Home

Agilent N4901 Serial BERT

1. Time Frequency 1 T Leakage and Windowing FFT assumes that the time record contains a representative section of an endless periodic signal It assumes that time records can be seamlessly concatenated If this is not the case a phenomenon called leakage occurs Agilent Serial BERT N4901 2 User s Guide January 2006 245 Leakage makes it impossible to detect minor adjacent spectral components The following two figures refer to a slightly disturbed sine wave The Appearance of Leakage Source Waveform Expected L pf fo 2fo 3fo 4fp errr A Time record Computed a spectrum af fo When you perform a Spectral Jitter measurement it is likely that some degree of leakage occurs The measurement therefore provides a choice of FFT windows that allow you to detect leakage and to reduce its impact An FFT window is a filter that sets the beginning and end of the time record smoothly to zero Such records can be seamlessly concatenated Windowing Input waveform Original sa VW Incorrectly sampled record MW Window function i Hanning window Assumed periodicity Time record Agilent Serial BERT N4901 2 User s Guide January 2006 Si eae eae q o A Correct spectrum f Spectrum A ee with leakage f Less leakage in spectrum f When you are using a window please note e No window removes leakage completely e Every window re
2. Pattern Generator Bios Self Test This test is checking all the modules chips and registers of the pattern generator more thoroughly than the Power Up test It may take about 5 minutes to complete Pattern Generator Power Up Messages This test is automatically run after switching on the instrument It mainly checks if all important modules chips and registers of the pattern generator are available and accessible Pattern Generator Auto Calibration This test runs a timing calibration that calibrates the clock to data alignment of the pattern generator It may take about 15 minutes to complete e Error Detector Bios Self Test This test is checking all the modules chips and registers of the error detector more thoroughly than the Power Up test It may take about 5 minutes to complete e Error Detector Power Up Messages This test is automatically run after switching on the instrument It mainly checks if all important modules chips and registers of the error detector are available and accessible e Error Detector Auto Calibration This test runs a timing calibration that calibrates the clock to data alignment of the error detector It may take about 15 minutes to complete Select All Unselect All Use the Select All button to select all tests in the Self Test Options dialog box After clicking the button it becomes the Unselect All button Clicking the button again will then unselect all tests and display the S
3. Understanding Error Detector Input Ports The error detector has the following input ports Clock In This port needs to be connected to a clock signal unless you use the error detector in Clock Data Recovery CDR mode where it derives the clock frequency from the data port See Clock Setup on page 107 for more details Data In and Data In This port is connected to the data signal and the inverted data signal Agilent Serial BERT N4901 2 User s Guide January 2006 101 e Gate In This port has two different functions In normal mode it can be used to enable and disable the error counter Errors are only counted when this signal is low The error counter can only count errors in 128 bit data blocks Thus the signal at the Gate In port must be low for at least 128 bits to enable the counter The following figure illustrates the effect of error gating Errors are only counted in the gray blocks Gate In Expected Data on Data Input A A only 128bit minimum signal granularity length not matched The burst sync mode is a special operating mode for measuring data in bursts of bits rather than one continuous stream of bits In this case the signal at the Gate In port controls the timing of synchronization and error counting for each burst Understanding Error Detector Output Ports The error detector provides several output ports that are used to connect it to other devices for further er
4. 600 0 mY sscale gt 0 448 UI 0 149 UI 0 149 UI 0 448 UI 0 746 UI Relative 0 597 Ul 0 299 UI 0 000 UI 0 299 UI 0 597 UI Eye Opening Reference The Eye Opening measurement returns the results in a graphical and in a numerical form The following sections provide explanations of the measured parameters and the display options that are specific to this measurement Additionally some information is provided to explain the theoretical background behind Properties that can be specified on the various tabs of the Properties dialog box Parameters Tab NOTE Ifyou modify the parameters on this page you have to rerun the measurement to update the results Set the criteria for moving to the next sample point Number of Compared Bits After this number of compared bits the measurement stops for the current sample point and moves to the next one The default is 1 million bits That means you can measure a bit error rate down to 10 6 one error per million A smaller number reduces the duration of the whole Eye Opening measurement A larger number increases the precision of the measured bit error rates Number of Errors After this number of errors the measurement stops for the current sample point and moves to the next one This allows you to speed up the measurement You can switch off this option if only the number of compared bits is important Agilent Serial BERT N4901 2 User s Guide January 2006 217 NOTE T
5. Procedures This section shows how to set up and perform an Eye Opening measurement As an example we measure the eye diagram of a shielded cable This requires the following steps Preparing the measurement see How to Prepare the Eye Opening Measurement on page 212 Executing the measurement see How to Execute the Eye Opening Measurement on page 213 Optimizing the view of the results see How to Optimize the View of the Results on page 214 Using the color bar see How to Use the Color Bar on page 215 Adding or changing colors see How to Add or Change Colors on page 215 Changing the BER threshold see How to Change the BER Threshold on page 216 Changing the BER range of a color see How to Change the BER Range of a Color on page 216 How to Prepare the Eye Opening Measurement To prepare an Eye Opening measurement to test a shielded cable 1 Disable the pattern generator outputs by pressing the OV Disable button in the PG Output Setup screen 2 Usea shielded cable to connect the pattern generator s Data Out port and the error detector s Data In port 3 Terminate all non connected pattern generator output ports with 50 Q 4 Switch to the Pattern panel and press Pattern Select Select an appropriate pattern for example a pure 2 15 1 PRBS for both the pattern generator and the error detector 5 For the pattern generator setup you need to specify the logic l
6. This helps to identify jitter sources and to reduce or eliminate their influence Prerequisites for Using the Spectral Jitter Measurement It is important that the initial sampling point is in optimum position which means in the middle of the eye opening because the Spectral Jitter measurement derives its starting sampling point from that point See also General Requirements on page 148 How the Spectral Jitter Measurement Works The Spectral Jitter measurement performs a capture and compare operation in the jitter region The number of compared and captured bits is adjustable 238 Agilent Serial BERT N4901 2 User s Guide January 2006 By default the analyzer sampling point is set to an offset of 0 5 analyzer clock periods or unit intervals UI from the optimum sampling delay Jitter width at threshold Threshold level aS Optimum sampling delay for the signal Sampling delay for the Spectral Jitter measurement adjustable Capture and compare means that a certain amount of incoming data is captured and also compared with expected data in real time The number of captured data points is adjustable While the measurement is running correct and incorrect data is captured The resulting records contain the captured data and the corresponding error information These records are automatically processed The error information is subject to a fast Fourier transform FFT FFT reve
7. disabled gt lt disabled disabled gt lt disabled 0 135 Ul 0 0141 Ul 1E 009 0 865 UI 0 251 Ul 182 Agilent Serial BERT N4901 2 User s Guide January 2006 The result graph shows you the points in time that have been investigated and whether the actual BER at these points was higher or lower than the BER threshold specified for the measurement The grey area in the middle indicates the uncertainty band see Explanation of the Fast Total Jitter Measurement on page 166 NOTE The test will fail if the BER floor is not well below the BER threshold specified for the measurement Explanation of the Numerical Results The Fast Total Jitter measurement result parameters are summarized in the following list Phase Margin The period of time where the bit error rate is lower than the Total Jitter BER Threshold Total Jitter Peak Peak Peak to peak value of the total jitter Calculated as the pulse period unit interval minus the Phase Margin at the Total Jitter BER Threshold Total Jitter Uncertainty The maximum of the uncertainties of both slopes Measured as the time between a point with a BER greater than the specified Total Jitter BER Threshold and the next point with a BER less than the specified Total Jitter BER Threshold left slope and vice versa right slope Total Jitter BER Threshold The BER threshold specified for the measurement Optimal Sample Point Delay The mean value of the left and rig
8. 2 8 Agilent Serial BERT N4901 2 User s Guide January 2006 Possible Cause Additional Information Agilent Serial BERT N4901 2 User s Guide January 2006 Measurement Problems Accumulated measurements can help you determine the cause of bit errors Bit errors can be caused by a variety of problems In addition instrument settings can affect how errors are measured and displayed The sync mode setting has the greatest effect Some phenomenon you might observe with the possible causes are listed below Constant Errors More Errored 0 s than 1 s Accumulated Results Accumulated Resuts interval Results Accurnulation Parameters Messurement Current Period Previous Period Bit Court 600 602 034 112 BB Error Ratio 16 ME3 Eroe Count 931 152 150 WD Eres 1 s Ratio ome Emceed t s Court 0 Hl Erceed 0 s Ratio 16ME3 Ereceed O s Court 61 152 150 A hardware failure may have occurred in your device The errors became constant at the point of failure Digital circuitry often has parallel architecture in which data lines are multiplexed in stages into one serial data line This example assumed that a PRBS pattern was going through digital circuitry that had a total multiplexing ratio of 256 1 The constant errors were produced from one data line that became stuck high in the first stage of multiplexing The data line that was stuck high produced errored 0 s 50 of the time with the PRBS pattern Therefore at the serial
9. Data Inverted Data Delay 0 1 Threshold TIP You can use the instrument s front panel knobs to adjust the sampling point It is recommended that you realign the sampling point each time you change the data output level or make changes to your device cables or anything that affects electrical path length Sampling Point Setup Procedures In most cases you will set up the sampling point automatically Even if you wish to make some manual adjustments it is recommended to start with automatically aligning the sampling point NOTE Before adjusting the sampling point make sure the pattern synchronization mode is set up properly See Pattern Synchronization Procedures on page 138 for details Agilent Serial BERT N4901 2 User s Guide January 2006 117 How to Set the Optimum Sampling Point Automatically 1 Click the Sampling Point Setup icon in the ED Setup menu 2 Select an appropriate BER threshold from the BER Threshold list 3 Press Auto Align to automatically set the optimum sampling point You can stop the search for the optimum sampling point at any time by pressing Cancel How to Adjust the Data Input Delay 1 Press the Sampling Point Setup icon in the ED Setup menu 2 Select the Avg 0 1 Threshold checkbox to use a good start point on the vertical voltage axis 3 Click Data Center to move the sampling point horizontally towards the center of the eye at the selected voltage You may click Cancel at any
10. For a sequence you can specify the following Choices are Immediate Sequence starts immediately after downloading it to the pattern generator Aux In High Aux In Low Sequence starts when Auxiliary Input is high or low Aux In Rising Edge Aux In Falling Edge Sequence starts with a rising or falling edge at the Auxiliary Input Command The Sequence Editor shows a Start button Sequence execution starts when this button is clicked A sequence can contain 1 to 4 blocks A field for entering and editing descriptive text A window that shows the present SequenceExpression for details see How a Sequence is Defined on page 51 You can edit the SequenceExpression in this field Any changes override the previous settings If you edit the SequenceExpression manually take care to adhere to the syntax and order of the keywords Remember that you must download the updated sequence to the pattern generator if your changes shall take effect Start Button The Start button of the Sequence Editor indicates that the sequence has been set up for manual start Click this button to start the sequence execution Agilent Serial BERT N4901 2 User s Guide January 2006 61 62 Agilent Serial BERT N4901 2 User s Guide January 2006 Setting up the Pattern Generator The Serial BERT s pattern generator generates an output signal based on a data pattern It has the following possibilities for generating an output signal
11. The following figure shows an example where the scale and the power values have been normalized to the power measured at 1 MHz 32 dB 1 Trace 65535 Points 24 dB 1608 Reference power 8 dB 0 dB 8 dB 16 dB 24 dB 32 dB 40 dB Noise Threshold 40 dB 48 dB 56 dB Logarithmic 100 00 KHz 10 00 MHz 1 00 GHz 1 00 MHz 100 00 MHz Terminal Optical Show Color Copied Bi gt All Agilent Serial BERT N4901 2 User s Guide January 2006 When you set the power values to relative this changes not only the graphical scale but also the calculated results You can thus calibrate the power values to any reference Noise Threshold The Noise Threshold can be moved with the mouse or your finger It can also be set on the View page of the Properties dialog This threshold is used to separate between total power and noise power All components below the Noise Threshold are considered noise When the threshold is changed the calculated Noise Power values change If desired change the Number of Top Frequencies to Show This determines the number frequency power pairs in the numerical section Up to 16 pairs can be calculated and displayed Graph Tab On the Graph tab you can use the several options to optimize the graphical display according to your needs Frequency Scale Choose between logarithmic and linear scale for displaying the frequencies Power Unit Display the power either i
12. gt Sample Threshold Resolution fi OmV Low Level 1 85V High Level f 350 mY IV Edge Resolution Optimization In this example we expect the signal voltages to be between 1 75 V and 0 95 V The Sample Threshold values proposed above cover this range well The Resolution is the distance between the measurement points when the threshold moves from the low to the high level A resolution of 10 mV results in 100 measured points per Volt Note that we have disabled the Edge Resolution Optimization 4 Press OK to close the Properties dialog box 5 Press the Start button to execute the measurement The measurement is run and the result window shows the bit error rates measured at 100 threshold levels from 1 85 V up to 850 mV 800 0 my Trace 202 Points 1410Y 1 2 Y 1 4 Y 16 Y I Logarithmic 1 000e 5 1 000e 3 none 1 000e 6 1 000e 4 1 000e 2 1 000e 188 Agilent Serial BERT N4901 2 User s Guide January 2006 How to Improve the Output Levels Display You can change the display of an existing measurement for example if you wish to see more details to investigate the graph This can be done on the Graph tab of the Properties dialog box 1 Press the Properties button If you have a mouse connected to your Serial BERT you can also click the right mouse button on the graph and select Properties from the context menu Switch to the Graph tab As an example select Show Measured Po
13. 1 00e 2 1 00e 3 1 00e 4 1 00e 5 1 00e 6 1 00e 7 Aelute 1 92 Ul ee 246 Ul he 2 40 Ul ee 2 64 Ul Toa 2 88 UI s 342 Ul atlas 3 36 Ul ae 3 60 Ul zan The following illustration shows a diagram with high resolution and edge resolution optimization For the areas of the diagram with edges additional sampling points are taken resulting in a much better display of the signal shape 41 00e 0 3 Traces 3888 Points Ul 2 500 ns All Errors 1 00e 1 1 00e 2 1 00e 3 1 00e 4 1 00e 5 1 00e 6 1 00e 7 koaa tme T 2 04 UI Seri 2 28 Ul aA 2 52 Ul AAF 2 76 Ul TAN 3 00 UI KAN 3 24 UI PE 3 48 UI ai 3 72 Ul Fast Total Jitter at BER This enables the Fast Total Jitter measurement Before enabling this measurement you need to know the BER floor of the device and to specify a BER threshold that is above that floor For details see Explanation of the Fast Total Jitter Measurement on page 166 Pass Fail Tab The Pass Fail tab of the Properties dialog box allows you to specify the criteria to decide whether the DUT passes or fails the test You can change pass fail criteria without rerunning a test The software only uses the criteria to rate the results of a measurement NOTE The pass fail criteria do not control measurement execution The measurement run will be completed even if the measurement fails for one or more of the criteria Agilent Serial BERT N4901 2 User s Guide January 2006 You can set pass fail limits individually for the output tim
14. Example At 100 MHz the pulse period is 1 100 x 107 s 1078 s 10 ns this is the unit interval UI For example a value in fact any parameter that is specified in time units of 37 ns is equivalent to 37 ns 10 ns 3 7 UI time values are expressed as multiples of the unit interval Agilent Serial BERT N4901 2 User s Guide January 2006 155 Measurements at different This makes it easy to analyze measurements at different frequencies frequencies If the measured phase margin is 8 ns at 100 MHz 1 UI 10 ns and 4 ns at 200 MHz 1 UI 5 ns The phase margin is 0 8 UI in both cases and it is immediately obvious that it does not depend on the system frequency The same is true for the delay resolution parameter in the DUT Output Timing Jitter and Eye Opening measurements If you specify the delay resolution as 0 01 UI the measurement software will always measure 150 points across the bathtub no matter what system frequency you choose only if edge optimization is switched off of course However if you specify the resolution in time the number of measured points will change inversely proportional to the system frequency doubling the frequency results in half the number of measured points Syntax Requirements All time related entries understand both time and unit interval notation regardless of the UI time selection made in the View tab of the Properties dialog box You can enter 17 s 5 ns or0 01 UI at any time On run
15. Explanation of the Result Display Unavailable Values Under certain circumstances some numerical results are not available This is indicated by lt invalid gt or lt not applicable gt in the numerical results table below the measurement graphic e lt invalid gt indicates that the value could not be calculated This is the case for example for phase margin when the BER threshold is set to high values and does not intersect with the bathtub curve e lt not applicable gt indicates that the value could be calculated however is not shown because quality criteria are not met This is the case for example for RJ DJ results when the r 2 value of one or both edges is lt 0 75 Even though RJ DJ values can be calculated in this case they are not shown because confidence in the results is too low Explanation of the Fast Total Jitter Measurement Results The Fast Total Jitter measurement provides both graphical and numerical results The example below shows a copied result and the display of measured points was enabled The results of the recent measurement are disabled because they refer to a standard bathtub measurement that provides many more results By actuating the Show indicator you can inspect both alternatively 1 Trace 129 Points Ul 141 642 ps All Errors gt TJ BER close to TJ BER TJ BER 0 71 Ul 0 42 Ul 0 14 Ul 0 14Ul 0 42 Ul Absolute 0 85 Ul 0 56 Ul 2 28 UI 0 00 Ul 0 28 Ul E sdisabled gt
16. Output Levels 198 Output Timing 180 Spectral Jitter 256 O Offset threshold 235 On Screen Keyboard 294 Optimal Sample Point Delay 180 183 223 Optimal Sample Point Threshold 223 Optimum Sampling Point 118 Oscilloscope Connections 16 Output blanking 76 Output Level Parameters 66 Output Levels measurement 184 Output Protection 65 Output Timing measurement 162 Output Timing Measurement Optimization 173 P Parameters Tab Eye Opening 217 Fast Eye Mask 234 Output Levels 192 Output Timing 173 Spectral Jitter 250 Pass Fail tab Eye Opening 218 Fast Eye Mask 236 Output Levels 194 Output Timing 174 Spectral Jitter 252 Pattern Creating New 36 Editing 37 Editor 40 57 Examples 29 from file 19 Hardware generated 19 Length 26 Load status indicators 44 Loading directly from Pattern Editor 39 Loading from file 38 mark density 47 memory based 19 Opening Existing 36 Agilent Serial BERT N4901 2 User s Guide January 2006 PRBN 21 PRBS 19 21 Properties 43 provided patterns 20 Repetition 28 Resolution 26 Saving 37 Selecting 44 Sequence 38 Sequence of patterns 51 software generated 20 46 48 Supported Types 24 Which Is Appropriate 22 Pattern Alternation Manually 94 With External Signal 94 Pattern Generator Auxiliary Input 65 Clock In port 64 Clock Out port 65 Data Out port 65 Delay Control Input 65 Error Add input 64 Ref Clock input 64 Trigger Out port 65 Pattern Mode 53 Peak Peak Noise 200 Per
17. With this option all errors are counted With this option only errors occuring in the block between the specified Start Bit and Block Length are counted Agilent Serial BERT N4901 2 User s Guide January 2006 143 Single Bit With this option only errors occuring at the specified position are counted The following legend applies for the above graphics Evaluated bits Not evaluated bits 144 Agilent Serial BERT N4901 2 User s Guide January 2006 TIP Audio Signals The use of warning tones is convenient when you want to monitor the instantaneous BER over a longer time without constantly watching the display With the Audio dialog box you can adjust the sound volume as well as some conditions when to play a sound is played Audio Signals Procedures To set up the warning sounds do the following 1 Press the Audio icon in the ED Setup menu 2 Check the Audio on checkbox to switch on the sounds 3 Use the Main Volume slider to adjust the volume After adjusting the slider a test tone is played at the new volume 4 Click the respective radio button if you want to play the sound whenever the BER exceeds the threshold specified in the BER Alarm Threshold field Enter an appropriate value in this field on All Error Rates greater than zero 5 Click OK to finish the audio setup You can always adjust the BER Alarm Threshold by clicking and dragging the small yellow marker in the BER bar in the top
18. e Providing a wide range of clock frequencies You can use the pattern generator s internal clock or an external clock for defining the frequency of the outgoing stream e Distorting the signal by adding jitter You can connect an external delay control device for example a function generator to add jitter to the generated signal e Adding errors to the output stream The Serial BERT can be set up to insert errors into the outgoing stream either internally according to an external signal or manually from the operator e Run time switching between two patterns You can set up two patterns and switch between them during runtime either automatically according to an external signal or manually e Suppressing the output stream The output signal can be suppressed according to an external signal The pattern generator also provides output ports that let you connect an external instrument such as an oscilloscope Agilent Serial BERT N4901 2 User s Guide January 2006 63 64 Input Ports Output Ports Agilent Input and Output Ports The pattern generator s input ports are used to set the pattern generator s clock frequency and to manipulate the output signal with respect to jitter error insertion and signal output The pattern generator s output ports are used to supply a clock signal and trigger for another device for example for the Serial BERT s error detector and an arbitrary data signal for testing yo
19. with conventional methods one usually needs to compare more than 10 bits for each sample point To measure a full eye opening this way with appropriate timely resolution takes time maybe days or weeks depending on the data rate and the probability of seeing one or no error in 10 bits is not higher than 37 percent If one would compare 10 bits for the same device the probability of seeing ten errors is even lower 12 but the probability of observing no error is almost zero The Fast Total Jitter measurement implements a method that reduces the measurement time considerably and provides a higher accuracy It is based on statistical and probability calculations The method was presented at the DesignCon 2005 and is described in Total Jitter Measurement at Low Probability Levels using Optimized BERT Scan Method included as 5989 2933EN pdf We will not go into the details but provide an introduction 166 Agilent Serial BERT N4901 2 User s Guide January 2006 The Uncertainty Band Quite often we do not need to measure the exact BER but can stop the measurement if we are sure that the BER is above or below a threshold In a jitter tolerance test for example we need just to assure that the device under test operates with a BER better than let us say 10712 whether the true BER is 1 1 x 107 or 2 7 x 107 5 is irrelevant To abort the measurement for a single point and proceed to the next we need two limits that tel
20. you will also find instructions for the update process CAUTION Before you update the instrument s software make sure that the software matches the instrument s hardware An update can be carried out even if the software does not match the hardware In this case the instrument may go into a corrupt state or not all features may be available The applicable hardware versions are listed with the software update version You can find the instrument s hardware version on a label at the rear panel General Notes to Updating the When updating the firmware server on the instrument make sure you Firmware Server select the Hardware installation option After the firmware server has been updated you are prompted to start a BIOS update Allow the BIOS to be updated the firmware server requires the modifications Do not interrupt the BIOS update Note that you must reboot the instrument after updating the firmware If you do not reboot the instrument the firmware may not have access to the instrument s hardware Agilent Serial BERT N4901 2 User s Guide January 2006 289 Configuring the Instrument Procedures Follow the instructions below to configure the instrument according to your personal needs Setting Date and Time Your instrument uses the date and time when e Saving files e Saving self test information e Saving calibration information To set the date and time of your instrument do the following 1 On the Uti
21. 4 Distort the output signal by adding jitter as described in Delay Control Input Procedures on page S6 5 Manually move the sampling point See Sampling Point Setup Procedures on page 117 for details 6 Make adjustments to your device This can be used to see how changes to your device affect the BER You can determine what adjustments improve or degrade the BER Instantaneous Measurements Reference This section describes all elements of the BER Results window and the various status indicators BER Results Window This window displays the actual BER or cumulative AccumBER BER results e The BER is the current BER calculated upon a period of 200 ms The AccumBER shows either the accumulated BER of the current accumulation or if no accumulation is running the results of the most recent accumulation This enables you to monitor real time BER behavior as you do things such as to manually adjust the sampling point to add errors or to make adjustments to your device Accumulated Actual Button Click this button to toggle between BER or AccumBER results Error Count Accum Error Count This area displays the actual error count or cumulative error count 260 Agilent Serial BERT N4901 2 User s Guide January 2006 NOTE The BER bar below the main menu bar also displays the actual BER You can drag the small yellow alarm threshold mark to change the BER Alarm Threshold Status Indicators BER bar
22. Elapsed Time BIT COUNT BIT FREQUENCY ED FREQUEN 0 1 2 496 08 2496 09 02 2 49E 08 2496 09 249E 03 2 49E 08 2496 09 2 49E 0 4 2 496408 2496 09 2 49E 05 2 49E 08 2496409 2 49E alaale gt on You can view and analyze the following information in the measurement log file 302 ERRORED 1 COUNT ERRORED 1 RATIO Agilent Serial BERT N4901 2 BIT COUNT BIT COUNT Cumulative ERROR COUNT ERROR COUNT Cumulative ERROR RATIO ERROR RATIO Cumulative ERRORED 0 COUNT ERRORED 0 COUNT Cumulative ERRORED 0 RATIO ERRORED 0 RATIO Cumulative ERRORED 1 COUNT Cumulative ERRORED 1 RATIO Cumulative SYNC LOSS ERROR SECONDS Cumulative PG CLOCK LOSS ERROR FREE SECONDS Cumulative ED CLOCK LOSS ERROR DECISECONDS Cumulative DATA LOSS ERROR FREE DECISECONDS Cumulative SYNC LOSS SECONDS Cumulative User s Guide January 2006 NOTE Saving Recalling BURST STATUS Cumulative BURST SYNC RATIO Cumulative TOTAL BURST COUNT Cumulative BAD BURST COUNT Cumulative Saving and Recalling Patterns You can save your pattern in either of the following ways Click the Save icon on the Pattern Editor toolbar If the pattern has not been saved earlier a file dialog box opens that allows you to navigate the file system and name your pattern Alternatively you can use the following procedure 1 From the File menu select Save 2 Then select Save Pattern in Editor to save the pattern with the current filename Or click Save
23. Eye Opening 221 Output Levels 197 Output Timing 178 Sampling Point Setup 123 BER vs Threshold Graph 195 Bias Tees 70 Bit Rate Different In Out Rates 14 Bit Rate Range Error Detector 108 Bracketing Approach 168 Burst Results 269 Burst Sync Mode 133 C Cancel button 122 CDR 107 Clock Data Recovery 107 Clock Falling Edge button 120 Clock In port Error Detector 101 Clock In termination 103 Clock Input 64 Clock Out port 65 Clock Rate Error Detector 112 Clock Rate Indicators 81 Clock Setup Error Detector 111 Clock Source External 80 Internal 80 Clock Termination 75 Color Tab 153 Complement 105 Confidence Levels 11 Connecting the DUT 13 Connections Amplifier 15 DCA 16 Different In Out Rates 14 Differential Inputs 14 External Data Source 17 Flip Flop 15 MUX DEMUX Pair 15 Oscilloscope 16 SONET SDH Receiver 16 Constant Errors 279 Contour plot 210 Creating New Pattern 36 Criteria for Moving to the Next Measurement Point 234 Crossover 68 D Data Center button 122 Data In port 101 Data Input Delay 118 Data input delay 120 Data Input Setup 105 Agilent Serial BERT N4901 2 User s Guide January 2006 DATA LOSS 276 Data Out port 65 Data Termination 75 Date and Time 290 dBER vs Threshold Graph 196 DCA Connections 16 Decimal Places Eye Opening 221 Fast Eye Mask 237 Output Levels 197 Output Timing 178 Decision Threshold 105 Delay 68 74 Delay Control Input 65 Deterministic Jit
24. In contrast to loading a pattern from file the download of a sequence requires an additional command Agilent Serial BERT N4901 2 User s Guide January 2006 55 User Defined Sequences Reference The elements of the Sequence Editor and the dialog boxes that can be opened from the Sequence Editor are described below Sequence Block Display A new sequence consists of one block that is infinitely repeated looped By default this block has a length of 512 bits and generates a Pause0 signal a continuous stream of zeros All this is shown on the display Edit r PauseO 512 The vertical waveform at the left hand side of the block indicates the setting of the pattern generator s Trigger Out port By default it generates a pulse a subrate of the present clock For information on how to change the trigger behavior see Sequence Trigger at Block Begin on page 57 Number of Blocks In the Blocks field you can directly increase or decrease the number of blocks range is 1 to 4 Edit PauseD 512 Edit PauseO 512 In this example the number of blocks has been set to three A None E block generates no pattern and is completely ignored indicated by a straight vertical blue line 56 Agilent Serial BERT N4901 2 User s Guide January 2006 Sequence Trigger at Block Begin After downloading the sequence to the pattern generator you can also change the behavior of the Trigger
25. It is recommended to use this prior to clock data alignments When used with the following patterns this function may provide a 0 1 threshold that requires no further adjustment e 2 n 1 PRBS patterns e Patterns with 50 mark densities large amplitudes and no excessive noise Agilent Serial BERT N4901 2 User s Guide January 2006 123 Sampling Point Display This area of the Sampling Point Setup window graphically illustrates the parameters of the current sampling point and data eye Click anywhere in the sampling point display to set the sampling point to that location Suggested eye A green area in the display indicates the voltage range specified in the Input Range fields for non differential modes Normal and Complement The Serial BERT is searching this range to find the eye when calculating the optimum sampling point Return to Results Click this button to return the 0 1 Threshold and Data Delay to the values measured during the last auto alignment The Data Inverted and Avg 0 1 Threshold checkboxes will be returned to their previous state Agilent Serial BERT N4901 2 User s Guide January 2006 Trigger and Aux Output Setting up the Error Detector Trigger and Aux Output The error detector has a Trigger Out and an Aux Out port that you can use to send signals to other external devices such as an oscilloscope The Trigger Out can be set up to send a trigger to an external device either according to the c
26. The error detector shifts the incoming data stream bitwise to match it to the expected data pattern A correct BER can only be measured with matching patterns Error accumulation You can specify whether a test runs for a specified time or until a specific number of errors has occurred This lets you carry out longer tests while logging the results to a file BER location mode You can specify whether all errors are counted or only the errors that occurred on a particular bit position or range of positions Agilent Serial BERT N4901 2 User s Guide January 2006 99 e Audio warnings You can set up the audio warnings so that the instrument beeps when a certain BER is exceeded Trigger output for external measurement instruments This allows you to connect other devices for further error analysis 100 Agilent Serial BERT N4901 2 User s Guide January 2006 Inputs and Outputs The error detector ED provides input and output ports for running tests and for connecting external equipment The error detector has the following inputs and outputs ps ooo Pattern Generator AUX ERROR TRIGGER IN ADD OUT CLK DATA DATA DELAY OUT OUT OUT CTRLIN Pattern Generator Error Detector can Daia Data Out pot O cioc Decision Data In Amplitude Offset Delay Amplitude Offset Threshold Delay
27. and displayed A Sync Loss is recognized when the BER is greater than the sync threshold This can be caused by a high error rate pattern misalignment or clock loss Choose the sync mode setting that is appropriate for the type of errors you anticipate e Automatic Sync with a sync threshold BER of 1E 3 is recommended for most applications With this mode selected the synchronization algorithm starts whenever the BER exceeds the threshold However it is not possible to make accurate BER measurements higher than the sync threshold Agilent Serial BERT N4901 2 User s Guide January 2006 129 e Manual sync can be used for synchronizing once confirming proper pattern alignment and then measuring BERs higher than the sync threshold This is useful for the following applications To monitor the integrity of clock signals You may wish to measure BERs that exceed the sync threshold to confirm clock slip To collect data for constructing eye contour information You may wish to move the sampling point to locations in the data eye that have BERs exceeding the sync threshold This mode doesn t allow the analyzer to automatically synchronize if the BER becomes greater than the sync threshold For example the analyzer will not re synchronize after momentary clock loss NOTE Adjusting the data input delay may cause momentary clock loss If you select Manual Sync mode this may also result in sync loss e Burst sync mode is a spe
28. as indicated in the table 198 Agilent Serial BERT N4901 2 User s Guide January 2006 Noise Results The level parameters are defined as follows e High Level The High Level is the mean of the upper dBER dTh distribution It is calculated as Y dBer threshold _ gt dBer Mean Low Level The Low Level is the mean of the lower dBER dTh distribution It is calculated as gt dBer threshold Mean gt dBer Mean Level The Mean Level is the middle between the High and Low Levels calculated as LowLevel HighLevel Mean Level 7 Amplitude The Amplitude is the difference between its High and Low Levels Threshold Margin The Threshold Margin is the distance between the upper and the lower BER curves at the position given by the BER Threshold setting The noise result parameters are summarized in the following table High Level Std Dev Low Level Std Dev Peak Peak Noise max Signal Noise Ratio RMS min Signal Noise Ratio Peak min Peak For some of these parameters pass fail limits can be set as indicated in the table Agilent Serial BERT N4901 2 User s Guide January 2006 199 The noise parameters are defined as follows High Level Std Dev The standard deviations are derived from the dBER dTh histogram The High Level Standard Deviation is calculated as SidDev lt l gt threshold Mean dBer gt dBer where Mean is the High
29. they are compatible with a variety of logic families With respect to Data Out Data Out has inverted logic e Clock Out and Clock Out The clock outputs serve as frequency bit rate references and can be set up so that they are compatible with a variety of logic families With respect to Clock Out Clock Out has inverted logic e Trigger Out This port allows you to connect a trigger for another device for example an oscilloscope The complementary outputs can be used when e additional output capability is needed for an instrument such as an oscilloscope or digital communication analyzer e your device requires differential inputs The pattern generator s Data Out and Clock Out ports must be terminated with 50 Q if they are not connected Understanding the Output Protection Algorithm The generator module offers a huge flexibility for external termination schemes and external termination voltages to address common technologies For details please refer to the Technical Data Sheet An internal protection algorithm continuously monitors the voltages of clock and data output It becomes active if the termination voltage is wrongly adjusted Agilent Serial BERT N4901 2 User s Guide January 2006 65 NOTE The protection algorithm is not active if the termination voltage is higher than 1 5 V The protection algorithm sets the output voltages to safe levels Vhign termination voltage 1 V Viow termination voltage 0
30. 0 mv 200 0 mv ooy 200 0 mY 400 0 mY 600 0 mY BER Threshold 1 000e 6 Logarithmic 1 000e 5 1 000e 3 1 000e 1 1 000e 6 1 000e 4 1 000e 2 1 000e Changing Properties If you change the measurement settings after the measurement has been run please note e Parameters that affect the data capture Changes on the Parameters tab take only effect if you run the measurement again To remind you that the present results have not been obtained with the modified settings and that you should repeat the measurement the result display shows a yellow bar e Parameters that change the display of the measured data Changes on the Pass Fail View Graph and Color tabs only affect the display of the results There is no need to repeat the measurement Variable Decision Threshold Method The method used by this measurement is commonly known as Variable Decision Threshold Method It provides a vertical analysis of the eye opening seen by the receiver This method allows you to determine more than just the actual levels The Output Levels measurement also calculates the Q factor a measure that describes the quality of the received signal and derived values These results can assist you in characterizing the device They can also enable you to predict very low bit error rates that would take a long time to be measured Agilent Serial BERT N4901 2 User s Guide January 2006 185 Output Levels Procedures This sectio
31. 02 04 06 08 delav adust lu delay adjust u 2 All measurement points that have BER between the BER Threshold and Minimum BER for RJ DJ Separation are transformed into Q space The Q factor describes the signal to noise ratio at the decision circuit It is described in Understanding the Q Factor Results on page 204 3 Linear regression is performed for both the left and right edges 4 The mean and sigma are calculated for both lines RJ is calculated as the mean of the two sigmas DJ is calculated as the period minus the difference of the two means 5 The estimated TJ is calculated Linear regression is used to extrapolate the bathtub curve to lower BER values The intersections of the resulting lines with the Residual BER for RJ DJ Separation are located 164 Agilent Serial BERT N4901 2 User s Guide January 2006 2 70e 10 2 60e 10 2 50e 10 2 40e 10 2 30e 10 2 20e 10 2 10e 10 2 00e 10 1 90e 10 1 80e 10 41 70e 10 1 60e 10 1 50e 10 1 40e 10 41 30e 10 1 20e 10 1 10e 10 1 00e 10 9 00e 9 8 00e 9 7 00e 9 6 00e 9 5 00e 9 4 00e 9 3 00e 9 2 00e 9 1 00e 9 Linear Relative The eye opening is calculated The estimated TJ is the period minus the width of the eye opening The illustration below shows a jitter curve where both RJ and DJ are present It also shows how the TJ peak to peak and RMS are calculated 41 Trace 79 Points Ul 333 333 ps All Errors 230 00 ps 21
32. 1 bit misalignment in this case 1E 3 1000 ones 1000 zeros Incoming data Retererce i 3 BER 1e 0 alignenert 1 ae frei ie 100 Retererce i alignment 2 i BER 2 5e 1 q 25 Reference alignment 3 BER 16 3 ibit ibit tht 1 bit 0 1 How Can You Tell if Your Synchronization is False You may suspect false synchronization under the following conditions e You are using a pattern other than PRBS and the error detector gains sync but it measures a constant fixed error ratio e You are using a pattern other than PRBS and the error detector gains sync but auto search functions Auto Align Clock Data Center 0 1 Threshold Center repeatedly fail In a false sync the sync threshold BER of 1E 3 may be met but eye edges at BER 1E 3 required by an auto search function may not be found This is because BERs less than 1E 3 do not exist within the data eye Agilent Serial BERT N4901 2 User s Guide January 2006 131 Possible error profile after fase synchronization Possible error profile after tue synchronization If you suspect a false sync try re synchronizing at a sync threshold BER lower than the fixed error ratio If sync is acquired without the problems listed above then your previous sync was false Your current sync should be on an exact pattern alignment NOTE While auto search functions are in progress the sync threshold BER is changed to the same value as the align
33. 100 00 KHz 10 00 MHz 1 00 GHz 1 00 MHz 100 00 MHz Terminal Electrical 248 Agilent Serial BERT N4901 2 User s Guide January 2006 How to Optimize the View of the Results After you have run a measurement the resulting graph and the calculated numerical values are displayed To improve the results you can change the measurement parameters 1 Press the Properties button to open the Properties dialog box 2 Use the different tabs in this dialog box to make the required settings Parameters tab These settings are used for data collection Changes require to run the test again See Parameters Tab on page 173 for details Pass Fail tab These settings determine whether the calculated results are recognized as passed or failed However a new test run is not required when doing changes here See Pass Fail Tab on page 174 for details View tab Graph tab and Color tab All settings on these tabs only affect the way the data is displayed You do not need to run the measurement again See View Tab on page 176 Graph Tab on page 178 and How to Change the Colors of the Graph on page 153 for details 3 Press OK when you have made all required changes to close the Properties dialog box Spectral Jitter Reference The Spectral Jitter measurement returns the results in a graphical and in a numerical form The following sections provide explanations of the measured paramete
34. 2006 91 Pattern Alternation This section explains the basic purpose of alternating patterns Two alternate patterns can be stored in one pattern file Note that you can also generate a sequence of up to four alternating patterns see User Defined Sequences on page 51 Standard and Alternating Patterns With the Serial BERT you have full control on the data stream that is sent by the pattern generator to your device under test You can use standard patterns to specify a test data pattern that is sent to your DUT repeatedly Or you can use alternating patterns to specify two patterns that are sent to the DUT alternately To implement alternating patterns the instrument s user pattern memory is split into two portions one for pattern A and one for pattern B The differences between these patterns are Standard patterns consist of one single pattern that is repeatedly sent Any arbitrary pattern up to 32 Mbit length can be set up However there are no further configuration possibilities This pattern is sent as long as the output port is enabled See Setting up Patterns on page 19 for more details e Alternating patterns consist of two patterns one of which is pattern A the A half the other is pattern B the B half Both patterns are of equal length each up to 16 Mbit The Alternate Pattern Control dialog box lets you control when which pattern is sent See the online Help for details When to Use Alte
35. 4 950459E 002 0 000000E 000 1 551503E 011 4 950459E 002 0 000000E 000 This file can be imported into spreadsheet applications such as Microsoft Excel for example Use the semicolons as delimiters for the rows or columns Version B 16 010306 Type DUT Output Timing Jitter Ul 2 50E 05 2 1 DataQ Delay BER All Compared Bits Errors All 3 60E 09 0 500044 1 00E 06 5 00E 05 3 73E 09 0 499992 1 00E 06 5 00E 05 3 85E 09 0 499947 1 00E 06 5 00E 05 3 98E 09 0 500007 1 00E 06 5 00E 05 154 Agilent Serial BERT N4901 2 User s Guide January 2006 Timing Unit Definitions The timing unit specifies the timebase for the measurements It is possible to switch between the unit interval UI or seconds for the time scale The timebase is set on the View page of the Properties dialog box The following figure and table shows the relationship between UI and seconds for a 500 MHz signal BER log scale An example of the measured range of time expressed in the different units is listed below Time rel 1 5ns Ons 1 5 ns abs 54 2 ns 55 7 ns 57 2 ns Unit interval rel 0 75 Ul 0 UI 0 75 UI abs 27 1 Ul 27 85 Ul 28 6 Ul The Unit Interval Unit interval values are a convenient way to express time values in a dimensionless form One unit interval is the equivalent of one clock period To convert a time value to a unit interval value divide it by the pulse period or multiply it by the clock frequency
36. 9 V e termination voltage remains unchanged As soon as a 50 load and the correct external termination voltage is detected again the algorithm automatically switches back to the user adjusted output levels These changes are neither visible in the user interface nor can they be queried by a remote command Example User settings with single ended operation 2 1 2 0 5 50 Qto 0 5V wrongly left open An open condition is detected and the protection algorithm sets the following levels Vhigh 0 5 V and Viow 0 4V Next the complement output Out is externally terminated with 50 Q to GND The protection circuit detects the correct external terminations and resets the voltages to the initial user settings Vhigh 2 Vand Viow 2 V Understanding the Output Level Parameters The following figure shows the parameters of a data or clock output signal 66 Agilent Serial BERT N4901 2 User s Guide January 2006 As shown in this figure the signal output levels have the following components e Viis the upper voltage level of the signal e Vj is the lower voltage level of the signal e Voe is the offset of the average voltage level from 0 V e Vampt is the amplitude of the signal When adjusting the output levels it is important to understand the concept of how the Serial BERT handles voltages Changing the Amplitude Typically during tests when you adjust the amplitude you want to keep the offset
37. All Errors 600 0 mv 400 0 mv 200 0 mv oov 200 0 mv 400 0 mY 600 0 mv sscale 0 448 UI 0 149 UI 0 149 UI 0 448 UI 0 746 UI Relative 0 597 UI 0 299 UI 0 000 Ul 0 299 UI 0 597 UI Terminal El Delay Agilent Serial BERT N4901 2 User s Guide January 2006 213 How to Optimize the View of the Results After you have run a measurement the resulting graph and the calculated numerical values are displayed To improve the results you can change the measurement parameters 1 Press the Properties button to open the Properties dialog box 2 Use the different tabs in this dialog box to make the required settings Parameters tab These settings are used while the measurement is running Changes require to run the test again See Parameters Tab on page 217 for details Pass Fail tab These settings determine whether the calculated results are recognized as passed or failed However a new test run is not required when doing changes here See Pass Fail Tab on page 218 for details View tab Graph tab and Color tab All settings on these tabs only affect the way the data is displayed You do not need to run the measurement again See View Tab on page 220 Graph Tab on page 222 and How to Change the Colors of the Graph on page 153 for details 3 Press OK when you have made all required changes to close the Properties dialog box 214 Agilent Serial BERT N4901 2 User s
38. Aux In Set the Internal Alternate Pattern Mode to Single Shot B and disable Aux In Set up the error detector so that the Clock Setup is set to Clock Data Recovery to get the error detector s clock from the incoming data stream Switch to the Analysis panel and then press the Error Location Capture icon Press the Start button to execute the measurement Status information is displayed in the lower section of the display Press Insert B button The two captured errors are illuminated red in the captured pattern grid In the result list the 1st Bit Error Location Capture indicates an error in bit 80 Use the error navigation buttons e g Next Error Prev Error to display further bit errors in the captured pattern Error Location Capture Reference The Error Location Capture measurement compares the expected pattern and the pattern that was actually received and shows the received pattern with the errored bits marked The following buttons help you to analyze the results Agilent Serial BERT N4901 2 User s Guide January 2006 221 e Hex Bin Choose between a binary or hexadecimal display of the pattern The current setting is shown in the status bar below the pattern window Note that in hexadecimal view the captured error can only be located as being one of a group of four bits To display which of the four bits is the errored bit you have to switch to the binary display e Exp Cap Expected Captured Togg
39. File menu click Open and then Open Instrument State This opens a dialog box which allows you to select the setup file you want to recall TIP Remember always to make a backup copy your setup files If your original files are damaged or lost you can restore the files from your backup copy NOTE The default pathname for user defined instrument states is C N4901A Settings Agilent Serial BERT N4901 2 User s Guide January 2006 301 Saving and Recalling Measurement Logs Measurement Logs are saved as CSV comma separated values files Follow the steps below to view results in a measurement log file 1 Copy the measurement log file from the Serial BERT to your PC By default log files are saved in the folder C N4901A 1log 2 Open a spreadsheet application on your PC 3 Import the measurement log file If your application has an import wizard you may need to indicate that data is delimited with commas 4 Once the file has been imported you may need to resize columns The imported log file should appear similar to the following example NOTE During measurement logging the Serial BERT logs data in ten second intervals Your log file may be missing up to the last ten seconds of data To avoid this condition accumulate for 10 seconds longer than desired m E I Cc D j E Description test Log Time 05 29 2000 12 14 58 334 Activation Mode Single Accumulation Measurement ELAPSEO_SECONDS Accumulation Limit 60
40. Inserting Errors To manually insert a single error into the output stream 1 Click the Error Add button at the top of the screen to insert one single bit error in the output stream Error Add l This button is active unless you define a fixed internal error rate The maximum number of errors that this function supports is one error per 128 bit block 2 Ifyou want to disable all other error insertion functions so that only manual error insertion is possible click the Error Add Setup icon in the PG Setup menu to open the Error Add Setup dialog box Jil 3 Select Off and click OK to close the dialog box 88 Agilent Serial BERT N4901 2 User s Guide January 2006 TIP Using an External Signal for Inserting Errors To configure the Serial BERT to insert a single error into the output stream according to an external signal 1 Connect an external instrument to the Error Add port The signals received at this port must be TTL compatible 2 Click the Error Add Setup icon in the PG Setup menu Jal 3 Select External Error Add and click OK to close the dialog box Now upon a rising edge of the signal at the Error Add port an error will be generated in the output stream by flipping a single bit within a bit block of 128 bits resulting in a maximum BER of 1 128 or 10 You can use an external signal to set up a bit error rate to fit your needs The frequency of the external signal can be calculated by multiplying the desi
41. Level of the terminal Low Level Std Dev The Low Level Standard Deviation is calculated as StdDev A threshold Mean dBer gt dBer where Mean is the Low Level of the terminal Peak Peak Noise The peak to peak Noise is calculated as Peak Peak Noise HighLevel LowLevel Threshold Margin Note that the Threshold Margin depends on the position of the BER Threshold Signal Noise Ratio RMS The RMS Signal to Noise Ratio is calculated as HighLevel LowLevel StdDev 1 StdDev 0 SNR RMS Signal Noise Ratio Peak Peak The peak to peak Signal to Noise Ratio is calculated as _ HighLevel LowLevel SNR PeakPeak po okPeakNoise Note that the Peak Peak Noise depends on the position of the BER Threshold 200 Agilent Serial BERT N4901 2 User s Guide January 2006 Q factor Results The Q factor result parameters are summarized in the following table Q Factor min Q Optimum Threshold min max Q Residual BER max Q High Level Q High Level Std Dev Q High Level Nr Points Q High Level R 2 Q Low Level Q Low Level Std Dev Q Low Level Nr Points Q Low Level R 2 For some of these parameters pass fail limits can be set as indicated in the table The numerical Q factor parameters are defined as follows Q Factor The Q factor is calculated as os ko Q 0 0 where u 9 is the mean level of the 1 and 0 rails respectively and 01 0 is the sta
42. Marware Generated Software Generated Patterns User Patterns 2 n 1 PRBS PRBS Zero Substitution Mark Density custom oeer i Ea T Fiber Chi l i 24n PRBS FDDI Gigabit ethernet 2MPRBS mark Density SDH SONET Industry Standard Pattern or modified Industry Standard Pattern See Example Patterns on page 29 for a complete listing of the example patterns Agilent Serial BERT N4901 2 User s Guide January 2006 What is PRBS The most common type of industry standard pattern is the Pseudo Random Binary Sequence PRBS PRBSs have a number of desirable properties including the ability to simulate random data with a balanced number of 1 s and 0 s They are referred to by their pattern lengths which are powers of 2 Longer pattern lengths give a better approximation of random data and provide a more rigorous test for the device However the pattern repetition time is significantly longer PRBS patterns provide a means to simulate the type of traffic that a system is likely to see random traffic They also enable you to stress your system to its limits because large patterns that include long strings of 1 s or 0 s are more stressful The patterns are easy to generate and can measure at high speeds because they are sent repeatedly and can be predicted Pattern synchronization for example is very fast When you compare the sequence with itself you get 50 errors in al
43. N bit Trigger Pattern e Bit Position If a memory based pattern is loaded in the pattern generator you have to specify the bit in the pattern that sends the trigger The bit position can be in the range 0 Pattern Length 1 For pattern lengths that are divisible by 512 such as 2 n PRBS patterns the trigger repetition rate is one pulse per pattern repetition Bit Pos Data Out Trigger Out 84 Agilent Serial BERT N4901 2 User s Guide January 2006 For pattern lengths that are not divisible by 512 such as some memory based patterns a trigger signal is sent once at the corresponding bit when the pattern reaches the 512 bit boundary If for example the pattern is 384 bits long a trigger signal is sent every 4th pattern repetition Bit Pos AD 1 A2 DataOut A AAJA AJAA Trigger Out If the entered number is greater than the pattern length the entry is adjusted to the pattern length 1 If you load a pattern that is shorter than the entry in this field the trigger will be set to the next to the last bit in the pattern e N bit Trigger Pattern This option is used for hardware generated PRBS patterns You can specify a pattern with n bit length A trigger pulse is sent when this pattern is found in the PRBS The trigger repetition rate is one pulse per 512 pattern repetitions Example Pattern length 2 7 1 127 bits N bit patt
44. Output Timing measurement You can also drag and drop the horizontal BER threshold in the graphical display to change this value e Min BER for RJ DJ Separation Lower limit of the BER range for RJ DJ separation Residual BER for Estimated Total Jitter BER level for which the estimated total jitter is calculated See Estimated Total Jitter on page 165 for details on how it is used Table Number Format You can select the number of Decimal Places to be displayed in the table Graph Tab On the Graph tab you can use the several options to optimize the graphical display according to your needs Timing Units Choose between Unit Interval and Seconds to select the timebase for the display s x axis Scale Choose between Logarithmic and Linear to select the scale for the display s y axis For example a DUT Output Timing measurement displayed on a linear scale may look like this 6 00e 001 1_Trace 207 Points Ul 2 000 ns All Errors 5 50e 001 5 00e 001 4 50e 001 4 00e 001 3 50e 001 3 00e 001 2 50e 001 2 00e 001 1 50e 001 1 00e 001 Pr 77 TT TT Tyr Br rr ee re ren erin paren cece ns 5 00e 002 Linear 0 50 UI 0 30 UI 010 UI 0 10 Ul 0 30 UI 0 50 UI 0 70 UI Relative 0 60 UI 0 40 UI 0 20 UI 0 00 UI 0 20 UI 0 40 UI 0 60 UI 178 Agilent Serial BERT N4901 2 User s Guide January 2006 Markers Zoom Show Measured Points To analyze the graphs at a particular point you can use the markers Additionall
45. Pattern in Editor as to open a file dialog box that allows you to save the pattern with a new name The default pathname for patterns is C N4901A Patterns Saving and Recalling Screen Images The instrument allows you to save an image of the current display screen You can save this image to the internal hard disk or a floppy disk You can later print this file or import it to a word processing or graphics editing program To save an image of the display screen do the following 1 In the File menu select Save 2 Select Save Screen Capture This opens a dialog box which allows you to save the image as a bmp 256 color bitmap file format file Follow the steps below to view screen images on an external PC 1 Copy the screen image file from the Serial BERT to your PC By default image files are saved in the folder C N4901A User 2 Open the file in Microsoft Paint or another graphics program Agilent Serial BERT N4901 2 User s Guide January 2006 303 Preset Instrument State Preset Instrument State Procedures To reset the instrument to the preset state select one of the following options Press the Preset button on the front panel or Select Preset in the File menu or Send an RST command to the GPIB controller 304 Agilent Serial BERT N4901 2 User s Guide January 2006 NOTE NOTE Self Test The self test checks specific hardware components for basic functionality At initialization
46. Pattern menu z 2 In the toolbar click the Open icon ow The Select User Pattern dialog box opens 3 Use this dialog box to locate and open the desired pattern You can select files of the following types 71630 BERT Pattern Files dat 86130A BitAlyzer Pattern Files ptrn ASCII Text Pattern Files txt See Supported Pattern Types on page 24 for descriptions of these files 36 Agilent Serial BERT N4901 2 User s Guide January 2006 Editing Patterns To edit a user pattern 1 Press the Edit Pattern on File icon in the Pattern menu A 2 If no pattern is loaded in the editor first open a pattern file as described in Opening Existing Patterns on page 36 or create a new one as described in Creating New Patterns on page 36 3 To change the pattern length or description or to switch between standard and alternating patterns press the Properties button The Pattern Properties dialog box appears 4 Define the pattern length whether it is alternate or standard and give it a description Click OK when finished 5 Edit the pattern as required For the function of the different tool buttons see Edit Pattern Window on page 40 Saving Patterns To save the current user pattern In the toolbar click the Save icon If the pattern has already been saved earlier the saved file is updated Otherwise the Windows standard Save As dialog box opens where you can define the file s p
47. Range on page 108 for more information 120 Agilent Serial BERT N4901 2 User s Guide January 2006 Auto Align Click this button to automatically set the optimum sampling point This routine will not stop if the optimum sampling point cannot be found If no optimum sampling point is found after a reasonable time you can click Cancel The status message bar will indicate if this function is unsuccessful The following settings may affect the result of the auto align function Data Inverted BER Threshold TIP Alternatively you can also press the Auto Align button on the front panel Sync Now Click this button to manually start a pattern synchronization See What Type of Synchronization Should You Use on page 129 for information on when you should use this function 0 1 THReshold Center Starts an auto search function that sets the 0 1 threshold to the optimum point of the incoming data eye on the vertical voltage axis without changing the data input delay This function can be used for determining the optimum 0 1 threshold for asymmetric data eyes or for patterns with an unequal mark density The status message bar will indicate if this function is unsuccessful This function clears the Avg 0 1 Threshold checkbox NOTE This function uses the alignment BER Threshold to determine the top and bottom eye edges Agilent Serial BERT N4901 2 User s Guide January 2006 121 Data Center This button starts a
48. Single Shot B With this option selected the Insert B button is enabled 3 Select OK when your selection is complete Agilent Serial BERT N4901 2 User s Guide January 2006 93 Manually Inserting Pattern B into the Output Stream You can insert a single shot of pattern B into the continuous stream of pattern A 1 Press the Alt Pattern and Aux In icon in the PG Setup menu to open the Alternate Pattern Control dialog box 2 Select Single Shot B and click OK to close the dialog box 3 During the test run press the Insert B button on the screen Insert B l This inserts a single occurrence of pattern B into the output stream of pattern A Inserting Pattern B via External Signal You can configure the Serial BERT to insert pattern B into the continuous output stream of pattern A according to an external signal 1 Connect an external instrument to the Aux In port The signals received at this port must be TTL compatible The granularity for the insertion of pattern B via external signal is 512 bits 2 Press the Alt Pattern and Aux In icon in the PG Setup menu to open the Alternate Pattern Control dialog box 3 In the Aux In section define whether the Serial BERT should insert pattern B in Level Sensitive or Edge Sensitive mode See Aux In on page 96 for more information on the different available modes 4 Click OK to close the dialog box 94 Agilent Serial BERT N4901 2 User s Guide January 2006 Su
49. The BER bar displays the BER calculated upon a period of 200 ms BER 1 145E 3 You can drag the yellow alarm threshold mark to change the BER Alarm Threshold Error Indicators These indicators inform you about the following errors e Bit error in the data stream located e Loss of pattern synchronization e Complete loss of data Loss of the error detector s clock signal Loss of the pattern generator s clock signal Errors SYNC DATA ED CLK PG CLE ees eel os LOSS For troubleshooting information see Solving Problems on page 275 Remote Indicator This indicator turns green when the instrument is under remote control pe Status Messages and More Button Status messages displays a one line instrument status or error message Click the More button to open an Error Message dialog box with the complete message This dialog box also contains a Help button that takes you to the respective section of the Online Help Agilent Serial BERT N4901 2 User s Guide January 2006 261 Elapsed Indicator This area displays the progress of the current accumulation period The accumulation period can be changed in the Accumulation Setup dialog box For more information see Accumulated Measurements Procedures on page 263 Elapsed 00 01 21 Pattern Types The PG and ED pattern types are displayed here for your reference Clock Rates The PG and ED measured bit rates are displayed here for your refer
50. This allows you to speed up the measurement You can switch off this option if only the number of compared bits is important NOTE The measurement moves to the next sample point when the first of the two criteria is reached Both numbers are ignored if the Fast Total Jitter measurement is selected Set the criteria for the sample delay e Resolution Specifies the time distance between sampling points A smaller value yields more sampling points in a unit interval You can enter the resolution in UI or ps ns s The timebase of the display is set on the View tab If the resolution used for the measurement is not high enough the bathtub curve does not clearly show the edges For example you may wish to change the Resolution from 0 01 to 0 005 and run the measurement again Optimization You can choose between Edge Resolution Optimization Fast Total Jitter at BER or None Edge Resolution Optimization Turns the resolution optimization on or off If this option is enabled the Serial BERT intelligently sets a resolution so that there are more sampling points at the edges This can greatly Agilent Serial BERT N4901 2 User s Guide January 2006 173 174 1 00e 0 1 00e 1 improve the results without dramatically increasing the duration of the test The following illustration shows a diagram with default resolution and no edge resolution optimization 3 Traces 453 Points Ul 2 500 ns All Errors
51. Used to test multiplexer demultiplexer MUX DEMUX pairs For accurate measurement results the timing of data signals between the MUX DEMUX pair must be set properly Agilent Serial BERT N4901 2 User s Guide January 2006 15 Diagram 6 Connections for a DCA or Oscilloscope PG only Used for measuring the output waveform of your device For the least amount of jitter the clock output signal may be used as a trigger for the oscilloscope Diagram 7 Connections for a SONET SDH Receiver PG only Often used with network equipment such as SONET SDH receivers The device shown may or may not require a clock signal 16 Agilent Serial BERT N4901 2 User s Guide January 2006 Diagram 8 Connections for an External Data Source ED only Used when the device is stimulated by another data source Illustrates the use of an oscilloscope in addition to the error detector DATA SOURCE NOTE Asin all setups the pattern sent to the error detector must be the same as the internal reference pattern of the error detector Dynamic patterns and or patterns with live payloads cannot be used NOTE Diagrams 1 through 8 can use an external clock reference connected to the Clock In port of the pattern generator if desired Agilent Serial BERT N4901 2 User s Guide January 2006 17 18 Agilent Serial BERT N4901 2 User s Guide January 2006 Setting up Patterns The purpose of data patterns is to simulate the type of data that
52. Version Format Description Count Length Data If you save a modified text format pattern to the same file name it is always saved in binary format If you wish to preserve the original text format file use the Save As file command and save the data to a different file Pattern File Specifications For your instrument to download a pattern file the pattern must be composed of at least 7 lines The first 6 lines start with a keyword to identify the field followed by an sign There should be no spaces around the sign and no blank lines Use the following keywords and examples to construct your ASCII file Note that the keywords must appear in the given order The version field is used to identify different file versions For this software version the field should be set to EPA 2 0 An exact match for this field is required for the pattern to load correctly Syntax Version EPA 2 0 The format field identifies whether the pattern is in dual hexadecimal or binary format Syntax Format lt Dual Hex Text Bin gt Hex and text formats are identical This is the creator s description of the pattern The data after the sign is read until the end of the line and used directly as the pattern description The maximum number of characters is 256 This identifies standard and alternating patterns Standard patterns have a value of 1 alternating patterns have a value of 2 Any other value is not valid Syn
53. a second shielded cable to connect the pattern generator s Clock Out port and the error detector s Clock In port 4 Terminate all non connected pattern generator output ports with 50 Q 5 Make sure to preset the instrument state before running the measurement Select Preset Instrument State from the File menu 6 Enable the pattern generator outputs by pressing the OV Disable button 7 Press Sync Now and then Auto Align to find the optimum sampling point Check that the synchronization and alignment were successful None of the error indicators should show red The resulting BER should be zero 226 Agilent Serial BERT N4901 2 User s Guide January 2006 How to Demonstrate the Error Location Capture Measurement To demonstrate the Error Location Capture measurement 1 Switch to the Pattern panel and press the Open button Load a Fiber Channel random data pattern from the demo patterns Demo gt Fiber Channel gt RPAT ptrn Press the Properties button and set the Pattern Type to Alternate To display both halves of the alternate pattern press the Alt pat view button Ensure also that the error detector tracks the pattern generator Insert two errors into pattern B To do so edit bits 80 and 100 of pattern B in overwrite mode Write the modified pattern to the pattern generator and the error detector by pressing the ToPG amp ED button For the pattern generator setup you need to specify the Alternate Pattern and
54. and the alignment were successful None of the error indicators should show red The resulting bit error rate should be zero How to Execute the Fast Eye Mask Measurement To run the Fast Eye Mask measurement 1 Switch to the Analysis panel and then press the Fast Eye Measurement icon 2 Press the Start button to execute the measurement The measurement is run and the result window shows the bit error rates measured at six measurement points Terminal Relative Time 0 4 UI 0 4 Ul 0 16 UI 0 16 Ul Voltage abs 224mY 224m 178 mv 222 mY 0 o o o o 22 4 mV 178 mY 178 mV 222 mV The Relative Time refers to the current sampling point The Voltages are the decision threshold voltages for measuring the bit error rate at this measurement point The voltages of the measurement points can be set as absolute voltages as offset voltages or as percentages How to Optimize the View of the Results After you have run a measurement the resulting numerical values are displayed To improve the results you can change the measurement parameters 1 Press the Properties button to open the Properties dialog box 2 Use the different tabs in this dialog box to make the required settings Parameters tab These settings are used for data collection Changes require to run the test again See Parameters Tab on page 234 for details Pass Fail tab These settings determine whether the calculated results a
55. constant This keeps the ideal sampling point within the eye The Serial BERT handles this by keeping Vo constant when Vampt iS changed Changing the Output Levels On the other hand you may want to adjust the output voltage level without changing the amplitude The Serial BERT handles this by keeping Vampt constant when either Vof Vni or Vio is changed Voltage Level Restrictions The Serial BERT cannot generate a signal that has voltage levels out of range If you try to enter a value for one parameter that would put another parameter out of limits the Serial BERT rejects the change This could happen for example if Vj is already at the minimum and you try to lower either of Vj or Vos or increase Vampt Agilent Serial BERT N4901 2 User s Guide January 2006 67 Understanding Delay and Crossover The Serial BERT provides the possibility of modifying the output data signal by varying the signal s delay to the clock signal and the signal s crossover Delay The exact time delay through a test setup can vary The delay function allows you to compensate for this by adjusting the frequency reference for the data outputs This varies the phase relationship of the data and clock outputs causes the data pulse to have a certain time delay after the clock pulse The higher the delay the greater the time difference between the clock signal and the data signal Data Out gt lt Delay Clock Out Crossover Crossover is the
56. correctly and that the error detector is already in synchronization 2 Press the Accumulation Setup icon in the ED Setup menu to check the settings made for the data accumulation See Error Accumulation Procedures on page 140 for details Agilent Serial BERT N4901 2 User s Guide January 2006 263 3 To begin accumulation press the Start Accum button on the front panel This causes the following events If Prompt for File Name has been selected during setup you will be asked to enter a name for the log file before accumulation begins The data display in the Accumulated Results window is reset The recording of data to the measurement log file is reset How to Stop Accumulated Measurements If you specified a single or repeat test the test runs until the specified end condition is reached If you selected a manual test it will run until you stop it manually Press the Stop Accum button on the front panel to stop the test immediately NOTE Errors are only counted when the signal at the Gate In port is low How to View the Results You can view data from accumulated measurements in several ways 1 Press the Accumulated Results icon in the Results menu to view the test results For easy comparison the results of the current and previous accumulation are listed here See Accumulated Measurements Reference on page 265 for details on the reported values 2 Examine the measurement log files on a
57. defined in a pattern file and loaded into the memory The Serial BERT reads the memory bit by bit and generates the output accordingly The following topics explain how the Serial BERT works with user patterns and also explain how user patterns are generated NOTE The way software generated patterns distorted PRBS PRBN are output is identical to that of user patterns The patterns are written to the memory and the output is generated from there Some of the information in these topics covers how the patterns are actually generated and thus also applies to software generated patterns When to Test with User Patterns Creating or editing patterns is necessary for quality testing because different patterns present different data loads to the device This can cause variations in the bit error ratio With user patterns you can define exactly the pattern used in the test Supported Pattern Types The Serial BERT can import the following pattern types e HP Agilent 71600 series patterns dat file type e Patterns saved as ASCII files by the Serial BERT or the Agilent 86130 ptrn and txt file type These are described in detail in Pattern File Specifications on page 25 The Serial BERT can save files as either dat ptrn or txt files NOTE Downloading a text format pattern into your instrument may take significantly longer than a binary pattern 24 Agilent Serial BERT N4901 2 User s Guide January 2006 NOTE
58. each test run e File Name Click this option if you want to enter a file name for measurement logs before starting the accumulation With this option you will overwrite the measurement log file with each new test run However in Repeat mode the results of all repetitions will be appended to this file and no information is lost e No Logging Click this option to turn off measurement logging Serial BERT will not save the test results Agilent Serial BERT N4901 2 User s Guide January 2006 141 Period The accumulation period cannot be specified for Manual tests Time Select this option to configure the error detector to accumulate bit errors for a specific period of time You can enter the desired time period in the Days Hr Hours Mn Minutes and Sec Seconds fields You may use the numeric keypad or front panel knob The minimum value is 1 second and the maximum value is 99 days 23 hours 59 minutes and 59 seconds When the selected time has elapsed the accumulation ends e Number of Errors Select this option to configure the error detector to accumulate bit errors until a specified number of errors has been measured The number of errors can be set to 10 100 or 1000 When the selected number of errors has occurred the accumulation ends after the next full second e Number of Bits Select this option to configure the error detector to accumulate bit errors until the specified number of bits have been examin
59. error detector so that the input range and the termination matches the pattern generator s levels Select Normal as the Active Input an Input Range of 1 3 V and set the Data Termination to 1 3 V 8 Also make sure that the error detector derives its clock from the incoming data stream using Clock Data Recovery See Clock Setup Procedures on page 111 for details 9 Press Sync Now and then Auto Align to find the optimum sampling point Check that the synchronization and the alignment were successful None of the error indicators should show red and the resulting BER should be zero How to Execute the Spectral Jitter Measurement To run the Spectral Jitter measurement 1 Switch to the Analysis area If the Spectral Jitter screen is not yet displayed press the Spectral Jitter icon 2 Press the Start button to execute the measurement The measurement software runs the measurement and displays the results A total of 65535 points are displayed for this measurement Numerical values are also displayed The values include Bit Error Rate Total Power and Noise power Additionally you can also configure up to 16 frequency power pairs The latter are sorted according to descending power contents These results are discussed in Explanation of the Numerical Results on page 180 1 Trace 65535 Points 0 009 0 0081 0 0072 0 0063 0 0054 0 0045 0 0036 0 0027 0 0018 0 0009 o Noise Threshold 0 0001 Logarithmic
60. from its file may fail if the referenced user patterns are not available A recalled sequence is not automatically downloaded to the pattern generator In sequence mode the function Error Dector Pattern tracks the Pattern Generator Pattern is automatically disabled You have to download the expected payload pattern to the error detector During initialization of the sequence pure zeros are sent Agilent Serial BERT N4901 2 User s Guide January 2006 53 User Defined Sequences Procedures You can create a new sequence save it in a file or load a saved sequence from its file Creating a New Sequence If you wish to create a new sequence 1 Choose Pattern and open the Sequence Editor mr By default the Sequence Editor shows one block that is automatically repeated It has a length of 512 bits and generates Pause0 data pure zeros 2 Click the Properties icon In the Properties window set the sequence start condition the number of blocks contained in the sequence 1 to 4 3 Click Ok to close the Properties window 4 Specify the data to be generated during execution of a block Click the Edit button of the block Choose from the list For example you can select a PRBS polynomial or browse for a pattern file A None block is completely ignored If necessary change the block length The block length of user patterns is derived from the file Remember that such patterns are auto
61. is higher than the Burst Sync Threshold the burst will be considered a bad burst CDR Settling Time End too late Synchronization Time Data Input wee T e e Pattern Synchronization Procedures To synchronize the incoming pattern to the expected pattern 1 Press Pattern Sync in the ED Setup menu 2 Select the synchronization mode Normal or Burst sync mode 3 If you selected Normal synchronization mode choose whether you want Automatic or Manual synchronization 4 Specify the Sync Threshold at which the pattern will be recognized at synchronized 5 In Manual synchronization mode use the Sync Now button to start the pattern synchronization This is recommended whenever you did changes to the pattern setup the voltage levels or the hardware connections for example altered cable lengths 6 Press OK to confirm your settings 138 Agilent Serial BERT N4901 2 User s Guide January 2006 Pattern Synchronization Reference The error detector can synchronize the data patterns in the following ways In Normal Sync Mode you can choose between Automatic Sync With this option selected the error detector constantly tries to synchronize the patterns when the BER threshold is exceeded Manual Manual synchronization can be selected for example if the signal delay is very unstable and you want to avoid that the re synchronization process affects the measurement results Burst Sync Mode Click this button t
62. larger external system Use of a modulated clock A frequency or delay modulated clock can be used to provide a small amount of jitter to the clock signal Use of a precision clock A precision clock with very low phase noise can be used to enhance the instrument s performance This is especially interesting for long term measurements Testing at bit rates lt 620 Mb s The Serial BERT s clock cannot be used for bit rates below 620 Mb s Bit Rate Range The Agilent Serial BERT N4901 2 provides bit rates from 150 Mbit s up to 13 5 Gbit s depending on the instrument s options However several specific properties and limitations need to be taken into account when working at low bit rates The limitations apply to the instrument according to the following hysteresis curve e If the bit rate falls below 615 Mbit s the limitations apply e If the bit rate exceeds 620 Mbit s the limitations no longer apply Agilent Serial BERT N4901 2 User s Guide January 2006 71 The following figure clarifies the behavior in the range between 615 Mbit s and 620 Mbit s Limitations do not apply in this range 620 615 Bitrate Mbit s Limitations apply in this range For the pattern generator the following rules apply e Below 620 Mbit s the pattern generator can only be operated with an external clock source because the internal clock source can only produce signals higher than 620 Mbit s The trigger output cann
63. left corner To switch the audio signals on and off you can also use the Audio On Off front panel button Agilent Serial BERT N4901 2 User s Guide January 2006 145 Audio Signals Reference The Audio dialog box contains the following elements Audio on Select this checkbox to enable the audio function Clear this checkbox if you want to switch off the audio function You can also use the Audio On Off button on the front panel to accomplish the same function Main Volume Click any point on the slider or drag it to change the main volume level You will hear a test tone at the new volume level Audio on BER Alarm Click this option to make the analyzer play warning tones when a specific BER is exceeded The BER Alarm Threshold is the threshold at which tones are produced Enter the BER threshold in the following format number E exponent For example 1 5E 3 TIP You can always adjust the BER Alarm Threshold by clicking and dragging the small yellow marker in the BER bar in the top left corner Tones on All Error Rates Click this option to make the analyzer produce warning tones when any BER greater than zero is measured The pitch will increase or decrease as the BER increases or decreases 146 Agilent Serial BERT N4901 2 User s Guide January 2006 Advanced Analysis The Serial BERT offers several different kinds of advanced measurements for various purposes DUT Output Timing Jitter This type of mea
64. new clock rate Delete Preset Button Select a preset clock rate in the Preset list and click this button to remove it from the list You can only delete presets that are user defined Clock Rate Indicators The PG Pattern Generator and ED Error Detector clock rate indicators in the status bar display their current clock rate ED Ck Rate 6 8731 MHz The ED clock rate is measured from the incoming clock signal or derived from the data signal Agilent Serial BERT N4901 2 User s Guide January 2006 81 Setting up the Pattern Generator Trigger Outpt Trigger Output The pattern generator s Trigger Out port can be used to send a reference signal or trigger to external devices like an oscilloscope or digital communication analyzer In pattern mode the pattern generator sends a trigger signal that is at least 32 bits long See also How the Serial BERT Generates Memory Based Patterns on page 27 for details In sequence mode the pattern generator can send a trigger signal whenever a block of the sequence starts or restarts See also Sequence Block Parameters on page 57 for details Several options are available for the trigger signal As an example you can send the trigger as a divided clock signal or as an indicator when the data pattern starts Trigger Output Procedures To set up the pattern generator s Trigger Output port 1 Connect the external instrument to the Trigger Out port 2 Click
65. of Signal Analysis Agilent Application Note 243 Publ No 5952 8898E Fibre Channel Methodologies for Jitter Specifications National Committee for Information Technology Standardization NCITS T11 2 Project 1230 Rev 10 June 1999 Yi Cai Bernd Laquai Kent Luchman Jitter Testing for Gigabit Serial Communication Transceivers IEEE Design and Test of Computers Jan Feb 2002 FF results The error record contains a bipolar rectangular signal If such a signal is periodic you can expect a spectrum as illustrated below p 4 gt requency 1 T Time The fundamental frequency and its harmonics appear 244 Agilent Serial BERT N4901 2 User s Guide January 2006 Such spectra have been measured with the Spectral Jitter measurement 1 Trace 65535 Points 0 27 0 24 0 21 018 015 012 0 09 0 06 0 03 0 Linear 3 00 MHz 5 00 MHz 7 00 MHz 9 00 MHz 11 00 MHz 2 00 MHz 4 00 MHz 6 00 MHz 8 00 MHz 10 00 MHz 12 00 MHz A logarithmic power scale shows the details 1 Trace 65535 Points 10 dB 20 dB 30 dB 40 dB i 50 dB 60 dB 70 dB 80 dB 90 dB 100 dB Linear 3 00 MHz 5 00 MHz 7 00 MHz 9 00 MHz 11 00 MHz 2 00 MHz 4 00 MHz 6 00 MHz 8 00 MHz 10 00 MHz 12 00 MHz bise Threshold 40 dB When the repetition period of the characteristic pattern in the error record increases you will also find the typical sine x over x decay of the spectral power
66. packaged into a Fiber Channel frame PRBS This directory contains PRBS patterns that have different lengths and mark densities e PRBS7_1 8 ptrn PRBS7_2 8 ptrn e PRBS7_1 2 ptrn e PRBS7_6 8 ptrn e PRBS7_7 8 ptrn e PRBS10_1 8 ptrn e PRBS10_2 8 ptrn e PRBS10_1 2 ptrn e PRBS10_6 8 ptrn e PRBS10_7 8 ptrn e PRBS15_1 8 ptrn e PRBS15_2 8 ptrn e PRBS15_1 2 ptrn e PRBS15_6 8 ptrn e PRBS15_7 8 ptrn e PRBS23_1 8 ptrn e PRBS23_2 8 ptrn Agilent Serial BERT N4901 2 User s Guide January 2006 31 32 TIP NOTE SDH CID SDH STM4 SDH STM16 PRBS23_1 2 ptrn PRBS23_6 8 ptrn PRBS23_7 8 ptrn PRBS7 1_1 2 ptrn The file names indicate how the patterns are configured For example PRBS23_6 8 ptrn has a pattern length of 2 23 8 388 608 bits and a mark density of 3 4 6 8 or 75 percent PRBS7 1_1 2 is an emulation of a hardware generated PRBS 2 7 pattern with a mark density of 50 percent This directory contains the following pattern SDH_STM 16_CID ptrn Pattern is a SDH STM 16 rate Consecutive Identical Digit CID pattern same length as a SDH frame with nine bytes of all 1s and nine bytes of all Os separated by random data This implementation uses one SDH frame to carry the CID pattern as follows A1 A2 and C1 bytes as per a normal frame followed by a block of bytes set to all 1s then the normal scrambler sequence 2 7 PRBS starting with 1111 111 Half way through the frame at row 5 column 40 an A1 A2 and C1 seq
67. s Guide January 2006 The following describes how changing the Gate In start and end points can affect the evaluation of bursts Gate In Start Starting too soon If the Gate In signal is applied too soon the error detector will either not be able to recover the clock or synchronize the pattern and will thus mark the burst as a bad burst thus increasing the bad burst count You can typically reduce the bad burst count by starting the gating period later CDR Settling Time Start too soon Synchronization Time Data Input Gate Input Settling Synchronization fail here 4 Attempted Settling Time ends here Starting too late If the Gate In signal is applied too late the Bit Count Time will be shortened and thus the burst sync ratio will be low which forces the measurement of more bursts to obtain the necessary level of reliability CDR Settling Time Start too late Synchronization Time Data Input End Margin o gt as ees Gate In End Ending too soon If the Gate In is deactivated too soon the same problems will occur as with starting the gate too late see previous point CDR Settling Time End too soon Synchronization Time Data Input Bit Count Time Margin ue TL Agilent Serial BERT N4901 2 User s Guide January 2006 137 Ending too late If the Gate In signal is deactivated too late the last received bits will not be part of the burst Thus the BER will increase If the BER
68. set threshold voltage must of course be within the input range Agilent Serial BERT N4901 2 User s Guide January 2006 107 The difference between the data path and the CDR path is that the comparator of the CDR is always single ended Thus this comparator always needs a threshold voltage that lies between the high and low levels of the incoming signal The differential threshold of the data path comparator has no relation to the single ended threshold of the CDR path comparator This means that in differential mode the two thresholds will be different and in single ended mode either normal and complement they will can be equal except during measurements The following figure shows a simplified block diagram It does not reflect the different input modes especially the differential case but it matches both single ended cases Comparator gt gt CDR Threshold Voltage ell Input Low Pass L_ DC tracking on off recovered Connector Filter switches Clock Threshold oa Voltage Data Processing Comparator gt P Bit Rate Range The Agilent Serial BERT N4901 2 provides bit rates from 150 Mbit s up to 13 5 Gbit s However several specific properties and limitations need to be taken into account when working at low bit rates The limitations apply to the instrument according to the following hysteresis curve Ifthe bit rate falls bel
69. testing Remote Error Indication REI alarm SDH_STM 16_MS AIS_alt ptrn Pattern for testing Alarm Indicator Signals AIS alarm SDH_STM 16_MS BIP_alt ptrn Pattern for testing frames with erred parity check sums B2 error SDH_STM 16_MS RDI_alt ptrn Pattern for testing Remote Defect Indication RDI alarm SDH_STM 16_OOF_alt ptrn Pattern for testing Out of Frame OOF alarm SDH_STM 16_RS BIP_alt ptrn Pattern for testing frames with erred parity check sums B1 error SDH STM64 SDH STM 64 ptrn A standard SDH STM 64 frame and payload Total pattern length of 1 244 160 bits Agilent Serial BERT N4901 2 User s Guide January 2006 33 SONETCID This directory contains the following pattern e SONET_STS 48_CID ptrn Pattern is a SONET STS 48 rate Consecutive Identical Digit CID pattern same length as a SONET frame with nine bytes of all 1s and nine bytes of all Os separated by random data This implementation uses one SONET frame to carry the CID pattern as follows Al A2 and C1 bytes as per a normal frame followed by a block of bytes set to all 1s then the normal scrambler sequence 2 7 PRBS starting with 1111 111 Half way through the frame at row 5 column 40 an A1 A2 and C1 sequence is inserted again followed by a block of bytes set to all Os then the normal scrambler sequence starting with 1111 111 again The length is 148 992 bits STS 48 corresponding to a time constant of about 4 msec SONET 0C
70. testing frames with erred parity check sums B3 error SONET_STS 48_P FEBE_alt ptrn Pattern for testing P FEBE alarm SONET_STS 48_S BIP_alt ptrn Pattern for testing frames with erred parity check sums B1 error For more information on the recommeded pattern for specific purposes see When to Use Which Pattern on page 22 User Patterns Procedures You can do the following tasks in the Edit Pattern window Creating New Patterns on page 36 Opening Existing Patterns on page 36 Editing Patterns on page 37 Saving Patterns on page 37 Loading Patterns to the Pattern Generator and Error Detector on page 38 Loading Patterns Directly from the Pattern Editor on page 39 Agilent Serial BERT N4901 2 User s Guide January 2006 35 Creating New Patterns To start a new pattern 1 Press the Edit Pattern on File icon in the Pattern menu A 2 In the toolbar click the New icon k This opens the Pattern Properties dialog box where you can Enter a Description for the pattern Specify the pattern s Length in Bits Specify the pattern s Pattern Type Standard or Alternate 3 After you have entered these parameters click OK The pattern is then opened in the Edit Pattern window 4 You can now continue with modifying the pattern according to your needs Opening Existing Patterns To open an existing user pattern 1 Press the Edit Pattern on File icon in the
71. the Serial BERT will automatically convert all entries to time values using the current unit interval So the UI is handy if you want to set values independently from the system frequency All voltage related entries understand 23 mV 0 01 V and so on All power related values can be entered as 50 mW 0 04 W 6 dBm and so on All dimensionless quantities understand decimal notations 10000000 0 0003 for example and scientific engineering notation 1e9 1 7e 3 for example Marker Definitions Markers are available for all measurements except of the Fast Eye Mask They make the analysis and tracing of the results more comfortable Markers can be switched on or off at any time when results are available in the graphical view For the DUT Output Timing Jitter measurement the software provides two types of markers 156 Agilent Serial BERT N4901 2 User s Guide January 2006 Linear Markers The following illustration shows the definitions for the markers and the values that can be derived from the marker position 4 00e 0 1 Trace 39 Points Ul 401 878 ps All Errors 1 00e 1 BER1 t 1 00e 2 1 00e 3 Markers Time BER 4 00e 4 Upper Right 0 443 Ul 2 6616 2 Lower Left 0 473 Ul 4 145e 7 1 00e 5 Delta 0 915 Ul 2 661 8 2 1 00e 6 BER2 t 1 00e 7 1 00e 8 Logarithmic 0 62 UI 0 37 UI 0 12 Ul 012Ul 0 37 Ul 0 62 Ul Relative 0 50 UI 0 25 UI 0 00 UI 0 25 Ul 0 50 Ul gt gt X1 X2 The following values are displayed for the
72. the error detector reference pattern e What logic family is your device compatible with The pattern generator has preset voltage levels to work with ECL SCFL LVPECL LVDS and CML Be sure to select the correct logic family See Understanding how the Serial BERT Uses Logic Families and Terminations on page 68 10 Agilent Serial BERT N4901 2 User s Guide January 2006 e What is the input termination voltage of your device Make sure it is the same voltage as displayed in the pattern generator termination list e What termination voltage is the output of your device designed to work with Make sure it is set for the error detector s data port termination See Data In Port Termination on page 103 e What bit rate do you need to test your device at The Serial BERT allows you to run tests from 150 Mb s up to 13 5 Gb s depending on the options with which you bought the instrument See Error Detector Bit Rate on page 111 for more information e Does your device require differential inputs As complementary outputs Data Clock Out and Data Clock Out may satisfy this requirement See Diagram 2 Connections for Differential Inputs on page 14 e Is the output of your device noisy or very small in amplitude You may need to manually set up a sampling point See When Should You Adjust the Sampling Point Manually on page 117 e Do you want to check the BER in bursts for example for loop testi
73. to a continuous spectrum the more the test signal looks like an analog noise source The benefit of this is that such a stimulus shows any frequency dependent effects in a device under test For example a clock recovery circuit designed to lock onto a single frequency may have a tendency to lock onto harmonics A long PRBS may uncover such undesirable behavior When to Use Which Pattern Selecting the appropriate pattern and test setup for your application is important Below is a list of recommendations that may be helpful e To simulate random traffic choose PRBS e For functional and alarm testing choose alternating memory patterns e If the test pattern requires a header choose one of the Example Patterns or use a custom pattern To stress lightwave transmitters and receivers vary the mark density of the pattern with a modulator or e o converter To induce baseline wander for margin testing vary the mark density and use alternating patterns e For pattern dependency testing choose one of the Example Patterns or use a custom user pattern vary the mark density e For eye diagram measurements and mask analysis use the Serial BERT with a digital communication analyzer or oscilloscope e For receiver sensitivity or eye contour measurements use the Serial BERT with a digital communication analyzer or oscilloscope e For a clock recovery circuit stress test vary the mark density or use CID Consecutive Identical D
74. to stress specific characteristics of your device For more information see Example Patterns on page 29 3 Ifyou want to select a different pattern for the error detector clear the checkbox Error Detector Pattern tracks the Pattern Generator Pattern The Error Detector Pattern tab appears Switch to this tab and repeat step 2 4 Click OK to finish The selected patterns are immediately loaded to the pattern generator and error detector The types of patterns that are currently loaded to the pattern generator and the error detector are always displayed in the lower left corner of the screen 38 Agilent Serial BERT N4901 2 User s Guide January 2006 Loading Patterns Directly from the Pattern Editor After you have finished editing a pattern in the pattern editor you can load this pattern directly to the pattern generator and error detector NOTE Keep in mind that your modifications in the Pattern Editor do not affect the pattern at the pattern generator or error detector until you download the pattern to the pattern generator error detector 1 First determine whether the pattern generator and error detector should use the same pattern Press the Pattern Select icon in the Pattern menu The PatternSelectForm dialog box opens 2 Select Error Detector Pattern tracks the Pattern Generator Pattern if you want both modules to receive the same pattern 3 Click OK to close this dialog box 4 The pattern editor now displays the f
75. used to test low frequency random jitter RJ and PLL tracking errors The pattern is composed of five logic highs followed by five logic lows This is repeated continuously and disparity rules are followed 11111 00000 11111 00000 binary e K28 5 ptrn This pattern is used to test Fiber Channel and Gigabit Ethernet elements It tests random jitter RJ and deterministic jitter DJ The pattern is composed of 10 bits followed by the complement of those 10 bits It contains isolated characters preceded by runs of the opposite character Disparity rules are followed 11 00 00 01 01 00 11 11 10 10 binary e JTPAT ptrn This is a Fiber Channel jitter tolerance pattern used to test receiver CDR to large instantaneous phase jumps 30 Agilent Serial BERT N4901 2 User s Guide January 2006 e D21 5 ptrn This pattern is used to test Fiber Channel and Gigabit Ethernet elements It tests random jitter RJ and the symmetry of logic transitions The pattern is composed of alternating logic highs and lows and is repeated continuously Disparity rules are followed 1010101010 1010101010 binary e CSPAT ptrn This Fiber Channel supply noise SPAT pattern is packaged into a Fiber Channel frame for use in system level tests e CRPAT ptrn This Fiber Channel Compliant RPAT pattern for jitter generation is packaged into a Fiber Channel frame e CJTPAT ptrn This Fiber Channel Compliant JTPAT pattern for jitter tolerance is
76. your device might receive in the real world Different patterns present different data loads to your device which can cause variations in the bit error ratio A bit pattern is sent from the pattern generator to your device At the same time the expected output pattern of your device is internally generated in the error detector to provide a reference Selecting a pattern is the first step in setting up a BER measurement The Serial BERT provides various patterns to fulfill most standard testing needs How does the Serial BERT Generate Patterns The Serial BERT has two possibilities for generating patterns either by directly generating them from the hardware or by loading patterns into the memory and generating the output according to the pattern in the memory These memory based patterns can be either user patterns file based or software generated patterns Hardware Generated Patterns The hardware generated patterns available with the Serial BERT are PRBS 2 n 1 patterns They provide long high speed random traffic patterns that can be used for testing a wide range of devices These patterns cannot be manipulated in any way Memory Based Patterns The memory based approach is more flexible It makes it possible to define any pattern within the physical limits of the memory for defining the output The Serial BERT provides two possibilities for defining the output by defining user patterns or by selecting a software generated patt
77. 0 0 mv 900 0 mv 1410Y oOOOOOOOO gt aav Zoom in A2V ASV A4 A5 ABV ATY 18V Linear The Q factor calculations are disabled and flagged as not applicable if the R value falls below 0 75 The R value can also be seen as an indicator of how well the noise distribution fits to Gaussian shape It will not fit for example if the received signal is dominated by cross talk or modal noise Agilent Serial BERT N4901 2 User s Guide January 2006 Eye Opening Advanced Analysis Eye Opening The Eye Opening measurement generates a three dimensional bit error rate BER diagram as a function of the sample delay and the sample threshold With this measurement the complete eye of the DUT output signal is measured The results comprise the voltage and timing of the eye opening and the optimum sampling point To get the result the sampling delay and the input threshold of the signal are shifted within 1 5 signal periods and a user defined voltage range The results comprise the voltage and timing of the eye opening and the optimum sampling point An example that illustrates the general principle of the Eye Opening measurement is shown in the figure below Threshold Optimum sampling point max 50 min 100 0 Start 1 5 Periods In detail the measurement is not that simple and more precise as it derives the contour of the eye from the bit error
78. 0 00 ps 190 00 ps 170 00 ps 150 00 ps 240 00 ps 220 00 ps 200 00 ps 180 00 ps 160 00 ps 140 00 ps Estimated Total Jitter The Estimated Total Jitter TJ allows you to predict the jitter expected for very low bit error rates that would take a long time to measure It is obtained by extrapolating the measured BER curves The TJ is estimated by extending the BER curves based on the points detected between the BER Threshold and the Minimum BER for RJ DJ Separation to the Residual BER for RJ DJ Separation level The estimated TJ is the period minus the width of the measured eye Agilent Serial BERT N4901 2 User s Guide January 2006 165 NOTE The following graphic is not available from the measurement software It is inserted here to show how the TJ period is calculated Extrapolated Bathtub log scale 10 I T T T T T I i 10 L f 4 3 BER Threshold 10 j 10 J 2 Min BER for RJ DJ amp 10 f Separation z i i 10 l J 10L i Residual BER for i Estimated TJ 10 J 10 1 Li 1 1 1 i 1 L J 0 8 0 6 0 4 0 2 0 2 0 4 0 6 as o delay adjust ui Explanation of the Fast Total Jitter Measurement The Fast Total Jitter measurement is an optimized method to determine the total jitter for devices that generate a very low error density BER well below 107 To measure not estimate the total jitter for a device with a BER of 10
79. 0e 2 1 00e 3 1 00e 4 1 00e 5 1006 6 1 00e 7 1 00e 8 Logarithmic 0 62 UI 0 37 UI 0 12UI 012UI 0 37 Ul 0 62 UI Relative 0 50 Ul 0 25 Ul 0 00 Ul 0 25 Ul 0 50 Ul Total Jitter Terminal View Tab The graph shows either the bathtub curve or the jitter distribution vs time The BER graph the bathtub shows the BER vs sample delay 1 00e 0 1 Trace 160 Points Ul 333 333 ps All Errors 1 00e 1 1 00e 2 1 00e 3 BER Threshold 1 0004 1008 4 1 00e 5 1 00e 6 1 00e 7 1 00e 8 1 00e 9 1 00e 10 Logarithmic 210 00 ps _ 150 00 i Ps Ps 210 00 BS Relative 180 00 ps 120 00ps 60 00 ps 0 00s 60 00ps 120 00 ps 80 00 ps 40 00 ps 90 00ps___ 30 00ps 30 00ps__ 90 00ps_ __ 150 00 The BER graph can be viewed in either linear and logarithmic view The logarithmic view is shown above The DUT Output Timing Jitter measurement calculates the jitter histogram as the absolute of the derivative of the measured bit error rate jitter dBER dt Agilent Serial BERT N4901 2 User s Guide January 2006 1 Trace 79 Points Ul 333 333 ps All Errors 2 60e 10 2 40e 10 2 20e 10 2 00e 10 1 80e 10 1 60e 10 1 40e 10 1 20e 10 1 00e 10 8 00e 9 6 00e 9 4 00e 9 2 00e 9 Linear 2 p Relative ET rE 30 00 210 00 p 190 00 p 170 00 p 150 00 p S S S S E S 220 00 ps 200 00 ps 180 00 ps 160 00 ps 140 00 ps NOTE Because the right hand slope of the tub does not provide additional informati
80. 0e 5 1 000e 3 1 000e 1 1 000e 8 1 000e 6 1 000e 4 1 000e 2 1 000e If Edge Resolution Optimization is enabled the software first scans the threshold range in relatively coarse steps and determines the areas where the signal produces no edges that means where the BER is nearly constant Then the algorithm uses the specified step width for measuring the regions where the BER changes and a much lower resolution about one tenth for measuring the regions where the BER is constant This reduces the number of data points to 145 as shown in the figure below 800 0 my Trace 145 Points 900 0 mY A0V AAV 12V ABV 14V ASV ABV AT A8V Logarithmic 1 000e 7 1 000e 5 1 000e 3 1 000e 1 1 000e 8 1 000e 6 1 000e 4 1 000e 2 1 000e As you can see from the figures above Edge Resolution Optimization does not sacrifice the precision of the measurement But it can reduce the measurement duration considerably especially if you compare a large number of bits at a low data rate Agilent Serial BERT N4901 2 User s Guide January 2006 193 Pass Fail Tab The Pass Fail tab of the Properties dialog box allows you to specify the criteria to decide whether the DUT passes or fails the test You can change pass fail criteria without rerunning a test The software only uses the criteria to rate the results of a measurement NOTE The pass fail criteria do not control measurement execution The measurement run will be c
81. 0e 9 1_Trace 84 Points Ul 8 000 ns All Errors 1 00e 1 1 00e 2 1 00e 3 1 00e 4 1 00e 5 1 00e 6 1 00e 7 Logarithmic 0 63 UI 0 38 UI 0 13 Ul 0413Ul 0 38 UI 0 63 Ul Relative 0 50 UI 0 25 Ul 0 00 Ul 0 25 Ul 0 50 UI As the bathtub borders are not uniform both have two edges the linear derivative the jitter will show two peaks 4 00e 10 1Trace 43 Points Ul 8 000 ns All Errors 1 00e 9 1 00e 8 1 00e 7 1 00e 6 Reeve oru CY ossu go OF ossu C5 osou OA Agilent Serial BERT N4901 2 User s Guide January 2006 159 Output Levels measurement If you switch to linear scale and enable the marker you can see its bell shape 1 Trace 43 Points Ul 8 000 ns All Errors ais Gaussian Marker Value cad Mu 0 551 UI 3 50e 9 Sigma 0 013 Ul 2 80e 9 Kappa 1 410e 0 2 10e 9 1 40e 9 7 00e 8 Linear 0 68 UI 0 63 UI 0 57 Ul 0 53 UI 0 47 Ul Relative 0 70 UI 0 65 UI 0 60 UI 0 55 UI 0 50 UI You can measure the random jitter distribution of each peak as well as the distance between the peaks which means the deterministic jitter You can also use the marker with logarithmic scale In this case it appears as a parabolic curve 1 00e 10 1Trace 43 Points Ul 8 000 ns All Errors 1 00e 9 Gaussian Marker Value Mu 0 528 Ul 10088 Sigma 0 003 UI 1 00e 7 Kappa 3 394e 1 1 00e 6 1 00e 5 Logarithmic 0 68 UI 0 63 Ul 0 57 UI 0 53 UI 0 47 UI Relative 0 70 UI 0 65 UI 0 60 UI 0 55 UI 0 50 UI A Gaus
82. 12 SONET OC 12 A standard SONET OC 12 STS 12 frame and payload Total pattern length of 77 760 bits SONET 0C48 All patterns are SONET OC 48 STS 48 frame standard pattern scrambler ON and 1010 payload e SONET_STS 48 ptrn Pattern contains no errors or alarms e SONET_STS 48_L AIS ptrn Pattern for testing Alarm Indicator Signals AIS alarm e SONET_STS 48_L BIP ptrn Pattern for testing frames with erred parity check sums B2 error e SONET_STS 48_L FERF ptrn Pattern for testing L FERF alarm SONET_STS 48_OOF ptrn Pattern for testing Out of Frame OOF alarm SONET_STS 48_P BIP ptrn Pattern for testing frames with erred parity check sums B3 error SONET_STS 48_P FEBE ptrn Pattern for testing P FEBE alarm SONET_STS 48_S BIP ptrn Pattern for testing frames with erred parity check sums B1 error 34 Agilent Serial BERT N4901 2 User s Guide January 2006 SONET OC48 alternate All patterns are alternate Pattern A is an error and alarm free SONET OC 48 STS 48 frame scrambler ON and 1010 payload Pattern B has either an error or alarm as per below SONET_STS 48_L AIS_alt ptrn Pattern for testing Alarm Indicator Signals AIS alarm SONET_STS 48_L BIP_alt ptrn Pattern for testing frames with erred parity check sums B2 error SONET_STS 48_L FERF_alt ptrn Pattern for testing L FERF alarm SONET_STS 48_OOF_alt ptrn Pattern for testing Out of Frame OOF alarm SONET_STS 48_P BIP_alt ptrn Pattern for
83. 2 400e 1 2 800e 1 3 200e 1 This graph shows the absolute values of the derivative of the bit error rates over the thresholds ABER dThreshold It visualizes the data that forms the basis for the calculations of the level and noise values The graph provides a special marker that allows you to estimate the data distribution by approximating it by means of a Gaussian normal distribution The Output Levels measurement provides a third graphical display the Q from BER versus Threshold graph This graph refers to the Q factor calculations For details on these calculations see QBER vs Threshold Graph on page 196 and Understanding the Q Factor Results on page 204 Output Levels Reference The Output Levels measurement returns the results in a graphical and in a numerical form The following sections provide explanations of the measured parameters and the display options that are specific to this measurement Additionally some information is provided to explain the theoretical background behind The properties of the Output Levels measurement can be specified on the various tabs of the Properties dialog box Agilent Serial BERT N4901 2 User s Guide January 2006 191 Parameters Tab NOTE Ifyou modify the parameters on this page you have to rerun the measurement to update the results Set the criteria for moving to the next sample point e Number of Compared Bits After this number of compared bits the measurement s
84. 4 56 Ul 4 80 Ul 2 04 Ul 2 28 Ul 2 52 Ul 2 76 Ul Absolute 1 68 Ul 1 92 Ul 246 Ul 2 40 Ul 2 64 Ul 2 88 Ul The following figure shows a typical jitter histogram with two peaks indicating the presence of random and deterministic jitter 1 Trace 79 Points Ul 333 333 ps All Errors 2 60e 10 2 40e 10 2 20e 10 2 00e 10 1 80e 10 1 60e 10 1 40e 10 1 20e 10 1 00e 10 8 00e 9 6 00e 9 4 00e 9 2 00e 9 Linear p Relative anne pee oo 210 00 p 190 00 PS a pg17020PS 150 00 p S S s 220 00 ps 200 00 ps 180 00 ps 160 00 ps 140 00 ps Agilent Serial BERT N4901 2 User s Guide January 2006 163 For a detailed explanation of the Fast Total Jitter measurement results see Explanation of the Fast Total Jitter Measurement Results on page 182 Understanding the Jitter Calculation After the output timing behavior of the DUT is measured the Serial BERT calculates the different jitter components Random and Deterministic Jitter To understand the RJ and DJ results it is helpful to first understand how the software generates the results 1 The bathtub curve is measured Bathtub log scale Bathtub inear scale 10 r r 3 0 5 mrn G r Foal l 10 L J 4 0 43 f 3 2 t H 04 10 f 0 354 10 i f o3 4 0p poe E f gez ss 810 O2r 10 3 015 10 f 01 10 0 05 t l 10 n n 1 4 E E E EEE ENESES 4 d 06 0 4 2 02 o4 06 EE 08 06 ET 0 2 o
85. 6 The u o and 0 9 values are calculated from a selected range of data points This calculation is correct if the noise distribution has Gaussian characteristics Then the bit error rate can be expressed as l 0 D BER D 5 ere ea 5 erfc where D is the decision threshold u 9 and 0 9 are the mean and standard deviation of the 1 and 0 rails and erfc x is the complementary error function This formula is the sum of two terms It considers the probabilities of deciding that a 0 has been received when a 1 was sent and that a 1 has been received when a 0 was sent For the following calculations the assumption is made that the BER is dominated by only one of the terms noted above depending on whether the threshold is closer to the 1 or 0 rail For the complementary error function 1 x 2 e 1 7 8 12 erfc x T e LTT an inverse logarithmic approximation exists 1 Log gt aren j 1 192 0 6681x 0 0162 x2 where x Log BER This function applied to the high level and low level data points yields new threshold vs value combinations In the area of low BER typically below 1075 these new data pairs should fit to two straight lines although a couple of assumptions and approximations have been made Agilent Serial BERT N4901 2 User s Guide January 2006 205 To determine the gradient and offset of these lines a linear regression is performed This is illus
86. 6 Points 800 0 mv 900 0 mV 10V AAV 32y 13v 14y ASV 1 Q factor I ABY ATV ABV Linear This graph illustrates the calculation of the Q factor 65 in the figure above as the best fit line through the calculated points This method makes the calculation reproducible For details see Q factor Results on page 201 Furthermore the following parameters can be set on the View tab of the Properties dialog box Calculate You can calculate measurement parameters for BER Threshold This is the bit error rate threshold at which the Threshold Margin is determined It is also the upper threshold for the Q factor calculations The BER Threshold is displayed in the BER vs Threshold graph There it can be positioned with the mouse or your finger if you are working directly on the Serial BERT Min BER for Q This is the lower threshold for the Q factor calculations Table Number Format You can select the number of Decimal Places to be displayed in the table Graph Tab On the Graph tab you can use the several options to optimize the graphical display according to your needs Scale Choose between Logarithmic and Linear to select the scale for the display s x axis The scale of the QBER vs Threshold graph is always linear Ratio Unit Ratios in the numerical results such as SNR can be displayed linearly or logarithmically in dB Agilent Serial BERT N4901 2 User s Guide January 2006 197 Mar
87. 8 mv 24 085 mV new results 5 To distinguish the two graphs change the color of one graph by tipping the red Color field You can now directly compare the graphical and numerical results of the two Output Levels measurements How to Rename Copied Measurements To rename a copied measurement for easy distinction 1 Highlight the copied measurement 2 Press Enter or choose Rename from the context menu 3 Change the name and press Enter How to Load Measurements To load previously saved measurements 1 Select Load Measurement from the Analysis menu 2 Use the file dialog box to locate the desired mcp file and press Open to load the measurement The respective screen of the Analysis panel opens and you can view the measurement Agilent Serial BERT N4901 2 User s Guide January 2006 151 How to Change Measurement Properties and the Graphical Display The various tabs in the Properties dialog box allow you to specify the parameters for the measurement The available options are not the same for all measurements See the Reference information of the respective measurement type for details To modify the measurement properties 1 Press the Properties button to open the Properties dialog box 2 Use the different tabs in this dialog box to make the required settings Parameters tab The settings made on the Parameters tab define how the data is collected If you make changes here you need to run
88. 93 on how to enable disable the touchscreen Enabling the LAN Connectors You can enable the LAN connection on your Serial BERT This will allow you to connect to network printers and remote servers with access to shared folders and files The Serial BERT has two RJ 45 LAN connectors on the rear panel They support 10 Base T and 100 Base T Ethernet networks using TCP IP and other Microsoft supported networking protocols The Serial BERT uses Microsoft XP Pro For detailed instructions on how to enable the LAN connectors please refer to the Microsoft XP Pro manual CAUTION Because your network settings are unique to your IT infrastructure Agilent Technologies will not be able to assist you with connecting your instrument to your network Please contact your network administrator or IT department for assistance For more information refer to the MS Windows resource kit available from Microsoft that is appropriate for your computer system CAUTION Do not connect the Serial BERT to a network that is configured to automatically install software on network devices Installing or overwriting files on the Serial BERT computer system may impact the operation of the instrument Please contact your network administrator or IT department to find out whether you have this type of network Agilent Serial BERT N4901 2 User s Guide January 2006 291 Configuring the Instrument Reference The Serial BERT has the following utilities for co
89. A AJAJAJAJA 28 Agilent Serial BERT N4901 2 User s Guide January 2006 CAUTION Implications of Using Memory Based Patterns There are several implications to the way Serial BERT handles memory based patterns a few of which are If you set up an alternating pattern where pattern B is a replica of pattern A except for the addition of one error and the pattern is an odd number of bits long the error will be repeated 512 times when B is sent because pattern B is sent 512 times before switching back to A If you want to send a periodic pattern you only have to define a single repetition For example if you want to send alternating 1 s and 0 s you only have to define the first two bits This pattern is automatically loaded 256 times to fill up a complete block of RAM If you set up a trigger to be generated whenever a pattern is sent the trigger is sent whenever all blocks that store this pattern have been sent and not every time the pattern is sent So for example if you have a standard pattern that is 384 bits long this pattern must be loaded into the memory four times to reach the 512 bit boundary A trigger would then be generated every 4 x 384 bits or 1536 bits Example Patterns The Serial BERT provides a set of example user patterns These patterns mimic real data packets and standard stress patterns The di
90. Agilent N4901 Serial BERT User s Guide ae Agilent Technologies Important Notice Agilent Technologies Inc 2006 Revision Revision 3 3 January 2006 Manual Part Number N4901 91030 Printed in Germany Agilent Technologies Herrenberger Stra e 130 D 71034 B blingen Germany Authors t3 medien GmbH Warranty The material contained in this document is provided as is and is subject to being changed without notice in future editions Further to the maximum extent permitted by applicable law Agilent disclaims all warranties either express or implied with regard to this manual and any information contained herein including but not limited to the implied warranties of merchantability and fitness for a particular purpose Agilent shall not be liable for errors or for incidental or consequential damages in connection with the furnishing use or performance of this document or of any information contained herein Should Agilent and the user have a separate written agreement with warranty terms covering the material in this document that conflict with these terms the warranty terms in the separate agreement shall control Technology Licenses The hardware and or software described in this document are furnished under a license and may be used or copied only in accordance with the terms of such license Restricted Rights Legend If software is for use in the performance of a U S Government prime contract or
91. Disable the pattern generator outputs by pressing the OV Disable button in the PG Output Setup screen 2 Usea shielded cable to connect the pattern generator s Data Out port and the error detector s Data In port 3 Terminate all non connected pattern generator output ports with 50 Q 4 Use a shielded cable to connect the pattern generator s Data Out port and the error detector s Data In port 5 Switch to the Pattern panel and press Pattern Select Select an appropriate pattern for example a pure 2 15 1 PRBS for both the pattern generator and the error detector 6 For the pattern generator setup you need to specify the logic levels and the bit rate Select ECL levels and a clock speed of 1250 MHz in this example This corresponds to a clock period of 800 ps See Setting up the Pattern Generator on page 63 for more information 7 Set up the error detector so that the input range and the termination matches the pattern generator s levels Select an Input Range from 2 V to 0 V Set the Data Termination to 2 V Set the Alignment BER Threshold to 1E 6 Set the Clock Setup to Clock Data Recovery to get the error detector s clock from the incoming data stream 8 Enable the pattern generator outputs by pressing the 0V Disable button 232 Agilent Serial BERT N4901 2 User s Guide January 2006 9 Press Sync Now and then Auto Align to find the optimum sampling point Check that the synchronization
92. Divided by n The trigger is set every n bits ai The trigger signal is sent every time when the complete pattern has been received This refers to the single pattern repetitions not the 512 bit blocks Clock Trigger Out Pattern Note that the length of the trigger signal is at least 40 bits and divisible by 4 Aux Out You have two options to set up the signal at the Aux Out port e Clock This option supplies the clock signal to the Aux Out port Data This option supplies the data signal via a comparator to the Aux Out port The comparator is controlled by the 0 1 threshold This lets you use an oscilloscope to determine if the 0 1 threshold is correctly set If the 0 1 threshold is set below or above the data eye the output at Aux Out will be constant high or low respectively 126 Agilent Serial BERT N4901 2 User s Guide January 2006 Hardware Generated Patterns Memory Based Patterns Pattern Synchronization The Serial BERT calculates bit error rates by comparing the received data with the expected data patterns To do this it needs to know where the start of the pattern is located in the data stream Introduction to Pattern Synchronization Pattern synchronization sync refers to aligning the incoming data pattern with the internal reference pattern This is accomplished in one of two ways For 2 n 1 PRBS patterns bits from the incoming data pattern seed the error detector s pattern generat
93. GUI If the measurement is running properly a certain BER should be displayed here In most cases the BER is not a constant value but slightly varying over time If no BER is displayed because there are no errors detected you can manually insert an error by pressing the Error Add button in the top right corner of the screen Error Add j When inserting an error the BER should increase for a short time and then return to the initial value How to Monitor the BER Results There are different ways how you can monitor instantaneous BER values 1 Listen to the BER warning tones See Audio Signals Procedures on page 145 for details 2 View the BER Results window which is described in BER Results Window on page 260 3 View the BER bar and errors indicator See Status Indicators on page 261 for details Agilent Serial BERT N4901 2 User s Guide January 2006 259 How to Analyze the BER Results To analyze the behavior of your device you can do the following modifications to the test setup and monitor the effect on the resulting BER 1 Add errors to the data stream See Error Addition Insertion Procedures on page 88 for details 2 Use alternating patterns to switch between different data streams See Pattern Alternation Procedures on page 93 for details 3 Adjust the parameters of the output port of the pattern generator as described in Input and Output Ports Procedures on page 71
94. Guide January 2006 How to Use the Color Bar The color bar at the right hand side of the diagram shows the assignment of BER thresholds to colors These colors can be changed In addition the color bar provides some useful options for controlling the display Color bar 41 Trace 9510 Points Ul 401 876 ps All Errors 600 0 mv 400 0 mv 200 0 mv Handle for the goy BER range 200 0 mv 400 0 mv 600 0 mv Slider for the BER lt scale gt 0 448 UI 0 149 UI 0 149 UI 0 448 Ul 0 746 Ul threshold Relative 0 597 Ul 0 299 Ul 0 000 UI 0 299 UI 0 597 Ul How to Add or Change Colors If you have a mouse connected to your Serial BERT you can add a color to the color bar or change a color From the context menu of the color bar choose from the following display options Gradient Add color To add an additional color to the color gradient at the cursor position The BER range for this new color will be assigned automatically Update Plots Dynamically Rainbow To display a large variety of colors for the bit error ranges Yellow Blue To display a color gradient from color 1 to color 2 for the bit er Green Blue ror ranges Red Green Black White White Black Agilent Serial BERT N4901 2 User s Guide January 2006 215 How to Change the BER Threshold The red slider indicates the present BER threshold To use another value as threshold Move the
95. J Estimated Total Jitter TJ The peak to peak RMS and mean values are displayed in the result table They can also be displayed graphically Jitter Measurement Parameters on page 180 describes how these components are calculated A dedicated Gaussian marker allows to investigate the jitter graph in detail and to measure the contribution of certain sections or jitter peaks to the overall results 162 Agilent Serial BERT N4901 2 User s Guide January 2006 You can set several signal parameters the values to be displayed and the output format according to your needs Fast Total Jitter Measurement Characteristics This measurement is an alternative to the standard jitter measurement It can be used to measure the total jitter at very low bit error ratios Whereas usual jitter measurements at a BER around 10 can take days due to the huge number of bits that has to be compared for each measurement point the Fast Total Jitter measurement can for example determine the total jitter at a BER of 107 in less than 20 minutes at a data rate of 10 Gbit s For details refer to Explanation of the Fast Total Jitter Measurement on page 166 Example Results The following illustration shows the resulting graph of a typical DUT Output Timing measurement 4 00e 0 1_Trace 55 Points Ul 2 500 ns All Errors 1 00e 1 1 00e 2 1 00e 3 1 00e 4 1 00e 5 1 00e 6 BER Threshold 1 000e 1 00e 7 Logarithmic 1
96. Out port for details see Sequence Trigger on page 84 In the example above the Trigger Out port has been set to Sequence Trigger mode and TrigOn is set for each block as shown by the vertical trigger indicator Edit Button of a Block To change the contents of a block click its Edit button This opens the Set Data Block Parameter dialog Sequence Block Parameters The Set Data Block Parameter dialog allows you to change the contents and the trigger generation of a sequence block Block Contents Choices are None The block may be present in the sequence but is completely ignored PauseO Pure zeros are generated when the block is executed You can change the block length if desired Pause Pure ones are generated when the block is executed You can change the block length if desired PRBS You can choose a PRBS of polynomial 2 n 1 The range of n is 7 10 11 15 23 or 31 You can change the block length if desired Pattern You can open a browser and search for the file that contains the pattern You can search for files with the extentions ptrn txt or dat If the pattern file contains two alternate patterns you can also specify whether the A or B pattern shall be generated see also Agilent Serial BERT N4901 2 User s Guide January 2006 57 Pattern File Specifications on page 25 If the pattern file contains just one pattern your choice is ignored You cannot change the block le
97. QUENCY ED FREQUENC 7 Elapsed Time 187 0 1 2 49E 08 19 0 2 2 49E 08 10 03 2 49E 08 11 0 4 2 49E 08 12 05 2 49E 08 You can view and analyze the following information in the measurement log file 2 49E 03 2 49E 09 2 49E 03 2 49E 08 2 49E 03 2 49E 2 49E 2 49E 2 49E 2 49E Agilent Serial BERT N4901 2 User s Guide January 2006 BIT COUNT ERROR COUNT ERROR RATIO ERRORED 0 COUNT ERRORED 0 RATIO ERRORED 1 COUNT ERRORED 1 RATIO SYNC LOSS PG CLOCK LOSS ED CLOCK LOSS DATA LOSS BIT COUNT Cumulative ERROR COUNT Cumulative ERROR RATIO Cumulative ERRORED 0 COUNT Cumulative ERRORED 0 RATIO Cumulative ERRORED 1 COUNT Cumulative ERRORED 1 RATIO Cumulative ERROR SECONDS Cumulative ERROR FREE SECONDS Cumulative ERROR DECISECONDS Cumulative ERROR FREE DECISECONDS Cumulative SYNC LOSS SECONDS Cumulative BURST STATUS Cumulative BURST SYNC RATIO Cumulative TOTAL BURST COUNT Cumulative BAD BURST COUNT Cumulative Agilent Serial BERT N4901 2 User s Guide January 2006 283 Solving Problems Other Messages Other Messages This section covers additional error messages that may occure when working with the Serial BERT Overheat Protection The Agilent Serial BERT N4901 2 is equipped with an overheat protection function to prevent it from overheat damage Caution Level Ifthe temperature of the error detector or pattern generator exceeds a certain threshold the Serial BERT displays the fo
98. Restore the Operating System on page 288 Note also that there is a recovery image on the hard disk See the Windows Help for instructions on restoring a recovery image Understanding BER Bit error ratio BER is the most fundamental measure of system performance It is the ratio of the number of errored bits to the total number of bits received 6 Agilent Serial BERT N4901 2 User s Guide January 2006 For example 1 error in 1000 bits corresponds to a BER of 1 1000 or 1 x 107 This would be displayed on the instrument as 1E 3 Refer to the following table to familiarize yourself with how BERs are displayed Notice that smaller exponents correspond to higher BERs and larger exponents correspond to lower BERs 1 1 1x10 1E 0 1 10 1x107 1E 1 1 100 1x107 1E 2 1 1000 1x107 1E 3 1 10 000 1x107 1E 4 1 100 000 1x107 1E 5 1 1 000 000 1x 10 8 1E 6 1 10 000 000 1x107 1E 7 Understanding the Serial BERT The Agilent Serial BERT N4901 2 also referred to as a Bit Error Ratio Tester BERT is a powerful instrument that enables you to analyze systems and components in the telecommunication and enterprise communication industries It is an essential tool for designing and troubleshooting communications systems high speed integrated circuits ICs and photonic components Its intuitive operation and leading edge performance will help you quickly verify error performance The Agilent Serial BERT N4901 2 has the necessary
99. Saving the current instrument state stores all the setup parameters for the instrument s hardware and software To save the setup with the current filename and directory path do the following In the File menu click Save and then Save Instrument State The current instrument state will be immediately saved 300 Agilent Serial BERT N4901 2 User s Guide January 2006 To save the setup with a new filename or directory path do the following In the File menu click Save and then Save Instrument State As This opens a dialog box which allows you to create subdirectories rename setup files or overwrite existing setup files NOTE Ensure your instrument state files have the extension btz If you do not specify an extension or use a different extension the instrument may not recognize the file as an instrument state file NOTE The default pathname for user defined instrument states is C N4901A Settings The number of instrument states you can save is limited only by the available space on the internal hard disk or floppy disk Each instrument state uses approximately 6 kB of disk space TIP Remember to always make a backup copy of your files If your original files are damaged or lost you can restore the files from the backup copy Recalling Instrument States Recalling the instrument state restores all the setup parameters for the instrument hardware and software To recall the setup do the following In the
100. T additionally by inserting a longer row of zeros in the data stream The length of the zero row can be freely specified up to the length of the pattern The following figure shows an example where a run of 10 zeroes is inserted into a PRBS 2 7 pattern Run of 0s ZRUN PRBN ZSUB Closing bit Mark Density Patterns The mark density of a pattern is the portion of bits that are ones A pure PRBS pattern contains an equal number of ones and zeros and thus has an average voltage in the center between high and low level Therefore the mark density of such a pattern is 1 2 Some devices may have problems processing data streams with different mark density as this results in other average voltages and thus different electrical conditions The ones mark density can be varied to put the pattern out of balance in a randomly distributed way This may be useful for systems that are AC coupled This helps to check for effects such as baseline wander To test these cases the Serial BERT provides PRBS patterns with the mark densities 1 8 1 4 1 2 3 4 and 7 8 Software Generated PRBS Procedures The recommended way to load the data patterns to the pattern generator and error detector is to use the Pattern Select Form dialog box See Loading Patterns to the Pattern Generator and Error Detector on page 38 for details Agilent Serial BERT N4901 2 User s Guide January 2006 47 Software Generated PRBS Referen
101. Trigger Output in the PG Setup menu a 3 Select the trigger pattern that you want to be generated See Trigger Output Reference on page 82 for descriptions of the available trigger signals NOTE Ifyou have downloaded a user defined sequence of patterns to the pattern generator the pattern related settings are ignored In this case you can either set up the Trigger Out port to generate a subrate of the present clock or to signal the beginning of a sequence block see also Sequence Block Parameters on page 57 for details Trigger Output Reference You can always generate a divided clock signal at the Trigger Out port The alternate trigger signals refer to patterns and are not generated in sequence mode 82 Agilent Serial BERT N4901 2 User s Guide January 2006 To support the generation of a trigger spike at the beginning of a sequence block the Trigger Output can be put into Sequence mode Clock Divided by n Select this option to send a trigger signal from the Trigger Out port at every nth clock pulse Note that the trigger signal itself consists of n 2 bits high followed by n 2 bits low For example Clock divided by 8 works as shown below ae Alternate Pattern Trigger Level Clock Trigger Out NOTE This option applies for alternating patterns only Select this option to set the trigger level high whenever alternate pattern B is sent The trigger output for 1024 bit patterns looks as foll
102. User s Guide January 2006 The detect word on which the error detector attempts to resync is chosen strictly by chance So if there are two instances of the detect word in the pattern the error detector has a 50 chance of selecting the correct one The more instances of the detect word exist in the pattern the higher are the chances for incorrect synchronization The software attempts in any case to identify a 48 bit pattern that occurs as seldom as possible in the pattern For very large patterns this can unfortunately take a very long time and the software ends the search if it expects that it would take longer to find an adequate detect word than it would to attempt to synchronize If the search for a detect word is ended the most unique detect word identified is used Further Considerations Patterns must always be synchronized in order to do accurate BER testing If patterns are out of alignment by just one bit errors can be as high as 50 5E 1 for PRBS patterns and 100 1E 0 for custom patterns By default the error detector is in automatic sync mode with a sync threshold of 1E 3 This setting is recommended for most applications and usually allows the synchronization function to be transparent requiring no attention However for special applications changes can be made to the sync mode and sync threshold What Type of Synchronization Should You Use The type of synchronization you use affects how errors are measured
103. al BERT modifies the output levels Agilent Serial BERT N4901 2 User s Guide January 2006 73 Vof This text field allows manual entry of the voltage level halfway between logic high and logic low the offset and displays the current value To modify the value click inside the text field and either Enter the desired value directly with the numeric keyboard e Use the knob by the numeric keyboard to fine tune the value TIP You can also adjust Vor with the Data Offset knob on the front panel See Understanding the Output Level Parameters on page 66 for information about how the Serial BERT modifies the output levels Vio This text field allows manual entry of the logic low voltage level and displays the current value To modify the value click inside the text field and either Enter the desired value directly with the numeric keyboard e Use the knob by the numeric keyboard to fine tune the value See Understanding the Output Level Parameters on page 66 for information about how the Serial BERT modifies the output levels Vampt This text field allows manual entry of the voltage amplitude and displays the current value To modify the value click inside the text field and either Enter the desired value directly with the numeric keyboard e Use the knob by the numeric keyboard to fine tune the value TIP You can also adjust Vampt With the Data Amplitude knob on the front panel See Understan
104. ally selected Limitations of the Auto Align Function When you start either of the functions Auto Align Data Center or 0 1 THR Center a window opens with a Cancel button These auto search functions are generally working normally when the BER bar is fluctuating If you do not see these signs of progress within a reasonable amount of time you may click Cancel Possible reasons that the Auto Align function does not find the optimum sampling point are The analyzer may have difficulty finding eye edges at the specified alignment BER threshold This could occur when the data eye is very small or there is a large amount of noise or jitter Incorrect synchronization could prevent the analyzer from finding eye edges There could be a problem with your measurement setup 116 Agilent Serial BERT N4901 2 User s Guide January 2006 When Should You Adjust the Sampling Point Manually You may want to manually adjust the sampling point in the following cases To characterize your DUT for example to see how the BER changes as the sampling point is moved To set a more accurate sampling point in data eyes that are very small The manual functions allow for finer adjustments of the sampling point than the Auto Align function does For example the manual 0 1 threshold voltage step size is smaller than the auto align voltage step size You can use the following funtions to manually adjust the sampling point e Avg 0 1 Threshold
105. als the spectral components and their power Several window algorithms are provided to reduce the influence of leakage Agilent Serial BERT N4901 2 User s Guide January 2006 239 Advanced Analysis 240 Spectral Jitter Jitter Distribution Over Time Jitter has a more or less characteristic distribution over time The histogram of pure random jitter shows its Gaussian distribution as illustrated in the following figure Deterministic jitter periodically adds and subtracts a delay to from the received signal Jitter caused by a square wave or on off signal has produced the following histogram Jitter caused by a triangle signal shows an even distribution Agilent Serial BERT N4901 2 User s Guide January 2006 Last but not least the histogram of jitter modulated by a sinusoid exhibits two significant peaks The jitter histogram of a real world signal shows most often a mixture of these characteristic distributions Measurement Results The Spectral Jitter measurement provides e A graphical view of power vs frequency This makes it easy to identify prominent spectral jitter components You can adjust the horizontal and vertical axes according to your needs e You can switch between linear and logarithmic scales e Graphical markers and the zoom function assist you when you are analyzing the graph visually e The numerical results include bit error rate total power and noise power They provi
106. an equal distribution of ones and zeros over time Otherwise the results are hard to predict and may be not reproducible Signal Processing If the error signal is obtained as explained above an analysis in the frequency domain reveals the absense or presence of deterministic jitter Dominant frequency components become visible and their contribution to the total jitter can be measured The modified error signal is subject to a fast Fourier transformation FFT FFT requires that the data record to be processed has a length that is a power of two 2 such as 217 oe 219 220 and so on From that data record the FFT calculates pairs of frequency power values The number of pairs is half the number of samples If you have chosen a record length of 2 which means 131 072 bits or 128 Kbit the result contains 65 536 pairs The maximum frequency is half the data rate used for the test Agilent Serial BERT N4901 2 User s Guide January 2006 243 The results are displayed in the Spectral Jitter measurement graphical and numerical result window About FFT For general information about the Fourier transformation and the special characteristics of the fast Fourier transform please refer to the standard literature For details see the following documents Frederic J Harris On the use of Windows in Harmonic Analysis with the Discrete Fourier Transform Proceedings of the IEEE Vol 66 January 1978 The Fundamentals
107. an only be set for a device that is DC connected CAUTION Selecting the wrong terminations may damage your device The Data In port is connected to a 50 Ohm load impedance or termination within the error detector Data termination refers to the voltage level at the end of this load The logic output from a device requires any connected equipment including the error detector to have a specific termination voltage Input Range Before you can synchronize the error detector to the incoming data stream you need to define the voltage range within which the eye is located Both the high and low level of the data signal must be within this range to find the eye NOTE The input voltage range is 2 V When you modify either the high or low voltage the other voltage is automatically adjusted NOTE See Sampling Point Setup Reference on page 119 for a complete description of all parameters of the Sampling Point Setup window 106 Agilent Serial BERT N4901 2 User s Guide January 2006 Clock Setup To measure the Bit Error Rate with the error detector the bit rate of the data stream must be known Depending on the options the instrument is delivered with you could use either an external clock source for the error detector for example the clock from the pattern generator or extract the clock signal from the incoming data CDR mode CDR mode does not work for all kinds of data patterns For example if the device under test s
108. analyzes the incoming data stream to detect the voltage level that is optimal for the derivation of the clock signal This function is recommended for all applications where a valid data stream is continuously arriving at the error detector For measurements with partly invalid data such as burst tests this function may cause a clock loss In such cases deactivate Auto Threshold and e enter a fixed voltage level manually or e press the Measure button to measure the voltage level of the incoming signal while a burst signal is applied Once the level is defined it remains fixed for the following measurements Note that the entered voltage level has to be within the input range of the error detector If the value lies outside this range the change is rejected and an error message is displayed Agilent Serial BERT N4901 2 User s Guide January 2006 113 Setting up the Error Detector Sampling Point Setup Sampling Point Setup This section provides basic information on the sampling point setup and eye diagrams How Does the Sampling Point Setup Work The sampling point of a data signal is defined by two values a point in time and a voltage level Each bit of the data signal is sampled at this point in time and in reference to this voltage level The point in time in reference to the clock signal is referred to as the data input delay and the voltage level is referred to as the 0 1 decision threshold The location of the
109. are calculated by the Serial BERT software and then loaded to the memory from which the necessary output is generated As such they are closely related to user patterns with the exception that they cannot be edited in the Pattern Editor window When to Test with Software Generated PRBS You can use these patterns for any application where random data and an even number of bits is required In addition you can vary the mark density or zero substitution to test the influence on the BER This allows potential problems like baseline wander for instance to be fully characterized How the Software Generates PRBS When you set up a software generated pattern the Serial BERT calculates the required pattern and loads this into the memory From here the memory is read bit by bit identically as for memory based patterns and the corresponding output is generated See also How the Serial BERT Generates Memory Based Patterns on page 27 Range of Software Generated PRBS Patterns The Serial BERT supports the following 2 n PRBS patterns 7 128 7 7 10 1 024 10 10 11 2 048 11 11 13 8 192 13 13 15 32 768 15 15 23 8 388 608 23 23 46 Agilent Serial BERT N4901 2 User s Guide January 2006 Zero Substitution Patterns A potential risk to bit errors are longer rows of zeros within a data stream The longest run of zeros in a 2 n PRBS consists of n zeros The Zero Substitution function can be used to stress the DU
110. ast Total Jitter measurement can reduce the measurement time by more than a factor of 40 Typical test times are approximately 20 minutes at 10 Gbit s and slightly more than one hour at 2 5 Gbit s for a measurement at the 107 BER threshold with an overall confidence level better than 90 Agilent Serial BERT N4901 2 User s Guide January 2006 169 DUT Output Timing Jitter Procedures This section shows how to set up and use the DUT Output Timing Jitter measurement As an example we measure the output timing and jitter behavior of a shielded cable This requires the following steps e Preparing the measurement see How to Prepare the DUT Output Timing Jitter Measurement on page 170 Executing the measurement see How to Execute the DUT Output Timing Jitter Measurement on page 171 e Optimizing the view of the results see How to Optimize the View of the Results on page 172 How to Prepare the DUT Output Timing Jitter Measurement To use the DUT Output Timing Jitter measurement to test a shielded cable 1 Disable the pattern generator outputs by pressing the OV Disable button in the PG Output Setup screen 2 Use a shielded cable to connect the pattern generator s Data Out port and the error detector s Data In port 3 Terminate all non connected pattern generator output ports with 50 Q 4 Use a shielded cable to connect the pattern generator s Data Out port and the error detector s D
111. ata In port 5 Switch to the Pattern menu and press Pattern Select Select an appropriate pattern for example a pure 215 1 PRBS for both the pattern generator and the error detector 6 For the pattern generator setup you need to specify the logic levels and the bit rate Select LVPECL levels and Gb Ethernet 1 06250 Gb s as clock speed See Setting up the Pattern Generator on page 63 for more information 7 Set up the error detector so that the input range and the termination matches the pattern generator s levels Select an Input Range from 1 V to 3 V Set the Data Termination to 1 3 V Set the Clock Setup to Clock Data Recovery to get the error detector s clock from the incoming data stream 170 Agilent Serial BERT N4901 2 User s Guide January 2006 8 Enable the pattern generator outputs by pressing the OV Disable button 9 Press Sync Now and then Auto Align to find the optimum sampling point Check that the synchronization and the alignment were successful None of the error indicators at the top of the user interface should show red and the resulting BER should be zero How to Execute the DUT Output Timing Jitter Measurement To run the DUT Output Timing Jitter measurement 1 Switch to the Analysis area If the Output Timing screen is not yet displayed press the Output Timing icon 2 Press the Start button to execute the measurement The measurement software runs the measurement and displays t
112. ath and name The available file formats to save the file in are e Pattern File ptrn in binary 8 bits byte hex 4 bits byte or 01 1 bit byte format e ASCII Pattern Files txt See Pattern File Specifications on page 25 for descriptions of these file types If you want to save the pattern file in one of the 12 memory slots use the file names from UPAT1 ptrn through UPAT12 ptrn respectively Agilent Serial BERT N4901 2 User s Guide January 2006 3 Loading Patterns to the Pattern Generator and Error Detector The recommended way to load the data patterns to the pattern generator and error detector is to use the Pattern Select Form dialog box 1 Press the Pattern Select icon in the Pattern menu The Pattern Select Form dialog box opens 2 Select one of the following pattern types for the pattern generator and follow the next required action as listed below User Pattern Browse for the desired pattern file Memory User Pattern Select memory slot 1 12 2 n 1 PRBS Select the pattern size 2 n PRBS Select the pattern size Mark Density Select the pattern size and the mark density ratio Zero Sub Select the zero substitution rate Sequence Enable this checkbox if you intend to download a user defined sequence Patterns generated with in a sequence are defined in the Sequence Editor NOTE Your instrument also contains user patterns that mimic real data packets and are designed
113. ation of the Numerical Results The measurement provides numerical results for The output timing measurement parameters are defined in the Output Timing Measurement Parameters following table A B 2 Optimal Sample The average of the left A and right B Point Delay bathtub BER threshold intersections Phase Margin The period of time where the bit error rate is B A lower than the BER threshold The A and B values are the left and right intersections of the bathtub curves with the BER threshold Obviously all values change if the BER threshold is modified The following illustration shows an example for a measurement BER linear log scale Optimum Sampling Point This graphical view is not available in the measurement software NOTE The jitter measurement parameters are defined in the following list Jitter Measurement Parameters All jitter measurement parameters except the Total Jitter RMS and NOTE Total Jitter Mean change with the BER threshold Total Jitter RMS The average of the left and right jitter histogram root mean squared values 180 Agilent Serial BERT N4901 2 User s Guide January 2006 Total Jitter Peak to Peak Peak to peak value for total jitter Calculated as the pulse period unit interval minus the Phase Margin Jitter Mean Mean value for total jitter Calculated as the weighted average of the left edge jitter histogram Random Jitter RMS The total jitt
114. ay s x axis Markers To analyze the graphs at a particular point you can use the markers Additionally you can display all related values for the markers in the marker readout Pressing the Reset Markers button will set the markers back to the default positions 150 0 my 1 Trace 13689 Po Ul 1 538 ns Tig y 1 000e 0 i 3 120 0 mY FENS Markers Time Voltage BER 90 0 mv En Upper Right 0 772 UI 90 0 mY 2 714e 1 F gt m Lower Left 0 113 UI 90 0 mY 2 500 1 7777 60 0 mV Detta 0885U 180 0m 24140e 2 T 1 A 30 0 mY 0 0 YV 30 0 mY 60 0 my Zoom Several zoom factors are available When you show the zoom graph you can also allow the zoom graph to track the mouse or your finger if you are working directly on the Serial BERT Show Measured Points If you want to see the points that have actually been measured choose Show Measured Points The MUI uses linear interpolation for the graph Show BER Threshold If you want to display the BER threshold contour line choose Show BER Threshold The BER threshold is set on the View tab or by the slider in the Color Bar NOTE If no slider for the BER threshold is displayed make sure that you have enabled the BER Threshold option on the View tab and that a suitable BER is entered 222 Agilent Serial BERT N4901 2 User s Guide January 2006 Explanation of the Numerical Results Additionally to the graphical results the measurement provides numerical resul
115. be a time consuming procedure With the Fast Eye Mask measurement of the Serial BERT pass fail information can be obtained within seconds This is achieved by measuring the bit error rate at a limited number of test points Up to 32 measurement points can be specified each defined by a sampling time relative to the actual sampling point which can be the optimum sampling point and a threshold voltage which is adaptive In practice six measurement points will often suffice to approximate the shape of the eye Six measurement points are preset by default The Fast Eye Mask can be run by using the appropriate SCPI commands making it possible to integrate the Serial BERT into a testing environment See the Programming Guide for details Example Results The Fast Eye Mask measurement measures the bit error rate at certain measurement points These points have to be positioned inside the expected eye opening Threshold voltage Measurement points A1 Time Actual optimum sampling point The results of a Fast Eye Mask measurement are displayed in numerical form Agilent Serial BERT N4901 2 User s Guide January 2006 Terminal Relative Time 0 4 Ul 0 4Ul 0 16 UI 0 16 Ul Voltage abs 224m 224m 178 mv 222 mY The rows Relative Time and Voltage define the positions of the measurement points The last row shows the measured bit error rates at these points The Relative Time of the mea
116. by a Gaussian normal distribution With a Gaussian marker you can measure the mean and standard deviation of the normal distribution see also The Gaussian Marker on page 157 and Graph Tab on page 197 QBER vs Threshold Graph This graph shows the extrapolation of the optimum Q factor and the optimum threshold level from a limited number of measured points The measured data points to be used for the calculation have to be within a contiguous BER range This range is defined by specifying the Min BER for Q lower threshold and the BER Threshold upper threshold Both thresholds can be set in the lower section of the View tab The Q factor can only be calculated if for both high and low lever rails two or more points fall within the defined BER range For reliable results use at least five measured points To ensure proper settings activate the Show Measured Points function on the Graph tab and switch to the BER vs Threshold graph Move the upper BER threshold marker vertical line so that a sufficient number of measured points is included for the calculation 800 0 my Trace 501 Points 900 0 mv AOV AAV 12V 13Y 14V ASV ABY ATV BER Threshold 2 194e 9 18v Logarithmic 1 000e 7 1 000e 5 1 000e 3 1 000e 1 1 000e 6 10008 4 1 000e 2 1 000e 196 Agilent Serial BERT N4901 2 User s Guide January 2006 From these points the following QBER vs Threshold graph is generated 1 Trace 1
117. by bad cables or connectors 280 Agilent Serial BERT N4901 2 User s Guide January 2006 Additional Information The output of your device may contain errors due to noise although a noisy waveform is not present at the output The noise may originate somewhere else within your device or system causing random bit errors to be passed on to other components Therefore your device may have a good output waveform and still transmit random bit errors Sync Loss Seconds Accumulated Results Error Free Seconds Errored Seconds Power Loss Seconds Syne Loss Seconds Possible Cause The sync loss seconds in this example may have been caused by high bursts of errors or momentary clock loss The clock signal from a clock recovery module may have momentarily ceased To confirm periods of momentary clock loss check the measurement log file Additional Information In this example the analyzer was in auto sync mode If the analyzer was in manual sync mode the results would not have been measured and displayed in the same way When the clock signal resumed the received and ED reference patterns could only be aligned if the synchronization functions were activated In manual sync mode this would not have happened because synchronization functions are only activated manually In manual sync mode sync loss and high BER would have been measured from the start of the momentary clock loss to the end of the accumulation period Example Log of Sy
118. cales and frame of the graphs default is black FreqkangesColor Color of the selected frequency ranges default is light yellow GridColor The color of the dashed grid lines default is gray PowerMarkerColor The color of the Noise Threshold marker default is red 2 Select the color that you want to change from the list of used colors to the left 3 Use the Color Set and the Color Palette to select the new color 4 To view the results of your changes press Apply This immediately updates the measurement window where you can see the new colors 5 Press OK to close the Properties dialog box Agilent Serial BERT N4901 2 User s Guide January 2006 153 Advanced Analysis Reference This section covers reference information such as definitions of important parameters and the descriptions of supplementary functions Exporting Result Data If you want to use the measurement results with other applications you can export the data to a file via Analysis Export Data The contents of the resulting file may look as follows Date 02 11 05 03 33 17 Version 1 0 Type TM Fast Eye Mask SB Electrical UI 9 6969E 011 Threshold Units Volts Delay Relative 1 1 SerialBERT Delay Threshold BER A11 3 878757E 011 1 504954E 001 2 497699E 001 3 878757E 011 1 504954E 001 2 499503E 001 1 551503E 011 3 504954E 001 4 922570E 001 1 551503E 011 3 504954E 001 4 952636E 001 1 551503E 011
119. can then paste it into any other application 3 Press the Format Options button to open the Export Logfile Format dialog box 4 Make further selections for the format of the exported data Press OK to close the dialog box 5 Press OK to export the data to the specified destination How to Compare Measurement Results If you intend to repeat a measurement with different settings you may wish to preserve the current results for easier comparison This can be done with all measurements except Error Location Capture by copying the results 1 In the lower left hand corner of the measurement window select the measured terminal by tipping on it 150 Agilent Serial BERT N4901 2 User s Guide January 2006 2 Select Copy from the Edit menu If you have a mouse connected to your Serial BERT you can directly copy the measurement via the context menu 3 Open the Edit menu or context menu once more and select Paste This inserts a copy of the chosen measurement results For an Output Levels measurement this looks as shown below 2 Traces 172 Points 600 0 mY 400 0 mY 200 0 mY oov 200 0 mY 400 0 m apne BER Threshold 1 000e 6 Logarithmic 1 000e 5 1 000e 3 1 000e 1 1 000e 6 1 000e 4 1 000e 2 1 000e ar fe 4 Repeat the measurement with modified parameters and observe the Terminal Electrical Mean Level Show Color Copied High Level Low Level 479 08 mV 24 085 mV 479 0
120. capabilities to fully characterize your digital device The instrument provides Excellent waveform performance e Outstanding rise times and jitter performance e A choice of logic family interfaces e Easy object oriented programming with the IVI COM interface e Remote commands similar to HP Agilent 71603B e A variety of ways in which to accumulate and view test results e A collection of example patterns e Convenient online documentation e Automatic setup of optimum sampling point e A variety of powerful advanced analysis tools N Agilent Serial BERT N4901 2 User s Guide January 2006 Which Features are Available The Serial BERT is available in different options They differ in the range of provided bit rates and in the available features The following options are available N4901B Bit rate range 150 Mbit s to 13 5 Gbit s N4902B Bit rate range 150 Mbit s to 7 Gbit s Note that the N4902B 7 Gbit s Serial BERT is upgradeable to an N4901B Serial BERT for 13 5 Gbit s To check the options of your Serial BERT press About in the Help menu The About dialog shows the installed options 8 Agilent Serial BERT N4901 2 User s Guide January 2006 Which Testis Appropriate Planning the Test Which Test is Appropriate These topics help you to decide on the appropriate test setup for your device Determining How to Test Your Device The Agilent Serial BERT N4901 2 can help you test the performance of componen
121. ce Use the Pattern Select dialog box to select the pattern s you want to write to the pattern generator and the error detector See Select Pattern Dialog Box on page 44 for details 48 Agilent Serial BERT N4901 2 User s Guide January 2006 Hardware Generated PRBS This section describes the basics of hardware generated PRBS 2 n 1 patterns When to Test with Hardware Generated PRBS Patterns You can use the Serial BERT s hardware generated PRBS patterns if you require simple random traffic for testing your device The 2 n 1 odd pattern lengths are not related to the divided clock rate When using an oscilloscope this will ensure that the random contents of the bit stream is sampled The spectral contents of PRBS 2 n 1 patterns and the special property of decimation may also be useful in many applications for example for multiplexer or demultiplexer testing How the Hardware Generates PRBS 2 n 1 PRBS is an inverted hardware generated pattern that is created by a series of shift registers with adjustable feedback The example below shows the register configuration for a 2 7 1 127 bit pattern inital The following table describes the operation of XOR ing two points for the different patterns 2 1 D D 1 0 inverted 210 1 D D 1 0 inverted 211 1 D1 D 1 inverted 215 1 D16 D14 1 0 inverted 223 1 D D18 1 0 inverted 231 1 D p28 1 0 inverted A
122. ce 0 95 95 e natural base number n number of bits examined without error b desired residual BER You may want to prove that your device has a certain residual BER If you know this desired BER and the desired confidence level you can calculate the number of bits that must be measured without error The number of bits in turn can be translated into a period of time at a certain bit rate That is what is done in the above table The following example shows how to calculate the number of bits if the desired BER b is 1E 10 and the desired confidence level C is 95 1 Change formula C 1 e nb to solve for the number of bits n n In 1 C b 2 Replace variables C and b n In 1 0 95 1E 10 3 Calculate the number of bits n In 1 0 95 1E 10 2 99 1E 10 2 99E 10 bits 4 Calculate the time period at the STM 16 0C 48 bit rate 2 48832 Gb s 2 99E 10 2 48832E 9 12 016 seconds 12 Agilent Serial BERT N4901 2 User s Guide January 2006 CAUTION NOTE Connecting the DUT This section provides information on how to connect your DUT for several common test scenarios The pattern generator s Data Out and Clock Out ports must be terminated with 50 if they are not connected The following procedure is recommended when setting up a test 1 If your DUT can handle 0 V press the OV Disable button in the PG Output Setup screen The pattern generator s Data Out and Clock Out outputs are se
123. centage threshold 235 PG CLK LOSS 276 Phase Margin 180 183 Power scale 254 PRBN 21 PRBS 21 PRBS Patterns 19 Printing 297 Problems Constant Errors More Errored 0 s than l s 279 Random Errors 280 Sync Loss Seconds 281 Properties dialog box DUT Output Timing Jitter 172 Eye Opening 217 Fast Eye Mask 234 Output Levels 191 Spectral Jitter 249 Pseudo Color Plot 210 Q Q Factor 201 Q High Level 202 Q High Level Nr Points 202 Q High Level R 2 203 Q High Level Std Dev 202 Q Low Level 203 Q Low Level Nr Points 203 Q Low Level R 2 203 Q Low Level Std Dev 203 Q Optimum Threshold 201 Q Residual BER 202 QBER vs Threshold Graph 196 Q Factor Calculations mathematics 204 Notes 207 R R 2 values 181 Random Errors 280 Random Jitter RMS 181 Ratios Graph 265 Relative Time 236 Requirements 10 Resolution Eye Opening 218 Fast Eye Mask 218 Output Levels 192 Output Timing 173 Results see Numerical Results Return to Results button 124 S Saving Patterns 37 Sequence creating 54 definition 51 loops 58 properties 61 recalling 55 saving 55 Sequence Editor 56 Toolbox 60 Sequence Mode 53 SequenceExpression 51 Sequences 51 Serial BERT introduction 6 Setting Date and Time 290 Signal Noise Ratio Peak Peak 200 Software generated patterns 20 46 48 SONET SDH Receiver Connections 16 Spectral Jitter measurement procedure 247 Standard patterns 20 Status Indicators 261 Supporte
124. cial operating mode for measuring data in bursts of bits rather than one continuous stream of bits For more information refer to Introduction to Burst Sync Mode on page 133 NOTE Ifthe error detector is in Manual Sync mode it is recommended that you keep an eye on the SYNC LOSS light at the top of the instrument There are various actions that can lead to loss of synchronization Check the SYNC LOSS any time you make changes to the instrument What is False Synchronization For patterns other than PRBS the error detector may gain sync at a point in the pattern that meets the sync threshold but is not the correct point where the internal reference pattern and the received data pattern match This is called false synchronization NOTE False synchronization cannot occur with PRBS patterns because a 1 bit misalignment would cause a measurement of 50 or more errors Thus the BER during a misalignment would always be greater than the sync threshold BER 130 Agilent Serial BERT N4901 2 User s Guide January 2006 For example consider a pattern of 1000 ones and 1000 zeros as shown in the following figure With reference alignment 1 the patterns are totally out of phase and the error detector is measuring 100 errors But as the reference moves closer to optimum alignment the percentage of errors gradually approaches zero reference alignment 2 and 3 For exact alignment the sync threshold must be set lower than the BER caused by a
125. cking touching the Test Box Using the On Screen Keyboard You can use the on screen keyboard to enter text in any text field It can be used along with the standard keyboard or keypad 1 Click in a text field to position the cursor in it 2 Press the On Screen Keyboard button to open the on screen keyboard 3 Click letters or numbers as desired For numbers you must add the unit by pressing the G n k m M u or x1 front panel button 4 Press the On Screen Keyboard button again to close the on screen keyboard 5 Continue with your work NOTE You can also access the on screen keyboard from the Utilities menu A check mark appears next to the menu entry indicating that it has been selected Using an External Monitor To switch between the Serial BERT s built in display and an external monitor do the following 1 Connect the external monitor with an appropriate power supply and with the VGA connector on the Serial BERT s rear panel 2 If the title bar of the Serial BERT user interface is not visible select Title Bar from the Utility menu 3 Minimize the Serial BERT user interface window 4 In the system tray next to the clock in the lower right corner right click the Display icon to open the menu with the available options 294 Agilent Serial BERT N4901 2 User s Guide January 2006 TIP NOTE SS 11 50 4m 5 Select Graphics Options then Output To and the desired output device Monitor for an e
126. coming Bits j x yes failed no O BER Bad Burst 1 lt no yes v l Total Burst 1 lt __ Incr Bit Counters Agilent Serial BERT N4901 2 User s Guide January 2006 133 NOTE Burst mode requires a complete measurement setup including an external input for the Gate In port This input is used for the sync alignment 1 Burst sync is started when the Gate In signal goes low indicating that a burst is arriving at the Data In port 2 Ifthe error detector is in CDR mode it first tries to generate a clock from the data A short time interval is required to stabilize the clock the CDR settling time Note that if the error detector is in CDR mode it is necessary to set a CDR threshold This is the BER threshold that the error detector uses to verify that the input pattern is synchronized See How does Clock Data Recovery Work on page 107 for more information If an external clock is used there is no delay The clock signal must however be continuous 3 The error detector can then synchronize to the incoming signal For PRBS the first received bits are used to seed the synchronization If there is an errored bit in this phase the synchronization fails bad burst For memory based patterns a unique 48 bit detect word is used for the synchronization This pattern should be available one time only in the pattern If the detect word is not
127. covery CDR 3 If you are using CDR you need to approximately specify the clock rate in the Clock Rate field with an accuracy of 0 01 4 If you are measuring bursts deactivate Auto Threshold and set the threshold manually You can also press the Measure button to measure the current DC level at the CDR input and set the measured value as CDR threshold 5 Clock rates that are frequently used can be saved as presets in the Preset list The Preset list always displays the presets within a certain frequency range Use the CDR Range selection list to display the presets of another range To select a clock rate from the Preset list click on it The clock rate appears in the Clock Rate field Clock Setup Reference The Clock Setup window contains the following elements Error Detector Bit Rate The bit rate of the data reaching the error detector is displayed in the lower right corner of the user interface fe Ck Rate 2 4883 GHz The bit rate of the error detector is set by the clock signal received by the error detector If the output bit rate of your device is not equal to its input bit rate the pattern generator and error detector will be operating at different frequencies Agilent Serial BERT N4901 2 User s Guide January 2006 111 Clock Mode The error detector supports two different clock modes External Clock Source If this mode is selected the Serial BERT uses the clock signal that is connected to the C
128. current marker positions Upper Right Time and BER of the upper right intersection of the displayed marker lines X2 and BER1 Lower Left Time and BER of the lower left intersection of the displayed marker lines X1 and BER2 Delta Distance between the two vertical marker lines on the time axis X2 X1 and distance of the two horizontal marker lines on the BER axis BER1 BER2 The markers can be moved to keep track of the data The Gaussian Marker A Gaussian marker is available for the following measurements DUT Output Timing Jitter measurement Output Levels measurement The Gaussian marker allows to measure the normal distribution of measurement results The distribution of random values is called normal if it can be described by the following formula _ py 20 f x 1 e ON2T Agilent Serial BERT N4901 2 User s Guide January 2006 157 This formula describes a bell shaped Gauss curve If u is zero and o varied you would get the curves illustrated in the figure below fix The height and position of a normal distribution can be specified in terms of two parameters u and o The parameter u is the mean the parameter o is the standard deviation The Gaussian marker shows such a curve Position height and width of this curve can be changed by dragging the handles and the actual parameter values are displayed The marker has three handles that can be dragged two for adjusting i
129. d Pattern Types 24 SYNC LOSS 277 Sync Loss Seconds 281 Sync Now button 121 Agilent Serial BERT N4901 2 User s Guide January 2006 T Termination Error Detector 106 Terminations Setup 71 Usage 68 Test Duration 11 Requirements 10 Test points Fast Eye Mask 230 Threshold absolute 235 offset 235 percentage 235 Threshold Eye Opening 223 Threshold Margin 199 Threshold Resolution High Level Eye Opening measurement 218 High Level Output Levels measurement 192 Low Level Output Levels measurement 192 218 Threshold Types 235 Time Eye Opening 223 Timing Unit Eye Opening measurement 222 Fast Eye Mask measurement 235 Total Jitter BER Threshold 183 Total Jitter Peak Peak 183 Total Jitter Peak to Peak 181 Total Jitter RMS 180 Total Jitter Uncertainty 183 Touchscreen Calibration 293 Off On 293 Trigger Out Error Detector 102 Trigger Out port 65 U Uncertainty Band 167 User pattern 19 User Defined Sequences 51 V Vampt 74 Variable Mark Density 47 Vhi 73 View Tab Eye Opening 220 Output Levels 195 Output Timing 176 Spectral Jitter 254 View tab 311 Fast Eye Mask 236 Vlo 74 Vof 74 Voltage of measurement point 236 X Xover 75 Z Zero SubstitutionPattern zero substitution 47 Zoom Function Eye Opening 222 Output Level 198 Output Timing 179 Spectral Jitter 255 312 Agilent Serial BERT N4901 2 User s Guide January 2006 Agilent Serial BERT N4901 2 User s G
130. de also frequency and power information about the dominant peaks in the spectrum e Absolute and relative power values are available Relative values can be normalized to the total jitter power or the power of a selected tone frequency bin e Pass fail limits can be set for the bit error rate total power and noise power and the allowed jitter power in user defined frequency regions Agilent Serial BERT N4901 2 User s Guide January 2006 24 Periodic Jitter and Bit Error Information When the incoming signal is sampled at the transition point periodic jitter manifests itself in the bit error record An example may be helpful to understand the phenomenon Let us assume we expect and correctly receive a simple 0 1 0 1 0 pulse signal If we would sample this signal one clock period earlier 1 UD we would see a bit error rate BER of 1 0 This is the maximum BER for this pattern Let us also assume this signal is overlaid by a periodic jitter source with sinusoidal characteristics Now we sample the incoming signal at the transition point 0 5 UI from the optimum sampling point The result is illustrated in the figure below Signal without jitter 1 0 E Pa Expected and correctly 0 4 0o 4 0o 1 o 14 0o 14 0o 4 0o 4 0o 4 0o 4 sampled data t i Periodical i sinusoid jitter I i i i i i H H H increasing right shift decreasing right shift increasing left shift decr
131. dialog box allows you to search for a specified bit sequence in the pattern If you click Newt the next occurrence of this bit sequence is highlighted The search pattern can be entered in either hex or binary format independently of the currently displayed format in the pattern editor You can continue editing the pattern while this dialog box is still open Looks for the next occurrence of the pattern from the current position to the end If there is no further occurrence the current selection does not change Click this button to close the dialog box Insert Block Dialog Box The Insert Block function provides an easy way to modify parts of the pattern or the entire pattern at once This can be used when setting up a new pattern It can also be used as an optional technique for editing existing patterns With the Insert Block dialog box you define the Range that is to be modified and an Action that will be done to this range Agilent Serial BERT N4901 2 User s Guide January 2006 43 44 Error Detector Pattern Tracks the Pattern Generator Pattern The available options for the Range are Select All Choose this option to select the entire pattern for editing Range Choose this option to select the range of bits specified by the bit positions entered in the From and To fields Fill From Cursor to End Choose this option to select the bits from the cursor location to pattern end Fill From Start to Cursor Choo
132. ding the Output Level Parameters on page 66 for information about how the Serial BERT modifies the output levels Delay This text field allows manual entry of data output delay and displays the current value in picoseconds 74 Agilent Serial BERT N4901 2 User s Guide January 2006 TIP NOTE To modify the value click inside the text field and either Enter the desired value directly with the numeric keyboard e Use the knob by the numeric keyboard to fine tune the value You can also adjust delay with the Data Out Delay knob on the front panel Xover This text field allows manual entry of the crossover percentage and displays the current value To modify the value click inside the text field and enter the desired value directly with the numeric keyboard or use the front panel knob to fine tune the value Data Polarity Inverted Select this checkbox to invert the logic of both data outputs Data and Clock Logic Levels You can quickly set up an output signal that is compatible with your device by selecting a logic family from the Logic Level list Select the logic family you want to use for the data and clock outputs When you select a family the corresponding signal voltage levels and the Termination for this family is automatically selected When you select Custom there are no changes in the expected terminations See Understanding how the Serial BERT Uses Logic Families and Terminations on page 68 fo
133. duces the spectral power e Results obtained from different devices can only be compared if the same window is used Spectral Jitter Procedures This section shows how to set up and use the Spectral Jitter measurement As an example we measure the spectral jitter behavior of a shielded cable This requires the following steps How to Prepare the Spectral Jitter Measurement on page 247 e How to Execute the Spectral Jitter Measurement on page 248 How to Prepare the Spectral Jitter Measurement To use the Spectral Jitter measurement to test a shielded cable 1 Disable the pattern generator outputs by pressing the OV Disable button in the PG Output Setup screen 2 Use a shielded cable to connect the pattern generator s Data Out port and the error detector s Data In port 3 Terminate all non connected pattern generator output ports with 50 Q 4 Enable the pattern generator outputs by pressing the OV Disable button 5 Switch to the Pattern menu and press Pattern Select Select an appropriate pattern for example a pure 215 1 PRBS for both the pattern generator and the error detector 6 For the pattern generator setup you need to specify the logic levels and the bit rate Select Logic levels LVPECL Clock speed Gb Ethernet 1 06250 Gb s See Setting up the Pattern Generator on page 63 for more information Agilent Serial BERT N4901 2 User s Guide January 2006 247 7 Set up the
134. e is to measure multiple data bursts in quick succession In this way you can accumulate data for all bursts together You may wish to accumulate the number of bits necessary for a statistically valid BER measurement Refer to Introduction to Burst Sync Mode on page 133 and Test Times and Confidence Levels on page 11 for details When the accumulation period lasts for the duration of all bursts the gaps in error counting while the Gate In signal was high will not be included in BER calculations The gaps in error counting will not be considered error free periods Agilent Serial BERT N4901 2 User s Guide January 2006 269 NOTE Ifthe gaps between data bursts are greater than 1 second or 100 ms in the measurement log file they may be visible in the accumulated results graph Burst Status The following states may be returned No Error No burst errors have occurred any of those listed below Gate Signal Too Long This error can occur if there are too many bits within a burst The limit is 4 Gbit At 13 Gb s this occurs roughly after 0 3 s slower frequencies have a higher gate in period This error has a higher priority than no unique 48bits No Unique 48 Bits Found For reliable synchronization a pattern must contain a unique 48 bit pattern the detect word If the current pattern does not have a detect word this error occurs If this status occurs the synchronization may be incorr
135. e lists the characteristics generally associated with some logic families All values are nominal ECL 0 95 1 7 750 2 SCFL 0 0 9 900 0 LVPECL 2 35 1 6 750 13 LVDS 1 425 1 075 350 1 250 CML 0 0 4 400 0 Terms e ECL Emitter Coupled Logic e SCFL Source Coupled FET Logic LVPECL Low Voltage Positive Emitter Coupled Logic PECL Lite e LVDS Low Voltage Differential Signaling e CML Current Mode Logic Agilent Serial BERT N4901 2 User s Guide January 2006 69 Why Incorrect Terminations Could Damage Your Device Choosing wrong terminations may cause your device to output voltage levels that are not as expected It may also cause excessive current or current flow in the wrong direction which can damage your device Note that an internal protection algorithm becomes active if both the termination voltage is wrongly adjusted and the termination voltage is below 1 5 V The protection algorithm sets the output voltages to safe levels Vhign termination voltage 1 V Viow termination voltage 0 9 V e termination voltage remains unchanged AC Coupling and Bias Tees The pattern generator s outputs are normally DC coupled even when AC termination is selected For this reason extreme caution must be taken when connecting your instrument to a device or test setup The diagram below shows a device that is AC coupled Notice that the capacitor is part of the test setup DUT Y
136. e menu 6 Make sure that the title of the dialog box is Format Floppy A Agilent Serial BERT N4901 2 User s Guide January 2006 287 E CAUTION Do not format any other drive but the floppy drive 7 Click Create an MS DOS startup disk 8 Click Start and follow the instructions NOTE The MS DOS startup disk only allows the system to boot into an MS DOS prompt The disk contains no additional tools and is not a substitute for Windows XP Setup Boot Disks How to Restore the Operating System 1 Turn off the instrument and connect a keyboard and mouse to the instrument See Connecting Peripherals on page 290 for details Insert the bootable diskette in the floppy drive Turn on the instrument On the command prompt type d and press ENTER a A Ww N Type recover press ENTER and follow the instructions 288 Agilent Serial BERT N4901 2 User s Guide January 2006 Configuring the Instrument Customizing the Instrument Configuring the Instrument This section covers information on the configuration of the Agilent Serial BERT N4901 2 Updating the Software New software features are offered for the Serial BERT from time to time you may be able to enhance the capabilities of your instrument by updating the software See the Agilent Serial BERT web site to see if updates are available for your instrument http www agilent com find SerialBERT Besides the latest versions of the user interface and firmware server
137. e output ports or change the termination voltage before adjusting the output levels See Understanding the Output Level Parameters on page 66 for information about how the Serial BERT works with output level parameters 72 Agilent Serial BERT N4901 2 User s Guide January 2006 The following topics tell you how to use the pattern generator s input ports e Bit Rate Procedures on page 79 Delay Control Input Procedures on page 86 e Error Addition Insertion Procedures on page 88 Pattern Alternation Procedures on page 93 Input and Output Ports Reference The pattern generator produces clock and data outputs that serve as frequency reference and device stimulus for the device under test NOTE The Serial BERT will not allow you to adjust a voltage beyond its limits The limit is determined by the Serial BERT s internal hardware If a limit is encountered the Serial BERT sends a message to the status bar For more information about voltage limits see the pattern generator s technical specifications Vhi This text field allows manual entry of the logic high voltage level and displays the current value To modify the value click inside the text field and either Enter the desired value directly with the numeric keyboard e Use the knob by the numeric keyboard to fine tune the value See Understanding the Output Level Parameters on page 66 for information about how the Seri
138. easing left shift A A A A r Y Bg T Signal with jitter 1 F r 0 TULL UU UL T T T T T T T T T T T T T T T T T T Captured data 1 0 1 0 1 0 1 0 1 1 o t 0 1 0 1 t Expected data 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 Error data 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 The jitter source moves the received signal to the right and to the left but the sampling point is fixed dotted lines show the undisturbed signal Because we are sampling at the transition point we get errors when the signal is shifted to the right and no errors when the signal is shifted to the left 242 Agilent Serial BERT N4901 2 User s Guide January 2006 NOTE The error signal shows a very characteristic and reproducible pattern The resulting BER for this kind of signal is 0 5 Other jitter frequencies yield different but also characteristic error patterns and also a BER which is half the maximum BER The maximum BER and hence the actual BER are pattern dependent Considering the BER bathtub curve the sampling point is in the middle of the descending line If random data is used the average bit error rate at the left hand side of the jitter region is 0 5 Sampling point BER 0 5 BER 0 25 BER 0 The bit error rate at the right hand side of the jitter region is 0 We therefore expect a bit error rate around 0 25 The Spectral Jitter measurement should only be used in conjunction with data that has
139. ect as could also the measured bit error rate There are standard patterns that may contain more than one instance of the used detect word Statistically every other burst would be correctly synchronized In this case it is recommended that you redefine the pattern This error can only occur with memory based patterns Unknown The status is unknown This can occur if accumulation has not been started or if Burst Sync mode has not been activated Burst Sync Ratio Percentage of time in burst while the Gate In signal is low that patterns are in sync and error counting is active A higher percentage is desirable because it indicates that more data is measured during each burst Bi Ti Burst Sync Ratio t Count Time CDR Settling Time Synchronization Time Bit Count Time Total Burst Count Total number of bursts received while in burst mode during the accumulation period 2 0 Agilent Serial BERT N4901 2 User s Guide January 2006 Bad Burst Count Total number of bursts where sync is never achieved during the accumulation period Bit Count The number of bits considered for the accumulation period is displayed here This may not equal the total number of bits sent from the PG Agilent Serial BERT N4901 2 User s Guide January 2006 2 1 Evaluating Results Bye Measurements Eye Measurements The purpose of eye measurements is to measure the eye height and width of the incoming data signal at specific alignm
140. ected It may also cause excessive current or current flow in the wrong direction which can damage your device NOTE The clock termination is set to 50 Ohm AC coupled and cannot be changed Agilent Serial BERT N4901 2 User s Guide January 2006 103 Inputs and Outputs Procedures To select the termination for the error detector NOTE Prerequisite You must know the termination voltage of the data signal that your DUT sends to the error detector 1 Inthe ED Setup menu press Sampling Point Setup to access the Sampling Point Setup window 2 Click the Edit button This opens the ED Input Setup dialog Differential Input Range 1 000 1 000 Termination 0 000 Edit l 3 In the Termination field enter the termination voltage that is appropriate for the incoming data signal You can use the front panel knob or the numeric keypad to enter the value E CAUTION Selecting the wrong termination may damage your device 4 Select the Data Inverted checkbox if your device inverts data You can now physically connect the DUT to the error detector 104 Agilent Serial BERT N4901 2 User s Guide January 2006 Data Input Setup Reference The ED Input Setup dialog is accessed from the Sampling Point Setup window It is used to set up the error detector s data input port Input The selection in this list defines how the signals arriving at the Data In and Data In connectors are interpreted The following op
141. ed The number of bits can be set from 1E7 to 1E15 When the selected number of bits have been exceeded the accumulation ends NOTE The actual test period may contain more bits than specified NOTE The accumulation period cannot be changed while accumulation is progress Browse Click this button to browse for the directory where you wish to save the measurement log files NOTE Changes to the parameters of the accumulation setup do not take effect until the start of the next accumulation 142 Agilent Serial BERT N4901 2 User s Guide January 2006 All Bits Block BER Location The Serial BERT can be run in BER location mode In this mode the errors are not counted on all bits of the data signal but on a particular bit or block of bits in the pattern You can specify which bit position s are considered during the measurement This allows for example to calculate the BER for the header bits or the payload of a data stream exclusively BER Location Procedures To set up the BER location mode do the following 1 Press the BER Location icon in the ED Setup menu 2 Select whether you want to locate errors on All Bits on a particular Block of bits or on a Single Bit in the pattern If you select Block or Single Bit enter the bit positions in the respective fields 3 Click OK to finish the BER location mode setup BER Location Reference The Bit Error Location Mode dialog box offers the following options
142. elect All button again Run Test After selecting at least one test use this button to start the test If you selected tests that need a longer time to complete a message box opens Confirm your selection by clicking OK A second message box is informing you while the test is still running and closes when finished 306 Agilent Serial BERT N4901 2 User s Guide January 2006 Message Buttons After the self test has finished the results Success or Failed are displayed next to each test If a test has failed you can use the Message button to open the BIOS Messages window with detailed error messages explaining the failure reason Agilent Serial BERT N4901 2 User s Guide January 2006 307 308 Agilent Serial BERT N4901 2 User s Guide January 2006 Index 0 1 Decision Threshold 119 0 1 Threshold 119 0 1 THReshold Center button 121 OV Disable button 76 10 MHz Ref Input 64 A Absolute threshold 235 AC Coupling 70 Accumulated Measurements Results 264 Running 263 Accumulated Results 267 Accumulation Parameters 269 Accumulation period 142 Adjust Output Levels 72 Advanced Analysis 147 Amplifier Connections 15 Amplitude 199 Auto Align 116 Auto Align button 121 Auto Threshold Error Detector 113 Aux In 65 Aux Out Error Detector 126 Auxiliary Out port 103 Average 0 1 Threshold 123 B Bathtub curve 162 BER 6 BER Results Analyzing 260 Monitoring 259 BER Results Window 260 BER Threshold
143. ence 262 Agilent Serial BERT N4901 2 User s Guide January 2006 NOTE Measurement Period NOTE Accumulated Measurements Accumulation refers to collecting measurement data over a specific period This can be used to create test scenarios that are reproducible and comparable Also you can let tests run over long times and then evaluate the results afterwards Furthermore you can use the log files to determine the location of occurring bit errors for further root cause analysis While accumulation tests are running you should not modify the measurement setup as the measured bit errors do not represent the performance of your DUT under real circumstances The period of time can be set by manual start and stop commands an absolute time setting or the time it takes to measure a specified number of bits or bit errors The accumulation period should be long enough to make a statistically valid BER measurement See Test Times and Confidence Levels on page 11 for details During measurement logging the Serial BERT logs data in ten second intervals Your log file may be missing up to the last ten seconds of data To avoid this condition accumulate for 10 seconds longer than desired Accumulated Measurements Procedures This section describes how to run accumulated measurements and how to analyze the results How to Run Accumulated Measurements To run accumulated measurements 1 Ensure that the instrument is set up
144. ends only blocks of ones and zeros there are no transitions in the data stream and the Serial BERT cannot recover the clock Also if you are testing bursts there are some special considerations for setting up CDR See the following sections for details How does Clock Data Recovery Work In CDR mode the CDR has to recover the clock from the incoming data To do this the hardware has to decide whether the voltage at the input connector is a logical 1 or 0 and then recover the clock from the detected transitions Because the regular threshold voltage is not only used to determine the optimum sampling for the data but also to perform measurements such as eye diagram or output level measurements it is not possible to use it for the clock recovery For this reason the clock recovery circuitry has it s own comparator for the incoming data This comparator also needs to know the threshold voltage 0 1 decision threshold The threshold voltage can be derived from the input signal via a low pass filter This will work fine for most applications But applications that do not provide a continuous data stream at the input for example any application using bursts cannot use this low pass filter because the threshold voltage will drift from the correct level when there is no input In such cases the threshold can be specified manually It is then no longer derived from the input signal see the following figure The manually
145. ent BER thresholds This information is displayed on a representation of an eye diagram Eye Measurements Procedures Eye measurements are made each time the Auto Align Clock Data Center or 0 1 Threshold Center functions are used To analyze the results 1 View the graph and table in the Eye Results window This window provides detailed information on the calculated eye See Eye Measurements Reference on page 272 2 Analyze the results by observing changes in the eye results due to the following adjustments Make changes to your device or measurement setup Run auto search functions at different BER thresholds This will allow you to see the eye margin at specific BERs and construct basic eye contour information 3 Change the sampling point in the Sampling Point Setup window See Sampling Point Display on page 124 for details Eye Measurements Reference Eye results are displayed for the Current Align and Previous Align process They are also graphically displayed on a representation of an eye diagram Eye Width and Height Eye width is the inside horizontal opening of the data eye in an eye diagram This measurement is made by the error detector after an Auto Align or Data Center The eye width at the current 0 1 decision threshold is displayed on a simulated eye diagram The left and right eye edges are determined by the user defined BER threshold Different BER thresholds will produce diff
146. er component with Gaussian distribution After transforming a contiguous range of measured points into Q space and performing a linear regression it is calculated as the mean of the sigmas of the two straight lines The contiguous range is limited by the the BER Threshold and the Min BER for RJ DJ Separation threshold Deterministic Jitter The total jitter component with non Gaussian distribution After transforming a contiguous range of measured points into Q space and performing a linear regression it is calculated as the period minus the difference between the means of the two straight lines Estimated Total Jitter A forecast of the expected jitter for very low bit error rates After extrapolating the measured BER curves it is calculated as the period minus the expected width of the eye opening No of points This is the number of points that has been measured between the BER Threshold and the Min BER for RJ DJ Separation threshold It is displayed for both slopes This number has to be greater than 2 for the RJ DJ and estimated TJ values to be applicable R values The R 2 values are calculated for both slopes of the bathtub curve They are a measure of how well the transformed points between BER Threshold and Min BER for RJ DJ Separation fit to the linear regression They have to be greater than 0 75 for the RJ DJ and estimated TJ values to be applicable Agilent Serial BERT N4901 2 User s Guide January 2006 181
147. erent eye width results 272 Agilent Serial BERT N4901 2 User s Guide January 2006 NOTE Eye height is the inside vertical opening of a data eye in an eye diagram This measurement is made by the error detector after an auto align or 0 1 threshold center The eye height at the current data input delay point is displayed on a simulated eye diagram The upper and lower eye edges are determined by the user defined BER threshold Different BER thresholds will produce different eye height results Upper eye edg f Eye height Most oscilloscopes measure eyes differently using statistical data Therefore the eye results of a Serial BERT and an oscilloscope may be different Agilent Serial BERT N4901 2 User s Guide January 2006 273 The table in the top half of the Eye Results window additionally lists the following values Eye Voltage Center One of two components of the sampling point the 0 1 decision threshold eye voltage center is displayed here e Eye Data Input Delay One of two components of the sampling point the data input delay is displayed here e Alignment Threshold The alignment BER threshold is displayed here This threshold was used to define the eye edges and the above eye results Delta Error Ratio The BER measured during the previous 200 ms is displayed here 274 Agilent Serial BERT N4901 2 User s Guide January 2006 Solving Problems This section provides information that can help you to tr
148. erial BERT N4901 2 User s Guide January 2006 161 DUT Output Timing Jitter This type of measurement is used to measure the timing and jitter behavior for a device under test DUT It uses a bit error rate BER measurement to evaluate the shape of the eye for the output signal of the DUT It also analyzes the jitter separates the random jitter and deterministic jitter components and estimates the total jitter A direct result is the determination of the optimum sampling point delay for receiving data from the DUT with maximum confidence DUT Output Timing Jitter includes the Fast Total Jitter measurement that can be used to measure the total jitter for devices which generate a very low error density in a reasonable time span Output Timing Characteristics The sampling point is swept automatically within a 1 5 clock period to generate a bathtub curve The resulting graph is centered around the optimum sampling point of the port In addition the results are available in a tabular view If a clock signal is defined the software measures the data to clock alignment and displays the absolute delay Jitter Characteristics The DUT Output Timing Jitter measurement calculates the jitter histogram as the absolute of the derivative of the measured bit error rate jitter dBER dt The jitter histogram allows to separate and calculate the different components making up the jitter e Random Jitter RJ e Deterministic Jitter D
149. ern User Patterns User patterns are file based editable patterns These may be patterns that you set up using the Serial BERT s Pattern Editor or any of the patterns delivered with the Serial BERT You can easily define your own patterns for any special requirements Agilent Serial BERT N4901 2 User s Guide January 2006 19 20 Software Generated Patterns The Serial BERT can also generate various PRBS 2 n patterns with a range of polynomes from 7 to 23 These patterns can be further modified by setting special mark densities ratio of 1 s to total bits or by defining the zero substitution Sequences It is possible to combine several patterns in a sequence A sequence consists of up to four blocks that can be looped Each block can generate a pause signal constant 0 or 1 a divided clock a 2 n 1 PRBS or a user pattern What Patterns are Available The Serial BERT offers various industry standard and customized patterns for many BER testing scenarios The following graphic illustrates what patterns are available on the Serial BERT Note that there are two main branches corresponding to the hardware generated and memory based patterns and the memory based patterns are further broken down into software generated patterns and user patterns Pattern Family Tree Memory Based Patterns
150. ern 1110001 Data Out 10111000100 10111000100 Trigger Out Shift Trigger Position Clicking the left and right buttons will move the bit trigger position back and forward by one bit Agilent Serial BERT N4901 2 User s Guide January 2006 85 Setting up the Pattern Generator lay Control Input Delay Control Input The Serial BERT features a wide bandwidth voltage controlled delay line in its pattern generator data path which can be used to delay the data relative to the clock The delay is controlled by an external signal connected to the Delay Ctrl In connector on the pattern generator front panel The delay control input is intended primarily for jitter tolerance testing where a data signal with a defined amount of jitter is needed Random jitter RJ can be created by applying high bandwidth white Gaussian noise periodic deterministic jitter DJ with sinusoidal SJ or rectangular triangle signals PJ The delay control input is most useful for the generation of high frequency low amplitude jitter Higher amounts of jitter at lower frequencies wander can be generated by supplying an FM modulated clock to the pattern generator s clock input Clk In in external clock mode The delay control input is equipped with overvoltage protection If the applied voltage is too high the input is automatically disabled and the following dialog appears Delay Control Protection Input Signal at the Delay Control Inpu
151. ern generator in output blanking mode If a logic low signal is supplied to the Aux In port pattern A is sent to the output port If a logic high is supplied no data is sent to the output port This function is the same as the output blanking function in the Alternate Pattern Control dialog box see Aux In on page 96 for details However this function also works with standard patterns In Sequence Mode When the generator is set up to generate a sequence this checkbox is disabled You can connect a signal to the Aux In port of the pattern generator The reaction on this signal for example start sequence execution or break loop is defined in the SequenceExpression For details see User Defined Sequences on page 51 76 Agilent Serial BERT N4901 2 User s Guide January 2006 Bit Rate Bit rate is defined as bits per second The pattern generator s clock rate sets the bit rate and serves as the frequency reference for the clock and data outputs your device and the error detector if it receives its clock from the pattern generator It can be generated internally or supplied from an external source When to Use an External Clock Source The Serial BERT s internal clock can be used for most testing purposes There are some circumstances however when an external clock source is required Synchronization with an external clock The Serial BERT can be connected to an external clock to allow it to run as part of a
152. ern s Total Length of the pattern in bits The available functions to change the view and to modify the pattern are described in Edit Pattern Window on page 40 GoTo Bit Dialog Box This function allows you to set the cursor to an arbitrary position in the pattern Enter the bit position address and click Go The cursor is placed in front of the character with the selected bit You can continue editing the pattern while this dialog box is still open 42 Agilent Serial BERT N4901 2 User s Guide January 2006 TIP Description Pattern Type Length in Bits Next Done Pattern Properties Dialog Box The Pattern Properties dialog box always opens when you create a new pattern It can also be used to view or change the attributes of the currently selected pattern You can use your instrument s software keyboard to enter text in the dialog box For more information see Using the On Screen Keyboard on page 294 This text field allows you to enter a description of the pattern s characteristics or purpose It can be up to 256 characters long Select Standard if you want to use this pattern as a standard pattern Select Alternate if you want to define two alternating patterns A and B This text field allows you to enter the pattern length To find out more about appropriate pattern lengths see Pattern Resolutions and Lengths on page 26 Find Pattern Segment Dialog Box The Find Pattern Segment
153. errors Errors if Os Expected To display the errors if 0 is expected but 1 received Errors if 1s Expected To display the errors if 1 is expected but 0 received 236 Agilent Serial BERT N4901 2 User s Guide January 2006 Grid You can set the Timing Unit for the display of the relative time of the measurement points on the Parameters tab Choose between Unit Interval and Seconds to select the unit Table Number Format You can select the number of Decimal Places to be displayed in the table Agilent Serial BERT N4901 2 User s Guide January 2006 237 Advanced Analysis Spe tral Jitter Spectral Jitter The Spectral Jitter measurement allows you to analyze the jitter inherent in the output signals of your device under test DUT as a function of the frequency This measurement can be used for investigating the behavior of the DUT for example to identify crosstalk effects Spectral Jitter Analysis The Spectral Jitter measurement provides a method for investigating the spectrum of the jitter that means its power distribution over frequency Deterministic jitter can be caused by repetitive data patterns It can also be caused by internal or external periodic effects The Spectral Jitter measurement detects even small periodic components that may be hidden in a high level of random noise It informs you about the frequencies of such components and measures their contribution to the total jitter
154. esponds to a clock period of 0 8 ns See Setting up the Pattern Generator on page 63 for more information 186 Agilent Serial BERT N4901 2 User s Guide January 2006 Set up the error detector so that the input range and the termination matches the pattern generator s levels Select an Input Range from 2 V to 0 V Set the Data Termination to 2 V Set the Alignment BER Threshold to 1E 6 Set the Clock Setup to Clock Data Recovery to get the error detector s clock from the incoming data stream Enable the pattern generator outputs by pressing the OV Disable button Press Sync Now and then Auto Align to find the optimum sampling point Check that the synchronization and the alignment were successful None of the error indicators should show red The resulting BER should be zero How to Execute the Output Levels Measurement To run the Output Levels measurement 1 2 3 Switch to the Analysis panel and then press the Output Levels icon Press the Properties button to open the Properties dialog box On the Parameters tab set the Sample Threshold parameters to suitable values that fit to the expected signal levels for example 1 85 V to 0 85 V Set the Resolution to 10 mV Agilent Serial BERT N4901 2 User s Guide January 2006 187 Properties Ea Parameters Pass Fail View Graph Color r Criteria for moving to next measurement point Number of Compared Bits IV Number of Errors
155. essage box that informs you about which module caused the problem Pattern Generator Bios Power On Errors al Previous instrument shutdown forced by critical temperature Offline Error Detector Bios Power On Errors Previous instrument shutdown forced by critical temperature Offline Agilent Serial BERT N4901 2 User s Guide January 2006 285 286 Agilent Serial BERT N4901 2 User s Guide January 2006 Customizing the Instrument The Serial BERT provides various utilities for adjusting the instrument to your personal preferences Restoring the System Although any bootable diskette created on another PC with either Windows 98 or Windows XP can be used for emergency repair of your system we strongly recommend to create the bootable diskette at your system Store the diskette on a safe place and use it to restore the operating system to its factory state SEE CAUTION When the operating system is restored all personal settings programs and user data will be lost Make sure that important data is saved to external media before restoring the system How to Create a Bootable Diskette 1 Minimize the Serial BERT user interface window by selecting Minimize GUI from the Utility menu 2 Place a blank diskette in the instrument s floppy disk drive 3 Click Start and then click My Computer 4 In the file browser select the symbol for drive A Do not open the drives folder only select it 5 Select Format in the Fil
156. evels and the bit rate Select ECL levels and a clock speed of 1250 MHz in this example This corresponds to a clock period of 0 8 ns See Setting up the Pattern Generator on page 63 for more information 212 Agilent Serial BERT N4901 2 User s Guide January 2006 6 Set up the error detector so that the input range and the termination matches the pattern generator s levels Select an Input Range from 2 V to 0 V Set the Data Termination to 2 V Set the Alignment BER Threshold to 1E 6 Set the Clock Setup to Clock Data Recovery to get the error detector s clock from the incoming data stream 7 Enable the pattern generator outputs by pressing the OV Disable button 8 Press Sync Now and then Auto Align to find the optimum sampling point Check that the synchronization and the alignment were successful None of the error indicators should show red The resulting BER should be zero How to Execute the Eye Opening Measurement To run the Eye Opening measurement 1 Switch to the Analysis panel and then press the Eye Opening icon 2 Press the Start button to execute the measurement The measurement is run and the display is continually updated The following illustration shows the eye diagram of the measured bit error rate and in the tabular view the calculated results for the Time Eye Opening the Threshold Eye Opening the Optimal Sample Delay and the Optimal Sample Threshold 41 Trace 9510 Points Ul 401 878 ps
157. ew Tab The graph shows either a Contour Plot a Pseudo Color Plot or only one curve for the selected bit error rate threshold Contour plot The contour plot shows discrete lines of equal bit error rate just like the contour lines on a map The color of a line indicates the respective BER value This graph is useful to visualize in which areas the BER changes a homogeneous BER field will give you no lines at all 150 0 my Trace 13757 Point E Ul 1 538 ns All Errors o 1 000e 0 120 0 mv 1 000e 1 90 0 mv 60 0 mv i 1 000e 2 Dam Fe S 4 1 000e 3 ooy gt 300m f N P ai 1 0006 4 60 0 mY 1 000e 5 90 0 mY 1 000e 6 120 0 m 150 0 mv E 1 000e 7 0 650 UI 0 390 UI 0 130 UI 0 130 Ul 0 390 UI 0 650 UI 0 520 UI 0 260 UI 0 000 UI 0 260 UI 0 520 UI Pseudo Color Plot This plot visualizes the BER by a continuous color gradient This is useful to get an immediate visual impression of the distribution of the BER in the eye diagram Ul 1 538 ns 150 0 my 1 Trace 13689 Points All Errors o 1 000e 0 120 0 mY 1 000e 1 90 0 mY 60 0 mY 1 000e 2 oe 1 000e 3 oov EN 1 000e 4 60 0 mv 1 000e 5 90 0 mv 1 000e 6 120 0 mv 150 0 mv mi 1 0000 7 0 130 Ul 0 130 UI 0 390 UI 0 850 UI 0 910 Ul 0 260 Ul 0 000 UI 0 260 UI 0 520 UI 0 780 Ul 1 040 Ul 220 Agilent Serial BERT N4901 2 User s Guide January 2006 Equal BER at BER threshold It shows the contour line at the BER thresho
158. f computational effort and hence time 250 Agilent Serial BERT N4901 2 User s Guide January 2006 Sample Point Offset By default the sampling point for the measurement is positioned 0 5 clock periods or UI ahead of the present analyzer sampling point If the sampling point is set in the middle of the eye this corresponds to the transition point of the incoming signal If you have set the sampling delay manually this is the time offset from that point This option allows you to fine tune the sampling point The unit is UI unit intervals This makes the setting independent of the present clock frequency Set the FFT window for the FFT Calculation The following windows are provided e Uniform e Hanning sometimes also called Hann e Hamming e Blackman The uniform window is no window at all This setting supplies the original error record to the FFT The other windows have the shapes illustrated in the figure below Uniform Hanning Hamming Blackman You can see from the figure that the Blackman window is the strongest filter Agilent Serial BERT N4901 2 User s Guide January 2006 251 The windows are based on the following formulas Uniform w n 1 0 Hannin won 0 5 1 eos 22 Hamming w n 0 54 0 46 cos 2 B lack ney w n 0 42 0 5 cos 22 0 08 cos NOTE Incase of leakage FFT windows improve the spectral reso
159. f the error detector must be set within the data eye The error detector uses the eye diagram to graphically display the location of the sampling point Additionally the appearance of the eye is useful for determining the quality of the data signal The height and width of the eye in the eye diagram are also displayed in the diagram Rise time overshoot and jitter are not displayed in the diagram What is an Eye Edge In the eye diagram the eye has 4 edges An eye edge is defined as a data input delay value or 0 1 threshold voltage value in the data eye that meets the following conditions The instantaneous bit error ratio BER is greater than or equal to the alignment BER threshold e An adjacent point has an instantaneous BER that is less than the alignment BER threshold Agilent Serial BERT N4901 2 User s Guide January 2006 115 Eye Edge Adjacent Point 4E 3 lt BER threshold BER threshold 1E 3 such as 1E 4 Sea 2 K f Vers Lys A AS Uke ay r ie Be oes ES n Aa A a WT ar e A y she ye ee Muir ye edge points oe S sbr 1 Eye Edge Point BER threshold such as 1E 3 1E 2 1E 1 BERs greater than BER threshold BER threshold When Should You Use the Auto Align Function Auto Align can be compared to the auto scale function of an oscilloscope It is the easiest way to set an optimum sampling point With one button press all sampling point settings are automatic
160. fail criteria without rerunning the test The software only uses the criteria to rate the results of a measurement NOTE The pass fail criteria do not control measurement execution The measurement run will be completed even if the measurement fails one or more of the criteria 218 Agilent Serial BERT N4901 2 User s Guide January 2006 You can set several pass fail limits For an explanation of the Eye Opening Pass Fail criteria see Measurement Parameters on page 223 i Parameters Pass Fail View Graph Color m Eye Opening Parameters IV Min Time Eye Opening pu M Min Threshold Eye Opening jv M Min Optimal Sample Delay fou Max Optimal Sample Delay fou IV Min Optimal Sample Threshold ov Max Optimal Sample Threshold jv Cancel Apply Help In the tabular view each of the calculated values will be marked with an icon if it failed the test The following illustration shows an Eye Opening measurement that has failed the criteria for the eye opening voltage 1 Trace 9510 Points Ul 401 878 ps All Errors 600 0 mY 400 0 mv 200 0 mv oov 200 0 mY 400 0 mY 600 0 mv scale 0 448 Ul 0 149 Ul 0 149 Ul 0 448 UI 0 746 Ul Relative 0 597 Ul 0 299 Ul 0 000 Ul 0 299 Ul 0 597 UI Threshold Eye Optimal Sample Time Eye i Opening Delay Opening Terminal Show Color Copied Agilent Serial BERT N4901 2 User s Guide January 2006 219 Vi
161. file 5 The log files will be saved in the Log Directory displayed at the bottom of the dialog box Use the Browse button to select a different directory 6 Click OK to apply your selections and close the Accumulation Setup dialog box For more information on how to run error accumulation tests see Accumulated Measurements Procedures on page 263 140 Agilent Serial BERT N4901 2 User s Guide January 2006 Error Accumulation Reference The Accumulation Setup window contains the following elements Activation Mode Manual Activate this option to configure the error detector to start and stop accumulation when the front panel buttons are pressed e Single Activate this option to configure the error detector to accumulate over one accumulation period and then stop Repeat Activate this option to configure the error detector to accumulate over repeated accumulation periods whereby one period follows another immediately There is no dead time between the end of one period and the start of the next NOTE Repeat will not create multiple measurement log files The data for all repeating periods will be saved to the same log file The accumulation mode cannot be changed while accumulation is running Measurement Log Prompt for File Name Click this option if you want to be prompted for a measurement log file name before each accumulation period begins With this option you can select a new file name for
162. found synchronization fails bad burst 4 If synchronization has been established the received data is analyzed until the signal at Gate In goes high indicating the end of the burst The data should continue arriving at Data In slightly longer than the Gate In signal 5 In case of failed synchronization the total burst counter and bad burst counter are incremented The bit counters total count error count etc are not incremented 6 At the end of the burst the data is analyzed the bits are counted as are the errors total errored Os errored 1s If the BER is higher than the Burst Sync Threshold in the Pattern Sync Setup dialog box the burst is considered a bad burst and the total burst count and bad burst count are incremented 7 For good bursts the total burst count is incremented and the bits are analyzed bits and errors counted The BER is also shown in the user interface 134 Agilent Serial BERT N4901 2 User s Guide January 2006 Timing of the Signal at Gate In The timing of the signal at Gate In is critical to the successful use of burst mode If the signal is not timed optimally the resulting BER will be either too high or the sync ratio good bursts total bursts will be too low The signal at Gate In has to start early enough for the error detector to get the clock the CDR settling time and synchronize to the pattern And it has to stop before the burst ends but should stay low as long as pos
163. fy setup problems ero esac The following list contains possible causes for setup problems sorted from the most severe to the least severe PG CLK LOSS This indicator turns red when no clock signal is detected at the PG Clock In port e External Clock Input is selected and the external clock is off or not connected Your cables or the external clock is faulty ED CLK LOSS This indicator turns red when no clock signal is detected at the ED Clock In port BERT connected looped back or to your device e The PG clock output is off or not connected e There is no connection to the ED clock input e Your cables are faulty BERT connected to your device There is no clock signal from your device e Your device is faulty DATA LOSS This indicator turns red when no data signal is detected at the ED Data In port BERT connected looped back or to your device e The PG data output is off or not connected 2 6 Agilent Serial BERT N4901 2 User s Guide January 2006 e There is no connection to the ED data input e The 0 1 threshold is not in the eye limits of the incoming data signal Use Auto Align or select Avg 0 1 Threshold e Your cables are faulty BERT connected to your device e There is no data signal from your device e Your device is faulty SYNC LOSS This indicator turns red when the measured BER is higher than the sync threshold BERT connected looped back or to your device The sa
164. g is a summary of the values of these parameters for the Serial BERT Bit Count Time Min 1536 bits Max 4 Gbits The optimal length is dependent on the BER threshold CDR Settling Time CDR mode 2 microseconds Non CDR mode 0 Synchronization Time Hardware PRBS 1536 bits Memory based patterns Min 15 kbits depends on when and how often the unique word occurs Valid after Gate CDR mode 1 5 microseconds Non CDR mode 1536 bits Gate Passive CDR mode 2560 bits or 1 5 microseconds whichever is larger Non CDR mode 2560 bits The following values can be derived from these parameters Begin Margin CDR Settling Time Synchronization Time End Margin Valid after Gate Bit Count Time Burst Length Begin Margin End Margin Gate Active Begin Margin Bit Count Time Optimizing the Timing There are three things to watch when optimizing the timing for burst sync mode e BER The bit error rate increases when the gate closes too late The duration of the signal should typically be reduced e Burst sync ratio The burst sync ratio is an indication of how much of the burst signal is in synchronization and can therefore be used for counting bit errors The higher the value the better Bad burst count The bad burst count counts the number of bursts that are invalid for example because synchronization failed or the BER of a synchronized burst exeeds the sync threshold 136 Agilent Serial BERT N4901 2 User
165. gilent Serial BERT N4901 2 User s Guide January 2006 49 Range of Hardware Generated PRBS Patterns PRBS Lengths The following table lists the hardware generated 2 n 1 PRBS pattern lengths that correspond to different n values This is before the pattern is finally inverted 7 127 7 6 10 1 023 10 9 11 2 047 11 10 15 32 767 15 14 23 8 388 607 23 22 31 2 147 483 647 31 30 Decimation of PRBS Patterns A special property of 2 n 1 PRBS patterns is that they can be demultiplexed into the same patterns at slower speeds with different phases This is also true for multiplexing The demux or mux must have n ports Hardware Generated PRBS Procedures The recommended way to load the data patterns to the pattern generator and error detector is to use the Pattern Select Form dialog box 9 See Loading Patterns to the Pattern Generator and Error Detector on page 38 for details Hardware Generated PRBS Reference Use the Pattern Select dialog box to select the pattern s you want to write to the pattern generator and the error detector See Select Pattern Dialog Box on page 44 for details 50 Agilent Serial BERT N4901 2 User s Guide January 2006 User Defined Sequences Setting up Patterns User Defined Sequences This section describes the basics of user defined sequences A sequence is created and maintained by means of the Sequence Editor that can be enabled from the Pattern window A se
166. he results The following illustration shows the bathtub curve of the measured bit error rate and in the tabular view the calculated results for the Optimal Sampling Delay the Skew and the Phase Margin 4 00e 9 1 Trace 39 Points Ul 401 878 ps All Errors 1 00e 1 1 00e 2 1 00e 3 1 00e 4 1 00e 5 1 00e 6 1 00e 7 Logarithmic 0 62 UI 0 37 UI 012 Ul o412UI 0 37 UI 0 62 Ul Relative 0 50 Ul 0 25 Ul 0 00 Ul 0 25 Ul 0 50 Ul Total Jitter RMS Show Color Copied F ate eee Terminal Delay Margin Agilent Serial BERT N4901 2 User s Guide January 2006 171 How to Optimize the View of the Results After you have run a measurement the resulting graph and the calculated numerical values are displayed To improve the results you can change the measurement parameters 1 Press the Properties button to open the Properties dialog box 2 Use the different tabs in this dialog box to make the required settings Parameters tab These settings are used for data collection Changes here require the test to be run again See Parameters Tab on page 173 for details Pass Fail tab These settings determine whether the calculated results are recognized as passed or failed However a new test run is not required when doing changes here See Pass Fail Tab on page 174 for details View tab Graph tab and Color tab All settings on these tabs only affect the way the data is di
167. he measurement moves to the next sample point when the first of the two criteria is reached Set the criteria for the sample delay Resolution Specifies the distance between sampling points The lower this value the more sampling points you have in a unit interval You can enter the resolution in UI or ps ns s The timebase of the display is set on the View tab The default is 0 01 UI that means 100 points per unit interval will be measured e Edge Resolution Optimization Turns the resolution optimization on or off If this option is enabled the Serial BERT intelligently sets a resolution so that there are more sampling points at the edges of the eye This can greatly improve the results without dramatically increasing the duration of the test Set the criteria for the sample threshold Resolution Specifies how many measurement points are taken within the sample voltage range Enter the value in mV Low Level Specifies the lower limit of the measurement voltage range Enter the value in mV This value should be slightly lower than the lowest expected signal voltage e High Level Specifies the upper limit of the measurement voltage range Enter the value in mV This value should be slightly higher than the highest expected signal voltage Pass Fail Tab The Pass Fail tab of the Properties dialog allows you to specify the criteria to decide whether the DUT passes or fails the test You can change pass
168. ht bathtub Total Jitter BER Threshold intersections Agilent Serial BERT N4901 2 User s Guide January 2006 183 Advanced Analysis tpt Levels Output Levels The Output Levels measurement allows you to characterize the behavior of the output levels of a device under test DUT The sampling delay is fixed The error detector s decision threshold is automatically swept within a user defined range A direct result is the determination of the optimum decision threshold level for receiving data from the DUT with maximum confidence Optimum sampling delay Three Available Views The Output Levels measurement provides three different graphical views to visualize the calculated results BER versus Threshold This graph shows the relationship between the analyzer decision threshold and the resulting BER It presents the raw data dB Histogram versus Threshold This graph shows the relationship between the analyzer decision threshold and the derivative of the bit error rate ABER dTh A Gaussian marker allows you investigate the peaks of this graph Q from BER versus Threshold This graph shows the extrapolation of the Q factor and the optimum threshold level from a limited number of measured points 184 Agilent Serial BERT N4901 2 User s Guide January 2006 Example Results The following illustration shows the BER versus Threshold graph of a simple Output Levels measurement 1 Trace 86 Points 600 0 mv 400
169. ick this icon to save the current sequence in its original file Save As Click this icon to save the current sequence in a new file A Save As dialog box opens for you to specify name and path for the file See Saving the Sequence in a File on page 55 for details IA To PG Click this icon to download the sequence from the Sequence Editor to E the pattern generator 10110 a From PG Click this icon to upload the present sequence from the pattern gen PG erator to the Sequence Editor This allows you to inspect and edit a sequence that has been loaded with a program Proper Click this icon to change the properties of the current sequence See ties Setting Sequence Properties on page 61 for details Break Click this icon to terminate an infinite loop that is set to manual 5 break condition Sequence execution continues with the next block j Resume Click this icon to interrupt and re initialize a running sequence If the sequence start condition is IMMediate the sequence restarts immediately If the sequence start condition is a signal at the Aux In port the se quence restarts upon the specified signal If the sequence start condition is Command sequence execution starts when the Start button is clicked 60 Agilent Serial BERT N4901 2 User s Guide January 2006 Sequence Start Condition Number of Blocks Description Sequence Expression NOTE Setting Sequence Properties
170. ields a BER resolution of 1078 3 Decrease the Resolution the threshold step size to 2 mV This gives us 500 steps per Volt 4 Press OK to confirm your changes 5 Press the Start button to repeat the measurement with the new parameters The measurement now takes more time than the previous but it is also much more precise 800 0 my Trace 1002 Points 900 0 mv A0V AAV 12V 13y 14y ASV 16V ATV BER Threshold 1 000e 6 18v Logarithmic 1 000e 7 1 000e 5 1 000e 3 1 000e 1 1 000e 8 1 000e 6 1 000e 4 1 000e 2 1 000e When you move the BER Threshold bar you will see that some of the calculated values change This is explained in Explanation of the Numerical Results on page 198 How to Use the Different Views To switch between the different views of the Output Levels measurement 1 Press Properties and switch to the View tab 2 Select dB Histogram versus Threshold This gives you another view of the measured data 3 Switch to the Graph tab and select Linear Scale The data remains the same but a linear scale makes it easier to see the distribution Agilent Serial BERT N4901 2 User s Guide January 2006 1 Trace 500 Points 800 0 m 900 0 mY 1 0 AAV 1 2 V 1 3 Y AAV ASV 1 6 Y 1417Y 1 8 Y Linear 2 000e 0 6 000e 0 1 000e 1 1 400e 1 1 800e 1 2 200e 1 2 600e 1 3 000e 1 a le A le le d le J i A E i 0 000e 0 4 000e 0 8 000e 0 1 200e 1 1 600e 1 2 000e 1
171. ight and width of the incoming data signal at specific alignment BER thresholds This information is displayed on a representation of an eye diagram Agilent Serial BERT N4901 2 User s Guide January 2006 257 Instantaneous Measurements With instantaneous measurements you can directly view the current BER on the screen In order to stress your device under test you can then for example add errors to the data stream or switch between different patterns and view the resulting BER Or you can adjust the output port parameters or the error detector s sampling point while the measurement is running and watch the resulting BER immediately Such measurements can be used to find out the benchmark data for example for a device prototype Test Methods You can analyze the behavior of your device by employing several test methods If the changes in the test setup affect the BER you can immediately notice it on the screen The resulting changes in the BER let you learn about the DUT s behavior and limitations Some examples of modifications that you can make are listed below e Add errors to the data stream You can for example perform a confidence test of your measurement setup The analyzer should recognize all errors that you add If it does not you may have a problem with your measurement setup See Error Addition Insertion on page 88 for details e Use alternating patterns to switch between different data streams For exa
172. igit patterns that contain long runs of 1 s and 0 s Alternatively use the Serial BERT or a DCA with an ITU T specified pattern for testing STM n systems This pattern is to verify the adequacy of timing recovery and low frequency performance of 22 Agilent Serial BERT N4901 2 User s Guide January 2006 STM n systems It consists of a user pattern that is comprised as follows The first row of overhead bytes for the STM n system all ones zero timing content high average signal amplitude pseudo random data with 50 mark density ratio This is followed by a repeat of the pattern overhead all zeros zero timing content low average signal amplitude and the PRBS For full details see ITU T Recommendation G 958 e For a regenerator test set up the bit rate according to the device specification TIP If your device converts the input pattern the error detector needs a different pattern than the pattern generator See Loading Patterns to the Pattern Generator and Error Detector on page 38 for details Agilent Serial BERT N4901 2 User s Guide January 2006 23 Setting up Patterns User Patterns User Patterns User patterns are user editable patterns that are written to the Serial BERT s memory for pattern generation see How does the Serial BERT Generate Patterns on page 19 for details User patterns can contain up to 32 Mbit of any arbitrary data pattern They may be generated by the Serial BERT software or
173. in addition to the ones listed in General Requirements on page 148 e Only memory based patterns with a unique 48 bit pattern detect word are allowed e No alignment features can run during error location capture Auto Align 0 1 Threshold Center Data Center Agilent Serial BERT N4901 2 User s Guide January 2006 225 e No other advanced measurement Output Timing Output Levels etc can be running e Error Location Capture can only run when the BER Location Mode is set to more than one bit for example all bits or a block with a length gt 1 Error Location Capture Procedures This section shows how to set up and perform an Error Location Capture measurement As an example first add a a couple of errors to an alternating pattern and then capture their position This requires the following steps e Preparing the measurement see How to Prepare the Error Location Capture Measurement on page 226 e Demonstrating the measurement see How to Demonstrate the Error Location Capture Measurement on page 227 How to Prepare the Error Location Capture Measurement To prepare an Error Location Capture measurement to detect the location of an errored bit in an alternating pattern 1 Disable the pattern generator outputs by pressing the OV Disable button in the PG Output Setup screen 2 Usea shielded cable to connect the pattern generator s Data Out port and the error detector s Data In port 3 Use
174. in the Alternate Pattern Control dialog box to enable the Insert B button on the main display Insert B When you press the Insert B button pattern B is inserted into the continuous output of pattern A Insert B pressed A0 AQ BO A0 Data Out Aj ajalalalalalalBIBIBIB AlAla A Aux In In this section you can enable the Aux In port to control the output pattern Aux In can receive signals that control alternate patterns or blank the output The signals received at this port must be TTL compatible The granularity for signals at this port is 512 bits The following options are available for Aux In Disable With this option selected the signal at the Aux In port will be ignored Level Sensitive If Aux In is logic high pattern B is sent If Aux In is logic low pattern A is sent Aux In DaaOut A A B A BBB Agilent Serial BERT N4901 2 User s Guide January 2006 Edge Sensitive Pattern A is sent out until a rising edge at Aux In is detected Then at the next complete repetition of pattern A a single occurrence of pattern B is inserted into the output Aux In Data Out e Output Blanking Pattern A is sent when the signal at Aux In is low Pattern B is not sent in this mode Aux In Data Out i pi ma A A A If Aux In is pu
175. inations Could Damage Your Device on page 70 5 Connect the DUT s input ports to the Serial BERT s output ports Agilent Serial BERT N4901 2 User s Guide January 2006 71 ee CAUTION Do not apply external voltages to the pattern generator outputs For more information see AC Coupling and Bias Tees on page 70 Output ports of the pattern generator that are not connected to another device must be terminated with 50 Ohm to prevent the Serial BERT from damage Note that for unconnected ports an internal protection algorithm automatically reduces the output voltages to safe levels Adjust Output Levels optional Data and clock offset and voltage levels can be adjusted This is typically done when you want to tune your BER measurement or stress the device You can use the knobs on the Serial BERT s front panel to fine tune the data and clock amplitudes and offsets If you want to set a specific value you can use the numeric keyboard To enter specific values for the data and clock outputs from the keyboard 1 Press PG Output Setup in the PG Setwp menu to access the PG Output Setup window 2 Click into the field you want to change 3 Enter the desired voltage using the numeric keyboard or the respective front panel knob NOTE Ifthe termination voltage is wrongly adjusted or the output ports are not connected an internal protection algorithm automatically sets the output voltage levels to safe levels Connect th
176. ing parameters and the jitter measurement parameters For information on the DUT Output Pass Fail criteria see also Output Timing Measurement Parameters on page 180 e For information on the Jitter Pass Fail criteria see also Jitter Measurement Parameters on page 180 Properties x Parameters Pass Fail View Graph Color m DUT Output Parameters 7 DUT Output Pass Fail Minimum Maximum I Phase Margin FE Optimal Semple Delay pu M DUT Output Parameters Jitter Pass Fail Minimum Maximum I Total Jitter RMS bu M Total Jitter Peak to Peak 0 2 U1 IV Fast Total Jitter Uncertainty foosur J Jitter Mean pu t S pu I Random Jitter RMS J Deterministic Jitter pu J Estimated Total Jitter pu NOTE Ifthe results of a Fast Total Jitter measurement are displayed only the appropriate parameters are compared and flagged e Phase Margin e Optimal Sample Delay e Total Jitter Peak to Peak e Fast Total Jitter Uncertainty Other pass fail limits may be enabled but are ignored In the tabular view each of the calculated values will be marked with an icon if it failed the test The following illustration shows a DUT Output Timing Jitter measurement that has failed the criterion for the phase margin Agilent Serial BERT N4901 2 User s Guide January 2006 175 176 Bit Error Rate Graph Jitter Histogram 4 00e 9 1_Trace 39 Points Ul 401 878 ps All Errors 1 00e 1 1 0
177. ints With this option selected all measured points are indicated in the graph with small squares 800 0 my Trace 202 Points 900 0 mv A0V AAV A2V 1413Y 1414Y 4145Y 16v ATV 1418Y BER Threshold 1 009e 6 Logarithmic 1 000e 5 1 000e 3 1 000e 1 1 000e 6 1000 4 1 000e 2 1 000e The headline of the graph tells the number of data points included in the display 101 This corresponds to the chosen threshold resolution of 100 steps per Volt With this number of measured points you do not get very precise results during the measurement but it is quickly finished Depending on your personal needs you can trade measurement precision against test time Several more options are provided on the Graph tab for you to adjust the display as you want For instance you can use markers for analyzing the graph or open a zoom window for viewing the details To find more information refer to Graph Tab on page 197 Agilent Serial BERT N4901 2 User s Guide January 2006 189 190 How to Change the Output Levels Properties In the example measurement we have set the focus on speed 100 threshold levels and 1 000 000 compared bits per measuring point You may wish to obtain more precise results 1 Press Properties and switch to the Parameters tab 2 Increase the Number of Compared Bits to 100 000 000 Remember One failure per 1 million bits yields a BER resolution of 107 One failure per 100 million bits y
178. ion of the optimum sampling point Agilent Serial BERT N4901 2 User s Guide January 2006 223 These parameters are derived as illustrated in the figure below Eye opening voltage Optimum sample point 224 Agilent Serial BERT N4901 2 User s Guide January 2006 Error Location Capture The Error Location Capture measurement allows to capture the position of an errored bit in a memory pased pattern The instrument saves the position of the errored bit and writes a bit sequence neighbouring the errored bit to a file This feature can be used to find rare or random errors A DUT could have problems handling long series of zeroes Error Location Capture can be used to locate the bit errors in such cases NOTE The measurement run is aborted by various actions like selecting a new pattern or starting synchronization or alignment Example Results The following figure shows the results of a typical Error Location Capture measurement 1110 1011 bali olabalie a e ablejape fel sbi at 40 1101 0011 0001 0111 0010 60 abaltala sbiayalie sbjalieha jojaliajal lel st alia 101 1000 0100 1011 1011 100 sa ee e ablojatio ajahabio a a eene 0011 1110 1011 0000 0101 ababjai a ahah shfa tojab shat aiaee Poelee 1101 0011 0001 0111 0010 1010 1000 1st Bit Error Location 80 Bit Error Count 128 Important Requirements In order to perform Error Location Capture measurements the following requirements have to be met
179. ios Graph This graph displays the delta errored 1 s ratio delta errored 0 s ratio and total delta error ratio at data points over the entire accumulation period The error ratios on the y axis are set to a range of 1E 0 100 errors to 1E 12 The accumulation period is on the x axis During accumulation data will appear to move from left to right on the ratios graph When the graph is completely filling the display the x axis time scale will double The data graph is then occupying only half of the display and will continue to move to the right again This will repeat until the accumulation period has ended Agilent Serial BERT N4901 2 User s Guide January 2006 265 266 Zooming NOTE NOTE Sizing Handles This graph can accumulate a large amount of data You can click any point on this graph to zoom into an area of BER data You can continue to click on the graph until you have zoomed to the maximum The point you click will appear in the middle of graph You can press the back button left arrow to zoom out again The zooming function cannot be used before the accumulation period has ended To view precise BER data for every point in time view the measurement log file While zoomed in you can move data left or right by dragging the graph or by using the front panel knob You can also zoom vertically by dragging the top and bottom sizing handles The current view can be m
180. is displayed here Error Free Milliseconds The number of milliseconds in which no errors were measured is displayed here e Errored Seconds The number of seconds in which errors were measured is displayed here e Errored Deciseconds The number of deciseconds in which errors were measured is displayed here 268 Agilent Serial BERT N4901 2 User s Guide January 2006 e Errored Centiseconds The number of centiseconds in which errors were measured is displayed here e Errored Milliseconds The number of milliseconds in which errors were measured is displayed here e Sync Loss Seconds The number of seconds in which the error ratio was greater than the sync threshold is displayed here Accumulation Parameters Elapsed Time The time elapsed during the accumulation period is displayed here Percent Complete For single and repeat accumulation modes the percentage of the accumulation period completed is displayed here to one decimal place For repeat mode this percentage is recalculated each time the accumulation period repeats Accumulation Mode The type of accumulation mode is displayed here The mode can be Manual Single or Repeat Accumulation Period For single and repeat accumulation modes the accumulation period is displayed here The period will be displayed in elapsed seconds error count or bit count depending on the selected method Burst Results The recommended application for burst mod
181. it rates below appr 620 Mbit s only the Error Location Capture measurement works Please refer to Bit Rate Range on page 108 for a detailed explanation of the limitations at low bit rates If the Serial BERT s pattern generator is used to send a data pattern to the DUT then also the following to conditions must be met e An appropriate pattern is selected The pattern generator is correctly set up For examples refer to the descriptions of the different measurements 148 Agilent Serial BERT N4901 2 User s Guide January 2006 Optimizing Parameters If the presented measurement results do not answer your questions adequately you can easily modify the setup parameters in the Properties dialog box When changing the measurement settings after the measurement has been run please note e Parameters that affect the data capture Changes on the Parameters page take only effect if you run the measurement again e Parameters that change the display of the measured data Changes on the Pass Fail View Graph and Color tabs only affect the display of the results There is no need to repeat the measurement Saving Measurements for Later Analysis You can save a completed measurement for later analysis This includes the settings you selected for the measurement and the measured data This allows you to load the measurement at a later stage and inspect all data in detail You can also change display parameters to evaluate this da
182. ition of errored bits in a memory based pattern The instrument searches for the first bit errored in the incoming bitstream and marks it in the pattern The address of the errored bits can be displayed after the error is located Agilent Serial BERT N4901 2 User s Guide January 2006 147 Fast Eye Mask The Fast Eye Mask measurement is first of all meant for production and screening tests It allows to determine very quickly whether the eye opening seen at the output signal of a device is within specifications that means within certain timing and voltage limits This is achieved by measuring the bit error rate at a limited number of test points Spectral Jitter The Spectral Jitter measurement allows you to analyze the jitter inherent in the output signals of your device under test DUT as a function of the frequency This measurement can be used for investigating the behavior of the DUT for example to identify crosstalk effects Before you can run any of these advanced measurements you have to properly set the required parameters After the test is complete you can view the results General Requirements For any measurement to yield reasonable results the following requirements apply The error detector must be correctly set up and synchronized with the pattern sent from the DUT output port Ideally the sampling point should be aligned to the optimum sampling point for example with the Auto Align function e For low b
183. kers To analyze the graphs at a particular point you can use the markers Optionally you can display all related values for the markers in the marker readout Pressing the Reset Markers button will set the markers back to the default positions Linear markers are available for the BER vs Threshold graph and the QBER vs Threshold graph A gaussian marker is available for the dBER vs Threshold graph See Marker Definitions on page 156 for details Zoom Several zoom factors are available When you show the zoom graph you can also allow the zoom graph to track the mouse or your finger if you are working directly on the Serial BERT Show Measured Points If you want to see the points that have actually been measured choose Show Measured Points NOTE Ifyou have a mouse connected to your Serial BERT you can access many parameters and display options conveniently from the context menu This looks as shown below PETC BER vs Threshold Display Options dBER vs Threshold gt QBER vs Threshold Log Scale Linear Scale Linear Ratio dB Ratio Properties v Show Measured Paints Explanation of the Numerical Results The result parameters are divided into three groups Level Results The level result parameters are summarized in the following table High Level min max Low Level min max Mean Level min max Amplitude min max Threshold Margin min For all parameters pass fail limits can be set
184. l 24 085 m wid If you compare the Low Level result with the limit we have set on the Pass Fail tab you will find that the measured result fails the upper pass fail limit for this parameter View Tab The graph shows either the BER vs Threshold the dBER vs Threshold or the QBER vs Threshold This graph shows the relationship between the analyzer decision threshold and the measured BER 800 0 my 1 Trace 501 Points 900 0 mv 1410Y AAV 12V 4143Y 1414Y A5V ABV ATV BER Threshold 1 007e 6 18V Logarithmic 1 000e 7 1 000e 5 1 000e 3 1 000e 1 1 000e 6 1 000e 4 1 000e 2 1 000e The BER considers all errors It is calculated as CC Errorls gt Error0s BER AllErrors total of Bits Agilent Serial BERT N4901 2 User s Guide January 2006 195 dBER vs Threshold Graph This graph shows the relationship between the analyzer decision threshold and the absolute values of the derivative of the bit error rate dBER dTh It is recommended to view this graph with a linear scale because a linear scale reveals the distribution more clearly 1 Trace 500 Points 800 0 mv 900 0 mv 10V AAV 12V 13v 14v ASV ABV 17y 18V Linear 000e 0 1 01 1 800e 1 Fi 2 200e 1 2 600e 1 3 01 2 000e 0 6 000e 00e 1 1 400e 1 00e 000e 1 0 000e 0 4 000e 0 8 000e 0 1 200e 1 1 600e 1 2 000e 2 400e 1 2 800e 1 3 200e 1 This distribution can often be approximated
185. l positions except when the pattern is exactly aligned Hardware generated PRBSs are non inverted PRBS 2 n 1 patterns that allow high speed measurement of random traffic They enable stressful testing of devices because large patterns contain long strings of 0 s and 1 s Software generated PRBSs are non inverted PRBS 2 n patterns They are also called PRBN patterns These patterns are useful for applications where the pattern length must be an even number The Serial BERT provides variations of software generated PRBS patterns allowing you to set a pattern s mark density ratio of 1 s and zero substitution These features make it easier for you to run specific stress tests on your DUT The following figure illustrates the PRBS based patterns available with the Serial BERT PRBS 2 n PRBS 2 n 1 PRBS Software Generated Hardware Generated Pure 2 n PRBS Mark 2 n PRBS 24n PRBS Density Zero Substitution Agilent Serial BERT N4901 2 User s Guide January 2006 21 PRBS Frequency Spectrum Digital non return to zero NRZ data has a sin x x 2 characteristic frequency spectrum The PRBS frequency spectrum follows this envelope with line spectra of spacing related to the sequence repetition rate The longer the sequence before it repeats the narrower the line spacing closer to a continuous spectrum Line spacing bit rate sequence length The closer
186. l us whether the BER is above or below the given threshold These limits have been calculated from the error probability density functions applicable to BER measurements The equations were solved for a level of confidence of 95 The calculations have led to the following table Min number Max numberof Maxnumber Min number of of errors compared bits of errors compared bits x 10 2 x 1012 1 0 05129 0 2 996 2 0 3554 1 4 744 3 0 8117 2 6 296 4 1 366 3 7 754 5 1 970 4 9 154 6 2 613 5 10 51 7 3 285 6 11 84 The BER threshold of 107 and hence the setting of the multiplication 012 factor to 1 is just an example by changing the exponent the table applies analog to other thresholds The following figure shows a plot of this table Agilent Serial BERT N4901 2 User s Guide January 2006 167 The Bracketing Approach D a N BER lt 1e 12 o Number of transmitted bits 10 ji LAL Number of errors Note that there is a gap where the BER is so close to 10 that we cannot decide For example if we compared 3 x 10 bits and got two errors a measured BER of 0 667 x 10712 we are in the uncertain white area on the graph In such a case we need to transmit more bits until the number of bits either reaches the upper limit 6 296 x 1013 or until we see more errors If the actual BER is very close to 107 however we are unable to apply a lower o
187. ld 150 0 my 4 Traces 53767 Points Ul 1 538 ns All Errors 4 000e 0 120 0 m 1 000e 1 90 0 mY 60 0 mv 1 000e 2 30 0 mY 1 000e 3 00y 30 0 mv 1 000e 4 60 0 mY 1 000e 5 90 0 m 1 000e 6 120 0 mY 150 0 mY 1 000e 7 0 910 UI 0 650 UI 0 390 UI 0 130 UI 0 130 UI 0 390 UI 0 650 UI 0 910 UI 0 780 UI 0 520 UI 0 260 UI 0 000 UI 0 260 UI 0 520 UI 0 780 UI 1 040 UI Furthermore the following parameters can be set on the View tab of the Properties dialog box Analyze You can analyze for e All Errors To display all errors e Errors if Os Expected To display the errors if 0 is expected but 1 received e Errors if 1s Expected To display the errors if 1 is expected but 0 received Calculate You can calculate measurement parameters for e 0 Errors To calculate the parameters for 0 errors BER Threshold To calculate the parameters for the given BER threshold The BER threshold influences all parameters of the Eye Opening measurement You can also drag and drop the horizontal BER threshold in the graphical display to change this value Table Number Format You can select the number of Decimal Places to be displayed in the table Agilent Serial BERT N4901 2 User s Guide January 2006 221 Graph Tab On the Graph tab you can use the several options to optimize the graphical display according to your needs Timing Units Choose between Unit Interval and Seconds to select the timebase for the displ
188. le the data view between the display of the expected data i e the data you defined on the error detector and the captured data i e the data that was captured at the error detector The current state is displayed in the title bar of the pattern grid e Run Select Toggle the data view between the current run and the previous run After a measurement run is successfully finished the display is automatically switched to the current run The current state is displayed in the title bar of the pattern grid e First Error Jumps to the first errored bit in the pattern e Prev Error Jumps to the previous errored bit in the pattern e Next Error Jumps to the next errored bit in the pattern Last Error Jumps to the last errored bit in the pattern Explanation of the Results The results are shown in various ways Graphic View The graphic view highlights the located bit errors The currently selected bit is highlighted in red all other errored bits are written in red Status Bar The status bar below the graphic view provides the following information Status The status indicates if the measurement is running or has ended If the measurement has ended it indicates if the measurement ended successfully or was aborted 228 Agilent Serial BERT N4901 2 User s Guide January 2006 Compare Pattern File Results Window Position The position of the current errored bit is shown Hex Bin This field indicates how the data is shown bi
189. lent Serial BERT N4901 2 User s Guide January 2006 297 Removing a Printer When you disconnect a printer from the instrument and no longer intend to use it you can remove the printer driver and increase the free disk space available on the instrument s internal hard disk drive To remove a printer driver 1 On the File menu click Print Setup then Configure Printer This opens the Printers dialog box 2 Click the printer driver you want to remove 3 In the File menu in the Printers dialog box select Delete A dialog box will appear to confirm the removal of the printer driver 4 Click Yes to confirm your selection Setting the Default Printer You can set the printer you use most often as the default printer Then all printouts will be sent to the default printer unless you explicitly select another printer To specify the default printer 1 On the File menu click Print Setup then Configure Printer This opens the Printers dialog box 2 Choose one of the currently mapped printers If the printer you want to use is not yet included in this list see Adding a Printer on page 297 3 On the File menu in the Printers dialog box click Set As Default 4 Click Close The default printer selection is applied and the dialog box closes NOTE The printer capabilities will vary depending on the printer you have installed and configured Printing a Hardcopy NOTE Before you can print you must first have the app
190. licable gt indicates that the value could be calculated however is not shown because quality criteria are not met This is the case for example for RJ DJ results when the r 2 value of one or both edges is lt 0 75 Even though RJ DJ values can be calculated in this case they are not shown because confidence in the results is too low Understanding the 0 Factor Results The variable decision threshold method used by the Output Levels measurement makes it possible to determine the Q factor of a signal NOTE The Q factor method and the related calculations have been published under Margin Measurements in Optical Amplifier Systems by Neal S Bergano F W Kerfoot and C R Davidson in IEEE Photonics Technology Letters Vol 5 No 3 March 1993 Mathematical Background Bit errors are caused by noise and the Q factor describes the signal to noise ratio at the decision circuit It is possible to calculate the Q factor from a limited number of measured BER vs threshold data points It is also possible to calculate expected bit error rates from the Q factor This is a method for predicting very low bit error rates typically below 107 that would take a long time to measure The Q factor is calculated as _ 4 Ho Q 0 0 where u 1 0 is the mean level of the 1 and 0 rails respectively and 910 is the standard deviation of the noise distribution on the 1 and 0 rails 204 Agilent Serial BERT N4901 2 User s Guide January 200
191. list of recommended patterns for your device and application see When to Use Which Pattern on page 22 e For a list of recommended connection diagrams for your device and application see Connecting the DUT on page 13 What are the Testing Requirements of Your Device Considering the requirements of your device will help you choose which instrument functions to use during the test setup and when making measurements e Do you need to use a custom test pattern If you have a long series of 1 s or 0 s in your custom pattern be sure to avoid false sync See What is False Synchronization on page 130 e Do you want to use alternating patterns Then you must first import or create a custom pattern that contains an A half pattern A and B half pattern B See Pattern Alternation on page 92 e Are you aware of how custom pattern size can affect the pattern output See Why the Serial BERT Repeats Memory Based Patterns on page 28 e Do you need more than two patterns or a mix of custom patterns and PRBS You can create a sequence A sequence can contain up to four blocks that can be looped Each block can hold a custom or standard pattern e Does your device change the input pattern it receives If your device inverts logic then select Data Inverted in the Sampling Point Setup window If your device changes patterns in other ways then you must know the expected output pattern and select it as
192. lities menu click Set Time and Date This opens the Date Time Properties dialog box 2 Set the date and time 3 Click OK Setting the GPIB Address Your instrument is set to a default GPIB General Purpose Interface Bus address You can change this address by doing the following 1 On the Utilities menu click Change GPIB Address This starts the GPIB Address Change dialog box 2 Select an address in the New GPIB Address list 3 Click OK The new address will be applied and the dialog box will close Each instrument must be set to a unique GPIB address to avoid multiple instruments transferring data at the same time NOTE The default address is 14 however addresses from 0 to 31 may be used if the default address is the same as another instrument s GPIB address Address 21 is usually reserved for the computer interface Talk Listen address and should not be used as the instrument address Connecting Peripherals You can connect an external monitor mouse and keyboard to the instrument on the Serial BERT s rear panel To connect a mouse or a keyboard you can use either the USB ports or the PS 2 ports 290 Agilent Serial BERT N4901 2 User s Guide January 2006 For more details refer to Using an External Monitor on page 294 and Using the On Screen Keyboard on page 294 NOTE Ifthe touchscreen is disabled the external keyboard and mouse are also disabled See Turning the Touchscreen Off On on page 2
193. lled to logic high and a standard pattern is selected no output pattern will be sent NOTE Due to the pattern generator behavior the patterns may be repeated up to 512 times before alternating See How the Serial BERT Generates Memory Based Patterns on page 27 for details Agilent Serial BERT N4901 2 User s Guide January 2006 97 98 Agilent Serial BERT N4901 2 User s Guide January 2006 Setting up the Error Detector The error detector analyzes an incoming bit stream compares it to the expected pattern and locates any inconsistencies The error detector requires the following settings to work correctly The expected pattern The error detector needs to know which data to expect so that it can detect bit errors e Correct input and output terminations to the DUT This ensures that your DUT cannot be damaged due to incorrect terminations e Correct clock frequency Required to recognize the bit rate in the data stream e Appropriate sampling point The sampling point defines where the error detector tries to differentiate between Os and 1s in the data stream This is necessary so that the error detector recognizes the data bits correctly e Synchronization to the incoming pattern The expected pattern must be synchronized to the incoming pattern so that the error detector can find any discrepancies The error detector provides the following functions to enable you to perform tests e Automatic pattern synchronization
194. llowing functions are available to insert errors in the bit stream Error Add Button The Error Add button in the top right corner is used to insert a single bit error in the data stream Error Add j This button is always active even if other error insertion mechanisms are running The maximum number of errors that this function supports is one error per 128 bit block Error Add Setup Dialog Box The Error Add Setup dialog box provides the following configuration options External Error Add With this option selected when a rising edge is detected at the Error Add port an error will be generated in the output stream by flipping a single bit within a bit block of 128 bits The signals received at this port must be TTL compatible 90 Agilent Serial BERT N4901 2 User s Guide January 2006 Internal Select this option to use internal error insertion functions e Error Rate Allows to select from a preset bit error ratio from 1 000E 02 to 1 000E 09 Errors will be added to the output pattern to produce this error ratio Bits Between Errors Allows you to enter the number of valid bits between two errors This is the inverse of the bit error ratio Off Select this option turn off all internal add rates and external error inputs This allows you to manually add single errors by pressing Error Add on the main display This flips a single bit in the output stream Agilent Serial BERT N4901 2 User s Guide January
195. llowing warning message respectively NOTE The temperature level is not yet critical However the Serial BERT may be damaged if operation is continued The message is intended to inform you to take the appropriate measures Save your current instrument settings shut down the Serial BERT and let it cool down before operating it again Critical Temperature If the Serial BERT is operated at this state for a longer time the temperature may reach a critical level In this case the following message is displayed Temperature Alert ED critical above normal Shutting down instrument Immediate startup not recommended Temperature Alert PG critical above normal Shutting down instrument Immediate startup not recommended The Serial BERT will be shut down your instrument settings will not be saved 284 Agilent Serial BERT N4901 2 User s Guide January 2006 Troubleshooting Overheating CAUTION If the Serial BERT has indicated overheating do the following e Shut down the instrument and let it cool 45 minutes to 1 hour e Reduce the environmental temperature e Make sure that the fans are running and the ventilation holes are not blocked If the problem continuously recurs contact Agilent support Do not operate the Serial BERT before it has cooled down again Otherwise your instrument may get damaged When you are switching on the Serial BERT after it was shut down due to a temperature error you will see a m
196. lock or according to the data pattern being generated The Aux Out can be set up to either send the error detector s clock signal directly from Clk In port or the error detector s clock signal recovered from Data In port or the input data passed over a comparator Trigger and Aux Output Procedures To set up the trigger and auxiliary outputs 1 Press Trigger and Aux Out in the ED Setup menu 2 For the Trigger select one of the two options Clock Divided by n Sends a trigger signal every 4 8 16 20 32 64 or 128 bits Pattern A trigger is sent every time the complete pattern was received 3 For the Aux Out select whether the Clock signal or the Data signal is supplied If Clock is selected the clock signal used by the error detector is output directly at Aux Out If Data is selected the input data is output over a comparator at Aux Out The comparator is controlled by the 0 1 threshold This lets you use an oscilloscope to determine if the 0 1 threshold is correctly set If the 0 1 threshold is set below or above the data eye the output at Aux Out will be constant high or low respectively 4 Click OK to finish the Trigger and Aux Out setup Agilent Serial BERT N4901 2 User s Guide January 2006 125 Trigger and Aux Output Reference The available options in the ED Trigger amp Aux Output window are Trigger You have two options to configure the signal at the Trigger Out port e Clock
197. lock In port Clock Data Recovery If this mode is selected the Serial BERT derives the exact bit rate from the incoming data stream To make this function work the approximate clock rate must be specified in the Clock Rate field with an accuracy of 0 01 Clock Rate This field is used to specify the approximate clock rate for CDR mode Serial BERT needs this value to be about 0 01 accurate to find the actual bit rate in the data stream Use the numeric keypad to enter the value followed by the unit that you can select with the G n M u k m and X1 keys on the front panel Alternatively you can select a value from the Preset list CDR Range Use this field to select the range of clock rate values that is displayed in the Preset list Preset List The Preset list contains clock rate values that are commonly used by the error detector It always displays the values that are in the range selected with CDR Range Click a preset to select it for the error detector clock rate Double click a preset to modify it Add Preset Use the Add Preset button to add a new preset to the Preset list You can then enter a value a unit and a short description for this new preset Delete Preset Use the Delete Preset button to delete the currently selected preset from the list 112 Agilent Serial BERT N4901 2 User s Guide January 2006 Auto Threshold If the Auto Threshold option is selected the Serial BERT continuously
198. lue returned to previous Status of Avg 0 1 Threshold checkbox returned to previous Data Center Canceled Data Delay value returned to previous 122 Agilent Serial BERT N4901 2 User s Guide January 2006 NOTE BER Threshold In this list select an alignment BER threshold that is appropriate for your application The alignment BER threshold is the pre defined threshold used by the Data Center 0 1 Threshold Center and Auto Align functions to define the edges of the data input eye in the time and voltage axes You may wish to change the threshold for the following reasons e Choosing smaller alignment BER thresholds will cause the auto search functions to set more accurate sampling points However if the BER threshold is set lower than the residual BER of the measurement the auto search functions will fail 1E 7 is the smallest BER threshold available Running the auto search functions and viewing eye results at different BER thresholds will allow you to see the eye margin at specific BERs You can also construct basic eye contour information Avg 0 1 Threshold Activate this checkbox to make the error detector continuously set the 0 1 threshold level to the average DC voltage level of the incoming data signal Manual 0 1 threshold changes are impossible while this function is selected This function is useful for providing a starting point for sampling somewhere within the data eye on the vertical voltage axis
199. lution FFT windows generally reduce the measured spectral power For an introduction to FFT windows see Leakage and Windowing on page 245 Pass Fail Tab The Pass Fail tab of the Properties dialog box allows you to specify the criteria to decide whether the DUT passes or fails the test You can change pass fail criteria without rerunning a test The software only uses the criteria to rate the results of a measurement NOTE The pass fail criteria do not control measurement execution The measurement run will be completed even if the measurement fails for one or more of the criteria Properties x Parameters Pass Fail View Graph Color r General I Bit Error Rate I Total Power I Noise Power oa TATE TT m Frequency Range F Frequency Range 1 F Frequency Range 2 F Frequency Range 3 F Frequency Range 4 P Frequency Range 5 F Frequency Range 6 J Frequency Range 7 J Frequency Range 8 Cancel i Apply Help 252 Agilent Serial BERT N4901 2 User s Guide January 2006 The following Spectral Jitter Pass Fail criteria can be selected for the pass or fail decision e Bit Error Rate e Total Power in dB e Noise Power in dB You can also define the pass fail criteria for each of the defined frequency ranges In the tabular view each of the calculated values will be marked with an icon if it failed the test The following illustration shows a Spectral Jitter mea
200. matically repeated until a multiple of 512 bits is reached Decide on Trigger On or Off This setting becomes effective if the Trigger Out port is set to Sequence Trigger trigger at block begin 54 Agilent Serial BERT N4901 2 User s Guide January 2006 5 Create loops if desired Click the out arrow of a block A loop condition box labeled INF appears Click the in arrow of the same or a previous block This closes the loop Click the condition box and choose the appropriate break condition 6 Download the sequence to the pattern generator and test it Fo 0110 gt 7 Save the sequence in a file for later re use Saving the Sequence in a File To save the present sequence under a new name 1 In the toolbar click the Save As icon 2 Select a suitable directory for example C N4901a Sequences 3 Enter a file name The filename extension is seq by default If the present sequence was loaded from a file and edited you can also click the Save icon In this case the saved file is updated Loading a Sequence From a File To recall a sequence from a file 1 In the toolbar click the Open icon Da a The Open Sequence dialog box opens 2 Use this dialog box to locate and open the desired sequence The browser shows all files with the suffix seq in the selected directory NOTE The sequence is shown in the Sequence Editor but NOT automatically downloaded to the pattern generator
201. ment BER threshold If you are using these functions and want to consistently re synchronize at a lower sync threshold you must set the alignment BER threshold to the same value as the sync threshold BER 132 Agilent Serial BERT N4901 2 User s Guide January 2006 Introduction to Burst Sync Mode The burst sync mode is a special operating mode for measuring data in bursts of bits rather than one continuous stream of bits The burst sync mode measures bit error rates for each burst of data after the error detector synchronizes to the incoming pattern The signal at the Gate In port controls the timing of synchronization and error counting for each burst This function is useful for example to analyze recirculating loop data With a recirculating loop you can simulate very long cable connections like transoceanic cables Data is sent into a looped cable drum and after a certain number of loops it is diverted into the error detector This way only a short burst of valid data reaches the Serial BERT followed by a long break Therefore the error detector needs to be synchronized at the beginning of every burst How Burst Sync Mode Works The following figure presents the basic order of events that make up a burst mode measurement It also illustrates how the instrument operates while in burst mode no co ose v CDR Settling Time Synchronization Time lt 3 Sync success Register In
202. mple if your device is designed to turn off in response to a high BER you can check if your device behaves as expected And you can observe what the BER is prior to data loss See Pattern Alternation on page 92 for details e Manipulate the signal being generated by the pattern generator as described in Input and Output Ports on page 64 e Stress the DUT by distorting the output signal by adding jitter as described in Delay Control Input on page S6 e Adjust the sampling point to see the effect on the BER See Sampling Point Setup on page 114 for details e Make modifications to your device setup This can be used to see how changes to your device affect the BER You can determine what adjustments improve or degrade the BER 258 Agilent Serial BERT N4901 2 User s Guide January 2006 NOTE Instantaneous Measurements Procedures This section describes how to run monitor and analyze instantaneous measurements Running Instantaneous Measurements The Serial BERT s error detector immediately starts calculating the bit error rate as soon as it receives a valid signal at the data ports Both the pattern generator and the error detector must be correctly set up and connected to the DUT To run a BER measurement 1 Check if all the values in the status line are correct pattern generator pattern and bit rate error detector pattern and bit rate 2 View the BER display in the top left corner of the
203. mpling point is incorrect Use Auto Align The data pattern is inverted Toggle Data Inverted or use Auto Align The ED pattern does not match the incoming data pattern Make the ED track the PG pattern or set ED and PG patterns independently the same Patterns have lost sync in manual sync mode Use Auto Align Sync Now or select Auto Sync mode BERT connected to your device e The data eye is very small If Auto Align does not correct sync loss set the 0 1 threshold manually e Your device has inverted the data pattern Toggle Data Inverted or use Auto Align e Your device has changed the data pattern Set the ED and PG patterns separately e If the fundamental BER of your device is higher than the sync threshold BER you can select a higher sync threshold BER Agilent Serial BERT N4901 2 User s Guide January 2006 2 7 Error This indicator turns red when errors are detected View the BER bar or BER Results to see the nature of the errors BERT connected looped back or to your device e Stable errors caused by the error add function Turn error add off e Stable errors caused by false sync Select a lower sync threshold BER e Variable and high errors may be caused by faulty connectors cables BERT connected to your device e Stable or variable errors can also be caused by your device Correct the problems with your device If the problems are not clear then accumulate measurement data and analyze results
204. n a logarithmic dB or linear scale Frequency Axis Range Show the entire frequency range or zoom in one part of it Markers To analyze the graphs at a particular point you can use the markers Additionally you can display all related values for the markers in the moavlrav vaand ait 28 dB 1 Trace 65535 Points 21 dB 114 dB 7 dB 0 dB 7 dB Zoom 4 21 dB 28 dB 35 dB 42 dB Logarithmic back to the zoom graph or your finger 100 00 MHz Agilent Serial BERT N4901 2 User s Guide January 2006 255 Explanation of the Numerical Results The measurement provides numerical results for e Bit Error Rate Total Power e Noise Power The Frequency and total power for up to the 16 frequencies with the highest total power You can set the number of top frequencies to be evaluated in the View tab see View Tab on page 176 for details 256 Agilent Serial BERT N4901 2 User s Guide January 2006 Evaluating Results The Serial BERT offers several different kinds of tests that can be run e Instantaneous Measurements This type of measurements is used to monitor the instantaneous BER during measurement setup experiments and adjustments e Accumulated Measurements This type of measurements is used to measure and log error performance over time Also you can capture the location of bit errors for further analysis e Eye Measurements This type of measurements is used to measure the eye he
205. n auto search function that aligns the data signal with the clock signal so that the error detector samples at the optimum point of the data eye in the time axis This automatically compensates for delays in the clock data paths preventing unnecessary errors The 0 1 decision threshold is not changed The status message bar will indicate if this function is unsuccessful NOTE Ensure that the received clock frequency is stable before using Data Center TIP The clock data alignment process time is pattern dependent and with some large user patterns the alignment can take several minutes If you encounter such a long time with a user pattern it may be possible to first perform clock data alignment on a pure PRBS pattern This generally does not affect alignment accuracy and can minimize measurement time This tip does not apply in cases of severe pattern dependent jitter or with devices that do not work with PRBS patterns NOTE This function uses the alignment BER threshold to determine the left and right eye edges Cancel Click this button to cancel the Auto Align 0 1 Threshold Center or Data Center functions while they are in progress The following parameters will be returned to their previous value or status e Auto Align Canceled Data Delay and 0 1 Threshold values returned to previous Status of Data Inverted and Avg 0 1 Threshold checkboxes returned to previous e 0 1 Threshold Center Canceled 0 1 Threshold va
206. n external PC This allows to access all data for detailed processing and analysis on a PC Also you can track the status of all indicators such as DATA LOSS and CLK LOSS during the accumulation period If you want to view results from an accumulation period prior to the current and previous test you can view the respective log file too How to Analyze the Results Accumulated measurements can help you determine the cause of bit errors Bit errors can be caused by a variety of problems In addition instrument settings can affect how errors are measured and displayed The sync mode setting has the greatest effect 264 Agilent Serial BERT N4901 2 User s Guide January 2006 NOTE Display Change Analyze the basic accumulated results and log files Possible observations that may occur are Errors became constant and remained constant for a period of time or number of errored 0 s was greater than errored I s see Constant Errors More Errored 0 s than 1 s on page 279 Errors were random see Random Errors on page 280 Sync loss seconds were measured see Sync Loss Seconds on page 281 Accumulated Measurements Reference The Accumulated Results window presents data from the current and previous accumulation period Data from the current and all previous accumulation periods is available in the measurement log files if correctly set up This window contains a graph and five tables Rat
207. n or hex format Length This field indicates the length of the captured data Note that the value here does not equal the length of the pattern The captured data is saved as an alternating pattern e Pattern A contains the expected data e Pattern B contains the errored data Os if the expected bits were also received 1s for errored bits To calculate the captured pattern XOR the bits from pattern A with the bits from pattern B The pattern description contains the first error the error count date and time The name of the pattern file is ELOC_RESULT_CURRENT ptrn for the current capture and ELOC_RESULT_PREVIOUS ptrn for the previous capture These patterns are saved under C N4901A Pattern on the machine with the firmware server The results are displayed in the window below the pattern You can compare the results of the Current Run with the results of the Previous Run Ist Bit Error Location Address of the first captured errored bit Bit Error Count Number of all errored bits captured during the measurement Agilent Serial BERT N4901 2 User s Guide January 2006 229 NOTE Fast Eye Mask The Fast Eye Mask measurement is first of all meant for production and screening tests It allows to determine very quickly whether the eye opening seen at the output signal of a device is within specifications that is within certain timing and voltage limits Measuring the eye openings with an oscilloscope used to
208. n shows you how to set up and perform an Output Levels measurement As an example we measure the output level behavior of a shielded cable This requires the following steps Preparing the measurement see How to Prepare the Output Levels Measurement on page 186 Executing the measurement see How to Execute the Output Levels Measurement on page 187 Improving the Display see How to Improve the Output Levels Display on page 189 Changing the Properties see How to Change the Output Levels Properties on page 190 Using the different views see How to Use the Different Views on page 190 How to Prepare the Output Levels Measurement To prepare an Output Levels measurement to test a shielded cable 1 Disable the pattern generator outputs by pressing the OV Disable button in the PG Output Setup screen 2 Usea shielded cable to connect the pattern generator s Data Out port and the error detector s Data In port 3 Terminate all non connected pattern generator output ports with 50 Q 4 Use a shielded cable to connect the pattern generator s Data Out port and the error detector s Data In port 5 Switch to the Pattern panel and press Pattern Select Select an appropriate pattern for this test We use a pure 2 23 1 PRBS segment 6 For the pattern generator setup you need to specify the logic levels and the bit rate Select ECL levels and a clock speed of 1250 MHz in this example This corr
209. nc and Clock Loss Data in this measurement log file confirms that there were periods of momentary clock loss M N P SYNC LOSS PG CLOCK LOSS DATA LOSS Agilent Serial BERT N4901 2 User s Guide January 2006 281 282 NOTE eth eh eth f d d f ad f ek et eh et ot ad Using the Measurement Log to Identify Problems Measurement Logs are saved as CSV comma separated variable files Follow the steps below to view results in a measurement log file 1 Copy the measurement log file from the analyzer to your PC By default log files are saved in the folder C N4901A 1log fi ed i ooooocooocoo ooooo0coo0ocooco oc c c e 0 0ceo 2 Open a spreadsheet application on your PC 3 Import the measurement log file If your application has an import wizard you may need to indicate that data is delimited with commas 4 Once the file has been imported you may need to resize columns The imported log file should appear similar to the following example During measurement logging the Serial BERT logs data in ten second intervals Your log file may be missing up to the last ten seconds of data To avoid this condition accumulate for 10 seconds longer than desired i A B c D E 1 Description test 2 Log Time 05 29 2000 12 14 58 334 3 Activation Mode Single 4 Accurnulation Measurernent ELAPSED SECONDS Accumulation Limi 60 000000 6 7 8 BIT COUNT BIT FRE
210. ndard deviation of the noise distribution on the 1 and 0 rails e Q Optimum Threshold The Q Optimum Decision Threshold is calculated as Ook Oko Optimum Decision Threshold 0 0 Agilent Serial BERT N4901 2 User s Guide January 2006 201 Q Residual BER The Q Residual BER is the expected BER at the Optimum Decision Threshold It is calculated as Q2 2 Q V2 opt BER Numbers below le 255 are expressed as zero Q High Level The Q High Level is the mean calculated from the linear regression curve for the high level data B Q High Level Std Dev The Q High Level Standard Deviation is the o Sigma calculated from the linear regression curve for the high level data Q High Level Nr Points This is the number of data points used for the calculation of the Q High Level value It depends on the setting of the BER Threshold and also on the setting of the Min BER for Q parameter The minimum for calculating Q factor values is two points It is recommended to use more than 5 points 202 Agilent Serial BERT N4901 2 User s Guide January 2006 Q High Level R 2 The R parameter is an indicator that shows how well the converted data points fit to the straight line It is calculated as 2 sy 2 n n 2x2 BX faya Gv The R parameter should be examined before trusting the Q values R Its maximum value is 1 0 It must be seen in conjunction with the number of data points F
211. ndows are U S registered trademarks of Microsoft Corporation Agilent Serial BERT N4901 2 User s Guide January 2006 Contents Planning the Test 5 Introduction to the Serial BERT 6 Which Test is Appropriate 9 Connecting the DUT 13 Setting up Patterns 19 User Patterns 24 Software Generated PRBS 46 Hardware Generated PRBS 49 User Defined Sequences 51 Setting up the Pattern Generator 63 Input and Output Ports 64 Bit Rate 77 Trigger Output 82 Delay Control Input 86 Error Addition Insertion 88 Pattern Alternation 92 Setting up the Error Detector 99 Inputs and Outputs 101 Clock Setup 107 Sampling Point Setup 114 Trigger and Aux Output 125 Pattern Synchronization 127 Error Accumulation 140 BER Location 143 Audio Signals 145 Agilent Serial BERT N4901 2 User s Guide January 2006 3 Advanced Analysis 147 DUT Output Timing Jitter 162 Output Levels 184 Eye Opening 209 Error Location Capture 225 Fast Eye Mask 230 Spectral Jitter 238 Evaluating Results 257 Instantaneous Measurements 258 Accumulated Measurements 263 Eye Measurements 272 Solving Problems 275 Setup Problems 275 Measurement Problems 279 Other Messages 284 Customizing the Instrument 287 Restoring the System 287 Configuring the Instrument 289 Configuring the Touchscreen 293 Printing 297 File Management 300 Preset Instrument State 304 Self Test 305 4 Agilent Serial BERT N4901 2 User s Guide January 2006 Planning the Test The foll
212. nds when the error ratio is better than 1 x 10 for 10 consecutive seconds These 10 seconds are considered part of the available time Percent Unavailability is the ratio of the unavailable seconds to the total gating period expressed as a percentage Agilent Serial BERT N4901 2 User s Guide January 2006 267 Severely Errored Seconds The ratio of the total number of available seconds with an error ratio worse than 1 x 10 to the total number of available seconds expressed as a percentage Degraded Minutes Severely errored seconds are discarded from the available time and the remaining seconds are grouped into blocks of 60 seconds Blocks which have an error ratio worse than 1 x 10 are called Degraded Minutes and Percent Degraded Minutes is the ratio of the total number of degraded minutes to the total number of 60 second blocks in the available time expressed as a percentage Incomplete blocks of less than 60 seconds are treated as complete blocks of 60 seconds e Errored Seconds The ratio of the errored seconds in the available time to the total number of seconds in the available time expressed as a percentage Interval Results Error Free Seconds The number of seconds in which no errors were measured is displayed here Error Free Deciseconds The number of deciseconds in which no errors were measured is displayed here Error Free Centiseconds The number of centiseconds in which no errors were measured
213. nfiguring the instrument Set Date Time Dialog Box This function opens the Date Time Properties dialog box of Windows Use this dialog box to set the date and time as well as the time zone of your location GPIB Address Change Dialog Box Your instrument is set to a default GPIB General Purpose Interface Bus address You can change this address by doing the following 1 On the Utilities menu click Change GPIB Address This starts the GPIB Address Change dialog box 2 Select an address from the New GPIB Address list 3 Click OK The new address will be applied and the dialog box will close Each instrument must be set to a unique GPIB address to avoid multiple instruments transferring data at the same time NOTE The default address is 14 however addresses from 0 to 31 may be used if the default address is the same as another instrument s GPIB address Address 21 is usually reserved for the computer interface Talk Listen address and should not be used as the instrument address 292 Agilent Serial BERT N4901 2 User s Guide January 2006 Configuring the Touchscreen Configuring the Touchscreen Procedures The following procedures help you to configure the touchscreen according to your personal needs Turning the Touchscreen Off On This function disables the instrument touchscreen external keyboard and mouse 1 On the Utilities menu click Touchscreen Off The Enable Touchscreen icon is placed in the upper
214. ng applications The Serial BERT provides capabilities that help you set up and run burst tests See Introduction to Burst Sync Mode on page 133 e Are you aware of how the sync mode setting can affect the way errors are measured and displayed See What Type of Synchronization Should You Use on page 129 Test Times and Confidence Levels A true BER measurement must be statistically valid Because it is not possible to predict with certainty when errors will occur your device must be tested long enough to have confidence in its BER performance The table below lists the test times required for different BERs at specific bit rates NOTE The test times in this table are valid for continuous measurements If measurements are made in bursts the test times will be longer continuous test time divided by the burst duty cycle times the burst sync ratio See Introduction to Burst Sync Mode on page 133 for background information Agilent Serial BERT N4901 2 User s Guide January 2006 11 The test times for 95 confidence level are 1E 14 8 4 hours 1 4 days 5 6 days 22 4 days 1E 13 50 minutes 3 3 hours 13 hours 2 2 days 1E 12 5 minutes 20 minutes 80 minutes 5 3 hours 1E 11 30 seconds 2 minutes 8 minutes 32 minutes 1E 10 3 seconds 12 seconds 48 seconds 3 2 minutes The formula for confidence level is as follows C 1 e nb Where C degree of confiden
215. ngth It shows the length of the pattern as stored in the file If the pattern length is not 512 bits or a multiple of 512 bits and this single block is looped the behavior depends on the loop end condition If the loop end condition is INFinite or determined by the Aux In port then the pattern is automatically repeated until 512 bits or a multiple thereof is reached In this case the duration of the block will be longer than indicated For example a pattern containing 17 bits will be repeated 512 times and the actual block length will be 8704 bits This has an impact on the trigger at block start Block Trigger If this checkbox is enabled a trigger spike will be output at the pattern generator s Trigger Out connector whenever the execution of the block begins provided that the Trigger Out port is set to Sequence Trigger mode see also Sequence Trigger on page 84 Loops Within a Sequence A loop defines the transition from the end of a block to the beginning of the same or a previous block It is not possible to jump into an existing loop It is also not possible to specify loops within loops exept the default overall loop For information on how to create a loop see Creating a New Sequence on page 54 f_n PauseD 512 Edit ASCIL patt 17 FE Edit PRBS 15 2048 58 Agilent Serial BERT N4901 2 User s Guide January 2006 Every loop has a delete button x and a loop condition button L
216. nly Patterns A and B in interleaved rows Patterns A and B in horizontally split windows If only a standard pattern is selected this icon has no function Pat Click this icon to open the Pattern Select Form dialog box For more ET Sel information see Loading Patterns to the Pattern Generator and Error Detector on page 38 TIP Ifyou are working on a remote PC you can use the keyboard shortcuts for Cut Copy and Paste Also you can use the clipboard function to copy strings in either binary or hex format to and from other applications When pasting them into the pattern in binary view a 1 sets a single bit in hex view a 1 sets 4 bits Agilent Serial BERT N4901 2 User s Guide January 2006 41 Pattern Editor Canvas and Status Bar The pattern editor canvas displays the pattern and offers you to edit it 0 0781 3580 30AD 4060 S4E0 w FoB 5406 SCOD 5654 OEOF 160 6579 8732 1098 3198 7186 20 C870 ADBE F10C 5620 ASD4 320 E006 SC40 0586 5408 S6A6 w0 SE40 DE6S 4F06 5406 1117 420 jep21 FOES DOES 6032 1976 so 203e 0498 70 eo 7x 5 The status bar below the editor canvas provides the following information from left to right Status Offline indicating that this is the display of the editor not the pattern generator nor the error detector e Current Position in bits of the cursor in the pattern e Editing mode Overwrite or Insert e Display mode Hex or Bin e Mark Density of the patt
217. nt delay value is checked for validity If the delay does not fall in the valid range it will then be moved into the valid range All subsequent changes to the sampling point delay will be checked for validity and rejected if invalid If you cannot find the optimum sampling point in the valid range of sampling points you can switch from the rising to the falling clock uary 2006 109 edge For this purpose activate the Clock Falling Edge checkbox on the Sampling Point Setup screen of the error detector setup 2ns jag P Clock 500 Mbit s Sampling Point Delay Rising Edge Sampling Point Delay Falling Edge Covered Range With this method you can still place the sampling point anywhere in the clock cycle to find the optimum sampling point even at low frequencies 3 226 ns i at I Clock 310 Mbit s L p q i Covered Range However if the bit rate falls below appr 310 Mbit s even with this method gaps occur in the range of possible sampling points i 6 66 ns i Clock 150 Mbit s 1 i L i t I i i i Covered Range with Gaps Gap Gap Gap Gap Gap As a result you might not be able to find the optimum sampling point if it falls in one of the gaps 110 Agilent Serial BERT N4901 2 User s Guide January 2006 Clock Setup Procedures To set the clock rate 1 Press Clock Setup in the ED Setup menu 2 Specify whether you are using an External Clock Source or Clock Data Re
218. nts for the measurement Agilent Serial BERT N4901 2 User s Guide January 2006 235 Relative Time and Voltage In the table at the bottom of the Parameters tab you can change measurement points Enter the Relative Time and Voltage according to the above settings Pass Fail Tab The Pass Fail tab of the Properties dialog box allows you to specify the criteria to decide whether the DUT passes or fails the test You can change pass fail criteria without rerunning a test The software only uses the criteria to rate the results of a measurement NOTE The pass fail criteria do not control measurement execution The measurement run will be completed even if the measurement fails one or more of the criteria The Fast Eye Mask measurement has only one pass fail condition The bit error rate If desired enable the BER Threshold and set a suitable threshold NOTE The BER Threshold will usually be 0 since you want the measurement to fail if a single bit was received in error The pass fail threshold applies to all measurement points An icon indicates all measurements where the bit error rate is higher than this threshold as shown in the figure below Terminal a1 2 3 4 Relative Time 0 0 4 Ul 0 160 UI 0 160 UI Voltage abs 1Y 1Y 700 000 my BerilBERT a a a View Tab The following parameters can be set on the View tab of the Properties dialog box Analyze You can analyze for e All Errors To display all
219. o enable the Burst Sync Mode Burst mode operates as follows For each burst of data BER measurements are only made after the error detector synchronizes to the incoming pattern The signal at the Gate In port controls the timing of synchronization and error counting for each burst The Sync Threshold defines when the error detector assumes the patterns to be synchronized The Serial BERT shifts the received pattern bitwise until the measured BER is below this threshold Agilent Serial BERT N4901 2 User s Guide January 2006 139 Error Accumulation The Serial BERT instantly starts measuring the BER when the error detector receives data However to perform tests that can be repeated and compared you can collect the measurement data in several ways and save it to a log file The available options are to test e for a specified time e until a certain number of errors occurred e until a specified number of bits have been sent e until you stop it manually Error Accumulation Procedures To set up the Serial BERT for error accumulation 1 Press the Accumulation Setup icon in the ED Setup menu 2 Select the desired Activation Mode for the type of test you want to run Manual Single or Repeat test 3 If you selected a Single or Repeat test you need to specify the accumulation Period certain Time Number of Errors or Number of Bits 4 Use the Measurement Logs area to specify whether you want save the results in a log
220. o select the desired file Memory User Pattern The Serial BERT supports 12 so called memory slots for storing user patterns The slots are stored in files named Upat1 ptrn to Upat12 ptrn To load a pattern that is stored in one of these memory slots choose this option and select the desired slot from the list 2 n 1 PRBS Select this option to use a hardware generated pattern of length 2 n 1 See How the Hardware Generates PRBS on page 49 for details 2 n PRBS Select this option to use a memory based pattern of length 2 n Mark Density PRBS Use this options to generate memory based 2 n PRBS patterns with specified mark densities Select the desired value from the list The available values are 1 8 1 4 1 2 3 4 and 7 8 Zero Sub PRBS Use the Zero Substitution function to insert a longer row of zeros into a memory based 2 n PRBS pattern The length of the zero row can be freely specified up to the length of the pattern Pattern Size You can select the pattern size according to your needs The pattern lengths available depend on the type of signal you have selected Sequence Enable this checkbox if you wish to use a user defined sequence The patterns used in a sequence are defined in the Sequence Editor see also User Defined Sequences on page 51 Agilent Serial BERT N4901 2 User s Guide January 2006 45 Software Generated PRBS Software generated patterns are PRBS patterns that can be modified They
221. old High level noise distribution Gaussian characteristics assumed Regions for Q factor measurements Low level noise distribution Gaussian characteristics assumed Optimum sampling delay To measure such small bit error rates you have to compare an adequate number of bits Depending on the DUT characteristics you may also have to use a very fine step size for the threshold variation Comparison of the Results On the other hand such a comparison may make it easier to understand the dependencies Agilent Serial BERT N4901 2 User s Guide January 2006 207 208 NOTE If you set the BER Threshold to a position that gives you only a few data points you will find The Q factor is relatively high e The R values are excellent greater than 0 95 e The mean levels and standard deviations returned by the Q factor calculations differ from the measured values The QBER vs Threshold graph may look as shown below 4 Trace 16 Ponts 300 0mv S00 0my AOV PERA 42y 13y 44y asy aey ary aey Linesr If you increase the BER Threshold to include more points you will find e The mean levels and standard deviations returned by the Q factor calculations approach the measured values e The Q factor decreases e The R values deteriorate The latter is caused by increasing nonlinearity This can be inspected in the QBER vs Threshold graph 1 Trace 31 Points 80
222. ollowing buttons in the toolbar If you chose that both the error detector and pattern generator use the same pattern Feet Loads the current pattern to both the tong pattern generator and error detector To PGED If you allowed different patterns for the two modules Fe Loads the current pattern to the 10110 pattern generator To PG 10110 Loads the current pattern to the error detector To ED Agilent Serial BERT N4901 2 User s Guide January 2006 39 User Patterns Reference The elements of the pattern editor and all dialog boxes that can be opened from the pattern editor are described below Edit Pattern Window The Edit Pattern window provides the following convenient pattern editing functions Toolbar EX Ctrl 5 New ce Open B Save Ec To PG To PG wgon To ED Teo To ED sam To PGED Cut Copy Paste Delete Undo Go To INS OVR Hex Bin Proper ties Find Click this icon to switch the toolbar from two rows to a single row ap pearance leaving more space for the pattern data Click this icon to create a new pattern See Creating New Patterns on page 36 for details Click this icon to open a pattern from a file See Opening Existing Patterns on page 36 for details Click this icon to save the current pattern If it was not previously saved a Save As dialog box opens for you to specify name and path for the file Click this icon to send the pa
223. ompleted even if the measurement fails for one or more of the criteria You can set pass fail limits individually for level noise and Q factor parameters To set the Level Pass Fail criteria see also Level Results on page 198 To set the Signal Noise Pass Fail criteria see also Noise Results on page 199 To set the Q Factor Pass Fail criteria see also Q factor Results on page 201 Your setup may finally look as shown below Properties Ea Parameters Pass Fail View Graph Color m Level Parameters M Level Pass Fail Minimum Maximum IV High Level fa Vv 300 mv M Low Level fasv a 68v IV Mean Level fray jiv I Amplitude fo y Jov py m Noise Parameters I Signal Noise Pass Fail Minimum Maximum I Signal to Noise Ratio RMS j Signal to Noise Ratio Peak Peak jp P Peak Peak Noise Jov MQ Factor Parameters I Q Factor Pass Fail Minimum Maximum I G Factor 0 I Residual BER jo F Optimum Threshold j y ov 194 Agilent Serial BERT N4901 2 User s Guide January 2006 BER vs Threshold Graph In the following figure you can see how errors are flagged 1 Trace 86 Points 600 0 mv 400 0 mv 200 0 mv ooy 200 0 mY 400 0 mY 600 0 mv BER Threshold 1 000e 6 Logarithmic 1 000e 5 1 000e 3 1 000e 1 1 000e 6 1 000e 4 1 000e 2 1 000e Show Color Copied High Level Low Level Mean Level I 430 91 mv 479 08 mV Terminal Electrica
224. on on the jitter the measurement s jitter display shows only the portion at the left hand side of the optimum sampling point The jitter histogram allows you to visually inspect the jitter components e Random Jitter RJ e Deterministic Jitter DJ Estimated Total Jitter TJ Jitter Measurement Parameters on page 180 describes how these components are calculated A Gaussian marker is available in this view to help analyze the jitter components See The Gaussian Marker on page 157 for details Furthermore the following parameters can be set on the View tab of the Properties dialog box Analyze You can analyze for e All Errors To calculate the BER values from all bits errors e Errors if Os Expected To calculate the BER values if 0 is expected but 1 received e Errors if 1s Expected To calculate the BER values if 1 is expected but 0 received Calculate You can calculate measurement parameters for e 0 Errors To calculate the parameters for 0 errors If this is selected RJ DJ separation is not available Agilent Serial BERT N4901 2 User s Guide January 2006 177 BER Threshold To calculate the parameters for the given BER threshold This is the BER level for which output timing numerical values phase margin skew etc are calculated It is also the upper limit of the BER range for RJ DJ separation The BER threshold influences some of the parameters of the DUT
225. oop Delete Button Clicking the delete button x removes the loop from the sequence Loop End Condition Button The initial default setting of the loop end condition is INF inite which means this loop will continue until the instrument is switched off Clicking the loop end condition button allows you to specify the loop end condition Choices are Counted The loop ends after the specified number of repetitions Sequence execution continues with the next block Infinite An infinite loop can be broken by Aux In High Aux In Low The loop ends when Auxiliary Input is high or low Sequence execution continues with the next block Aux In Rising Aux In Falling The loop ends when a rising or falling edge at the Auxiliary Input is detected Sequence execution continues with the next block Manual The loop ends when the Break button of the Sequence Editor is clicked Sequence execution continues with the next block NOTE The pattern generator reacts on the end condition as soon as the pattern has finished Agilent Serial BERT N4901 2 User s Guide January 2006 59 Sequence Editor Toolbox The Sequence Editor toolbox provides the following functions New Click this icon to create a new sequence This opens the Properties k dialog See Creating a New Sequence on page 54 for details Open Click this icon to open a sequence from a file See Loading a Se quence From a File on page 55 for details Save Cl
226. or Other options are e AuxInHi AuxInLo Sequence starts when Auxiliary Input is high or low e AuxInRising AuxInFalling Sequence starts with a rising or falling edge at the Auxiliary Input e Manual Sequence starts when the Start button is clicked Block Options Default is pattern PO Pause0 a length of 512 bits and Trigger On Other options are e Patterns P1 Pausel undistorted PRBS a user pattern from a file or a fraction of the clock pulse divider range 2 127 e Length Must be 512 bits or a multiple thereof Shorter or longer user patterns are automatically repeated until the next 512 bit boundary is reached see also How the Serial BERT Generates Memory Based Patterns on page 27 Trigger Refers to the pattern generator s Trigger Out port If the Trigger Out is set up to generate a divided clock pulse the block setting is ignored But if the Trigger Out is set to Sequence Trigger and the block to TrigOn a trigger pulse is generated whenever the block starts or restarts No trigger is sent if the block is set to TrigOff See Trigger Output Procedures on page 82 for information on how to set the mode of the Trigger Out port LOOP Options Default is one overall endless loop for the whole sequence e Bw Bm Define the start and end block of a user defined loop Loops always define the transition from the end of a block to the beginning of the same or a previous block It is not p
227. or causing it to generate a precisely aligned internal reference pattern For software generated and user patterns a 48 bit pattern from the pattern is used as a detect word Optimally this detect word should be unique within the entire pattern The error detector searches for this detect word within the incoming data stream and uses the point in the data stream as a reference and compares all following bits with the pattern If the measured BER is better than the synchronization BER the error detector is synchronized There are thus three possible outcomes for a synchronization e Single instance of the detect word in the data stream Expected Incoming Pattern Bit stream Detect Word a 48 bi ae Detect Word Correct Sync 48 bits Detect Word 48 bits C Agilent Serial BERT N4901 2 User s Guide January 2006 127 e Multiple instances of the detect word with correct synchronization Expected Incoming Pattern Bit stream Reference Point Correct Sync e Multiple instances of the detect word with false synchronization Expected Incoming Pattern Bit stream ee __ lie Reference Point False Sync If the error detector attempts to synchronize on the incorrect detect word the BER will be unacceptably high and if automatic synchronization is selected the error detector attempts another resync 128 Agilent Serial BERT N4901 2
228. or example Two data points always fit perfectly well but the resulting Q factor calculations are not reliable On the other hand 50 data points may reveal a poor R value This tells you that the linearization is prone to errors If the R value falls below 0 75 the Q factor calculations are not applicable Q Low Level The Q Low Level is the mean calculated from the linear regression curve for the low level data B Q Low Level Std Dev The Q Low Level Standard Deviation is the o Sigma calculated from the linear regression curve for the low level data Q Low Level Nr Points This is the number of data points used for the calculation of the Q Low Level value It depends on the setting of the BER Threshold and also on the setting of the Min BER for Q parameter The minimum for calculating Q factor values is two points It is recommended to include more than 5 points Q Low Level R 2 See the definition of Q High Level R 2 Agilent Serial BERT N4901 2 User s Guide January 2006 203 Unavailable Values Under certain circumstances some numerical results are not available This is indicated by lt invalid gt or lt not applicable gt in the numerical results table below the measurement graphic e lt invalid gt indicates that the value could not be calculated This is the case for example for phase margin when the BER threshold is set to high values and does not intersect with the bathtub curve e lt not app
229. ossible to jump into an existing loop It is also not possible to specify loops within loops exept the default overall loop 52 Agilent Serial BERT N4901 2 User s Guide January 2006 Loop end condition Default is an INFinite endless loop Other options are AuxInHi AuxInLo Loop continues until Auxiliary Input is high or low AuxInRising AuxInFalling Loop continues until Auxiliary Input receives a rising or falling edge Manual Loop continues until the Break button is clicked Counted loop Loop repeats a specified number of times Sequence Mode Characteristics When a sequence is executed the pattern generator is in Sequence Mode This mode differs from usual Pattern Mode Pattern related triggers cannot be generated In sequence mode the Trigger Out port can produce either a continuous pulse a fraction of the current clock pulse or a trigger spike at the beginning of each block that is set to TrigOn The duration of that spike depends on several conditions and is therefore undefined Software generated PRBS with polynomial 2 n cannot be referenced in a sequence If several PRBS patterns shall be included all have to use the same polynomial Every sequence has to be downloaded to the pattern generator before it can be executed Every sequence can be stored in a file and recalled from that file However the SequenceExpression does not include user patterns Recalling a sequence
230. ot be set up to trigger on certain pattern positions or pattern sequences If this option is enabled for example in the user interface the trigger is sent once for each pattern but the bit position cannot be specified The option to trigger on the divided clock signal is supported as usual e There are restrictions to the available clock to data delay values The pattern generator can vary the clock to data delay only within a range of 750 ps to 750 ps relative to the clock signal For frequencies above 666 Mbit s this range is sufficient to cover the complete clock cycle 1 unit interval For lower frequencies the valid data delay range is smaller than the clock cycle The pattern generator cannot generate signals with a delay outside this range Therefore the data delay cannot be set to all values within the clock cycle Clock 500 Mbit s Valid Data Delay Valid Range for i i Data Delay i i 78 Agilent Serial BERT N4901 2 User s Guide January 2006 If the clock frequency falls below this threshold the current data delay value is checked for validity If the delay does not fall in the valid range it will be moved into the valid range All subsequent changes to the data delay will be checked for validity and rejected if invalid Bit Rate Procedures You can use an external clock source or the Serial BERT s internal clock to control the bit rate To set the bit rate 1 Ifthe external clock so
231. ou can use an external bias network also known as a bias tee to power your device You must ensure however that the network is oriented correctly If it is not damage may occur to your device or instrument 70 Agilent Serial BERT N4901 2 User s Guide January 2006 CAUTION The diagram below shows a bias tee that is positioned correctly Notice that the pattern generator s outputs are protected by the blocking capacitor DC Bias DUT Output Bias Tee Input and Output Ports Procedures You have to do the following to set up the pattern generator s output ports Setting Logic Levels and Terminations Before you can start sending signals to your device you have to set the logic levels and terminations 1 Press PG Output Setup in the PG Setup menu 2 In the Data Logic Level list select the logic family that is appropriate for the data input of your device 3 In the Clock Logic Level list select the logic family that is appropriate for the clock input of your device When you modify the Data Logic Level or Clock Logic Level the instrument automatically sets the appropriate voltage levels and the Termination voltage 4 If necessary change the data or clock termination voltage by entering a new value You can also use the front panel knob to change the value The logic level automatically changes to Custom Selecting the wrong terminations may damage your device See Why Incorrect Term
232. oubleshoot the Serial BERT in case any problems occur Setup Problems You may run into the following problems when setting up the Serial BERT Front Panel not Working If the front panel is not working and a hardware wizard calls on you to install the driver for an unknown USB device cancel the hardware wizard and reboot the instrument This might happen if the front panel controller is not initialized correctly at boot up Incorrect Error Detector Input Voltages The error detector s protection circuit disconnects the input amplifier if the incoming signal does not lie within the defined input range In this case a dialog box will open that lets you define the input range sample point 0 1 threshold termination voltage and state When you click OK the input amplifier will be reenabled If the input signal is still out of range for any of these parameters the error detector will be immediately disconnected again Error Detector Does not Count Bits If the analyzer does not count any bits press Sync Now This might happen after bad glitches at the clock input Agilent Serial BERT N4901 2 User s Guide January 2006 2 5 DLL Alignment Failed If an error message containing the text DLL alignment failed is displayed increase or decrease the bit rate by 1 bit s See Bit Rate Procedures on page 79 for details on setting the bit rate The instrument provides several error indicators to help you identi
233. oved vertically if you click between the handles and drag them up or down Agilent Serial BERT N4901 2 User s Guide January 2006 Accumulated Results In the Accumulated Results table the following values are listed Bit Count The number of bits considered for the accumulation period is displayed here This may not equal the total number of bits sent from the PG Error Ratio The ratio of the error count to the bit count is displayed here to 4 significant digits Error Count The number of errored bits measured during the accumulation period is displayed here Errored 1 s Ratio The ratio of the errored 1 s count to the bit count is displayed here to 4 significant digits Errored 1 s Count The number of logic 1 s measured as logic 0 s during the accumulation period is displayed here Errored 0 s Ratio The ratio of the errored 0 s count to the bit count is displayed here to 4 significant digits Errored 0 s Count The number of logic 0 s measured as logic 1 s during the accumulation period is displayed here G 821 Measurements Availability The ratio of the available seconds to the total gating period expressed as a percentage Unavailability The error ratio is calculated over 1 second timed intervals during the gating period An unavailable period begins when the error ratio is worse than 1 x 10 for 10 consecutive seconds These 10 seconds are considered part of the unavailable time The unavailable period e
234. ow 615 Mbit s the limitations apply e If the bit rate exceeds 620 Mbit s the limitations no longer apply 108 Agilent Serial BERT N4901 2 User s Guide January 2006 The following figure clarifies the behavior in the range between 6 15 Mbit s and 620 Mbit s Limitations do not apply in this range 620 615 Bitrate Mbit s Limitations apply in this range For the error detector the following rules apply Agilent Serial BERT N4901 2 User s Guide Jan For low frequencies you cannot use the automatic data alignment functions Auto Align and Data Center Instead you need to align the error detector manually For instructions please refer to the Online Help go to the Error Detector Setup section select Sampling Point Setup and switch to the Procedures information There are restrictions to the available sampling point delay values The error detector can vary the sampling point delay only within a range of 0 ps to 1 6129 ns relative to the clock signal For frequencies above 620 Mbit s this range is sufficient to cover the complete clock cycle 1 unit interval For lower frequencies the maximum sampling point delay is smaller than the clock cycle Therefore the sampling point cannot be set everywhere within the clock cycle Clock 500 Mbit s Valid Sampling Point Valid Range for i i Sampling Point 0 1 6129 ns If the clock frequency falls below this threshold the current sampling poi
235. owing topics provide some information that can help you in planning tests with the Serial BERT Introduction to the Serial BERT explains the basics of bit error rates BER and BER testing with the Serial BERT e Which Test is Appropriate provides you with some basic information about the types of testing you can do with the Serial BERT e Connecting the DUT tells you how you should connect your DUT to the Serial BERT Agilent Serial BERT N4901 2 User s Guide January 2006 5 Introduction to the Serial BERT These topics explain some of the basics of BER Serial BERT and BER testing with the Serial BERT Important Information about the Pattern Generator Outputs The pattern generator output ports must be terminated with 50 Q if they are not connected e Data Out e Data Out e Clock Out e Clock Out Termination of output ports improves the test performance Important Information Regarding Security The Serial BERT is a PC based instrument with a standard Windows operating system As such it is subject to the same security protection measures as any other PC See the Microsoft web site for more information regarding data security http www microsoft com security default mspx Every user must have administrator privileges to run the Serial BERT firmware Otherwise the firmware does not have access to the instrument s hardware It ist recommended that you create a recovery disk for the instrument See How to
236. ows DataOut A A B A BB A Trigger Out NOTE This pattern must be at least 1024 bits long Alternate Pattern Trigger Pulse Select this option to send a trigger pulse whenever the pattern being sent changes A to B or vice versa For patterns with a length divisible by 512 there is one trigger pulse sent for every pattern repetition Data Out A B A B B meroa L If the length of the pattern is not divisible by 512 a trigger is sent whenever the pattern reaches a 512 bit RAM boundary Agilent Serial BERT N4901 2 User s Guide January 2006 83 Sequence Trigger This checkbox becomes available after a user defined sequence has been downloaded to the pattern generator Click this checkbox to switch the Trigger Out port from Clock divided by n to Sequence mode In Sequence mode the Trigger Out can generate a spike whenever the execution of a block starts or restarts Whether that happens for a particular block or not is defined for each block individually in the SequenceExpression See also Sequence Block Parameters on page 57 Pattern Trigger Position NOTE This option is not available for alternate patterns Select this option to send a trigger signal that is synchronized to a certain position in the pattern Depending on the type of pattern currently loaded in the pattern generator you have to specify the trigger position either according to Bit Position or
237. ppressing the Outputs via External Signal You can configure the Serial BERT to suppress the output according to an external signal 1 Connect an external instrument to the Aux In port The signals received at this port must be TTL compatible The granularity for the insertion of pattern B via external signal is 512 bits 2 Press the Alt Pattern and Aux In icon in the PG Setup menu to open the Alternate Pattern Control dialog box 3 In the Aux In section select Output Blanking See Aux In on page 96 for more information on the different available modes 4 Click OK to close the dialog box Pattern Alternation Reference The Serial BERT offers the following options for alternating between pattern A and pattern B NOTE These functions are only available if Alternate Pattern is selected in the Pattern Properties dialog box Continuous A Select this option to send out only pattern A repeatedly Data Out A AA A A A A A A A A A Continuous B Select this option to send out only pattern B repeatedly B0 B1 B2 DataOut B B B B B B B B B B B B Agilent Serial BERT N4901 2 User s Guide January 2006 95 96 Alternate AB Select this option to alternately send out patterns A and B A B A B Data Out Single Shot B Select this option
238. quence consists of up to four blocks Each block can generate a pause signal constant 0 or 1 a divided clock signal a 2 n 1 PRBS or a user pattern Single or multiple blocks can be looped An overall loop restarts the sequence after it has come to its end When to Use a Sequence You may wish to test a device that uses a certain protocol for processing data For example the device might expect synchronization data a preamble payload data and a suffix All this can be provided by a user defined sequence How a Sequence is Defined The sequence is defined by a SequenceExpression which is formulated in its own language checked by the Sequence Editor The SequenceExpression specifies e the sequence start and break conditions e the blocks their contents and trigger output e the loops You can inspect the contents of the SequenceExpression in the Properties dialog of the Sequence Editor The SequenceExpression uses the following keywords Version optional Description optional Start optional Block repeated for each block numbered Loop repeated for every loop not numbered Agilent Serial BERT N4901 2 User s Guide January 2006 51 Example of a SequenceExpression Version 1 0 Start IMM Block 1 PRBS11 1024 TrigOn Block 2 C N4901A Pattern Upat10 ptrn Block 3 P0 512 TrigOff Loop B1 B1 2 Start Options Default is IMM immediately after downloading the sequence to the pattern generat
239. r additional information Data and Clock Terminations Use the Data Termination and Clock Termination checkboxes to toggle between true differential DC coupling and center tapped termination In center tapped mode you can specify the termination voltage The Data Termination and Clock Termination fields are used to set the termination voltage When you select a Data or Clock Logic Level the terminations are set accordingly It is reeommended that you keep the default values of the terminations Agilent Serial BERT N4901 2 User s Guide January 2006 75 See Understanding how the Serial BERT Uses Logic Families and Terminations on page 68 for additional information OV Disable eae CAUTION Before clicking this button ensure that your device s inputs tolerate ground potential Use this button to clamp both the pattern generator s output ports data and clock to ground You can see the status of the outputs in the display area of the pattern generator Note that you can also use the buttons below the display to enable or disable this function for data and clock separately This function can be useful in case you want to connect or disconnect a device to any of the output ports to prevent any damage Output Blanking In Pattern Mode When the generator is set up to generate a pattern use this checkbox to activate the Aux In port of the pattern generator You can then connect a signal to this port and operate the patt
240. r upper limit to the BER no matter how many bits we transmit Whether such a test fails or passes depends entirely on the application Once we are able to decide for a measured point whether its BER is above or below the BER threshold we can determine the total jitter at the intersection of the BER threshold with the bathtub curve Since we are unable to find a single point on the slope where the BER is exactly 10712 we search for an interval that brackets the point at which the BER is equal to 107 This is illustrated for the left hand slope in the following figure Agilent Serial BERT N4901 2 User s Guide January 2006 10 10210 10 1 1012 108 16 4 12 a x x10 w We do not need to know the exact BER values at x and x It is sufficient to assure that BER x is greater than 1071 and BER x is lower than 107 at a confidence level of 95 The algorithm then assumes that x for the left hand slope of the bathtub curve is in the middle of the bracketing interval After repeating the procedure to determine Xp for the right hand slope it calculates the total jitter peak to peak like in the standard timing jitter measurement Measurement Duration The duration of a Fast Total Jitter measurement depends on e the BER threshold e the bit rate e the sample delay step size e the contribution of random jitter e the contribution of deterministic jitter Compared to a conservative bathtub measurement the F
241. rate one out of every 512 bits was measured as an errored 0 1 256 50 1 512 In addition the inverse of the error ratio was a multiple of 2 In this example the analyzer could have been in manual sync or auto sync mode with 1E 2 as the sync threshold BER If the sync mode settings were different the results would not have been measured and displayed in the same way 279 The BER in this example approx 2E 3 was greater than the default sync threshold BER 1E 3 If the analyzer was in the default auto sync mode it would have activated the synchronization functions and continually shifted the position of the reference pattern in an attempt to find a lower BER BER 1 953E 3 NOTE The accumulated results table displays the total accumulated error ratio Look in the measurement log file to see the error ratio during the period of constant errors Example Log of ConstantErrors The error ratio during periods of constant errors was 0 00195313 This corresponds to 1 953E 3 2 49E 09 2 49E 09 486000 0 00195313 2 49E 409 2 49E 409 486000 0 00195313 2 49E 09 2 49E 09 486000 0 00195313 2 49E 09 2 49E 09 486000 000195313 2 49E 09 2 49E 409 486000 0 00195313 2 49E409 2 49E409 486000 000195313 2 49E 09 2 49E 09 486000 0 00195313 2 49E 09 2 49E 09 486000 000195313 Accumulated Results t l f 15697 679 Possible Cause Random errors may have been caused by a noisy waveform Noise can be caused
242. rates that have been measured Agilent Serial BERT N4901 2 User s Guide January 2006 209 Three Available Views The Eye Diagram can be visualized in three different ways Contour Plot The contour plot shows discrete lines of equal bit error rate just like the contour lines on a map The color of a line indicates the respective BER value This graph is useful to visualize in which areas the BER changes a homogeneous BER field will give you no lines at all Voltage V Delay ns e Pseudo Color Plot This plot visualizes the BER by a continuous color gradient It uses different colors for the regions between the lines of equal BER This is useful to get an immediate visual impression of the distribution of the BER in the eye diagram Threshold V Delay ns 210 Agilent Serial BERT N4901 2 User s Guide January 2006 Equal BER at BER threshold This graph displays only one curve for the chosen bit error rate threshold Voltage V Delay ns Example Results The following figure shows the graphical result of a typical Eye Opening measurement 120 0 mV 90 0 mv 60 0 mv 30 0 mv oov 30 0 mV 50 0 mV 90 0 mV f 120 0 mv e aca Une asoom PRR SS Ss PERU et pba Nee vic a escale gt 0 650 UI 0 390 UI 0 130 UI 0 130 UI 0 390 UI 0 650 UI Relative 0 520 UI 0 260 UI 0 000 UI 0 260 UI 0 520 UI Agilent Serial BERT N4901 2 User s Guide January 2006 211 Eye Opening
243. re recognized as passed or failed However a new test run is not required when changes are made here See Pass Fail Tab on page 236 for details Agilent Serial BERT N4901 2 User s Guide January 2006 233 View tab All settings on this tab only affect the way the data is displayed You do not need to run the measurement again See View Tab on page 236 for details 3 Press OK when you have made all required changes to close the Properties dialog box Fast Eye Mask Reference The Fast Eye Mask measurement returns the results in a numerical form The following sections provide explanations of the measured parameters and the display options that are specific to this measurement Properties that can be specified on the various tabs of the Properties dialog box Parameters Tab Note that if you modify the parameters on this page you have to rerun the measurement to update the results Set the criteria for moving to the next sample point e Number of Compared Bits After this number of compared bits the measurement stops for the current sample point and moves to the next one The default is 1 million bits That means you can measure a bit error rate down to 10 one error per million A smaller number reduces the duration of the whole Fast Eye Mask measurement A larger number increases the precision of the measured bit error rates e Number of Errors After this number of errors the measurement stops for
244. rectory for the example user patterns is C N4901A Patterns Demo You can access this folder from the Serial BERT s Start menu N4901A Data gt Pattern Any changes you make to the patterns in this folder are non recoverable there is no backup folder To avoid inadvertent changes it is recommended that you copy the demo files to a different directory before loading them Agilent Serial BERT N4901 2 User s Guide January 2006 29 This folder contains the following subdirectories with example patterns FDDI FDDI_Jitter ptrn The Fiber Distributed Data Interface FDDI data dependent jitter test pattern is used for testing FDDI components or physical links The pattern is 1280 bits long and is transmitted continuously during the test by repeating the pattern The sequence causes a near worst case condition for inter symbol interference and duty cycle base line wander e FDDI_Wander ptrn The FDDI baseline wander test pattern is used to test FDDI components for the effects of a change in the average DC level of the signal The first 45 000 bits are of the code group 01010 the last 45 000 bits are of the code group 10101 Fiber Channel Gigabit Ethernet SPAT ptrn This is a Fiber Channel supply noise stimulus test pattern e RPAT ptrn This is a Fiber Channel random data pattern e K28 7 ptrn This pattern is used to test Fiber Channel and Gigabit Ethernet elements It has a large low frequency content that is
245. red bit error rate by the clock frequency If for example you want a bit error rate of 3x10 and you are running a test at 12 5 Gb s the frequency of the external signal is 3x107 x 12 5x109 3750 Hz Using the Serial BERT to Insert Errors To configure the Serial BERT to automatically insert errors in the output stream to achieve a desired bit error rate 1 Click the Error Add Setup icon in the PG Setup menu Jal 2 Select Internal 3 Use one of the following options to define the Error Insertion Rate Error Rate allows you to select a BER from a list Bits between Errors allows you to specify how many regular correct bits are sent without errors before inserting the next error 4 Click OK to close the dialog box Agilent Serial BERT N4901 2 User s Guide January 2006 89 TIP To find out how your DUT reacts on very small bit error rates set up the pattern generator to enter errors once every 10 bits and run a longer accumulative test You can then find the DUT s true error rate by calculating the difference between the bit error rate set up in the pattern generator and the accumulated bit error rate found by the error detector NOTE Ifyou set up too high an error rate the error detector will not be able to synchronize to the incoming pattern When setting up an error rate always make sure that the synchronization threshold is higher than the bit error rate Error Addition Insertion Reference The fo
246. right hand corner of the instrument display wi 2 Click the Enable Touchscreen icon to enable the touchscreen keyboard and mouse again Calibrating the Touchscreen In normal usage you should not need to calibrate the touchscreen display But if you feel that it needs to be realigned with the video display do the following 1 Ifthe title bar of the Serial BERT user interface is not visible select Title Bar from the Utility menu 2 Minimize the Serial BERT user interface window 3 In the system tray next to the clock in the lower right corner right click the Touchscreen Settings icon to open the menu with the available options Q wl 11 47 AM 4 Select Calibrate then Device 1 and Normal A blank screen with a target appears 5 Click the target to start the calibration When the calibration is complete the calibration window will close Agilent Serial BERT N4901 2 User s Guide January 2006 293 Configuring the Double Click You can adjust the behavior of the touchscreen display to simulate the double click action of a typical personal computer or trackball To configure the touch settings 1 Select Display Utility from the Utility menu This opens the Configuration Utilities dialog box 2 Click Display to open the Display Area Configuration dialog box 3 Click the Timed Double Touch slide bar to adjust the amount of time that can pass as you touch the display screen You can preview your adjustment by cli
247. rnating Patterns Alternating patterns can be used To systematically insert errors in the bit stream at particular positions The error detector expects pattern A only If you set up pattern B so that there are only minor differences when you run the test the error detector will detect the changed bits as errors 92 Agilent Serial BERT N4901 2 User s Guide January 2006 To see how long it takes your DUT to settle You can set up an easy to process output stream as the standard pattern and a difficult output stream as the alternate pattern for example all Os for the standard pattern and alternating 1s and Os as the alternate pattern You can then track how long it takes for your DUT to recover after the alternate pattern has been sent Pattern Alternation Procedures The Alternate Pattern Control dialog box controls how the output stream switches between the standard pattern and the alternate pattern You have the following possibilities for setting up alternating patterns Setting Up a Periodical Output To configure the pattern generator to generate periodical data output 1 Press the Alt Pattern and Aux In icon in the PG Setup menu to open the Alternate Pattern Control dialog box 2 Select one of the following options Continuous A Only pattern A is output repeatedly Continuous B Only pattern B is output repeatedly Alternate AB The output stream switches between pattern A and pattern B
248. ropriate printer driver installed The printer properties and capabilities will vary depending on the printer you have installed and configured for the instrument To print a hardcopy of the instrument s display to the default printer Click Print from the File menu on the main menu bar 298 Agilent Serial BERT N4901 2 User s Guide January 2006 NOTE Configuring Printer Properties You can change the properties of a selected printer 1 On the File menu click Print Setup 2 Click Configure Printer This opens the Printers dialog box 3 In the File menu in the Printers dialog box select from the following options Server Properties Click this option to change such items as printer margins spooler warnings and alarms and port configurations Properties Click this option to change such items as pre determined access times security items and LAN protocols Document Defaults Click this option to change such items as pages per sheet print quality paper select and portrait vs landscape layouts 4 Select Close in the File menu of the Printers dialog box The new settings will are applied and the dialog box closes The printer properties and capabilities will vary depending on the printer you have selected Printing Problems You should check the following when your printer fails to print 1 Check if the printer is turned on and the printer cable is securely attached to both the printer and the in
249. ror analysis Trigger Out Use this port to send a trigger signal to another connected device such as an oscilloscope Error Out This signal can be used to trigger an external instrument to help in error analysis When any number of errors are located within a 128 bit data block a single trigger pulse is sent If continuous errors are detected the error out signal would be a clock signal n Errors within 5 z sg bc gaaion Eron pocong aani yrs g Data IN A A Error Out ANAM single error marking 102 Agilent Serial BERT N4901 2 User s Guide January 2006 e Aux Out This port can be used to output the clock signal or the data signal to another connected device The following figure shows how the clock signal is directed to Aux Out in Clock mode CLOCK MODE CLK IN AUX OUT I CDR Recovered CLK DATA IN Comparator 7 gt e internal Signal 0 1 Threshold Analyzer Data In Port Termination To ensure a valid setup and to protect the devices from damage proper termination must be specified for both Data In connections You can specify the termination by entering the termination voltage in the respective field CAUTION Selecting the wrong termination may damage your device Why Can Wrong Terminations Damage Your Device Choosing wrong terminations may cause your device to output voltage levels that are not as exp
250. rs and the display options that are specific to this measurement Additionally some information is provided to explain the theoretical background behind the jitter calculations Properties that can be specified on the various tabs of the Properties dialog box Agilent Serial BERT N4901 2 User s Guide January 2006 249 Parameters Tab NOTE Ifyou modify the parameters on this page you have to rerun the measurement to update the results Set the data acquisition parameters e Acquisition Depth Suitable numbers can be chosen from the drop down list The Acquisition Depth determines the length of the time record used for the FFT A smaller number reduces the duration of the Spectral Jitter measurement and its precision A larger number increases the measurement duration but also the frequency resolution of the measured spectral components The frequency resolution of the measurement is data rate Acquisition Depth For example If you have a data rate of 2 5 GHz and an Acquisition Depth of 128 Kbit the frequency resolution is 19 0735 kHz The relations are illustrated in the following figure 1000000 1 GBitis 2 5 GBit s 100000 4 7 7 De reneass a a a eon 3 125 GBit s i i i i 10 3125 GBit s D O S Frequency Resolution Hz 16KBit 32KBit 64KBit 128KBit 256KBit 512KBit 1MBit 2MBit 4MBit SMBit 16MBit 32MBit Acquisition Depth NOTE A high Acquisition Depth requires a high degree o
251. s necessary to send complete patterns This applies to user patterns alternating patterns memory based PRBN patterns mark density patterns and zero substitution patterns Therefore when the Serial BERT loads a pattern into its internal memory it repeats the pattern as necessary until the pattern reaches a block boundary For example if a 1023 bit pattern is loaded this is repeated in the RAM 512 times and thus occupies 1023 blocks Or if a 192 bit pattern is loaded this is repeated 8 times and occupies 3 blocks as indicated in the following figure This applies to all memory based patterns available to the Serial BERT Block 0 bit 192 bit 384 ee BE nnn Block1 pit 64 bit 256 bit 448 Cl Dee E Gm ti Block 2 bit 128 bit 320 R E D CEERRREMN During test run the Serial BERT always sends out all blocks of a pattern In the case of the 192 bit pattern mentioned above the pattern is repeated 8 times when sent out in effect making it a 1536 bit pattern How the Serial BERT Handles The same behavior of repeating a pattern to fill up complete blocks Alternating Patterns also applies to alternating patterns In this case each half of the pattern is loaded into a separate location in the RAM When an alternating pattern is generated first all blocks of pattern A are sent then all blocks of pattern B are sent and so on A0 A1 A2 BO B1 B2 AO A1 A2 A AJA AJAJA A A B B B BiB B BIB AJA
252. sampling point is the decision factor as to whether the incoming bits are identified as logic 0 s or 1 s To measure the accurate bit error ratio at the input port false readings of logic 0 s or l s must be avoided Therefore the sampling point must be set to the optimum location within the data eye The functions within the Sampling Point Setup window allow you to e Prepare the error detector for the incoming data signal regarding the connector termination e Graphically display the eye diagram in terms of voltage input delay and a BER threshold e Adjust the location of the sampling point What is an Eye Diagram An eye diagram provides a way to view all possible transitions between adjacent bits of a data stream by overlaying them on the display of a high speed oscilloscope It is typically produced by triggering the oscilloscope with a synchronous clock signal 114 Agilent Serial BERT N4901 2 User s Guide January 2006 When the oscilloscope is triggered by a clock signal data patterns will not be examined Instead if persistence is set high enough the shape of an eye will be observed The eye is bounded by overlaid logic 1 and 0 voltages top and bottom and multiple 0 to 1 and 1 to 0 transitions left and right ae p TANA l m n l O 2 l wy I I a EES I Different portions of the bit pattern Eye diagram Overlay of all portions of the bit pattern The sampling point o
253. se this option to select the bits from pattern start to the cursor The available options for the Action are 0101 0101 Choose this option to fill the range with alternating highs and lows Invert Bits Choose this option to invert the high low logic in the range All Zeros Choose this option to fill the range with all logic lows All Ones Choose this option to fill the range with all logic highs Status Bar The types of the patterns that are currently loaded to the pattern generator and the error detector are always displayed in the left half of the status bar at the bottom of the screen PG Ptm 297 1 PRBS ED Ptm 291 1 PRBS Select Pattern Dialog Box This dialog box lets you define the pattern you want to use for testing It contains the following elements With this checkbox you can determine whether both pattern generator and error detector use the same data pattern If you clear this checkbox the Error Detector Pattern tab appears where you can select a different pattern for the error detector Agilent Serial BERT N4901 2 User s Guide January 2006 The checkbox is automatically cleared when you load a user defined sequence to the pattern generator See Sequence Mode Characteristics on page 53 for details Pattern Types You can choose between the following types of patterns User Pattern from File With this option you can load a pattern that is stored in a file Use the Browse button t
254. sian Marker introduction 157 Output Levels 196 Output Timing 162 GPIB Address 290 Graph Tab Eye Opening 222 Output Levels 197 Output Timing 178 Spectral Jitter 255 H Hardware generated patterns 19 High Level 199 High Level Std Dev 200 I Input Range 310 Error Detector 106 Internal Clock Source 80 Interval Results 268 Introduction to the Serial BERT 6 invalid 182 Inverted Data Error Detector 105 Pattern Generator 75 J Jitter example 159 fast total 166 Jitter Mean 181 L LAN Connectors 291 Level Results 198 Linear Markers 157 Loading Pattern 38 Pattern directly from Pattern Editor 39 Logic Families 68 Logic Levels Setup 71 Low Level 199 Low Level Std Dev 200 M Manual Pattern Alternation 94 Manually Inserting Errors 88 Mark Density 47 Markers Gaussian 157 introduction 156 Linear 157 Mean Level 199 Measured Points 222 Measurement Log 141 282 Measurement points defaults 231 Memory based Patterns 19 Min BER for Q 197 More Errored 0 s than 1 s 279 MUX DEMUX Pair Connections 15 N No of points 181 Noise Threshold 255 Normal 105 not applicable 182 Number of Compared Bits Eye Opening 217 Fast Eye Mask 234 Output Levels 192 Output Timing 173 Number of Errors Eye Opening 217 Fast Eye Mask 234 Output Levels 192 Output Timing 173 Number of Top Frequencies to Show 255 Number of Valid Points 235 Numerical Results Eye Opening 223
255. sian marker is used when the dBER vs Threshold Graph is displayed This graph shows the relationship between the decision threshold and the absolute values of the derivative of the bit error rate dBER dTh A linear scale reveals the distribution more clearly than a logarithmic scale see ABER vs Threshold Graph on page 196 In the example below Mu and o Sigma will be the same as the Level and Standard Deviation results calculated by the measurement 1 Trace 500 Points 800 0 mv 900 0 mv A0V AAV 1 2 V 1 3 Y 1 4 Y ASV 1 6 Y 1417Y 1 8 Y Linear 20i o 000e 0 000e 0 4 000e 160 Gaussian Marker Value Mu ATV Sigma 16 7 mY Kappa 1 018e 0 6 000e 0 1 000e 1 1 400e 1 1 800e 1 2 200e 1 2 600e 1 3 000e 1 0 8 000e 0 1 200e 1 1 600e 1 2 000e 1 2 400e 1 2 800e 1 3 200e 1 Agilent Serial BERT N4901 2 User s Guide January 2006 But your dBER distribution may also look like this 1 Trace 316 Points 1 0 900 0 mv 800 0 mv 700 0 mv 600 0 mv 500 0 mV 400 0 mv 300 0 mv 200 0 mv 100 0 mv oov Linear 3 000e 0 9 000e 0 1 500e 1 2 100e 1 2 700e 1 3 300e 1 3 900e 1 4 500e 1 5 100e 1 5 700e 1 6 300e 1 0 000e 0 6 000e 0 1 200e 1 1 800e 1 2 400e 1 3 000e 1 3 600e 1 4 200e 1 4 800e 1 5 400e 1 6 000e 1 The measurement will calculate the Level and Standard Deviation results from all data points The marker allows you to measure the u Mu and o Sigma of the individual peaks Agilent S
256. sible without hitting the maximal length limitation to extend the time where bits can be counted The following figure illustrates how the duration of the Gate In signal can influence the bit count time the time where bits are actually counted Valid after Gate CDR Settling Time Synchronization Time Data Input lt i Begin Margin End Margin lt yo Bit Count Time Gate Input k 4 lt rie gt i Gate Active i Gate Passive Gate Active is the interval in which the Gate In signal is active The Bit Count Time is the part of the burst from which the bits can actually be counted The remainder of the burst is covered by the Begin Margin and End Margin The CDR Settling Time is the time that the error detector requires in CDR mode to get the clock from the data stream The Synchronization Time is the time the error detector requires to synchronize to the pattern This time depends on the pattern type PRBS or memory based The Valid after Gate indicates the minimum length of data that has to arrive at Data In after the Gate In signal is deactivated Interval between Bursts There is no requirement how long a signal should be available at Data In before gating starts the data should simply be stable A pause is however required after the gating to allow the error detector to evaluate the results This time is known as Gate Passive Agilent Serial BERT N4901 2 User s Guide January 2006 135 The followin
257. slider with the mouse or your finger if you are working directly on the Serial BERT This is particularly useful if you have set the display to show Equal BER at BER Threshold By moving the slider you can investigate the eye openings at destinct BER thresholds NOTE If no slider for the BER threshold is displayed make sure that you have enabled the BER Threshold option on the View tab and that a suitable BER is entered How to Change the BER Range of a Color The bit error ranges are set automatically but you can change these areas To do so Move the handles of a color with the mouse or your finger if you are working directly on the Serial BERT This is particularly useful if you have set the display to show the Pseudo Color Plot By default you may see a graph like the one below 1 Trace 9510 Points Ul 401 878 ps All Errors 4 000e 0 600 0 m 1 000e 1 400 0 mv 4 000e 2 200 0 mv 1 000e 3 oov 1 000e 4 200 0 mv 1 000e 5 400 0 mv 1 000e 6 600 0 mY 1 000e 7 E sscale gt 0 448 UI 0 149 UI 0 149 UI 0 448 UI 0 746 UI Relative 0 597 UI 0 299 UI 0 000 UI 0 299 UI 0 597 UI 216 Agilent Serial BERT N4901 2 User s Guide January 2006 Now you may wish to know more precisely what happened between yellow and red After extending the range of yellow such information becomes available 41 Trace 9510 Points Ul 401 876 ps All Errors n 600 0 mv 400 0 mv 200 0 mY 0 0v 200 0 mv 400 0 mY
258. splayed You do not need to run the measurement again See View Tab on page 176 Graph Tab on page 178 and How to Change the Colors of the Graph on page 153 for details 3 Press OK when you have made all required changes to close the Properties dialog box DUT Output Timing Jitter Reference The DUT Output Timing Jitter measurement returns the results in a graphical and in a numerical form The following sections describe the results They provide explanations of the measured parameters and the display options that are specific to this measurement e Additionally some information is provided to explain the theoretical background behind the jitter calculations TIP You have access to the user documentation including User s Guide over the instrument s Start menu Start gt Programs gt N4900 Series Documents Properties that can be specified on the various tabs of the Properties dialog box 172 Agilent Serial BERT N4901 2 User s Guide January 2006 Parameters Tab NOTE If you modify the parameters on this page you have to rerun the measurement to update the results Set the criteria for moving to the next sample point Number of Compared Bits After this number of compared bits the measurement stops for the current sample point and moves to the next one Number of Errors After this number of errors the measurement stops for the current sample point and moves to the next one
259. strument s printer connector 2 Check if the printer driver for your printer is installed 3 Check if the printer port is configured correctly For more information on printer problems see the printer s operation manual Agilent Serial BERT N4901 2 User s Guide January 2006 299 File Management File Management Procedures This section provides information on saving and recalling instrument states log files patterns and screenshots You can save the files on the local disk Note that the local disk must not be re partitioned in a Windows 2000 environment You can also plug an USB stick into the rear of the instrument for saving and recalling files No driver is required because Windows has a built in driver that supports USB memory drives In the file browser you can access the USB stick via the Removable Disk drive To avoid loss of data stop the USB device and wait until the LED light of the USB stick is off before unplugging the USB stick To avoid loss of data when removing the USB stick 1 Minimize the GUI 2 In the system tray next to the clock in the lower right corner right click the Safely Remove Hardware icon A window containing a list of the USB devices appears 3 Click on the popup message Safely Remove Mass Storage Device 4 Confirm the acknowledgement message Save to remove hardware 5 Check that the LED light of the USB stick is off and unplug the USB stick Saving Instrument States
260. subcontract Software is delivered and licensed as Commercial computer software as defined in DFAR 252 227 7014 June 1995 or as a commercial item as defined in FAR 2 101 a or as Restricted computer software as defined in FAR 52 227 19 June 1987 or any equivalent agency regulation or contract clause Use duplication or disclosure of Software is subject to Agilent Technologies standard commercial license terms and non DOD Departments and Agencies of the U S Government will receive no greater than Restricted Rights as defined in FAR 52 227 19 c 1 2 June 1987 U S Government users will receive no greater than Limited Rights as defined in FAR 52 227 14 June 1987 or DFAR 252 227 7015 b 2 November 1995 as applicable in any technical data Safety Notices CAUTION A CAUTION notice denotes a hazard It calls attention to an operating procedure practice or the like that if not correctly performed or adhered to could result in damage to the product or loss of important data Do not proceed beyond a CAUTION notice until the indicated conditions are fully understood and met WARNING DANGER A WARNING notice denotes a hazard It calls attention to an operating procedure practice or the like that if not correctly performed or adhered to could result in personal injury or death Do not proceed beyond a WARNING notice until the indicated conditions are fully understood and met Trademarks Windows and MS Wi
261. sured points refers to the actual sampling point If synchronization and auto alignment were successful the current sampling point is the optimum sampling point Instead of UI one Unit Interval is equal to one system clock period the relative time can also be specified in seconds The Voltage is the decision threshold voltage at this measurement point The voltages of the measurement points can be set as absolute voltages as offset voltages or as percentages This is done on the Parameters page of the Properties dialog By default the six measurement points are symmetrically placed as illustrated in the figure below Threshold voltage Analyzer threshold Time UI Actual sampling point These settings can be changed and up to 32 measurement points can be defined NOTE Critical areas at the error detector are generally close to 0 5 UI Agilent Serial BERT N4901 2 User s Guide January 2006 231 Fast Eye Mask Procedures This section shows how to set up and perform a Fast Eye Mask measurement As an example we measure the eye of a shielded cable This requires the following steps How to Prepare the Fast Eye Mask Measurement on page 232 How to Execute the Fast Eye Mask Measurement on page 233 How to Optimize the View of the Results on page 233 How to Prepare the Fast Eye Mask Measurement To prepare a Fast Eye Mask measurement to test a shielded cable 1
262. surement is used to measure the timing and jitter behavior for a device under test DUT It uses a bit error rate BER measurement to evaluate the shape of the eye for the output signal of the DUT It also analyzes the jitter separates the random jitter and deterministic jitter components and estimates the total jitter A direct result is the determination of the optimum sampling point delay for receiving data from the DUT with maximum confidence DUT Output Timing Jitter includes the Fast Total Jitter measurement that can be used to measure the total jitter for devices which generate a very low error density in a reasonable time span Output Levels The Output Levels measurement allows you to characterize the behavior of the output levels of a device under test DUT The sampling delay is fixed The error detector s decision threshold is automatically swept within a user defined range A direct result is the determination of the optimum decision threshold level for receiving data from the DUT with maximum confidence Eye Opening The Eye Opening measurement generates a three dimensional bit error rate BER diagram as a function of the sample delay and the sample threshold With this measurement the complete eye of the DUT output signal is measured The results comprise the voltage and timing of the eye opening and the optimum sampling point Error Location Capture The Error Location Capture measurement allows to capture the actual pos
263. surement that has failed the criterion for the Bit Error Rate 1 Trace 65535 Points 32 dB 24 dB 16 dB 8 dB 0 dB 8 dB 16 dB 24 dB 32 dB 40 dB Noise Threshold 40 dB 48 dB 56 dB Logarithmic 100 00 KHz 10 00 MHz 1 00 GHz 1 00 MHz 100 00 MHz Top Freq al Terminal Optical Noise Power eo 46 664 dB 30 176 dB Agilent Serial BERT N4901 2 User s Guide January 2006 253 204 Power scale Absolute vs relative values View Tab The graph shows the noise versus frequency The View tab provides you with various options for analyzing the data and setting frequency ranges for evaluation The linear power scale is well suited for identifying large peaks in the spectrum The View tab allows you to change the power scale from linear to dB The dB scale is a logarithmic scale The dB scale allows you to examine the whole power range The following figure shows an example of a double logarithmic display 1 Trace 65535 Points 24 dB 32 dB Noise Threshold 40 dB a 56 dB 64 dB 72 dB 80 dB 88 dB 96 dB 104 dB Logarithmic 100 00 KHz 10 00 MHz 1 00 GHz 1 00 MHz 100 00 MHz The figure above shows the absolute power scale and values as calculated by the FFT You can also calculate relative values Relative values can be normalized to either the total power True Relative or to the power of a selectable frequency bin Relative
264. t gt 256 kbit 16 bit gt 1 Mbit gt 512 kbit 32 bit gt 2 Mbit gt 1 Mbit 64 bit gt 4 Mbit gt 2 Mbit 128 bit gt 8 Mbit gt 4 Mbit 256 bit gt 16 Mbit gt 8 Mbit 512 bit How the Serial BERT Generates Memory Based Patterns The basic concept of how the Serial BERT generates memory based patterns is relatively simple The Serial BERT has 32 Mbits of internal RAM memory used to store data patterns The RAM is organized in 64k of 512 bit blocks During test run the pattern generator reads this memory bit by bit and generates the corresponding voltage at the output port 1 high 0 low The error detector compares the incoming signal voltage data bits with the bits in the memory and counts the errors The Serial BERT has actually three RAM memory locations where patterns are stored in the Pattern Editor memory in the pattern generator and in the error detector When you modify patterns with the pattern editor the patterns in the pattern generator and error detector are not automatically updated For this you need to write the pattern to the pattern generator and or the error detector There the pattern is saved in the RAM of the respective module Agilent Serial BERT N4901 2 User s Guide January 2006 27 Why the Serial BERT Repeats Memory Based Patterns To generate output signals the Serial BERT can send only complete 512 bit blocks of data To maintain predictability of the output signal it i
265. t is out of hardware specification Please adjust signal levels Enable Delay Control Input again pane You can enable the input after reducing the voltage to safe level Delay Control Input Procedures To set up a jitter tolerance test 1 Press PG Output Setup in the PG Setup menu 2 Disable the Delay Control In port by clearing the Delay Control Input checkbox 3 Connect the device that supplies the jitter signal to the Delay Control In port 86 Agilent Serial BERT N4901 2 User s Guide January 2006 4 Enable the port again by selecting the Delay Control Input checkbox TIP While monitoring the BER you can use this checkbox to switch jitter on and off and view the effect on the BER Delay Control Input Reference This checkbox in the PG Output Setup window is used to disable or enable the Delay Control In port where you can supply an external signal to add jitter to the output signal Agilent Serial BERT N4901 2 User s Guide January 2006 87 Error Addition Insertion To test error correction algorithms alarms and other functions that are imbedded in the data pattern you can insert logic errors flipped bits into the pattern The instrument provides several options in the pattern generator s Error Add dialog box for inserting error bits manually or automatically Error Addition Insertion Procedures You have the following options for inserting errors into the output data stream Manually
266. t Hex Description This is a simple standard pattern Count 1 Length 128 Data 01 02 03 04 10 20 40 80 ff 00 ff 00 ff 00 ff 00 Version EPA 2 0 Format Text Description This is a simple alternating pattern Count 2 Length 96 Data 01 02 03 04 10 20 40 80 ff 00 ff 00 Data 01 02 03 04 10 20 40 80 00 ff 00 Pattern Resolutions and Lengths The Serial BERT has 32 Mbits 8 MB RAM for storing patterns As discussed in How the Serial BERT Generates Memory Based Patterns on page 27 all patterns are loaded into the RAM until a 512 bit boundary is reached If the pattern is an odd number of bits long it is loaded into the RAM 512 times Because of the restricted memory space there are also restrictions to the pattern lengths Agilent Serial BERT N4901 2 User s Guide January 2006 For example a user pattern that is 256 001 bit bit long would have to be loaded into the RAM 512 times to reach the 512 bit boundary Such a pattern would occupy more than 128 Mbits in the RAM The Serial BERT handles such patterns by rounding them off to the nearest length that would fit in the memory For the above example the length would be rounded off to 256 000 bits This rounding off factor is known as the pattern s resolution The following table lists all pattern resolutions lt 64 kbit lt 32 kbit 1 bit gt 64 kbit gt 32 kbit 2 bit gt 128 kbit gt 64 kbit 4 bit gt 256 kbit gt 128 kbit 8 bit gt 512 kbi
267. t to 0 V This is indicated in the PG Output Setup screen 2 Connect the DUT as necessary 3 Terminate any non connected Data Out and Clock Out ports normal and complementary 4 Ifthe outputs are disabled press the OV Disable button to enable them To avoid damaging your device set up the devices properly before making connections ESD can damage or destroy electronic components Coaxial cables with both ends unconnected may store electrostatic charges Before connecting any coaxial cable of this sort to a device or instrument momentarily short the center and outer conductors of the cable When making connections ensure the proper use of a grounded resistor isolated wrist strap Agilent Serial BERT N4901 2 User s Guide January 2006 13 14 Connect your device according to the diagrams below Diagram 1 Connections for Different In Out Data Rates Often used for multiplexers or when the data rate received by the device is not equal to the data rate sent from the device LEGEND C Clck D Date Diagram 2 Connections for Differential Inputs Application is the same as in Diagram 1 Illustrates how to connect to a device requiring differential signals Agilent Serial BERT N4901 2 User s Guide January 2006 Diagram 3 Connections for an Amplifier Often used for amplifiers Diagram 4 Connections for a Flip Flop Often used for flip flops Diagram 5 Connections for a MUX DEMUX Pair
268. ta even more thoroughly NOTE When saving the measurement the following settings are not saved e selected pattern e settings of the pattern generator e settings of the error detector To save these settings see Saving Instrument States on page 300 Besides saving a measurement as a file you can also export the data to any spreadsheet application Agilent Serial BERT N4901 2 User s Guide January 2006 149 Advanced Analysis Procedures This section describes some basic functions that apply to all measurements If you want to store the measurement results for later use you can either save them as an mcp file to be opened in the Serial BERT software or as a txt file for spreadsheet applications How to Save Measurements To save a measurement as an mcp file for later use in the Serial BERT user interface 1 After your measurement has finished and the results are displayed select Save Measurement from the Analysis menu 2 Select a proper file name and location and press the Save button How to Export Measurements To export the measurement data into a txt file for later use in external spreadsheet applications 1 After your measurement has finished and the results are displayed select Export Data from the Analysis menu 2 Inthe Export dialog box select the path and file name for the txt file Alternatively you can activate the Clipboard checkbox to copy the data to the clipboard In this case you
269. tax Count lt 1 2 gt This is the length of the pattern in bits For a standard pattern of 128 bits for example the value is 128 For an alternating pattern it is the length of one of the patterns only The Data tag identifies the start of the pattern data itself For standard patterns there is only one Data tag Alternating patterns require two Data tags Note that the actual data starts on the line after the tag Syntax Data lt CR LF gt lt data contents gt Agilent Serial BERT N4901 2 User s Guide January 2006 25 26 Standard Pattern in Text Alternating Pattern in Text In dual format each character represents one bit only the numbers l and 0 are allowed In hex or text format each character represents four bits only the characters 0 to 9 and A to F are allowed not case sensitive In binary format each character represents eight bits all ASCII characters are allowed Example X in binary format or 88 in hex or text format produce the same bit sequence as 10001000 in dual format In hex or text format the data must be separated by one white space The white space can be a tab a space or a new line In hex or text format For pattern lengths that are not multiples of 8 bits the data has to be padded to provide the correct number of hex characters The extra bits are ignored ASCII Pattern Files Examples Version EPA 2 0 Forma
270. ter 181 Different In Out Rates Connections 14 Differential 105 Differential Inputs Connections 14 Double Click Configuration 294 DUT Connecting 13 DUT Output Timing Jitter measurement 162 E ED CLK LOSS 276 Edge Resolution Optimization Eye Opening 218 Output Levels 192 Output Timing 173 Edit Pattern Window 40 57 Editing Patterns 37 Equal BER at BER Threshold 211 Error Accumulation 141 Error Add button 88 Error Add input 64 Error Detector 1 0 threshold 105 Auto Threshold 113 Aux Out port 103 Clock In port 101 Data In port 101 Data In setup 105 Error Out port 102 Gate In port 102 Trigger Out port 102 Error Indicator 278 Error Insertion Constant Rate 89 External Signal 89 309 Manually 88 Error Location Capture measurement 225 Error Out port 102 Estimated Total Jitter 181 Example of an Output Level Measurement Changing Properties 190 Improving the Display 189 Running the measurement 187 Example Patterns 29 External Clock Source 80 External Data Source Connections 17 Eye Diagram 114 Eye Edge 115 Eye Height 272 Eye Opening measurement 209 Eye Width 272 F False Synchronization 130 Fast Eye Mask measurement 230 Fast Total Jitter at BER 174 Fast Total Jitter measurement 162 Explanation 166 Numerical results 183 Result Display 182 File Management 300 Flip Flop Connections 15 G G 821 Measurements 267 Gate In port 102 Gate In signal timing 135 Gating 76 Gauss curve 158 Gaus
271. that you are measuring at the edges of the data eye Pattern synchronization can be lost during sampling point adjustment If Auto Sync is enabled the Serial BERT will resynchronize the patterns automatically If the manual synchronization mode is selected you need to synchronize the error detector by clicking the Pattern Sync icon and then the Sync Now button Sampling Point Setup Reference Depending on the options of your Serial BERT some of the following functions may not be valid for your instrument See Introduction to the Serial BERT on page 6 and Which Features are Available on page 8 for a description of the available options The Sampling Point Setup window provides the following elements 0 1 Threshold The 0 1 decision threshold is the reference voltage level for the evaluation of the incoming data signal At the sampling point the data signal voltage ist compared to the 0 1 decision threshold Signals greater than this threshold will be measured as a logic 1 signals below this threshold will be measured as a logic 0 0 1 Threshold cannot be used while Avg 0 1 Threshold is selected Agilent Serial BERT N4901 2 User s Guide January 2006 119 The 0 1 Threshold field allows the manual entry of the 0 1 threshold and displays the current value in millivolts You can adjust the 0 1 threshold in the following ways Click inside the text field and use the numeric keypad or front panel knob to enter a val
272. the measurement again Pass Fail tab On the Pass Fail tab you can specify the conditions that must be met by the DUT to pass a test Changes here do not require a new test run The data will be newly calculated and the results will be displayed View tab Graph tab and Color tab With the functions on these tabs you can modify the graphical display of the results at any time without rerunning the measurement This allows you to analyze the results in different ways and with different preferences 3 After you have made your changes on the various tabs press Apply to employ the settings without leaving the Properties dialog box Or press OK to confirm the modifications and close the Properties dialog box 152 Agilent Serial BERT N4901 2 User s Guide January 2006 How to Change the Colors of the Graph You can customize the colors of the graphical display This function is common to all measurements 1 In the Properties dialog box switch to the Colors tab Properties x Parameters Pass Fail View Graph Color Properties Color Set Standard Colors x Color Palette FreqRangesColor GridColor Ti PowerMarkerColor Depending on the current measurement type you can change BackColor The background color of the graphs default is white BERMarkerColor The color of the BER Threshold indicator default is red ForeColor The foreground color of the s
273. the current sample point and moves to the next one This allows you to speed up the measurement You can switch off this option if only the number of compared bits is important NOTE The measurement moves to the next sample point when the first of the two criteria is reached 234 Agilent Serial BERT N4901 2 User s Guide January 2006 Set the criteria how the sampling points of the eye are displayed e Timing Unit Choose between Unit Interval and Seconds to set the timing values for the measurement points on the Parameters tab This setting does not affect the display of results Threshold Type Choose between Absolute Offset or Percentage to set the thresholds for the measurement points Absolute You specify the thresholds for the measurement points as absolute voltages Offset You specify the thresholds for the measurement points relative to the threshold voltage of the actual sampling point Percentage You specify the thresholds for the measurement points as a percentage of the current eye opening voltage of the error detector Threshold voltage 0 4 0 16 0 16 0 4 Actual sampling point Depending on the quality and characteristics of the eye opening the resulting 50 threshold may deviate from the decision threshold defined in the loaded setting Number of Valid Points Change the Number of Valid Points if you wish to use less or more measurement poi
274. the following self tests are automatically run e Pattern Generator Power Up Test Error Detector Power Up Test The current system state is saved and the instrument is re initialized before performing the selected tests The saved system state is restored upon completion of the self test When Self Test is selected from the Utility menu the Self Test Options dialog box opens In this dialog box you can select additional self tests that will run when you click Run Test During accumulation or while auto search functions are in progress Self Test is not available The pass fail results of the self test are displayed in the Self Test Options dialog box In case of failed tests you can view detailed error messages in a separate window Self Test Procedures To run a self test 1 Select Self Test from the Utility menu This opens the Self Test Options dialog box 2 Select the desired tests and click the Run Test button 3 After the self test has finished the results are displayed next to each selected test 4 Ifa test failed you can view detailed error messages by clicking the respective Message button Self Test Reference This section contains information about the Self Test Options dialog box used for setting up and running a self test and for viewing the results Agilent Serial BERT N4901 2 User s Guide January 2006 305 Self Test Options The Self Test Options dialog box offers to select the following tests
275. time 4 Now you can use the numeric keypad or front panel knob to manually adjust the value in the Data Delay field The Sample Point in the eye diagram moves horizontally as you change the value TIP While adjusting the data input delay you can monitor the BER bar on the analyzer display or listen to the BER warning tones Higher BERs may indicate that you are measuring at the edges of the data eye NOTE Pattern synchronization can be lost during sampling point adjustment If Auto Sync is enabled the Serial BERT will resynchronize the patterns automatically If the manual synchronization mode is selected you need to synchronize the error detector by clicking the Pattern Sync icon and then the Sync Now button 118 Agilent Serial BERT N4901 2 User s Guide January 2006 TIP NOTE NOTE NOTE How to Adjust the 0 1 Decision Threshold 1 Press the Sampling Point Setup icon in the ED Setup menu 2 Clear the Avg 0 1 Threshold checkbox 3 Click 0 1 Threshold Center to move the sampling point vertically towards the center of the eye at the selected delay You may click Cancel at any time 4 Now you can use the numeric keypad or front panel knob to manually adjust the value in the 0 1 Threshold field The Sample Point in the eye diagram moves vertically as you change the value While adjusting the 0 1 threshold you can monitor the BER bar on the analyzer display or listen to the BER warning tones Higher BERs may indicate
276. tions are available Differential If differential mode is selected both input ports need to receive a signal The actual data signal is measured as the voltage difference between the two incoming signals e Normal In normal mode only the Data In port receives the data signal the Data In port is inactive e Complement In complement mode only the Data In port receives the data signal the Data In port is inactive Data Inverted Activate this checkbox to invert the polarity of the error detector reference pattern This function is required if your device inverts data Threshold Only available if the input is set to normal or complement mode Choices are e Averaging The 0 1 threshold voltage is derived from the input signal via a low pass filter This will work fine for most applications e Manual input Enter a value for applications that do not provide a continuous data stream at the input for example any application using bursts because the averaged 0 1 threshold voltage will drift from the correct level when there is no input The manually set threshold voltage must of course be within the input range Agilent Serial BERT N4901 2 User s Guide January 2006 105 Termination In this field enter the termination voltage that is appropriate for the incoming data signal This selection should be made before the device is connected to the analyzer If the input is set to differential mode a termination voltage c
277. tops for the current sample point and moves to the next one e Number of Errors After this number of errors the measurement stops for the current sample point and moves to the next one This allows you to speed up the measurement You can switch off this option if only the number of compared bits is important NOTE The measurement moves to the next sample point when the first of the two criteria is reached Set the criteria for the sample threshold Resolution Specifies the distance between the sampling points The lower this value is the more sampling points you have in the selected voltage range The minimum step width is hardware dependent At the time being the minimum is 1 mV e Low Level This is the lower end of the measured voltage range High Level This is the upper end of the measured voltage range Edge Resolution Optimization Turns the resolution optimization on or off If this option is enabled the Serial BERT intelligently sets a resolution so that there are more sampling points at the edges This can greatly improve the results without dramatically increasing the duration of the test 192 Agilent Serial BERT N4901 2 User s Guide January 2006 In the following example a step width of 2 mV was used This has led to 502 data points per Volt 800 0 my Trace 502 Points 900 0 mv A0V AAV 12V A3V A4V 45V 4146Y A ATV 18V BER Threshold 3 000e 3 Logarithmic 1 000e 7 1 00
278. trated in the figure below Threshold Gradient 1 0 K Gradient 1 69 Q factor A straight line can be expressed as Y A BX where Y is the inverse error function of BER and X is D the decision threshold The following calculations are performed for the high and low level data ZX LY XY Uyay X 7 x 2 eA x A _ LY xX A _ B where n is the number of respective data points The results of the linear regression are displayed in the QBER vs Threshold graph 4 Trace 16 Points 800 0 mv 900 0 mV 10V AAV A2v 13v 14v 15v ABV 17V 1418Y Linear 4 12 20 28 38 44 52 60 68 76 84 206 Agilent Serial BERT N4901 2 User s Guide January 2006 This graph shows two straight lines The intersection of these lines marks the Q factor and the Q optimum Threshold Mathematically the standard deviation and mean values are calculated as oy ees B H B This calculation leads to the values of u 9 and 6 o Notes on the 0 Results You can specify the range of data points used for these calculations by means of the Min BER for Q and BER Threshold parameters see also View Tab on page 195 Use of the Q factor Calculations For example you can calculate the Q factor from relatively few data points with measured bit error rates between le 10 and le 5 The calculation tells you the optimum decision threshold and also the BER expected when the error detector uses this thresh
279. trument The instrument supports parallel and serial printers as well as USB and LAN connected printers To connect your printer to the instrument 1 Turn off power to the instrument and printer before connecting 2 Connect the printer to the instrument with the cable that came with your printer Once you have connected the printer to the instrument you will need to add the printer driver before you can use the printer Adding a Printer Use the following procedure to add a printer to the instrument 1 On the File menu click Print Setup then Add Printer This opens the Add Printer Wizard dialog box 2 Click My Computer for a local printer installation or click Network printer server for a printer connected to another server Before you can add a network printer you must connect to a Local Area Network LAN In addition your login must provide you with access to the network printers that you wish to add Please contact your network administrator or IT department for assistance 3 Follow the instructions in the Add Printer Wizard until you have selected a local printer and configured the printer port 4 If your printer is not in the list of available printers you will need a copy of the printer driver Select a printer from the displayed list and click the Have Disk button The instrument will then prompt you to put your printer driver disk into the floppy disk drive drive A 5 Follow all other instructions Agi
280. ts Measurement Parameters The measurement parameters are defined in the following List Time Eye Opening This is the maximum extension of the BER threshold contour line in sample delay direction eye width Note This value is different from the horizontal extension of the BER threshold bounding box Threshold Eye Opening This is the maximum extension of the BER threshold contour line in sample voltage direction eye height Note This value is different from the vertical extension of the BER threshold bounding box Optimal Sample Delay This is the sample delay coordinate of the center of a bounding box around the BER threshold contour line Optimal Sample Threshold This is the sample voltage coordinate of the center of a bounding box around the BER threshold contour line Unavailable Values If you make a measurement that results in an incomplete eye no closed contour line at the BER threshold the numerical results cannot be calculated In this case the software will issue a warning message If this happens you should first of all check the settings of the low and high decision threshold voltages Understanding the Result Parameters All result parameters are calculated from the BER threshold contour line and its bounding box So all parameters change with the BER threshold The result display of the Eye Opening measurement shows the maximum eye opening time the maximum eye opening voltage and the posit
281. ts and systems for high speed digital transmission equipment An important parameter of a digital system is the rate at which errors occur in the system The analyzer will compare the bits sent to your device with the bits received from your device It will record the number of bit errors and display results such as the bit error ratio BER In addition accumulated results can be analyzed to help you determine what caused the bit errors The analyzer can be set up to work with your device in many different ways To determine the best measurement setup you must first decide how you want to test your device Before actually setting up and making measurements you may want to consider the following questions e What is the purpose of your test e What are the testing requirements of your device What is the Purpose of Your Test Does the purpose of your test fall under one of the following categories e To monitor instantaneous BER during measurement setup experiments and adjustments See Instantaneous Measurements on page 258 To measure and log error performance over time See Accumulated Measurements on page 263 To measure the eye height and width of the incoming data signal at specific alignment BER thresholds See Eye Measurements on page 272 Agilent Serial BERT N4901 2 User s Guide January 2006 9 The following references may also be helpful in determining how to test your device Fora
282. ts width one for changing its height and position The following figure shows an exemplary DUT Output Timing Jitter measurement result with a Gaussian marker 1 Trace 43 Points Ul 8 000 ns All Errors pss Gaussian Marker Value Mu 0 528 UI 3 50e 9 Sigma 0 004 UI 2 80e 9 Kappa 3 849e 1 2 10e 9 1 40e 9 7 00e 8 Linear 0 68 UI 0 63 UI 0 57 UI 0 53 Ul 0 47 Ul Relative 0 70 UI 0 65 UI 0 60 UI 0 55 UI 0 50 UI 158 Agilent Serial BERT N4901 2 User s Guide January 2006 The marker readout provides the following information Mu u Mean DUT Output Timing Jitter measurement The position of the marker center on the time scale Output Levels measurement The position of the marker center on the vertical thresh old scale Sigma o Standard deviation The RMS value of the marked area Kappa K Linear scaling factor DUT Output Timing Jitter A Gaussian marker is used when the jitter graph is displayed It is Measurement most useful if deterministic jitter is present In this case the jitter graph shows more than one peak see also Estimated Total Jitter on page 165 The jitter values Mean RMS and Peak Peak refer to the whole jitter The Gaussian marker allows to measure the normal distribution of individual jitter peaks You can thus determine how parts of the jitter histogram contribute to the overall jitter values You may have measured a bathtub curve like the one shown below 1 0
283. ttern from the editor to the pattern generator Click this icon to send the pattern from the editor to the error detector Click this icon to send the pattern from the editor to the pattern gen erator and the error detector These functions follow Microsoft Windows conventions Click this icon if you want to jump to an arbitrary bit position See GoTo Bit Dialog Box on page 42 for details Click this icon to toggle between the insert and overwrite editing mode With nsert mode you insert bits at the cursor position shifting the rest of the pattern to the right With Overwrite mode you overwrite the bits at the cursor position Click this icon to toggle between viewing the pattern in hex or binary mode Click this icon to change the properties of the current pattern See Pattern Properties Dialog Box on page 43 for details Click this icon to open the Find Pattern Segment dialog box and search for a specified segment in the pattern See Find Pattern Seg ment Dialog Box on page 43 for details 40 Agilent Serial BERT N4901 2 User s Guide January 2006 Insert Click this icon to insert a block of data in the pattern at the current cursor position See Insert Block Dialog Box on page 43 for details Select Click this icon to select all the bits in the pattern All B E rane Alt Pat Click this icon repeatedly to view alternating patterns in the following iig View ways Pattern A o
284. ue Click anywhere in the sampling point display The sampling point will be moved to that location Turn the Decision Threshold knob on the front panel A small window appears displaying the threshold value This knob can be used at any time while in any window Data Delay The data input delay defines the point in time in reference to the clock signal at which the incoming data signal is measured Specifically it is the time delay from the active clock edge to the time at which the data is actually sampled This field allows the manual entry of the data input delay and displays the current value in picoseconds This delay can be set as high as 1 bit period or 10 ns 10 000 ps whichever is less You can adjust the data input delay in the following ways Click inside the text field and use the numeric keypad or front panel knob to enter a value Click anywhere in the sampling point display The sampling point will be moved to that location Turn the Data In Delay knob on the front panel A small window will appear displaying the delay value This knob can be used at any time while in any window Clock Falling Edge Activate this checkbox to move the sampling point from the rising edge to the falling edge of the clock signal This function is useful when the maximum data input delay adjustment is not large enough to move the sampling point to the desired location This occurs at very low frequencies only See Bit Rate
285. uence is inserted again followed by a block of bytes set to all Os then the normal scrambler sequence starting with 1111 111 again The length is 148 992 bits STM 16 corresponding to a time constant of about 4 msec SDH STM 4 ptrn A standard SDH STM 4 frame and payload Total pattern length of 77 760 bits All patterns are STM 16 SDH frame standard pattern scrambler ON and 1010 payload SDH_STM 16 ptrn Pattern contains no errors or alarms SDH_STM 16_HP BIP ptrn Pattern for testing frames with erred parity check sums B3 error SDH_STM 16_HP REI ptrn Pattern for testing Remote Error Indication RDI alarm Agilent Serial BERT N4901 2 User s Guide January 2006 e SDH_STM 16_MS AIS ptrn Pattern for testing Alarm Indicator Signals AIS alarm e SDH_STM 16_MS BIP ptrn Pattern for testing frames with erred parity check sums B2 error e SDH_STM 16_MS RDI ptrn Pattern for testing Remote Defect Indication RDI alarm e SDH_STM 16_OOF ptrn Pattern for testing Out of Frame OOF alarm e SDH_STM 16_RS BIP ptrn Pattern for testing frames with erred parity check sums B1 error SDH STM16 alternate All patterns are alternate Pattern A is an error and alarm free SDH STM 16 frame scrambler ON and 1010 payload Pattern B has either an error or alarm as per below SDH_STM 16_HP BIP_alt ptrn Pattern for testing frames with erred parity check sums B3 error SDH_STM 16_HP REI_alt ptrn Pattern for
286. uide January 2006 313 Copyright Agilent Technologies 2006 Printed in Germany January 2006 Nagy 91 030 ae Agilent Technologies
287. ur device Pattern Generator AUX ERROR TRIGGER IN ADD OUT CLK DATA DATA DELAY OUT OUT OUT CTRLIN Pattern Generator pata pata Data Out CTR Sock Decision Data In Amplitude Offset Delay Amplitude Offset Threshold Delay Error Detector af a The pattern generator has the following inputs and outputs Understanding the Pattern Generator s Input Ports The pattern generator provides the following input ports e 10 MHz Ref In This connector allows you to connect a 10 MHz reference clock Clock In Allows you to connect a clock signal from another device e Error Add Allows you to connect a device that generates a trigger for inserting a single bit error in the output stream Agilent Serial BERT N4901 2 User s Guide January 2006 NOTE e Delay Control In Allows you to connect a signal for generating jitter in the data outputs e Aux In The signal at the Aux In port can be used to switch between patterns A and B or to blank suppress the output signal When a user defined sequence is executed it can also be used to control the sequence execution Understanding the Pattern Generator s Output Ports The pattern generator provides the following output ports e Data Out and Data Out The data outputs serve as device stimuli and can be set up so that
288. urce has a frequency of 10 MHz connect the clock source to the pattern generator s 10 MHz Ref In port If the clock has another frequency connect it to the Clock In port 2 Click the Bit Rate Setup icon in the PG Setup menu as 3 Select the appropriate clock source Internal Clock Source This setting uses the internal clock oscillator 10 MHz Reference Clock This locks the clock generator to an external 10 MHz reference External Clock Source This connects the external clock directly to the clock generator 4 Select a clock rate For Internal Clock Source you can either select a clock rate from the Presets list or enter a clock rate in the Value and Units field For 10 MHz Ref Clock Source the clock rate connected to the 10 MHz Ref In port must be 10 MHz The selected clock rate applies to the pattern generator This is also the clock rate generated at the pattern generator s Clock Out port If the error detector receives its clock from the pattern generator it runs at the same clock rate Agilent Serial BERT N4901 2 User s Guide January 2006 79 Bit Rate Reference The Bit Rate Setup window contains the following elements Clock Source Clock source can be one of the following e Internal The clock generator uses the internal oscillator e 10 MHz Ref The clock generator uses the 10 MHz Reference that must be connected to the 10 MHz Ref Input e External The pattern generator uses the e
289. voltage level where the overlapped rising and falling edges of the logic levels intersect This adjustment varies the widths of the logic highs and lows The following figure shows examples of crossover at 50 80 and 20 50 80 20 Understanding how the Serial BERT Uses Logic Families and Terminations The clock and data inputs of your device have load impedances or terminations Clock and data terminations refer to the voltage levels at the end of these loads These voltage levels are related to the logic family of your device Your device should usually only receive signals that are compatible with its termination voltage and logic family 68 Agilent Serial BERT N4901 2 User s Guide January 2006 When you change a logic family for the data or clock outputs the termination voltage normally associated with the logic family is automatically selected This is assumed to be the termination of your device and will be used to determine the actual output voltage levels of the pattern generator CAUTION If you manually change the termination a warning message is shown indicating that the termination is not typically appropriate for the logic family If you answer No the warning message closes with no changes to the terminations If you answer Yes the data termination voltage is applied and the Logic Level is changed to CUSTOM See also Why Incorrect Terminations Could Damage Your Device on page 70 The following tabl
290. xternal clock that must be connected to the Clock Input You can choose between Automatic and Manual Mode Automatic and Manual Mode The Manual checkbox is available if you have chosen an external clock source Clock Source fexternal Z Manual In automatic mode the external clock frequency is continually measured When the external clock frequency changes the bit rate changes as well In manual mode you can enter the expected frequency or press the Measure amp Set button Measure amp Set requires that the external clock generator is connected and active Its frequency is measured once and determines the bit rate When the external clock frequency changes PG Clock Loss is signaled Value and Units This text field allows manual entry of the clock rate value Click inside this text field and use the numerical keypad or front panel knob to enter a clock rate value Complete the entry with a frequency unit by pressing the G n k m M u or X1 front panel key 80 Agilent Serial BERT N4901 2 User s Guide January 2006 Preset List The Preset list contains clock rate values that are commonly used by the pattern generator Click a preset to select it for the pattern generator clock rate Double click a preset to modify it Add Preset Button Click this button to open the Enter Bit Rate dialog box and add a new clock rate to the preset list In this dialog box you can enter a value unit and text description of the
291. xternal monitor Notebook for the built in display Use the same procedure to switch back to the built in display Configuring the Touchscreen Reference This section describes the options of the on screen keyboard Click Set to access the Options menu Move OSK Window Click this to enable the arrow keys on your actual keyboard They can then move the software keyboard You can always move the On Screen Keyboard window by clicking its title bar and dragging it Additionally you can switch between different views of the on screen keyboard The following table lists the displayed buttons for the different views Qwerty Full X X X X Qwerty Simple X X X Qwerty Minimum X X Block Full X X X X Block Simple X X X Block Minimum X X Minimum X Qwerty options mimic the layout of an actual keyboard Block options present the letter characters in alphabetical order Agilent Serial BERT N4901 2 User s Guide January 2006 295 TIP Ifyou select one of the larger on screen keyboard displays such as Qwerty Full or Block Full the keyboard may cover most of the instrument display If this occurs continue to make your keyboard entry then either close the on screen keyboard or change to a smaller display 296 Agilent Serial BERT N4901 2 User s Guide January 2006 NOTE Printing Printing Procedures This section provides information on setting up and using a printer Connecting a Printer to the Ins
292. y you can display all related values for the markers in the marker readout Pressing the Reset Markers button will set the markers back to the default positions 4 00e 0 1Trace 107 Points Ul 2 500 ns All Errors 1 00e 1 1 00e 2 1 00e 3 1 00e 4 Markers Time BER 1 00e 5 Upper Right 0 501 UI 3 240e 1 1 od A 1 booe 6 Lower Left 0 470 Ul 5 784e 6 UQE p Delta 0 970 Ul 3 240e 1 imaz ey a Or ee ee es ee Logarithmic 0 60 UI 0 36 UI 0 12 UI 012U1 0 36 UI 0 60 UI Relative 0 48 Ul 0 24 Ul 0 00 Ul 0 24 Ul 0 48 Ul O72UI Several zoom factors are available When you show the zoom graph you can also allow the zoom graph to track the mouse or your finger if you are working directly on the Serial BERT 4 00e 0 1Trace 107 Points Ul 2 500 ns All Errors 1 00e 1 1 00e 2 1 00e 3 1 00e 4 1 00e 5 1 00e 6 BER Thresholl poer EL E A a E a a a E a l Logarithmic 0 60 UI 0 36 UI 0 12 Ul 0412UI 0 36 UI 0 60 Ul Relative 0 48 UI 0 24 UI 0 00 Ul 0 24 Ul 0 48 Ul 0 720 If you want to see the points that have actually been measured choose Show Measured Points 4 00e 0 1_Trace 55 Points Ul 2 500 ns All Errors 1 00e 1 1 00e 2 1 00e 3 1 00e 4 1 00e 5 1 00e 6 BER Thresholg wor C Aka ae ay e a a a e a l Logarithmic 0 60 UI 0 36 UI 0 12 Ul 0412UI 0 36 UI 0 60 Ul Relative 0 48 UI 0 24 UI 0 00 uI 0 24 Ul 0 48 Ul 0 720 Agilent Serial BERT N4901 2 User s Guide January 2006 179 Explan

Agilent N4901 Serial BERT

Contents

Download Pdf Manuals

Related Search

Related Contents