis Options The following table lists the OPTION command options specific to HB analysis Table 16 HB Analysis Options
436. ture VCO IEEE Trans Solid State Circuits May 2002 pp 653 656 gt J van der Tang P van de Ven D Kasperkovitz and A Roermund Analysis and design of an optimally coupled 5 GHz quadrature LC oscillator IEEE Trans Solid State Circuits May 2002 KX pp 657 661 gt F Behbahani H Firouzkouhi R Chokkalingam S Delshadpour A Kheirkhani M Nariman M Conta and S Bhatia A fully integrated low IF CMOS GPS radio with on chip analog image rejection IEEE Trans Solid State Circuits Dec 2002 ae pp 1721 1727 k Setup for Harmonic Balance Analysis xk Oscillation Frequency 1575 MHz GPS L1 frequency Amplitude 5 Volts peak to peak zero to 5V xx Vdd 2 5 V kk HSPICE Simulation Options option delmax 1n ACCURATE LIST NODE kk HSPICE RF Simulation Options option sim_accuracy 10 kk option savehb a hbs loadhb a hbs option POST param Vtune 2 0 Failures vtune 1 param Cval 0 2p Dae ete Ns a ec a iy Ne Vtune vc gnd DC Vtune Vdd vdd gnd 2 5 First oscillator section Low Q resonator with Vdd at center tap of inductors Rla IP ri 100k S These R s set the Q Rib ri IN 100k L1 IP vdd 16 5nH HSPICE RF User Guide Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 5 Using HBOSC Analysis fora CMOS GPS VCO L2 vdd IN 16 5nH Ccl IP gnd Cval I to Q Cc2 IN gnd Cval I to Q Differential fets M1 IP IN cs gnd NMOS 1 0 35u w
437. ubcircuits within your HSPICE netlist or HSPICE RF Use the ENDS statement to terminate a SUBCKT statement Use the Eom statement to terminate a MACRO statement Use X lt subcircuit name gt the subcircuit call statement to call a subcircuit that you previously defined ina MACRO or SUBCKT command in your netlist where lt subcircuit_name gt is the element name of the subcircuit that you are calling This subcircuit element name can be up to 15 characters long Usethe INCLUDE statement to include another netlist as a subcircuit in the current netlist Node Naming Conventions Nodes are the points of connection between elements in the input netlist You can use either names or numbers to designate nodes Node numbers can be from 1 to 999999999999999 node number 0 is always ground HSPICE or HSPICE RF ignores letters that follow numbers in node names When the node name begins with a letter or a valid special character the node name can contain a maximum of 1024 characters In addition to letters and digits node names can include but NOT begin with the following special characters Ko fa H Ly lt gt 1 ioc Node names that begin with one or more numerical digits cannot contain brackets for example 123 r55 Whereas node names that begin with alphabetic character may contain brackets for example n123 r55 HSPICE RF User Guide 75 Z 2007 03 Chapter 4 Input Netlist and Data Entry Inp
438. uency swept input frequency values 285 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Periodic Time Dependent Noise Analysis PTDNOISE 286 File Description Noise source contributions are listed sequentially and are controlled by the PTDNOISE command line parameters listtime listfreq listcount listfloor and listsources MEASURE Syntax and File Format The syntax for MEASURE PTDNOISE is MEASURE PTDNOISE meas name STROBEJITTER onoise freq sweep Only STROBEJITTER can be specified However this will result in all of the parameters listed below being output to the msnptn file MEASURE PTDNOISE allows for the measurement of these parameters integnoise time point tdelta value slewrate and strobed jitter Parameter Units Description strobejitter sec Calculated from the noise voltage integrated over the frequency range specified by frequency_range divided by the slewrate at the same node s at the time point specified by time_value While only STROBEJITTER can be specified all of the parameters listed below are also output to the msnptn file integptdnoise V Total ptd voltage noise in v z integrated over a frequency range specified by frequency_range at the time point specified by time_value The value is stated as a voltage V timepoint sec Time point at which the ptdnoise and slewrate are calculated tdelta value sec TDELTA value used to calculate
439. uide Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Multitone Harmonic Balance Transfer Function Analysis HBXF Input Syntax HBXF out_var lt freq sweep gt Parameter Description out_var Specify i 2 _port_elem or V n1 lt n2 gt freq_sweep Frequency sweep range for the input signal also referred to as the input frequency band IFB or fin A sweep of type LIN DEC OCT POI or SWEEPBLOCK Specify the nsteps start and stop frequencies using the following syntax for each type of sweep LIN nsteps start stop DEC nsteps start stop OCT nsteps start stop POI nsteps freq_values SWEEPBLOCK BlockName Specify the frequency sweep range for the output signal HSPICE RF determines the offset frequency in the input sidebands for example f1 abs fout k f0 s t f1 lt f0 2 The f0 is the steady state fundamental tone and f1 is the input frequency Output Syntax This section describes the syntax for the HBXF PRINT and PROBE statements PRINT and PROBE Statements PRINT HBXF TYPE NODES ELEM HSPICE RF User Guide 289 Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses Multitone Harmonic Balance Transfer Function Analysis HBXF 290 PROBE HBXF TYPE NODES ELEM Parameter Description TYPE TYPE can be one of the following TFV existing source TFI placeholder value for the current source attache
440. ulate the DC leakage current of a design hierarchy For example POWERDC lt keyword gt lt subckt_namel gt HSPICE RF User Guide 415 Z 2007 03 Chapter 16 Advanced Features POWER DC Analysis This statement creates a table that lists the measurements of the AVG MAX and MIN values for the current of every instance in the subcircuit This table also lists the sum of the power of each port in the subcircuit You use the SIM _POWERDC_HSPICE option to increase the accuracy of operating point OP calculations Or for even higher accuracy in operating point calculations you use the SIM_POWERDC_ ACCURACY option For syntax and description of this statement and options see POWERDC OPTION SIM_POWERDC_ACCURACY or OPTION SIM_POWERDC_HSPICE in the HSPICE and HSPICE RF Command Reference Power DC Analysis Output Format Leakage Current Result Subckt Name XXX Instance Name Port Max A Min A Avg A Total Power Max W Min W Avg W NOTE Power Sum Ii Vi Subckt Name XXX Instance Name Port Max A Min A Avg A Total Power Max W Min W Avg W Example global vdd vss powerdc all xl inl midl inv x2 midl outl inv subckt inv in out mn out in vss vss nch mp vdd in out vdd pch ends end Output 416 HSPICE RF User Guide Z 2007 03 Chapter 16 Advanced Features POWER Analysis Leakage Current Result Subckt Name Top Level Instance Name Port Max A Min A Avg A x1 Ro
441. ulation results Default 0 Keyword that specifies the inductance calculated by a polynomial HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Passive Elements Parameter Description c0 cl Coefficients of a polynomial in the current describing the inductor value cO is the magnitude of the Oth order term c1 is the magnitude of the ist order term and so on NT turns Number of turns of an inductive magnetic winding mname Saturable core model name See the Passive Device Models chapter in the HSPICE Elements and Device Models Manual for model information In this syntax the inductance can be either a value in units of henries an equation a polynomial of the current or a magnetic winding Required fields are the two nodes and the inductance or model name f you specify parameters the nodes and model name must be first Other parameters can be in any order f you specify an inductor model see the Passive Device Models chapter in the HSPICE Elements and Device Models Manual the inductance value is optional Example 1 In the following example the L1 inductor connects from the coilin node to the coilout node with an inductance of 100 nanohenries L1 coilin coilout 100n Example 2 The Lloop inductor connects from node 12 to node 17 Its inductance is 1 microhenry and its temperature coefficients are 0 001 and 0 Lloop 12 17 L lu TC1 0 001 TC2 0 Example 3 The Lcoil inducto
442. upport the following parameter sweeps with the same syntax as standard HSPICE LIN DEC OCT DATA POl You can also use the MONTE keyword for a Monte Carlo analysis or the OPTIMIZE keyword for optimization Generating Output Files HSPICE RF generates a table of simulation outputs f the output is text the default the text is put into a lis file If you specify OPTION POST then HSPICE RF generates simulation output in a format suitable for a waveform display tool The default output format for transient analysis in HSPICE RF is the same as in HSPICE the trO file format For additional information see Standard Output Files in the HSPICE Simulation and Analysis User Guide The Synopsys interactive waveform display tool CosmosScope can display both the text simulation results and binary output within the X window environment All output functions PRINT PROBE MEASURE and so on can use power output variables in the form p devicename just as in HSPICE You can also use the power keyword Larger output files from multi million transistor simulations might not be readable by some waveform viewers Options are available that enable you to limit the output file size See Limiting Output Data Size on page 403 for more information HSPICE RF User Guide Z 2007 03 HSPICE RF Output File Types Table 3 shows the output file extensions that HSPICE RF analyses produce The base file name of ea
443. urces and must match the period specified in the SN command Driven Phase Frequency Example This example demonstrates the Shooting Newton based analysis of a driven phase frequency detector Extracted portions of the input file are presented below The complete phasefreqdet sp input file for this example is located in the following directory installdir demo hspicerf examples During the analysis the phase of the input signal is swept between 0 degrees and 90 degrees using five equally spaced steps This enables us to measure the phase detector gain at the output load We use the SN command syntax that specifies the expected period of the steady state waveforms as a frequency 0 5GHz and the time resolution as the number of harmonics 127 HSPICE RF User Guide Z 2007 03 Chapter 3 HSPICE RF Tutorial Example 7 Using Shooting Newton Analysis on a Driven Phase Frequency Circuit and a Ring Oscillator A periodic time varying AC noise analysis based on the Shooting Newton algorithm is performed using the SNNOISE command The SNNOISE analysis requires an output node v Ifin where the noise is to be measured an input noise source Vref which serves as the reference for the noise computation and a frequency sweep for the noise analysis Optionally an index term can be defined The index term specifies the output frequency band at which the noise is evaluated For this case we want to evaluate the low frequency noise of the phase
444. ut Netlist File Composition 76 If you use braces in node names HSPICE or HSPICE RF changes them to brackets You cannot use the following characters in node names lt blank gt You should avoid using the dollar sign after a numerical digit in a node name because HSPICE assumes whatever follows the symbol is an in line comment see Comments and Line Continuation on page 71 for additional information It can cause error and warning messages depending on where the node containing the is located For example HSPICE generates an error indicating that a resistor node is missing Ri 1S 2 1k Also in this example HSPICE issues a warning indicating that the value of resistor R1 is limited to 1e 5 and interprets the line as R1 2 1 withouta defined value R1 2 1 1k The period is reserved for use as a separator between a subcircuit name and a node name subcircuitName nodeName f a node name contains a period the node will be considered a top level node unless there is a valid match to a subcircuit name and node name in the hierarchy The sorting order for operating point nodes is a Z L7 By x y Using Wildcards on Node Names You can use wildcards to match node names wildcard matches any single character For example 9 matches 92 9a 9A and 9 x wildcard matches any string of zero or more characters For example e If your netlist includes a resistor n
445. v x2 2 3 inv x3 3 out inv x4 out 4 inv macro inv in out mn out in 0 0 nch w 10u l 1lu mp out in 1 1 pch w 10u l 1lu com param multl 1 polycd agauss 0 0 06u 1 xl polycd sigma 0 06u nactcd agauss 0 0 3u 1 xwn nactcd sigma 0 3u pactcd agauss 0 0 3u 1 xwp pactcd sigma 0 3u toxcd agauss 200 10 1 tox toxcd sigma 10 vtoncd agauss 0 0 05v 1 delvton vtoncd sigma 0 05 vtopcd agauss 0 0 05v 1 delvtop vtoncd sigma 0 05 rshncd agauss 50 8 1 rshn rshnced sigma 8 rshpcd agauss 150 20 1 rshp rshpcd sigma 20 level 28 example model model nch nmos level 28 lmlt mult1 wmlt mult1 wref 22u lref 4 4u xl x1 xw xwn tox tox delvto delvton rsh rshn model pch pmos level 28 lmlt mult1 wmlt mult1 wref 22u lref 4 4u xl xl xw xwp tox tox delvto delvtop rsh rshp 1d 0 08u wd 0 2u acm 2 ldif 0 hdif 2 5u rs 0 rd 0 rdc 0 rsc 0 rsh rshp js 3e 04 jsw 9e 10 4 4 transient with sweep tran 20p 1 0n sweep sigma 3 3 5 meas s delay trig v 2 val vref fall 1 targ v out val vref fall 1 meas S power rms power transient with Monte Carlo tran 20p 1 0n sweep monte 100 meas m delay trig v 2 val vref fall 1 targ v out val vref fall 1 meas m power rms power probe tran v in v 1 v 2 v 3 v 4 end 372 HSPICE RF User Guide Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Worst Case and Monte Carlo Sweep Example Transient Sigma Sweep Results The plot in Figure 39
446. verage Table 28 Sigma Deviations Type Parameter Slow Fast NMOS XL RSH DELVTO TOX XW PMOS XL RSH DELVTO TOX XW HSPICE RF selects skew parameters based on the available historical data that it collects either during fabrication or electrical test For example HSPICE RF collects the XL skew parameter for poly CD during fabrication This parameter is usually the most important skew parameter for a MOS process Figure 29 is an example of data that historical records produce HSPICE RF User Guide 355 Z 2007 03 Chapter 14 Statistical and Monte Carlo Analysis Worst Case Analysis 356 Figure 29 Historical Records for Skew Parameters in a MOS Process Fab Database Run PolyCD 101 0 04u 102 0 06u 103 0 03u pop lt 3 sigma 2 sigma 1 sigma i Mean XL value Using Skew Parameters Figure 30 shows how to create a worst case corners library file for a CMOS process model Specify the physically measured parameter variations so that their proper minimum and maximum values are consistent with measured current IDS variations For example HSPICE can generate a 3 sigma variation in IDS from a 2 sigma variation in physically measured parameters Figure 30 Worst Case Corners Library File for a CMOS Process Model pop SS Slow Corner Skew Parameters Extracted Skew Parameters TT Typical Corner S
447. w in the time domain as t Equation 23 v t i v t Taw y t T v T d t i t 0 i v t represents the resistive currents from nonlinear devices qrepresents the charges from nonlinear devices y represents the admittance of the linear devices in the circuit m i represents the vector of independent current sources vis a variable that represents the circuit unknowns both node voltages and branch currents and f v t is an error term that goes to zero when Kirchoff s current law is satisfied Transforming this equation to the frequency domain results in Equation 24 F V I V QQ V Y V I 0 Note Time differentiation is transformed to multiplication by jaterms which make up the Q matrix in the frequency domain The convolution integral is transformed to a simple multiplication The Y matrix is the circuit s modified nodal admittance matrix All terms above are vectors representing the circuit response at each analysis frequency The following equation shows the vector of complex valued unknowns in the frequency domain for a circuit with K analysis frequencies and N unknowns Equation 25 V Ya 0 Va 1 lt Va K 1 Vo 0 Vy E HSPICE RF finds the unknown vector V which satisfies the system of nonlinear equations shown in the equation above This is done via the Newton Raphson technique by using either a direct solver to factor the Jacobian matrix or an indirect solver The indirect solve
448. wave Computer Aided Engineering Volume 1 No 1 pages 22 37 1991 R Gilmore and M B Steer Nonlinear Circuit Analysis Using the Method of Harmonic Balance A Review of the Art Part Il Advanced Concepts International Journal of Microwave and Millimeter wave Computer Aided Engineering Volume 1 No 2 pages 159 180 1991 V Rizzoli F Mastri F Sgallari G Spaletta Harmonic Balance Simulation of Strongly Nonlinear Very Large Size Microwave Circuits by Inexact Newton Methods MTT S Digest pages 1357 1360 1996 S Skaggs Efficient Harmonic Balance Modeling of Large Microwave Circuits Ph D thesis North Carolina State University 1999 R S Carson High Frequency Amplifiers 2nd Edition John Wiley amp Sons 1982 S Y Liao Microwave Circuit Analysis and Amplifier Design Prentice Hall 1987 Y Saad Iterative Methods for Sparse Linear Systems PWS Publishing Company 1995 J Roychowdhury D Long and P Feldmann Cyclostationary Noise Analysis of Large RF Circuits with Multitone Excitations IEEE Journal of Solid State Circuits volume 33 pages 324 336 March 1998 10 A Demir A Sangiovanni Vincentelli Analysis and Simulation of Noise in Nonlinear Electronic Circuits and Systems Kluwer Academic 1998 HSPICE RF User Guide 295 Z 2007 03 Chapter 10 Large Signal Periodic AC Transfer Function and Noise Analyses References 296 HSPICE RF User Guide Z 2007 03 11 S p
449. width and length PJ Periphery of junction unitless for LEVEL 1 diode and meters for LEVEL 3 diode Overrides PJ from the diode model If you do not specify PJ HSPICE or HSPICE RF calculates it from the width and length specifications WP Width of polysilicon capacitor in meters for LEVEL 3 diode only Overrides WP in the diode model Default 0 0 LP Length of polysilicon capacitor in meters for LEVEL 3 diode only Overrides LP in the diode model Default 0 0 WM Width of metal capacitor in meters for LEVEL 3 diode only Overrides WM in the diode model Default 0 0 LM Length of metal capacitor in meters for LEVEL 3 diode only Overrides LM in the diode model Default 0 0 HSPICE RF User Guide Z 2007 03 Chapter 6 Testbench Elements Active Elements Parameter Description OFF Sets the initial condition for this element to OFF in DC analysis Default ON IC vd Initial voltage across the diode element Use this value when you specify the UIC option in the TRAN statement The IC statement overrides this value M Multiplier to simulate multiple diodes in parallel The M setting affects all currents capacitances and resistances Default 1 DTEMP The difference between the element temperature and the circuit temperature in degrees Celsius Default 0 0 W Width of the diode in meters LEVEL 3 diode model only L Length of the diode in meters LEVEL 3 diode model only You must specify two nodes
450. within a subcircuit HSPICE RF multiplies the product of both levels Do not assign a negative value or zero as the M value S Scale Parameter To scale a subcircuit use the S local scale parameter This parameter behaves in much the same way as the M parameter in the preceding section OPTION hier_scale value OPTION scale value X1 nodel node2 subname S valueM parameter The OPTION HIER_SCALE statement defines how HSPICE RF interprets the S parameter where value is either m 0 the default indicating a user defined parameter or 1 indicating a scale parameter The OPTION SCALE statement defines the original default scale of the subcircuit The specified S scale is relative to this default scale of the subcircuit The scale in the subname subcircuit is value scale Subcircuits can originate from multiple sources so scaling is multiplicative cumulative throughout your design hierarchy RF User Guide 89 Chapter 4 Input Netlist and Data Entry Using Subcircuits 90 xl ay inv S lu subckt inv in out x2 a b kk S 1m ends In this example HSPICE RF scales the X1 subcircuit by the first s scaling value 1u scale Because scaling is cumulative X2 a subcircuit of X1 is then scaled in effect by the S scaling values of both X1 and X2 1m 1u scale Using Hierarchical Parameters to Simplify Simulation You can use the hierarchical parameter to simplify simulations An example is shown in the following
451. y in a region close to the carrier Generally you will want to use this algorithm if you interested in phasenoise close to the carrier and do not need to determine a noise floor NLP computation time is almost independent of the number of frequency points in the phasenoise frequency sweep Periodic AC Algorithm The periodic AC PAC algorithm is valid in a region away from the carrier and is slower than the NLP algorithm The PAC algorithm is used for getting phasenoise in the far carrier region and when you need to determine a noise floor The computation time for the PAC algorithm is approximately linearly dependent on the number of frequency points in the phasenoise frequency sweep If you are using the PAC algorithm you should try to minimize the number of points in the sweep Another issue is that the PAC algorithm becomes more ill conditioned as you approach the carrier This means that you may have to generate a steady state solution with more harmonics to get an accurate simulation as you get closer to the carrier So if you find that the PAC is rolling off at close in frequencies you should rerun HB analysis with a larger number of harmonics Although typically you will not see improvements in PAC accuracy beyond more than about 100 200 harmonics HSPICE RF User Guide Z 2007 03 Chapter 9 Oscillator and Phase Noise Analysis Phase Noise Analysis Early in your testing the best way to verify that NLP and PAC are giving ac
452. y oscillator has amplitude and phase fluctuations we can write as Equation 36 v t A t cos pt o f In the preceding equation A t is the time varying amplitude for the noisy oscillator ft is the time varying phase for the noisy oscillator is the frequency of oscillation In most applications the phase noise is of particular interest because it represents frequency fluctuations about the fundamental which you cannot HSPICE RF User Guide Z 2007 03 Chapter 9 Oscillator and Phase Noise Analysis Phase Noise Analysis remove These fluctuations are random processes and are typically expressed in terms of their power spectral density For most oscillators the phase noise is a low frequency modulation that creates sidebands in the oscillators spectrum about For example the following equation represents a simple sinusoidal variation in the phase Equation 37 v t Acos He 0 psina is the peak phase deviation specified as 0 Ao w Aq is the peak angular frequency deviation For Op 1 the following equation approximates the output 0 Equation 388 v t Alcoscan 08 o Wp t COS apn That is when the peak phase deviation is small the result is frequency components on each side of the fundamental with amplitude 2 B Op The Single Sideband Phase Noise L fm is the ratio of noise power to carrier power in a 1Hz bandwidth at offset m 2xf which
453. ysis for an Amplifier References to sources must use SI units in conjunction with the previous equation to convert from dBm to Watts The colon is used as a labeling convenience Second a voltage source element is used as a two tone power source by setting the power flag and a source impedance of 50 ohms is specified The HB keyword is used to identify the amplitude interpreted in Watts with the power flag set phase harmonic index and tone index for each tone Vin rfind gnd dc 0 power 1 z0 50 50 Ohm src HB Pin W 011 tone 1 HB Pin W 0 1 2 tone 2 Third the HB command designates the frequencies of the two tones and establishes the power sweep using the dBm power variable The intmodmax parameter has been set to 7 to include intermodulation harmonic content up to 7th order effects HB tones 900MEG 910MEG nharms 11 intmodmax 7 SWEEP Pin dBm 50 0 0 0 2 0 Last the HSPICE RF ability to specify specific harmonic terms is used in the PRINT and PROBE statements to pull out the signals of particular interest Notice the three different formats PRINT HB P Rload This reference dumps a complete spectrum in RMS Watts for the power across resistor Rload PRINT HB P Rload 1 0 This reference selectively dumps the power in resistor Rload at the first harmonic of the 1st tone PRINT HB P Rload 2 1 This reference selectively dumps the power in resistor Rload at the 3rd intermodulation product frequency 890 M

HSPICE RF User Guide - RFIC Group @ Fudan University

Contents

Download Pdf Manuals

Related Search

Related Contents