Undersökning och utveckling av multi
Contents
1. Passband 500 1000MHZ 2 dB Bandwidth 5 of f snter Lower Stopband 0 450MHZ 50 dB Upper Stopband 1200 3000MHZ 50 dB Size 249 x 137 x 50 mm Input Connector Type SMA Female Output Connector Type SMA Female G GPIB Digital Interface R 24 VDC Tuning By Infrared Sensor Tuneable Band Pass Filter DSBT 1500 3000 2 O O GRI LP Passband 1500 3000MHz 2 7 dB Bandwidth 2 Of foenter Lower Stopband 0 1450MHz 50 dB Upper Stopband 3105 6000MHZ 50 dB Size 187 x 73 x 50 mm Input Connector Type SMA Female Output Connector Type SMA Female G GPIB Digital Interface R 24 VDC Tuning By Infrared Sensor Amplifier AMF 7D 00101800 30 10P Specifications at 23 C Frequency 0 1 to 18 GHz Gain 55 dB min Gain Flatness 2 dB max Noise Figure 3 dB max Noise Temperature 288 6 K max VSWR In 2 2 1 max VSWR Out 2 2 1 max P1dB Out 10 dBm min Voltage 15 V nom Current 425 mA nom Outline Drawing 121681 Operating Temp 40 to 75 C Ee taarg ee 4 H ar NN 0 Uc L di ss ze EK Gaass OUTUNE ULTRA WIDE BANI AMF 4X3 STAGE D AMPL c 33502 121681 C gt m e CK 121681 Rev C Attenuator 47 40 33 Fixed Coaxial Attenuators Model 47 Medium Power N or 3 5mm Connectors2. Start frequency and Stop Frequency The selected filter will be displayed in the Filter window along with if the filter is tuneable its tuneable centre frequency Finally the chosen attenuation or amplification will be displayed in the Gain labelled window The user is then prompted to confirm his selected options for the measurement If the user made a mistake or wishes to change the selection selecting no in the prompt allows the user to make new choices and then press Select Configuration again If the user is content with his selection he selects yes and the Start measurement button becomes activated allowing the user to press it The picture below Figure 19 displays an example of a user confirming a selected configuration 26 LP 0 0 5 Ghz Tune BP 0 5 1 Ghz BP 1 1 5 Ghz Tune BP 1 5 3 Ghz HP gt 3 Ghz External component Select Configuration Figure 19 Confirming a configuration The different selections and options will now be explained in more detail From the Select Band dropdown menu the user is able to choose between predetermined bands or if the user does not find the choice he is looking for he can select the Custom option from the dropdown menu allowing him to specify the start and stop frequency for the band he wants to measure After pressing the Select Configuration after s
3. 1915 1930 1995 FDD 1900G B33 1900 1920 1900 1920 TDD B34 2010 2025 2010 2025TDD A band B35 1850 1910 1850 1910 TDD B36 1930 1990 1930 1990 TDD B37 1910 1930 1910 1930 TDD B38 2570 2620 2570 2620TDD D band B39 1880 1920 1380 1920TDD F band B40 2300 2400 2300 2400 TDD B41 2496 2690 2496 2690 TDD Sprint Table 1 Bands and frequencies He also mentioned that the more interesting bands for the spurious emission measurements are B1 B3 B7 B8 B12 and B13 UL uplink corresponds to the receiver frequencies for that particular band while DL downlink is the transmitter frequencies To settle what was needed of the filter box input from Ericsson was necessary to define the requirements A requirement specification was written based on those requirements See Appendix A Various datasheets for filters were studied to determine if they fulfill the requirements for the planned filter box To make the filter box even more general tuneable filters were investigated as they would allow sweeping across frequencies Using tuneable filters would in theory allow for a more general configuration that can be used for a wider variety of measurements and not only the OOB TX measurements which is the main purpose of the filter box In addition the total amount of components needed for the filter box is likely to be reduced by using tuneable filters All components which are going to be used for the planne
4. CHALMERS Unders kning och utveckling av multi filterbox For matningar pa spurioser Examensarbete inom hd gskoleingenj rsprogrammen Dataingenjor och Elektroingenjor DAVID WAHLBERG LUCAS WIMAN Institutionen for Signaler och system CHALMERS TEKNISKA HOGSKOLA G teborg Sverige 2014 Unders kning och utveckling av multi filter box F r m tningar p spurioser Examensarbete inom H gskoleingenj rsprogrammen Dataingenj r och Elektroingenj r DAVID WAHLBERG LUCAS WIMAN DAVID WAHLBERG amp LUCAS WIMAN 2014 Department of Signals and Systems Chalmers University of Technology SE 412 96 G teborg Sweden Telephone 46 0 31 772 1000 Foreword We would like to thank Michael Johansson for the opportunity to work with this bachelor thesis We would also like to thank our supervisors G ran Hult and J rgen Ankarberg who have helped us during the work Thanks to Samir Kilim and Martin Hedin for help with programming and theoretical questions We would also like to extend our thanks to Jonas Noreus for his help with questions on how to write this thesis David Wahlberg amp Lucas Wiman G teborg 2014 Summary This report describes the theoretical design of a flexible filter box for Ericsson at Lindholmen which purpose is to be able to be used in the measurement of spurious emissions which are unwanted interfering frequencies outside radio frequency bands The filter box needs over a wide frequency range
5. OR RETURN Vdc COMMON 0 NOT REQUIRED Power Interface SOLDER TERMINAL TOP OF SWITCH CONN MS3112E XX XXP OR EQUIV TOP OF SWITCH CUT OFF POWER CIRCUIT LATCHING ONLY TTL LOGIC CUT OFF PWR CIRCUIT amp TTL LOGIC LATCHING ONLY Intermodulation LOW PASSIVE INTERMUDULATION SUPERIOR PASSIVE INTERMODULATION 9 Special Options optional leave blank if not required IP HIGH TEMP HIGH POWER CONNECTOR MS EXTENDED TEMP RANGE 54 DEG C TO 90 DEG IMSP HIGH TEMP HIGH POWER CONNECTOR AND EXTENDED TEMP RANGE Special options Intermodulation PIM Options Power Interface Common Polarity f MOMENTARY MOMENTARY WITH SUPPRESSION DIODE MOMENTARY WITH INDICATOR MOMENTARY WITH SUPPRESSION DIODE AND INDICATOR OTHER SPECIFY DC 6 0 GHz DC 26 5 GHz DC 20 0 GHz OTHER SPECIFY 24 32 Vdc 28 0 NOMINAL OTHER NOT REQUIRED D SUB MINIATURE SERIES TOP OF SWITCH OTHER OTHER SPECIFY CUSTOM ORDER NO OTHER OPTIONS REQUIRED ULTRA LOW PASSIVE INTERMODULATION STANDARD SWITCH CHARTER ENGINEERING INC INTERNET http www ceiswitches com EMAIL sales ceiswitches com PAGE 41 Cables SUCOFORM_141_ CU SUCOFORM SUCOFORM_141_CU 141 CU PE Cable design 1 2 3 1 2 3 Fig 1 SUCOFORM_141_CU Fig 2 SUCOFORM_141_CU_PE 1 Centre conductor Solid silver plated copper wire 0 95 mm 2 Dielectric Solid PTFE 2 95 mm 3 Outer conductor Tin soaked copper braid 3 58
6. power The filter box needs to be of reasonable size and cost if it is going to be of interest for future development A necessary amount of switches to accommodate the filters is required 4 Original Support for bands B1 B3 B4 B7 and B8 Base 5 Original Support for Band B12 and B13 Wanted 6 Original Support for other bands Extra K Original Digitally tuneable filters Wanted 8 Original One port for output Base 9 Original External component support filters etc Base 10 Original Support for measuring down to 115dBm Base SEM level 11 Original SNR greater than 5 dB Base 12 Original Support for up to 10W CW Base 13 Original Reasonable size Base 14 Original Reasonable cost Wanted 15 Original Support for TX OOB spurious emission Base measurements 16 Original Support for RX OOB spurious emission Wanted measurements 17 Original Support for other measurements Extra 18 Original Switches for switching between Base filters amplifiers 19 Original Support for switching to a signal path with Base high attenuation Software requirements The application will be written in Agilent VEE The application will need to be able to select what filters that are going to be used for the measurement The ability to switch between different amplifiers in the application is also required The addition of preset configurations for different band is desired to si
7. the highpass filter which is tied with the tuneable filters for the worst IL at 1dB 1 2589 1 1 1220 1 1 1220 1 1 9953 1 Bor ebben 0 8912 0 8912 0 7943 0 8912 0 7943 0 8912 0 8912 0 7943 0 8912 0 8912 1 1220 1 3 6465 0 8912 0 7943 0 8912 0 8912 316227 766 Conversion to dB 10l0g3 6465 5 62dB As seen by the equations there is a notable difference in attenuation between best and worst case even without taking the cabling into consideration However to investigate the cable influence on attenuation an average of 30cm cable between stages was assumed which including input port to the first switch and the last switch to output totals 210cm Using the formula for attenuation provided by cable manufacturer Huber Sunher See Formula 7 20 Attenuation a f bxf dB m Formula 7 Attenuation formula Where f is frequency in GHz and a and b are coefficients specified for the cable For the SUCOFORM_141_CU cables considered for the box a 0 355 b 0 04 At 0 5Ghz and 210cm 0 355 0 5 0 04 0 5 0 27dB m 2 1m 0 27dB m 0 567dB At 13GHz and 210cm 0 355 0 513 0 04 13 1 8dB m 2 1m 1 8dB m 3 78dB In the worst case scenario of noise factor for the filter box the value is 5 62dB and the worst case scenario for cables is 3 78dB Adding these values gives a total of 9 4dB attenuation Assuming the filter box is receiving a signal of 115 dBm it is then coupled
8. 10dB coupler to lower its maximum power to 40dBm is needed while also lowering the SEM to 125dBm With the carrier at 40dBm using cables and switches specified to 160dBc PIM at 43dB CW would result in around 129dBm passive intermodulation products which is too much for the 125dBm signal to be consistently measurable due to the SNR being lower than 5dB Measurements on the strongest CWs would therefore need additional attenuation to lower the PIM levels In contrast using cables and switches with 130dBc PIM would result in 99dBm PIM products which is not low enough to measure the weak SEMs even with further attenuation This leads to the 160dBc PIM components being required before the carrier has been attenuated by a filter After the signal has passed through a filter the carrier should be attenuated to 10dBm or less which means a 130dBc PIM level should be more than sufficient for the remaining components see calculations below 100W 50dBm The signal is coupled by 10dB 40dBm 10W 10W Signal passes through a cable or switch resulting in PIM 130dBc PIM cable gt 99dBm IM products 22 160dBc PIM cable gt 129dBm IM products Signal passes through a filter with a minimum of 50dB attenuation in the stopband 40dBm 50dB gt 10dBm 0 1 mW 0 1 mW Signal passes through a cable or switch resultning in PIM 130dBc PIM switch gt 149dBm IM products 160dBc PIM switch gt 179dBm IM products 3 3 9 Siz
9. Should the filter box with the proposed components be insufficient for a measurement one or two external components can be connected one instead of a filter and one instead of the amplifier or attenuator 3 3 6 Visualisation of the filter box After the decision on components was complete a schematic of the filter box was created See Figure 15 A connection without a component between the two last switches is included in case no gain is necessary for the measurements Furthermore J rgen Ankarberg noted that it is good to have the option of a straight path through the box though the type of switches used can have a maximum of six input ports so no such path exists between the first two switches It is however possible to connect a cable instead of an external component thus the option of a non modified path exist eeng gt LP Filter a Tunable BP HH Filter L_ a L FF se L__sh_y I it Luch BP Filter po Switch I Switch Switch gt p Switch gt 40d8 gt LM attenuator Tunable BP A Filter ji gt HP Filter Figur 15 Filter box configuration 19 3 3 7 Noise and attenuation calculations The noise figure of passive components is equal to their attenuation i e the attenuation is equal to the insertion loss of ea
10. WEIGHT 175 g 6 oz maximum PHYSICAL DIMENSIONS T 103 7 4 1 7 21 0 827 16 0 63 DIA OLA MAX 41 15 1 52 DIA MAX DIN A DIMA 14 0 0 55 13 2 0 52 3 5mm Female 3 5mm Male DIMA 24 1 0 95 19 0 0 75 NOTE All dimensions are given in mm inches and are maximum unless otherwise specified MODEL NUMBER DESCRIPTION We Attenuation Connector Options 1st digit is input side Wm D I 2 dat bo opda Basic Model Number 5305 Spectrum Drive Frederick MD 21703 7362 TEL 301 846 9222 800 638 2048 Fax 301 846 9116 web www aeroflex com weinschel email weinschel sales aeroflex com Revision Date 9 30 2012 Switches H6P 140138 H6P 140139 and H4P 140139 fn OC 20 20 40 40 80 20 124 124 18 0 18 0 28 5 H3 H6 SERIES SE E SP3T SP6T Ka DC 26 5 GHz 147 DIA THRU 4 MOUNTING HOLES 2 010 SQ CONTACT ACTUATOR SWITCHING SPEED ACTUATOR VOLTAGE ACTUATOR CURRENT POWER HANDLING ENVIRONMENTAL OPERATING TEMP SINE VIBRATION RANDOM VIBRATION UFE ENCLOSURE CONTACT CONNECTOR SHELL AVAILABLE OPTIONS INDICATOR CIRCUITRY SUPPRESSION DIODE TTLLOGIC FAILSAFE LATCHING MOMENTARY 20 mSEC MAX 28 0 12 0 15 0 24 0 Vdc CONTACT CEI 200 WCW 1 0 GHz STANDARD 300 WCW 1 0 GHz OPT P 40 DEG CTO 90 DEG C STANDARD 54 DEG C TO 90 DEG C OPT MS 20G Sms 20GSms 1 000 000 CYCLES ALUMINUM ELECTROLESS NICK
11. addition the outline of an automatic measurement algorithm for the filter box was developed The algorithms purpose is to ensure instruments do not break during measurements using the filter box Abstract Den h r rapporten beskriver den teoretiska designen av en flexibel filterbox f r Ericsson Lindholmen vars syfte r att anv ndas f r m tningar av spurioser som r o nskade st rande frekvenser utanf r radiofrekvensband Filterboxen m ste ver en vid frekvensr ckvidd kunna f rst rka svaga spurioser f r att g ra dem m tbara samt att d mpa b rv gen s att instrumenten som anv nds tillsammans med filterboxen inte g r s nder Rapporten beskriver de tillv gag ngs tt som anv nts f r att unders ka om komponenter f r filterboxen klarar de brus storlek och kostnadskrav som st llts Komponenterna studerades enbart i teori med hj lp av datablad och information fr n distribut rer eftersom det hade varit f r dyrt att best lla in alla komponenter och testa dem Id n med att anv nda filter med inst llbar frekvens i filter boxen f r att t cka s manga frekvensband som m jligt och samtidigt minska m ngden filter som kr vs f r boxen har unders kts Baserat p unders kningarna av brus och storlek utvecklades en filterboxkonfiguration som teoretiskt s tt klarar brus storlek och intermodulationskraven Men m ngden passiv intermodulation en typ av distortion i passiva icke linj ra element fr n alla komponenter
12. base for the filter box and search for filters complimenting them Having covered all bands except for two very low priority ones within the range of both tuneable filters only three more filters would be needed for measurements in the full range of DC 13GHz a lowpass filter for DC 0 5Ghz See Figure 11 a bandpass filter for 1 1 5Ghz See Figure 12 and a highpass filter for 3 13GHz See Figure 13 Considering where the bands lie each filter was quoted to a manufacturer The following were the specifications sent to the manufacturer with additional comments on why these specifications were requested All filters able to handle 10W CW This is the highest power components need to handle if a 10dB coupler is used Lowpass Passband insertion loss 0 500MHz 2dB Stopband attenuation 650 3000MHz 50dB 15 DC 500MHz 2dB 650MHz 3000MHz 50dB Figur 11 Requested Lowpass filter frequency response The passband covers the frequencies up to the where the first tuneable bandpass filter operates the stopband has some margin up to 699MHz which is where band B12 starts Bandpass Centre frequency 1250MHz Bandwidth 500MHz Passband insertion loss 1000 1500MHz 2dB Stopband attenuation 0 920MHz 50dB Stopband attenuation 1600 3000MHz 50 dB 1000MHz 1500MHz 500MHz 2dB 1600MHz 920MHz 50dB 50 dB Figur 12 Requested Bandpass filter frequency response Passband covers the frequencies between the tuneable
13. e EXPRESS 800 542 4457 Features 4 Designed to meet environmental requirements of MIL DTL 3933 Rugged injection molded connectors Specifications NOMINAL IMPEDANCE 50 Q FREQUENCY RANGE dcto 18 0 GHz MAXIMUM DEVIATION OVER FREQUENCY Nominal ATTN dB Deviation dB 3 6 0 75 10 20 0 75 30 40 1 00 MAXIMUM SWR Frequency GHz 10 20 30 40 dB de 8 1 20 8 12 4 1 25 12 4 18 1 35 POWER RATING mounted horizontally 50 watts average unidirectional to 25 C ambient temperature derated linearly to 5 watts 125 C Note 3 dB model can handle 100 Watts average unidirectional 1 kilowatt peak 5 usec pulse width 2 5 duty cycle Maximum power rating into output port is 10 Watts average POWER COEFFICIENT lt 0 0003 dB dB watt TEMPERATURE COEFFICIENT lt 0 0004 dB dB C TEMPERATURE RANGE 55 C to 125 C NEROFLEX WEINSCHEL dc to 18 0 GHz 50 Watts M ROHS TEST DATA Swept data plots of attenuation and SWR from 50 MHz to 18 GHz CONNECTORS Type N connectors per MIL STD 348 interface dimensions mate nondestructively with MIL C 39012 connectors 3 5mm Connectors mate nondestructively with SMA per MIL C 39012 2 92mm and other 3 5mm connectors Opti Descistion Opti Descripti 1 3 5mm Female 3 Type N Female 2 3 5mm Male 4 Type N Male CONSTRUCTION Black finned aluminum body stainless steel connectors with gold plated beryllium copper contacts
14. filter box so these terms needed to be investigated Another issue is so called passive intermodulation which is difficult to estimate without physical components See the theoretic reference for more information regarding Passive Intermodulation 3 3 Filter box design In order to make a filter box as general as possible it will need to cover all frequency bands of interest to Ericsson The bands are located at very different frequencies from as low as 700 MHz up to 2700 MHz therefore filters with many different frequency ranges will be required Furthermore since the spurious emission measurements are conducted out of band it will be necessary to measure from 0 Hz up to as high as 12 75 GHz The table below Table 1 sent in an email from J rgen Ankarberg represents various frequency bands tested at Ericsson 12 UL MHz DL MHz Band BO 890 915 935 960 FDD 900P B1 1920 1980 2110 2170 FDD 2100 B2 1850 1910 1930 1990 FDD 1900 B3 1710 1785 1805 1880 FDD 1800 B4 1710 1755 2110 2155 FDD AWS B5 824 349 869 894 FDD 850 B6 830 840 875 885 FDD B7 2500 2570 2620 2690 FDD 2600 B8 880 915 925 960 FDD 900E B9 1749 9 1784 9 18449 18799FDD 17 1800 B10 1710 1770 2110 2170 FDD B11 1427 9 1452 9 1475 9 1500 9 FDD 1500 B12 698 716 728 746 FDD 700 B13 777 787 746 756 FDD 700 B14 788 798 758 768 FDD 700 B17 704 716 734 746 FDD 700 B20 832 862 791 821 FDD 800DD B25 1850
15. kunde inte best mmas med enbart teori P g a detta s r intermodulationsfr gan fortfarande os ker En applikation skriven med programspr ket Agilent VEE utvecklades f r att l gga till m jligheten att styra den t nkta filterboxen med en dator Eftersom det inte fanns n gra fysiska komponenter att testa med applikationen s fick stubbar anv ndas stubbar r ers ttningskod f r att simulera komponenternas uppf rande Detta medf r att applikationen beh ver fortsatt arbete s den kan integreras med framtida fysiska komponenter Ut ver programmet designades grunderna f r en automatisk styrningsalgoritm f r filterboxen Dess syfte r att f rs kra att inga instrument eller komponenter g r s nder under m tningar med filterboxen Abbreviations BW Bandwidth CW Constant Wave dBc decibel in relation to carrier DUT Device under test LNA Low noise amplifier IL Insertion loss IM Intermodulation OOB Out of Band PIM Passive Intermodulation RS Requirement Specification RX Receiver SEM Spurious Emissions SNR Signal to Noise Ratio TX Transmitter VR Verification Report VS Verification Specification Table of Contents a IHR SI UCT ION assess sesies sisnesessavncganis anene ee nees oases kasus bensen e EAE E E E ENN 1 1 1 Background EE 1 e E 2 E DG NIMIMATIONS i1icses sect tce venetceedeesdcronatersen EA R EE EErEE TN RE EO E EE 2 124 COP O E eE nns nrns o
16. loss return loss etc For this report the most important S parameters will be Su and S2 which are used for insertion loss and gain 2 1 3 Noise In electronic circuits there is thermal noise which is generated from components in the circuit when the components have a higher temperature than 0 Kelvin 3 This noise will negatively affect a signal e g the signal from a radio transmitter Signal to Noise Ratio abbreviated SNR is the ratio between signal and noise power and is usually expressed in decibels A high SNR is important to differentiate the signal from the noise 4 SNR is affected negatively by the Noise Factor of components Noise Factor is the ratio of input SNR to output SNR expressed in linear Instead of Noise Factor Noise figure is often specified for components Noise Figure 10log Noise Factor Formula 2 Noise Figure to Noise Factor conversion Pignal SNRap 10 logio Psignal aB Proise aB P noise Formula 3 SNR in dB 2 1 4 Decibel milliwatts dBm Decibel milliwatt dBm is a power measurement relative to 1 milliwatt and is expressed as x 10log Formula 4 Watt to dBm And x P 1mW x 1010 Formula 5 dBm to Watt Where P is the power in Watt and x is the power in dBm It is used to measure the absolute power because of its capability to express both large and small values in a small form dBm is used in radio microwave and fiber optic work 5 OdBm equals ImW 30dB
17. mm Electrical cable data Impedance 50 Ohm Operating frequency 33 GHz Capacitance 92 pF m Velocity of propagation 71 Time delay 4 7 ns m Nom attenuation coefficienta 0 3631 coefficientb 0 0377 Max attenuation coefficienta 0 3995 coefficientb 0 0415 Max operating voltage 1 9 kVrms Min screening effectiveness up to 18 GHz 100 dB Attenuation calculation amp 5 aVf GHz b f GHz dB m General cable data SUCOFORM_141_CU_PE 22511639 SUCOFORM_141_CU 22511635 ltem number Additional jacket Colour Diameter Weight 4 0 kg 100m 65 165 C IEC 60332 1 UL 1581 1080 VW 1 47 kg 100m 40 85 C Temperature range Flammability passed 8 mm 40 mm Y12 Min bending radius static Min bending radius repeated Suiteble connectors cable group Drewing Suitable connectors see cable group Y12 please refer to page 59 ff 50 HUBER SUHNER microwave cables and assemblies SUCOFORM_141_CU 141 CU PE Cable attenuation Nominal values 25 C ambient temperature Attenuation dB m Frequency GHz Power Handling Maximum values 40 C ambient temperature and sea level C W Power Watts 600 300 0 3 6 9 12 15 18 21 24 27 30 33 Frequency GHz HUBER SUHNER microwave cables and assemblies 51
18. not be commonly known for the reader and provides a better explanation 2 1 Theoretical terms This section describes various terms and theory used in the project 2 1 1 Spurious emissions Spurious emissions are unwanted radio frequencies outside the channel The spurious emissions can consist of harmonic emissions parasitic emissions intermodulation products and frequency conversion products Spurious emissions can appear in the bandwidth and outside of the bandwidth of a band 1 2 1 2 S parameters S parameters which are also known as scattering parameters are used to quantify how radio frequency energy propagates through a multi port network S parameters provide a means to accurately describe the properties of more complex networks as black boxes For example when a radio frequency enters on one port some part of the signal is reflected back out of the same port some of it scatters and propagates through other ports some of it might become amplified and some of it dissipates as heat or electromagnetic radiation The S parameters of a two port network are generally known as Su voltage reflection coefficient of the input port Si2 reverse voltage gain Sa forward voltage gain S22 voltage reflection coefficient of the output port These parameters are often placed within a matrix see Formula 1 2 E e S21 S22 Formula 1 S parameter matrix These parameters are used to calculate gain insertion
19. of the signal is read then it either lowers the frequency until it is no longer considered strong or switches on the amplifier and measures at the frequency if the signal is considered weak Finally the filter tunes to the band stop frequency At this point the other main branch of the algorithm will start running When the other branch of the algorithm starts running the frequency of the filter is above the band and a test checks whether or not the tuned frequency is close to the band from the upper side If it is attenuation is switched on and another test keeps checking if the frequency is in band and increases the frequency until it is no longer within the band Following this the filter box switches to an empty signal path and checks if the signal is strong or not If it is the frequency is increased until the signal is no longer considered strong otherwise the algorithm jumps back to the check which determines if the frequency is close to the band or not If the filter is tuned to a frequency far above the band the filter box switches on the amplifier and measures Afterwards a test checks if the filter has tuned to its max frequency If it has not the frequency is increased and measured until the test validates the frequency as max Otherwise the measurement ends and the user make a new selection for the application 4 RESULTS AND CONCLUSIONS Presented in this chapter are the results of the research and work outlined in the process c
20. the sequence output pin has sent a container to the objects sequence input pin When an Agilent VEE program is running if there is more than one thread in the program they will all execute at the same time Creating a user interface in Agilent VEE can be done by switching between panel and detail view and specifying the properties of each object if they are going to be used for the panel view The panel view represents the user interface and what the operator will be able to see once the application is running See figure 9 and 10 The user can control the program with various graphical objects representing knobs sliders and buttons Each object in the panel view has a corresponding object in the detail view 13 10 LOSE H lsa TR SSES File Edit View Debug Flow Device User Inle fave Tvee Jabeury Juawinigsur Ei Jasmcig Palga guepumd F Ready ExecMode VER PROF MOD WEB tes cp 2014 05 09 Figure 9 VEE Application detail view User I 1 oe x File Edit View Debug Flow Device System I O Data Display Tools Database Excel Window Help Led sl Thus GAIA Jabeuew Juan eut ES asmo Palqo 8 uoun af Ready ExecMode VEER PROF MOD WEB Cer S il EI G B lamm L Figure 10 VEE Application panel view 11 3 PROCESS This chapter describes the various methods and procedures use
21. 1 Ghz BP 1 1 5 Ghz Tune BP 1 5 3 Ghz HP gt 3 Ghz External component Ep E Sei E Figure 21 Selecting filter centre frequency The Start measurement button as of currently only shows a message box informing the user that the measurement will now start 3 4 4 Outline for measurement algorithm In addition of coding a basic user interface for the planned filter box an outline for an automatic switching during measurement algorithm was created The purpose of this algorithm is to plan future functionality for the application in VEE by allowing the program to control the tuneable filters and the switching of signal paths in the filter box If the application has an automatic measurement function it would allow the user to save time by allowing the program to do the majority of the switching of components As of currently the application would require the user to make several measurements while studying the spectral analyser and power meter to investigate whether attenuation or amplification is required This algorithm is planned to be used only when the tuneable filters are used for a measurement The picture below Figure 22 depicts a flowchart of how the measurement algorithm could work 29 freq d than band lt Freq Increased lowered freq No Above band No S No Figure 22 Outline for measurement algorithm In gen
22. EL PLATED PER MIL C 26074 CLASS 4 EXCEPT LOW OR ULTRALOW IM MODELS ALUMINUM BLUE BERYLLIUM COPPER GOLD PLATED PER MIL G 45204 CRES PASSIVATED PER QQ P 35 OR BRASS ELECTROLESS NICKEL PLATED PER MIL C 26074 CUT OFF POWER CIRCUIT D SUB MINIATURE SERIES MS CONNECTOR OR EQUIV MIL SPEC LOW ULTRALOW AND SUPERIOR INTERMODULATION IM REFERENCE FOR DIMENSIONS ONLY ALSO AVAILABLE WITH A D SUB OR MS CONECTOR 2 80 DIA 1 062 DIA B C SOLDER TERMINALS SWITCH WITH TTL LOGIC OR PWR CONNECTOR THIS DIMENSION TO BE 2 60 MAXIMUM CONNECTOR SMA FEMALE MIL C 39012 7 PLACES G CHARTER ENGINEERING INC PAGE 40 INTERNET http www ceiswitches com EMAIL sales ceiswitches com PART NUMBER REFERENCE H3 H6 SERIES PART NUMBER amp ORDERING INFORMATION DO O O O O O O Q c Series H3 H6 Actuator Frequency Voltage 1 Switch Series H3 SP3T SMA H4 SP4T SMA H5 SP5T SMA H6 SP6T SMA 2 Actuator FAILSAFE FAILSAFE WITH SUPPRESSION DIODE FAILSAFE WITH INDICATOR FAILSAFE WITH SUPPRESSION DIODE AND INDICATOR LATCHING LATCHING WITH SUPPRESSION DIODE LATCHING WITH INDICATOR LATCHING WITH SUPPRESSION DIODE AND INDICATOR vzzr roan Frequency Bandwidth DC 13 0 GHz DC 2 0 GHz DC 12 4 GHz DC 4 0 GHz DC 18 0 GHz 10 14 Vdc 12 0 NOMINAL 13 18 Vdc 15 0 NOMINAL 20 30 Vdc 24 0 NOMINAL 5 Common Polarity 1 POSITIVE COMMON 2 NEGATIVE
23. F E pdf 22 May 2014 2 S parameters microwaves101 com 2014 Online Available http www microwaves101 com encyclopedia sparameters cfm 22 May 2014 3 B Molin Analog Elektronik Edition 1 6 Malm Studentlitteratur AB 2001 4 D M Pozar Microwave Engineering 4th ed Hoboken NJ Wiley 2012 5 Dong J Network Dictionary Saratoga Javvin Technologies 2007 E book Available books24x7 6 Poole I Passive Intermodulation PIM Distortion radio electronics com 2014 Online Available http www radio electronics com info rf technology design passive intermodulation pim basics tutorial php 22 May 2014 7 PIM Power Levels A Brief Tutorial Summitek Instruments summitekInstruments com 2014 Online Available http en youscribe com catalogue tous professional resources pim power levels a brief tutorial 532632 11 June 2014 8 Tuneable filters klmicrowave com 2007 Online Available http www klmicrowave com tuneable php 22 May 2014 9 Poole I RF directional coupler basics tutorial radio electronics com 2014 Online Available http www radio electronics com info rf technology design coupler combiner splitter rf directional coupler design basics tutorial php 22 May 2014 10 Huang B TTL Logic levels learn sparkfun com 2014 Online Available https learn sparkfun com tutorials logic levels ttl logic levels 2 June 2014 11 RF Switc
24. aces which allow the instrument to be controlled remotely from the PC 16 The idea of using a separate bandreject filter for each band was discarded due to the following reasons Size cost and inflexibility Regarding size it is of interest to keep the box as compact as possible Using one filter solely for the purpose of covering a single band significantly increases the number of filters needed compared to using a tuneable filter to sweep over frequencies thus covering several bands When it comes to cost it is assumed more filters results in a higher cost even though tuneable filters are individually more expensive Finally the use of non tuneable filters would result in a configuration where the only variable part is the externally connected components which decreases flexibility of the filter box By looking at different data sheets for an assortment of tuneable filters the tuneable notch filters were also rejected as the sweeping range was too small therefore many different filters would still 14 be needed Furthermore while the centre frequency is tuneable the bandwidth is not so it could not match all bands within the sweeping ranges for some bands it was too narrow whereas for others it was too wide The bandwidth of the bands varies between 1OMHz and 75MHz while the 34B BW for the tuneable notch filters in the relevant frequency range varies between 6MHz and 27MHz i e to attenuate the carrier more than 4 cascaded filt
25. by 10dB resulting in 125dBm From the components and cables the signal is further attenuated to 134 4dBm At this point the signal needs to be amplified in order to be distinguishable The 55dB amplifier increases the signal to 79 4dBm which is above the noise floor at 90 dBm With these values the SNR can be established SNRap 79 4 90 10 6 dB As can be seen from this calculation the signal is 10 6 dB stronger than the noise at the worst case scenario with a signal at 115 dBm and the total attenuation of 9 4 dB As noted above these values are just to acquire an idea of what attenuation to expect from the cabling and will most likely differ slightly when compared to reality Below is a graph displaying how the spurious emissions signal power changes as it passes through the filter box When the strictest spurious emissions signal strength is unmodified their level is at 115 dBm See Figure 16 21 SEM Signal strength behaviour SEM Noise floor Figur 16 SEM Signal strength behaviour 3 3 8 Intermodulation Components used in these calculations were specified at 43dBm carrier while calculations are made at 40dBm therefore the PIM levels will be slightly better compared to the specified level The weakest spurious emission that is to be measured has a power of 115dBm and the strongest carrier is 50dBm CW 50dBm is too much for most components in the filter box so coupling a carrier with a
26. ch component 18 Using this information together with the already provided insertion loss and gain of the amplifier the total attenuation and SNR of the selected components can be calculated to examine if the components match the requirements The cascaded attenuation for the filter box is calculated with Friis formula for noise factor See Formula 6 Pee p pop total STT Gi GG E Formula 6 Friis formula for noise factor Where F is the noise factor and G is the gain of a component The attenuation of the box will vary depending on the frequency a higher frequency generally equals a higher attenuation Cables can have a large impact on attenuation especially at high frequencies Since cable lengths between components are not known no exact calculations can be made Presented here are the best case and worst case scenarios for unwanted attenuation without regards to cabling Best case scenario Frequency at 1 1 5GHz means the switches operate at their best insertion loss IL 0 2 dB while also allowing the use of the filter with the lowest IL 0 3dB the non tuneable bandpass filter 1 0715 1 1 0471 1 1 0471 1 1 9953 1 Po 10414 a E 0 955 0 955 0 9333 0 955 0 9333 0 955 0 955 0 9333 0 955 0 955 1 0471 1 2 4545 0 955 0 9333 0 955 0 955 316227 766 Converting Noise Factor to Noise Figure Noise Figure 10log2 4545 3 9dB Worst case scenario Frequency at 13GHz results in the worst IL 0 5dB for the switches and using
27. d filter box will contribute with an amount of noise and attenuation which will need to be calculated for the entire system The cost for each component will also need to be taken into account As the final system configuration will need to be reasonably sized approximations on size had to be made For the different components contact was established with manufacturers and distributors in order to inquire about the prices and further specification of the components 3 3 1 Filter box specification The requirements for the filter box were determined by discussing with the supervisor at Ericsson and reading the various documents related to measurements To be able to measure signals at 13 115dBm with a SNR of gt 5dB the signal first needs to be amplified to 85 dBm or higher as the noise floor of the spectrum analyser used at the product design section is around 90dBm at a resolution bandwidth of 100kHz Additionally the PIM of components will have to be taken into account so that the SEM from the DUT can be differentiated from the intermodulation IM products caused by the filter box The received constant wave can have an average power of up to 100W which is more than most components can handle To prevent breaking any components a coupler will have to be used to lower the signal power to 10W or less before it enters the filter box All components should have an impedance of 50Q to minimize reflections caused by unmatched impedances Frequ
28. d in the project 3 1 Collection of basic data First of all research had to be made to develop a better understanding of the various measurements which the filter box was supposed to be included in To accomplish this several reports were studied The documents were provided by supervisor Jorgen Ankarberg who is working with software and hardware integration at the product design section of Ericsson These documents consist of VS VR and RS reports which describe the requirements and the procedure for the tests For each frequency band it is of great importance that the transmission does not interfere with other frequencies outside the specified bandwidth The test case has different requirements depending on the frequency band 14 Ericsson measures spurious emissions for frequencies from DC up to around 13 GHz The requirement for the amount of spurious emissions can vary depending on the band and what frequency interval that is being measured Some intervals might have a stricter requirement while others allow a more relaxed requirement The strictest requirement is around 115 dBm which is a very small amount of allowed spurious emissions 3 2 Fundamental theory To be able to design the filter box the necessary background theory was necessary to learn According to supervisor at Chalmers Goran Hult it is important to know how S parameters work J rgen Ankarberg at Ericsson noted the importance of noise and attenuation in regards to the
29. e calculations After deciding on the components to use in the box based on technical attributes such as IL and gain size was taken into account Height width and length of all components were noted and they were regarded as blocks for the purpose of determining the total space required to fit them inside a box At first they were simply considered to be arranged in accordance to the visualisation in Figure 15 except with no space in between components for air or cables Since all components are placed in one layer the height is equal to the height of the tallest component which is the switch See appendix B for component specifications Height 51 054mm For width every stage was calculated separately and the highest value was used Width switch 51 054mm Width filters 73 137 12 7 19 09 78 76 320 55mm Width Amplifiers and attenuators 31 496 41 15 72 646mm As seen from the calculations above the filter widths give the total width of the layer The length was calculated by simply adding the lengths of the longest component in every stage resulting in Length 69 342 249 69 342 69 342 141 7 69 342 668 068mm As shown by the calculations placing the components in this manner results in a box with the measurements of 51 054x320 55x668 068mm yielding a total volume of 10 93 litres An existing filter box in the lab was measured to 220x450x650mm or 64 35 litres Since the existing box is roughly 6 times as
30. e data can be a scalar number an array of 500 strings or simply nil which is an empty container As a container of data arrives to the input pin of an object it is called pinging and when the container comes out at the output pin of the object the output pin is activated There can also be pins at the top and bottom of an object These pins are called the sequence input and sequence output pins The sequence input pin at the top of an object if connected holds the execution of an object until the pin receives an empty container i e the sequence input pin gets pinged The sequence output pin at the bottom of an object when connected becomes activated once the object and all data propagation have finished executing See figure 8 mee Sequence input pin HA Input pins Output pins Sequence output pin Figure 8 Object pins The flow of data in Agilent VEE applications moves from left to right and then from top to bottom from each object In detail it works the following way First the data input pins and sequence in pins if connected are pinged Secondly the object executes Thirdly the output pins are activated Fourthly the data propagates from left to right Finally the sequence out pin is activated to show that the object has finished executing and propagates an empty container from top to bottom This output pin does not always need to be connected to something but can be used to make sure an object starts to execute only after
31. eing excluded for frequencies over 3GHz without the use of an external filter 17 3 3 3 Selecting amplifiers According to supervisor J rgen Ankarberg it was likely for the filter box configuration to use an amplifier in the signal path to make the signal readable Typically an amplifier is placed first in a cascaded system to increase SNR compared to amplifying at a later stage However for the filter box if the amplifier is placed so it amplifies the signal before the carrier is attenuated the other components would be destroyed The spectrum analyser at Ericsson has a noise floor around 90dBm to be able to measure SEMS as low as 115dBm they have to be amplified first Considering the signal is coupled by a 10dB coupler before entering the filter box the SEMS are down to 125dBm Insertion loss and cables attenuate the signal further resulting in a power around 135dBm depending on path and frequency For a SNR over 5dB an amplification of 50dB or more is required To leave some room for error an amplifier with 55dB gain was selected The insertion loss of the amplifier is relevant as it affects the attenuation however the amplifiers found with sufficient gain to fit the requirement had quite high insertion loss compared to amplifiers with lower gain so the insertion loss had to be accepted regardless 3 3 4 Selecting switches To choose the signal path through the filter box high frequency switches are needed By reading excel she
32. electing the Custom option the user gets a prompt where he first enters the starting frequency of the band followed by a prompt for the stop frequency The picture below Figure 20 shows an example of the user entering a start frequency for the custom band selection 21 untitled lel a ramm n Select attenuation amplification Gain 55 dB D ee LP 0 0 5 Ghz Real64 Input si Tune BP 0 5 1 Ghz Enter band starting frequency BP 1 1 5 Ghz Se Tune BP 1 5 3 Ghz HP gt 3 Ghz Ok Cancel eee Rust eer ST Select Configuration Figure 20 Entering a custom band The Select attenuation amplification menu allows the user to choose between using an amplifier of 55 dB an attenuator of 40 dB a straight signal line or simply an external component such as a different amplifier or attenuator For the selection of filters a vertical slider is used to let the user choose what filter that is going to be used for the measurement Two of the selectable filters are tuneable when either is chosen and confirmed with the Select Configuration button the program will prompt the user to enter a value which sets the centre frequency for the tuneable filter The picture below Figure 21 shows a user entering a value for the selected tuneable filter 28 CJjuntited St LP 0 0 5 Ghz Tune BP 0 5
33. encies measured vary from 0 Hz up to 12 75 GHz with the relevant bands in the 700 MHz to 2700 MHz range Using the frequencies for the previously mentioned range a set of three ranges were written down Table 2 in order to get a better perspective of the specifications needed for tuneable filters The start number was obtained using the band with the lowest starting frequency within that range while the stop number represents the highest frequency The bandwidth BW is the minimum and maximum bandwidths for the given range A list of requirements for the filter box was specified using the assumptions above See Appendix A for the filter box requirements Start Stop BW Mhz 699 960 10 35 1710 2170 45 75 2500 2690 70 Table 2 Possible filter ranges 3 3 2 Selecting filters Using the table of possible filter ranges Table 2 a number of different filters were found at K amp L Microwave 15 Among these tuneable band reject and bandpass filters were included The tuneable filters are able to sweep over frequencies while the non tuneable ones cover specific bandwidths Many of the filter datasheets stated that they were able to be digitally controlled which according to Martin Hedin and Samir Kilim at Ericsson would work with Agilent VEE but instrument drivers would need to be written for the tuneable filters in order to control them as the filters are not included in standard libraries of Agilent VEE Instrument drivers are interf
34. eral this algorithm is meant to ensure that the measurement for spurious emissions with the filter box proceeds without breaking any of the equipment used for the measurement This includes switching on the attenuator or amplifier where it is needed during the measurement Furthermore the algorithm makes sure measurement is conducted for all frequencies below and above the selected band The algorithm consists of two main branches one where the tuned filter frequency is below the start frequency of the band Shown to the left in the picture and one where the tuned filter frequency is above the stop frequency of the band Shown to the right in the picture When the tuned filter frequency is below the centre frequency of the band the algorithm works in the following way First a test checks if the tuned filter frequency is close to the start of the band If it is not it is determined safe to use the amplifier then measure and lastly the tuneable filter increases its frequency and repeats the test until the frequency is considered close to the band Once considered close to the band the attenuator is switched to circumventing the risk of breaking the spectrum analyser should the tuned frequency be in band Following this the algorithm reads 30 the current frequency to determine if it is in band or not If it is the frequency is repeatedly lowered until it is no longer in band The algorithm then switches to an empty signal path and the strength
35. ers would be needed for a particular band With both types of notch filters out of the picture tuneable bandpass was investigated Sweeping the bandpass filter outside of the band with the band in the filters stopband would allow for more flexibility as it does not rely on the filter having a specific bandwidth With this idea and the much wider frequency ranges of the bandpass filters only a few filters would be needed to cover most if not all bands in the requirement specification In addition to covering as many bands as possible the filters also needed to fulfil the requirements for stopband attenuation attenuating the carrier so that the SEM can be measured by the spectrum analyser as well as being able to measure as close to the carrier as possible With a stopband attenuation of 50dB enough to bring a 10W carrier down to 0 1mW the tuneable filters which were looked into should handle the first requirement if the filter is considered ideal However in reality no filter is ideal Additionally a drawback with having the flexibility of tuning is that the shape factor of a tuneable filter is not as good as a non tuneable filter The reason for this is that tuneable filters as of currently have a lower order than normal non tuneable filters This affects the stopband negatively and measuring close to the carrier might become a limitation Regardless it was decided that using two tuneable filters for 0 5 1GHz and 1 5 3GHz respectively as a
36. ets with existing filter box configurations provided by Ericsson a few options and manufacturers were found Four switches two with six input ports and two with four input ports are required for the envisioned setup These ports connectors should ideally match the connectors of the components ensuring that no loss of signal power due to mismatched connector types occurs Since the signal path through the box regardless of frequency and power passes all switches they need to handle the full frequency range DC 13GHz as well as 1OW CW An additional property to consider is that PIM levels should be as low as possible When investigating the control of switches Samir Kilim at Ericsson stated that the switches were able to be integrated with Agilent VEE to allow digital control Including TTL logic and cutoff power circuits would improve usability and lower power consumption He also noted that adding an indicator to the switch would help the testers verify that the correct switch is activated When contacting a seller on these matters each of the added attributes would increase the price of the switches but even with all options selected it would still be within reason 3 3 5 Additional components The use of couplers will be required for some measurements but they will be connected to the input port of the filter box This makes it possible to use components with lower power handling than 100W as the coupler attenuates the power of the signal bef
37. filters range The lower stopband includes Band 8 uplink however the filter is not steep enough to include the downlink of the band which stops at 960MHz Upper stopband at 1600MHz leaves ample room for band 3 that start at 1710MHz Note that band 11 and 21 is inside the passband of this filter These are the two bands previously mentioned as low priority and are not critical for the main application of the filter box 16 Highpass Passband insertion loss 3000 13000MHz 2dB Stopband attenuatino 0 2700 MHz 50dB 3000MHz 13000MHz 2dB l 2700MHz 50dB Figur 13 Requested Highpass filter frequency response The passband covers the remaining frequencies up to 13GHz and the stopband works for all relevant bands the closest being the downlink of band 7 at 2690MHz The specifications sent covers all required bands for TX but misses out on the RX of band 8 it also includes all other bands except for 11 and 21 If measurements are to be made on those bands an external filter will have to be connected The specifications returned from the manufacturer matched the requested specifications with the exception of the highpass filter which had the following frequency response See Figure 14 3000MHz 13000MHz 2dB 2500MHz 50dB Figur 14 Actual Highpass filter frequency response Having the stopband at 2700MHz was not possible and was changed to 2500MHz resulting in support for B7 TX 2500 2570MHz and RX 2620 2690MHz b
38. h definitions amp terminology ceiswitches com 2014 Online Available http www ceiswitches com DEFINITIONS html 4 June 2014 12 History of GPIB ni com 2012 Online Available http www ni com white paper 3419 en 22 May 2014 13 Agilent VEE 9 32 agilent com 2013 Online Available http cp literature agilent com litweb pdf 5989 9833EN pdf 22 May 2014 14 Verification specification for RUS and RRUSO1 B2 B3 and B9 Transmitter Performance 15 K amp L Microwave klmicrowave com 2007 Online Available http www klmicrowave com 6 June 2014 16 Instrument Driver Overview agilent com 2008 Online Available http www home agilent com agilent editorial jspx cc SE amp lc eng amp ckey 1461187 amp id 1461187 5 June 2014 17 RF cables hubersuhner com 2014 Online Available 33 http www hubersuhner com en Products Radio Frequency Cables 5 June 2014 18 Stiles J Noise Figure of Passive Devices itte ku edu 2005 Online Available http www ittc ku edu jstiles 622 handouts Noise 20Fi gure 200f 20Passive 20Devices pdf 22 May 2014 19 Stubs and Drivers www jmu edu 2007 Online Available https users cs jmu edu bernstdh web common help stubs and drivers php 2 June 2014 Personal references 20 J rgen Ankarberg Ericsson 21 Samir Kilim Ericsson 22 Martin Hedin Ericsson 34 APPENDICES Appendix A Filter box requirement specifica
39. hapter The conclusions made are also included as is the difficulties that were encountered during the project Possible future plans for development are also present in this chapter 4 1 Result The theoretical outline of a more general filter box with tuneable filters was designed The filter box based on the choice of filters is able to support spurious emission measurements for bands B1 B3 B4 B12 and B13 However for B8 only RX is supported due to the bandpass filter having an inadequate lower stopband Furthermore because the highpass filter could not meet the requirements for stopband frequency support for B7 will require an external filter From the noise calculations the box is able to handle the requirement of a 5dB signal to noise ratio even in a worst case scenario The calculations of the size proved that the box would be of acceptable size for placement in a 19 rack When it comes to the cost of all so far established components i e cables switches filters attenuator and amplifier the price seems reasonable compared to the guidelines received from Ericsson A basic user interface in Agilent VEE has been created which simulates the behavior of switching between components and bands to measure In addition the outline of an algorithm for automatic control during measurement using tuneable filters has been created 31 4 2 Discussion This section covers the discussion of the result and interesting topics It is po
40. icrowave Amplifier AMF 7D 00101800 30 10P Miteq Attenuator 47 40 33 Aeroflex Weinschel Switch 6port H6P 140138 Charter Engineering Switch 6port H6P 140139 Charter Engineering Switch 4port H4P 140139 Charter Engineering Switch 4port H4P 140139 Charter Engineering Cables SUCUFORM_141_CU Huber Suhner Lowpass filter 7L120 510 T1500 0 O 3 0 dB Cut off Frequency 510 MHz Passband Insertion Loss 1 500 MHz 0 9 dB Stopband Attenuation 650 3000 MHz 53 dB Return Loss 14 0 dB 1 5 1 VSWR Filter Size 116 84 x 12 70 x 12 70 mm Input Connector Type SMA Female Output Connector Type SMA Female Highpass filter 11SH10 3000 T13000 0 O 3 0 dB Cut off Frequency 3000 MHz Passband Insertion Loss 3000 13000 MHz 2 0 dB Stopband Attenuation 2500 MHz 50 dB Return Loss 9 54 dB 2 0 1 VSWR Filter Size 44 45 x 25 40 x 12 70 mm Input Connector Type SMA Female Output Connector Type SMA Female Bandpass filter 13ED50 1250 T500 0 O Centre Frequency 1250 MHz 3 0 dB Bandwidth 500 MHz Insertion Loss 0 3 dB Stopband Atten DC 920 MHz 50 dB Stopband Atten 1600 3000 MHz 50 dB Return Loss 11 7 dB 1 7 1 VSWR Filter Size 138 27 x 78 76 x 19 02 mm Input Connector Type SMA Female Output Connector Type SMA Female Tuneable Band Pass Filter DSBT 500 1000 5 O O GRI LP
41. imulate the start of a measurement 19 The features of this program are the ability to choose which filter and amplifier to be used for the signal path for the measurement and the ability to tune the tuneable filters All selectable components in the user interface matches the ones discussed in previous chapters 25 The picture below Figure 18 shows the application running and what options that is available for the user LJ l Bag E se D gea LP 0 0 5Ghz Tune BP 0 5 1 Ghz BP 1 1 5 Ghz Tune BP 1 5 3 Ghz HP gt 3 Ghz External component Figure 18 Filter box control interface In the program the user is able to select what band filter and attenuation he wants to use for the measurement from three dropdown menus Compared to the flowchart in the previous section these options do not need to be chosen in a specific order A picture with components representing the inside of the filter box is used to inform the user where his selections take place in the filter box marked with red arrows To start a measurement the user must have selected a band a filter and attenuation if any followed by pressing the Select Configuration button to confirm his choices When pressed all options selected will be displayed in their respective labelled windows The selected band along with its start and stop frequency will be displayed in Band
42. ion on some switches that allows it to cut off the DC current on the actuator after switching thus saving energy 11 Figure 4 Revolver switch 2 2 4 GPIB GPIB also known as General Purpose Interface Bus is a bus developed by Hewlett Packard in the 1960s It is used to connect and control programmable instruments and is widely used as a standard worldwide The control of the instrument can be implemented in different ways either with an external controller such as a GPIB USB interface See Figure 5 or a GPIB Plug in Controller card 12 Figure 5 GPIB USB interface 2 3 Agilent VEE Agilent VEE Visual Engineering Environment is a visual test programming environment used for tests measurement and instrument control applications See Figure 6 The programming language is graphical and flow driven A developer can connect graphical high level objects to each other and the way these objects are connected controls the flow of the application File Edit View Debug Flow Device System VO Data Display Tools Database Excel Window Help DSU EE OO E usa ee Aba Onn ORs BrmMOGax esatew egam El a oh el Gh be by Ge Mp Program Explorer vax TA Untitled e TED Properties vax Real64 Slider Slider M 1 X le ShowTileBar True Title Real64 Slider Title BackColor E Slider Title Font Object Title Text Title ForeColor J Object Title T E Behavior AutoExecute False Breakpoint False EnableEd
43. iting True Wait For Input False DI Design Declared Name Declared Scope Localto Context 2 Title Title string of the object Amplitude 0 727 Amplitude Time Span 20m Num Points 256 Figure 6 Example VEE application Objects are the components used to create an application in VEE Every single object represents programming lines and instructions used in more common programming languages such as C or Java An object can once it executes do an assortment of functions such as displaying a graph or executing MATLAB script In addition a VEE program may contain user defined objects which are objects the user has created These user defined objects can also contain several existing objects In the Agilent VEE programming environment an object can be shown in two different modes Icon view and Open view Icon view is where the object is minimized to cover as little space as possible on the screen while the Open view is used to further inspect the object to get a better idea of what it contains and its instructions See figure 7 7 EES sp Function Cosine e Frequency 200 Amplitude 1 DcOffset Func Phase BEE In Time Span pom Num Points ee Icon view Open view Figure 7 Object views When objects are connected they create a thread Data flows between objects in packages called containers The container can hold data in various forms For example th
44. librate and they are extremely sensitive to external influence and the position of the cables Switching from one radio unit to another could lead to the failing of the measurement setup resulting in several hours of work being spent setting up the measurement again A way to possibly avoid this problem would be the development of a multi filter box with several sets of filters controlled by a computer See Figure 1 The product design section at Radio Design Center Lindholmen is interested in a general multi filter box used for specific out of band measurements As of currently these kinds of measurements are done at system verification However because of time constraints at product design there are occasions when there is no time to wait for system verification to do the measurements Therefore it would be of interest for the product design section to have their own filter box for measurements External extoma SE A a LNA Ctrl SW1 SW3 Connected to PC Figure 1 Early sketch of filter box 1 2 Purpose The purpose of this report is to determine if it is possible to create a flexible filter box for verification measurements on radio units for radio base stations The filter box should be able to attenuate a carrier signal to a level where a spectrum analyser can measure very weak signals out of band and also amplify spurious emissions the very weak signals in case that they are under
45. m equals 1W 40dBm equals 10W and so on Calculating power with dBm is simple since most components specify parameters such as insertion loss or gain in dB For example an amplifier with 20dB gain will amplify a 20dBm signal to 40dBm 20dBm 20dB 40dBm In linear using the same numbers an amplifier with 100 gain will amplify a 0 1W signal to 10W 100 0 1W 10W 2 1 5 Passive Intermodulation PIM Passive intermodulation is a form of intermodulation distortion that produces interference in high frequency circuits This can for example affect the wanted signal from a radio transmitter by hiding it or distorting it Passive intermodulation occurs when two or more signals exist in a passive and non linear device or element The signals meld and mix with each other generating a completely different signal 6 The PIM level is related to the carrier power an increase of 1dB carrier power approximately results in a 3dB higher PIM level 7 2 2 Electronic components This section describes the components used for this project 2 2 1 Tuneable Filters For the purposes of this project all electronic filters will be categorised as high frequency components A tuneable filter allows the user to control a bandpass or bandreject filter by tuning where to start rejecting or letting through desired frequencies This can be controlled by turning a small knob on the device or controlling the filter digitally using an optional remote digital inte
46. mplify usage of the application A graphical representation of the current configuration would further increase user friendliness A descriptive user manual is important so the user can understand how the application works 21 Original Adjustable frequency range by switching Base and or tuning filter 22 Original Adjustable amplification by switching Base amplifier 23 Original Preset for bands B1 B3 B4 B7 and B8 Wanted 24 Original Preset for Band B12 and B13 Extra 25 Original Preset for other bands Extra 26 Original Easy to use interface Base SE Original Graphical representation of settings Wanted 28 Original User Manual Wanted 29 Original Application written in Agilent VEE Base Appendix B Components and datasheets The following list is a compilation of the components that was chosen to match the requirements laid out in the filter box design chapter as well as their datasheets if available The filter specifications were provided by Peter Perman consequently no proper datasheets are included Component type Product name Manufacturer Lowpass filter 7L120 510 T 1500 O O K amp L Microwave Highpass filter 11SH10 3000 T 13000 O O K amp L Microwave Bandpass filter 13ED50 1250 T500 O0 O K amp L Microwave Tuneable bandpass filter SBT 500 1000 5 N N K amp L Microwave Tuneable bandpass filter SBT 1500 3000 5 N N K amp L M
47. neath the noise floor of the spectrum analyser This investigation is meant to give a good theoretical framework for the possibilities of physically constructing the filter box should it be proven reasonable This includes costs size and various component requirements for the filter box 1 3 Delimitations In depth software simulations of the filter box will not be investigated due to cost and time constraints For the same reasons the filter box will only be constructed in theory The filter box s main purpose is to measure out of band spurious emissions other applications may be possible but they are not the focus of the investigation 1 4 Scope This investigation shall determine the requirements for the various measurements and the restrictions they put on the system used for measurement and the filter box The filter box shall not limit the system used for measurement in terms of signal to noise ratio Additionally possible components and layouts for the filter box will be investigated with costs and limitations in mind Also the graphical programming language Agilent VEE will be used to program a way to control the planned filter box 1 5 Disposition This report is written with each title in the method chapter running in chronological order of what was conducted The result chapter covers the final result discussion and possible future plans for the project 2 FRAME OF REFERENCE This chapter describes various terms that may
48. ore it enters the filter box An attenuator will be used with the automatic measure algorithm presented in 3 4 4 to prevent the spectrum analyser from breaking when tuning the filters Additionally according to Jorgen Ankarberg a test setup for spurious measurements in band usually features a path with around 50dB attenuation so having the choice of using an attenuator instead of an amplifier can be useful to make the filter box more flexible thus a 40 50dB attenuator will be included in the filter box To connect all components with each other radio frequency coaxial cables are going to be used In a fixed setup like the filter box semi rigid cables can be useful They generally offer better electrical 18 properties than flexible cables but at the expense of not being formable 17 Even so the flexible type has previously been used in filter boxes for specific bands at Ericsson and consequently should offer satisfactory properties Moreover since no design and layout of the inside of the box is confirmed only the flexible kind will be considered as they can be used almost regardless of how the inside is configured It is important that the cables have the right connectors hence they will have to be specified after all other components are chosen so everything matches Also of importance is the PIM generated by the cables it cannot be higher than the lowest requirement for SEM from the DUT otherwise the PIM might be mistaken for SEM
49. range of a spectrum analyser can prevent measurements on the most extreme cases 50dBm carrier and 115dBm SEM Even with the carrier attenuated by 50dB down to OdBm the difference is 115dB which might be too much for spectrum analysers to handle Using filters with better stopband attenuation alleviates this problem It is hard to determine how difficult it is to implement the automatic measurement algorithm in Agilent VEE It is possible the algorithm can be improved and adjusted and is likely to be changed because of the shape factor issue Given more time it is possible that other component choices exist which fulfils the requirements and are cheaper than the ones we have picked Maybe it is possible to adjust the parameters of the tuneable filters further thus circumventing the shape factor issue 4 3 Future Work Investigate other components to provide support for B7 and TX of B8 Further development of the Agilent VEE user interface application Prepare its integration with physical components and solve possible bugs Write a manual for the Agilent VEE user interface application Investigate if there are cheaper and or better alternatives to components 32 Investigate if there are better tuneable filters with better shape factor Investigate the intermodulation of the components REFERENCES 1 Spurious Emissions itu int 1997 Online Available http www itu int dms pubrec itu r rec sm R REC SM 329 7 199707 S PD
50. rder to further understand the language In addition the programming software and the exercises were provided by Ericsson 3 4 2 Flowchart for the application Ul For the programming of the user interface a flowchart was drawn in order to understand how the flow of the application was going to work See Figure 17 The user needs to be able to select which band that is going to be used for the configuration Additionally the ability to change the configuration of the filter box by choosing the signal path is necessary This is done by selecting which filter external component to be used along with attenuation gain external components Once a configuration has been selected the user is prompted if the configuration is correct If confirmed the user is able to start the measurement Otherwise the user can make a new selection for the configuration 24 Select Band Select Filter E Select Gain Yes Measure 0 H Figure 17 Flowchart for the application 3 4 3 Programming control for the filter box With the help of the flowchart a user interface application was developed for the filter box However due to the fact that no physical filters were available to be programmed stubs were instead used Stubs are stand ins for not yet implemented modules in the code which in this case would be the filters amplifiers and switches since there were no physical components to apply the control to In addition a stub was used to s
51. rface 5 such as GPIB 8 2 2 2 Directional Coupler Couplers are passive devices used to split a portion of the power travelling in one transmission line out through another connection or port The coupler is a four port device consisting of an input port an output port a coupled port and an isolated port See Figure 2 The main line is usually between ports I and 2 and is suited to carry higher power levels while port 3 and 4 are better suited for lower powers because they are intended to only carry a small proportion of the main line power A coupler does not have set ports any port of the 4 can be the input port resulting in the directly connected port being assigned as the transmitted port the adjacent port becomes the coupled port and the diagonal port becomes the isolated port 9 For example a 10dB coupler See Figure 3 attenuates the power by 10db Input port a Coupled Isolated port port Figure 3 Directional Coupler 2 2 3 Switches All switches this report refers to are electromechanical switches suited for RF and microwave applications See Figure 4 In this report it is assumed the reader has the basic knowledge of how a switch works However the terms TTL and cut off power circuit might come across as unfamiliar TTL logic stands for transistor transistor logic and is used to enable the status of the switch to be controlled by the level of the TTL logic input 10 A cut off power circuit is an opt
52. spacious it was assumed the chosen components would most likely fit in a similar case even including cables and room for ventilation Note that this configuration does not have any power supply as it had not been looked into at this point nor does it include any controller for the switches The existing box is mounted in a 19 rack so keeping the width the same would allow the new box to be easily fitted in the same type of rack Worthy of note 23 is that the filter box in the lab protrudes a bit though the backside of the rack which is 600mm deep According to J rgen Ankarberg at Ericsson it is not a problem if the box protrudes a little but keeping the filter box within the rack dimensions is preferable 3 3 10 Filter box cost After establishing contact with manufacturers and distributors the costs for the components were written down and calculated The prices will not be documented in the report as they are considered confidential by Ericsson 3 4 Agilent VEE Programming This part of the method chapter covers the development of the user interface for the control of the filter box The application was written in the Agilent VEE programming language 3 4 1 Learning the basics The programming language Agilent VEE was studied by reading and solving problems in the form of small exercises to teach the user how the basic functions work and what the language is capable of Experienced programmers of VEE were consulted at Ericsson in o
53. ssible that more environmental friendly components exist for creating this filter box but it was not a matter we investigated very deeply It is uncertain if these components would have sufficient requirements However it can be something to investigate for future work Since the drawback from using tuneable filters results in worse shape factor it might not be possible to measure as close to the carrier as desired Working around it by attenuating the signal more results in worse SNR which in many cases would make the SEM undistinguishable from the noise However it is questionable if using non tuneable filters for the box would be better Since they are not tuneable a large amount of filters would need to be used if the same amount of bands is going to be covered And a higher amount of filters is guaranteed to increase the price and size of the filter box If more filters are the answer using the existing filter boxes for separate bands is probably better than building a bigger box with all filters Another problem is that it is hard to account for all passive intermodulation without having components to test PIM needs to be measured on a real system since it is dependent on how clean connectors are and if they are thoroughly connected In addition only the switches and cables have specified levels of PIM consequently intermodulation of the other components are unaccounted for and would have to be investigated before purchase The dynamic
54. t heceeiteainavann E Hae b bR E A E 3 s M6 hs of ee gerne me en seder gens NAS HR ae er ee ANSE SRS Snor rn RER SEA nere 3 2 FRAMIE OF REFERENCE eet hee e eege Ee eege 3 2 1 Theoretical terms E 3 2 2 Electronic componens eeren mre ere er os err eee eer er SR 5 GE We we 8 Bi PROCESS aoso 5vslorsetnssvsdsend enn ssd kr ESA bs Hee bd Ene sie ed v n Eje EE KEES rr verb 12 SI G NSENON OF BASIC EE 12 SEENEN 12 3 3 Filter box design EE 12 3 4 Agilent VEE Programming vectsrveisieconeestecenssatecevipotecwuesncane EErEE EE RE EENE EEEREN TEE E EE 24 4 RESUITSAND CONCLUSIONS cceccscvcsesacanetcncsunettocns necesesunete e E aE Oa E E E EER 31 eh 31 4 2 DISCUSSION E 32 4 EE VM OTK eege 32 REFERENCES sordir sirr gees GAEREN tant s s skara bsr EENEG 33 APPENDICES geet omaenese ee lesen eege eebe ee eegene dee SNS SNR 1 Appendix A Filter box requirement spechficatton rr rna nn r rr rr rna 1 Appendix B Components and datasheets kee 1 1 INTRODUCTION This chapter contains the background purpose delimitations and the scope of the project 1 1 Background Ericsson uses filters for verification measurements on radio units to radio base stations At the product verification section specific filter boxes are used for measurements on different radio bands which make their setup and configuration trivial However at the design integration section these measurements take a lot of time each set of measurements can take hours to setup and ca
55. tion This appendix contains the requirements for the filter box Interfaces The switches and tuneable filters in the box will be controlled by a PC connected by GPIB The box will have extra ports for connecting external components for other measurements to increase flexibility An output port for connecting a spectrum analyser to measure signals 1E Original GPIB for communication between Base filter box and PC 2 Original Ports for connecting external Base modules 3 Original Port for spectrum analyser Base Functional requirements for filter box The filter box has to be able to measure TX OOB spurious emissions on bands B1 B3 B4 B7 and B8 Support for band B12 and B13 is desired and if possible support for other bands as well Ideally the filters will be digitally tuneable to keep the number of filters needed low In order for the signal to be distinguishable it is necessary for the PIM level generated by the components to be 130 dBm or less with a 40dBm carrier This requirement makes it possible to measure signals at 115 dBm which is the strictest requirement for OOB measurements with a SNR of at least 5dB To cover all desired measurements the ability to handle a constant wave of 10W is required For signals to be measurable by the spectrum analyser the carrier needs to be attenuated by at least 50 dB by the filters Otherwise the spectrum analyser will in all likelihood break due to the carrier
56. to amplify low power spurious emissions to make them measureable and attenuate the carrier signal so the instruments used in combination with the filter box does not break The report describes the procedure of investigating viable components for the filter box which fits with the requirements in terms of noise size and cost of each component These components were only studied in theory using datasheets and information from distributors as it would be very expensive to order them for the purpose of testing The idea of using tuneable filters as components for the filter box in order to cover a wider variety of frequency bands and lower the amount of necessary filters was looked into Based on the investigation of noise and size a filter box configuration was developed In theory it meets the noise size and intermodulation requirements however the amount of intermodulation a form of distortion in passive non linear elements from all components cannot be determined in theory and could therefore be incorrect An application written in the programming language Agilent VEE was developed to add the possibility of controlling the filter box with a computer However since there were no actual physical components to test with the application the use of stubs was required stubs are replacements for real components to simulate the components behaviour As a result the application itself requires further development so it can integrate with physical components In
Download Pdf Manuals
Related Search