2 - John Miles KE5FX
Contents
1. 44 0100 0100 10dB unchanged 1 6 05 0000 0101 20dB unchanged C200 41 0100 0001 30dB lt 30dB was 040430 now become 10 20 0 gt 0000 0100 04 20 40 0100 0000 40dB 40dB was 1040430 now become 0 0 40 gt 0100 0001 41 C215 01 0000 0001 50dB 50dB was 0 20 30 now become 10 0 40 gt 0100 0000 40 C21E 00 0000 0000 60dB lt 60dB was 10 20 30 now become 0 20 40 gt 0000 0001 01 PATCH THE ROM WITH THIS 1 2 45 CIEC 44 ClF6 05 C200 04 C20A 41 C215 40 C21E 01 Image files for the 492 1 6 ROMs are available in both original and modified form at http www ko4bb com in the ROM Images and Drivers Tek 49x firmware and notes directory The attenuator control table is likely to appear at a different offset in other firmware versions including those for other instrument models so you should verify the location of the original byte array before making any changes The firmware for models later than the original 492 496 is checksummed so it will be necessary to update the checksum after applying any binary patches See Firmware checksum calculation for more information Appendix Firmware checksum calculation When modifying the firmware in instruments later than the original 492 496 models it s necessary to calculate and store a new checksum word to avoid error messages at bootup time The following program is a small Win32 console application that will display the current and new2. GPIB di sconnect void ds cal voi d GPIB CTYPE type connection type if type GC NI 488 printf Error This test requires a National Instruments GPIB interface it printf will not work with Prologix adapters n return C8 C 1024 Cc 0 64 532 k 125 32 il 0 for 532 i l i lt 10 i for 532 j l lt 100 j C ilsj k 225 i 142100 if i gt 25 id cf_cal void SI GS WP e C 1001 KE IT printf nConnect calibrator then press any key when test complete GPIB_write INIT REF 20 SPAN 2M SIG while kbhit GPIB wri wri id 101 sens cal void GPIB write INIT FREQ 10M SPAN 100K di sconnect rite AE 5M SIG WAI T rite TUNE 5M SI G WAIT id lo2 range cal voi d printf VnPress any key when test complete while kbhit borg writ e TUNE 2M SIG WAIT GPIB write TUNE 2M SI G MIT getch DE printf Done n id 102 sens printf nPress key when test complete while kbhit GPIB write TUNE 2K 516 WAI T TUNE 2K SI G WAIT getch printf Done n any key when test complete DE nCenter marker on screen then press any key to continue da void shutdown voi d GPIB disconnect mai n 32 argc C8 argv printf nThis program assists w
3. MOTHER BD 829 MHz DIPLEXER 2 qom pg imzo Siek 9sok 41 15v Sla 6 v 15 m 7 12V R1019 816 rios la 39K Ls A R1029 R1025 3 5 25K 59 63 10 4 12v itis 2v R1015 12 f 11 59 25 2 Veins 2 plas 12 Ria 1015 sor 1 25K Ogas 2022 R2025 oua 10 45V DV R3015 10 0 PROGRAMMABLE BIA lt p 03019 R4016 R3014 B Q352 1S HEAT SINKED TO 1K 15K THE BOARD MOUNT mos EXTRUSION THEREFORE IT 15 PART LISTED AS A quM b 4 01013 cont PT Ten PART OF A04 ASSEMBLY 094753 pi oo cion 1 A vices 670 9051 00 947 POF is LO DRIVER 2221 233 1st LO DRIVER Symptom Unstable amplitude response with slow continuous variation Amplitude instability in the newer units e g later production 494APs as well as 495Ps with serial numbers after BO30000 is becoming all too common thanks to Tek s eagerness to adopt surface mount technology at a time when component quality wasn t yet up to par with traditional through hole tech If you own one of these early 1990s vintage analyzers chances are good that your logamp board in particular already looks like this The black crud in the highlighted area doesn t originate from the resistors or from the soldering or assembly process Rather it s caused by outgassing of corrosive vapors from one or more of the 10 uF 50V electrolytics nearby Any corroded traces should be cleaned or brid
4. So I then wondered where that 10 dB was going It turned out that the levels were as specified in the manual until the output of the 10 dB gain steps amp The manual says 3 dBm but the best I could do with the Amp gain cranked wide open was 5 dBm Just to make sure that it wasn t a bad board I swapped the one from my 494AP and got the same result I noticed while I had the 41 VR module open that the VR input on the 495P is much different than that of the 494AP It has an additional gain adjust pot and more circuitry It appears to be for setting the 60 dB gain For the 494AP the Band 1 leveling adjustment is set fully CCW With the 495P things may be handled differently I noticed that the Band 1 leveling had previously been set to about mid range That is where I left it The VR module in my 494AP is 644 0645 02 agrees with the 494AP manual The VR module in my 495P is 644 0776 01 VR module 2 may be fully compatible but there appear to be some differences with VR module 1 What is even more puzzling is that the 494AP manual references 644 0645 02 as being for the 495P Somewhere down the line it changed This particular 495P B031241 late 1993 production probably has more in common with a 2795 than an early 495P Along the same lines the later 497s post B020000 are relabeled 494As as far as l ve been able to tell They can be recalibrated for use up to 21 GHz When upgrading the VR assemblies to bring a 492AP
5. and R1032 to align the two signals at midscreen Appendix Attenuator control firmware patch RF attenuator failures can be difficult to repair due to the lack of available replacement parts It s possible to replace the attenuator with a three section unit from any of several other spectrum analyzer models with similar frequency coverage using appropriate circuitry tohandle any differences in solenoid voltages or latching characteristics But unless the replacement attenuators switched elements correspond to the 10 20 and 30 dB pads used in the 49x series some glue logic may be needed as well You can use almost anything from a few discrete transistors to an FPGA to modify the behavior of the Z axis RF Interface assembly or you can take the easy way out and hack the firmware Below Luis Cupido offers some notes on modifying the V1 6 492 firmware to allow it to drive the 10 20 40 dB attenuators found in many analyzers from HP and other manufacturers The original table below contains the attenuator control bytes for each attenuator setting addr value att settings 1 2 45 OdB 44 gt 10 C1F6 05 20dB C200 41 30dB C20A 40 40dB C215 01 50dB C21E 00 60dB The bit meanings are as below bit 7 6 5 4 3 2 1 0 is lt x gt lt 20dB gt x lt gt lt gt lt 30dB gt x lt 10dB gt Original and modified byte values 1 2 45 0100 0101 unchanged
6. should always be done with the covers in place especially given the possibility of leakage from 10 MHz GPS clocks in many labs and ham shacks these days If your analyzer doesn t meet its noise floor specifications with a 50 ohm terminator at the input jack this is a good thing to check Symptom Span division does not match the value displayed on the CRT Check the blue epoxy encapsulated ceramic capacitors on the Span Attenuator board for DC resistance or shorts These are used as RF bypasses across the TTL switch lines to help keep digital bus noise out of the LO control loop If troubleshooting leads you to suspect either U1042 or its driver logic be sure to rule out the capacitors first Symptom Noisy or defective cooling fan Failure of the high performance tubeaxial fan is common on higher mileage 49x 275x analyzers The OEM fan is an exotic part which is unobtainable today Especially on the newer analyzers with 10 Hz resolution avoid the temptation to hack in a cheap CPU fan as a substitute Chances are it won t put out enough air to cool the entire analyzer safely Additionally the fan is mounted in close proximity to sensitive IF circuitry so both EMI and mechanical vibrations will appear as excessive composite noise or sidebands at close in offsets It s a good idea to run a phase noise plot before and after replacing the fan to make sure you haven t degraded the instrument s performance watching carefully for spurs When con
7. to support a 496 496P or 492P Use caution when toggling DIP switches found in these analyzers especially the GPIB switch on the back panel These switches sometimes fail when actuated for the first time in 20 years leading to difficult to diagnose problems Software applications Several measurement applications were marketed by Tektronix for the P suffix GPIB programmable analyzers in the 49x 275x line allowing users to take screenshots automate basic RF measurement tasks and control equipment located at remote sites Beginning with the 494P the instruments firmware could drive pen plotters directly However resurrecting obsolete Tektronix controllers and plotters is not usually worthwhile unless you re faced with the need to recreate or support a specific measurement system Many commonly used spectrum analysis applications have been recreated on modern PC operating systems Commercial applications that support the 49xP 275xP spectrum analyzers include PrintCapture from F amp F SoftTools and various utilities from Aphena including SoftPlot RemoteControl and Plottergeist On the freeware side the KE5FX GPIB Toolkit by the author of this document provides a large collection of Windows applications for Tektronix Advantest and HP Agilent spectrum analyzers including full C C source code The GPIB Toolkit includes utilities for screen capture elapsed time spectrum recording with waterfall displays phase noise measureme
8. Tektronix 49x 275x 279x Series Spectrum Analyzer Service Notes John Miles KE5FX john miles io November 30 2015 This public domain document may be reproduced and distributed freely The latest edition is available at http www ke5dfx com 49x_notes pdf Errata and helpful notes on calibration service procedures Page numbers refer to 494 P service volume 1 070 4416 00 but in most cases there are equivalent pages sections in all other 49x 275x manuals 3 Deflection amplifier gain and frequency response p 5 6 Step 3b Apply a 500 kHz signal In later editions e g 494A P volume 2 070 5561 00 this was corrected to 500 Hz Step 3b and set Triggering to Int I ve found that at least some models need to be left in FREE RUN mode to trigger on the specified video waveform The 49x 275x deflection amplifier underwent several revisions throughout production including a complete redesign in 1988 Regardless of your instrument model refer to the manual change information in the 494 A manual 070 5560 00 if your instrument has one of these newer subassemblies Don t forget to disable digital storage before adjusting the X Y deflection gain Recalibrate the digital storage subsystem immediately after any adjustments are made 5 Frequency Control System Calibration p 5 9 Step e When observing the markers either turn digital storage off or follow the procedure in 13 Digital storage calibration first recommend calibr
9. X Loop from C002 to FFFF inclusive 11 CPX 0000 11 BNE L_DBC4 11 unsigned short sum 0 for int i start 2 lt end i unsigned short C sum amp 0x8000 1 0 sum lt lt 1 sum sum image i imegetetartell sum unsigned short updated unsigned short amp image start printf New checksum X n n updated if updated original printf Checksumis unchanged n return 0 fflush stdin printf Press y to update 9s any other key to exit n argv 1 char ch getch ch y ch Y ILE out fopen argv 1 wb if out NULL printf Error could not open for writing n argv 1 return 1 if fwrite image 32768 1 in 1 printf Error could not write to 9sVn argv 1 return 0 The source code above can be compiled under Windows by saving it as chksum cpp and running cl chksum cpp with your environment configured for access to the Visual Studio command line tools It can also be compiled under Linux MS DOS or MacOS with few if any changes The program operates on a specified 16KB bank within a 32 KB binary EPROM image bin file After modifying the bin file as needed to alter the model name attenuator control table or other properties run the chksum exe program and tell it which bank 0 or 1 to update After updating the checksum you can write the bin file to a 27256 3 or faster EPROM and install it in the analyze
10. agnets to make one stronger magnet At the other end of the wire I made another small loop for a securing screw that I ran through an unused screw hole on the side frame near the fan The wire was then bent to center the magnets at the center of the hub I 6th inch off the face I think the wire I used was actually a lead free solder for plumbing very malleable but not as soft as lead No more buzz No obvious downside but certainly not tested to 21 GHz either I m not a RF guy so I can t say if a strong magnetic field located near the fan will cause a performance hit For certain the fan vibration was a problem posted by cap on Mon Jan 10 2011 6 04 pm This is a pretty common way for the fans to fail so cap s suggestion could be quite helpful A small stationary magnet should not cause any problems when mounted in that vicinity assuming the fan continues to run OK in its presence Of course strong magnets should be kept well away from the YIG tuned devices on the RF deck Symptom Failure of front panel CENTER FREQUENCY control The CENTER FREQUENCY control s optical encoder uses a 2mm incandescent bulb which can be replaced by a type 7153 lamp 5V 115 mA Mouser Electronics part number 606 CM7153 Symptom Excessive phase noise Check ESR on the two aluminum electrolytics on the YIG oscillator interface board These capacitors are paralleled for low ESR so one of them will need to be temporarily disconnected in order to meas
11. al locations were left alone In earlier logamp assemblies the order is as specified left to right on the component side INPUT REFERENCE LEVEL R1012 LIN MODE BALANCE R1025 OUTPUT REFERENCE LEVEL R1030 LOG GAIN R1037 LOG FIDELITY R1060 also called log linearity in some manuals In later assemblies the correct control order is LIN MODE BALANCE H1025 LOG GAIN R1037 OUTPUT REFERENCE LEVEL R1030 INPUT REFERENCE LEVEL R1012 LOG FIDELITY R1060 HNMODE BALANCE INPUT R1012 ir reU 1 2 LOG FIDELITY R106 Step h Return the step attenuator to 0 dB Display should be full screen 6 dBm if not readjust the signal generator output for 6 dBm This confusing step was modified in the later 494A P manual step i but never actually fixed The overall goal is to achieve and maintain fullscreen reference level response with 6 dBm at the input to the logamp It s not clear what Tek s intention was when documenting this part of the calibration procedure but it appears the intent is for the vertical position control to be adjusted to maintain a fullscreen indication with 6 dBm If someone has a better interpretation please email me and help keep this document up to date VEL Note that the first generation 492 496 logamp modules used 0 dBm as a reference level rather than 6 dBm If your signal generator has a calibrated attenuator HP 8656 8657 8642 8662 you obvio
12. an with later ones where more resilient ferrules were used If your RF attenuator seems balky check Q3028 on the Z axis RF interface board and see the Manual Change Information page for reference M65465 page 308 of the 494A P volume 2 PDF A70CR3028 was added to protect Q3028 from reverse voltage induced by the attenuator coils 49x 275x analyzers with the preselector and external mixer features have several coax switches in the front end signal path but the switches have been very reliable in my experience Bias tees and diplexers are more likely to be faulty Before condemning the first mixer in the 21 GHz analyzers for instance make sure the DC bias path from the LO driver board to the mixer bias termination is OK The front end mixers in the 1 8 GHz analyzers were operated at zero bias but the 21 GHz models used a circuitous bias path with several potential points of failure Excerpts from the 494AP schematics appear below with the first mixer bias path highlighted in green P O A23 ASSY C1010 C1013 829 MHz 1 0 1 0 2nd CONVERTER LO SECTION RION 180 1 2 Guz LOW PASS R1012 o Pan 100 Ox Teml 11015 t1012 cron x 4504 1 0 I 1 7 FLIS d2 829 T Leros 55 I 7 an 1010 41011 81015 r te MIXER 111 gt 9mH 200 20 Bias 1 1012 1016 L 10 7 4 crore 1019 670625691 570 2 82 J 082 J 082 A23
13. ating digital storage immediately after the deflection amp gain so you aren t confused in any subsequent steps by storage registration errors This is reflected in later manuals such as the 494A P s Step e 3 apply 0 2 us time marks from the time mark generator to the RF input Any 5 MHz sinewave source is OK for this especially if it s strong enough to create some additional harmonic distortion in the front end Select 100 kHz RBW to yield narrower markers Step f Max Span Dot Don t bother to change the resistor based on the response at the left end of the display as the manual specifies Change the resistor only if the dot marker is substantially out of alignment at higher frequency markers e g at 1 8 GHz in step f 3 The 1 8 GHz marker may be too weak to see neighboring markers are OK for this test as well Figure 5 10 Adjustment and test point locations for calibrating the frequency control system p 5 11 TP1073 is located on the Span Attenuator circuit board not the 1st LO Driver circuit board as shown in the 494 P and earlier manuals 8 Log Amplifier Calibration p 5 14 This procedure involves various interacting control settings ambiguous instructions and printing errors Don t adjust your logamp unless you actually need to Symptoms such as excessive amplitude drift will not be corrected by calibration alone While Tektronix continually upgraded their instruments to use the latest modules in production the same can
14. checksums for a given binary ROM image file and allow you to update the file if desired This program is also available at the http www ko4bb com site in the ROM Images and Drivers Tek 49x firmware and notes Tek 497p 9 7 directory i ncl ude lt stdio h gt incl ude conio h i main int argc char argv printf n printf Tektronix 49x ROM checksum utilit 1 00 of 20 May 04 n printf John Miles KE5FX jmil es pop net jon printf e I ee I e IA ee ee eee eee ee eee ee ee eee eee eee in if argc lt 3 E 0 amp amp argv 2 0 1 printf nUsage chksum filename bin bank n BE Bank 0 for first 16K bank in image 1 for second 16K bank n n return 1 static unsigned char image 32768 FILE in fopen argv 1 rb if 1 NULL printf Error could not find s n argv 1 return 1 if i image 32768 1 in 1 printf Error could not read S n argv 1 return 1 fclose in int start 0 int end 16383 if argv 2 0 1 start 16384 end 32767 unsigned short original unsigned short amp image start printf Previous checksum X n original 11 T Original 6802 checksum al gorithm 1 CLR A Init sum 0 A MSB B LSB T CLR B L DBC4 ASL B 16 011 sum lt lt 1 ROL A Shift MSB out to Carty 1 ADC 00 16 bit add of 00 plus carry gt sum 11 ADC A 00 1 N
15. echnology but they do an outstanding job of covering topics that are relevant to users of 1980s to 1990s vintage instruments Other Tektronix application notes describe the use of accessories such as waveguide mixers and tracking generators as well as specific measurement techniques Links to online copies of these will be added to future versions of this document as they become available Appendix Typical phase noise baseline plots The traces below were acquired with PN EXE from the GPIB Toolkit measuring the output of a Wenzel crystal oscillator whose phase noise is substantially lower than the analyzers own Typical examples of the 494AP 494P and 492P spectrum analyzers were compared Minimum RF attenuation was set to 0 dB in all three cases and 20 dB of carrier clipping was used 60 dBc Hz 70 dBc Hz 80 dBc Hz dBc Hz 100 dBc Hz 118 dBc Hz 120 dBc Hz 130 dBc Hz 140 dBc Hz 166 Hz 1 kHz 18 kHz 166 kHz 1 MHz Trace Carrier Hz Carrier dBc Hz at 10000 Hz Clin dB instrument S N B010388 100 666 660 494 S N B 16845 100 000 660 8 00 494P S N 055692 100 000 660 8 00 492P The 494P appears noisier close the carrier but this is actually a consequence of its broader 30 Hz RBW specification compared to the 494AP s 10 Hz minimum bandwidth The 30 Hz filter skirt and not the LO phase noise is responsible for the observed performance at offsets below ab
16. ely 8 inch pounds with the bezel still in place This procedure will bind the CRT with a safe amount of pressure allowing it to undergo normal handling jolts without overstressing either the CRT face or the plastic implosion shield It s not necessary to follow any instructions in the manual regarding removal of the old CRT from its shield cladding There s no reason to do this unless your replacement CRT didn t come with its own shield Troubleshooting topics General notes on power supply service Regulation tolerances for the low voltage supply buses are specified in service wlume 1 and are not usually a problem However the 100V and 300V supplies are frequently below their specified values If any 66 kHz ripple or sagging is observed all of the miniature 2 2 uF 200V axial lead electrolytic capacitors in the power supply should be replaced These are inexpensive high failure rate parts so they should be replaced in any event if the power supply is otherwise disassembled for service recommend checking the ESR on any replacement high voltage electrolytics as well Several new capacitors in this voltage range have proven defective when obtained from surplus sources High ESR in the filter capacitors on the lower voltage rails 17V 5V has been reported asa cause of excessive power supply temperature so when servicing the power supply you should check ESR on all electrolytics as a matter of habit Check heat sink fasteners for tig
17. er DC charge Other bench practices associated with ESD hazards In the 1 8 GHz analyzers 495P 496P and their 275x equivalents it s easy to substitute an off the shelf mixer for the Tektronix front end mixer The Mini Circuits ZX05 83LH is a good choice that will maintain or even exceed the original mixer s conversion gain and IMD performance It s necessary toreverse the RF and IF ports when installing an aftermarket DBM The mixers in most of the 21 GHz analyzers seem to use Tektronix part numbers in the 119 1017 xx family Note that 492s with 8 GHz coverage option WJ use 119 1353 xx parts instead One difference between 119 1017 xx and 119 1353 xx seems to be the presence of a DC block at the input to the former 21 GHz mixer It s been reported by Luis Cupido CT1DMK that the beam lead diodes can be replaced with either BAT30 low barrier parts from Infineon or MSKN709 medium barrier diodes from Metelics Low barrier Metelics parts should also be usable as may similar diodes from MA COM Luis has found that the better known HCSC9101 diodes need more LO power than the analyzer has available Although the mixer and its diode carrier are easy to work with on the bench replacing the diodes on the carrier is another matter Gap welding capabilities are required for the gold ribbon to substrate attachment while gap reflowing is used to attach the ribbons to the diode holder although the latter operation may be performed with a small s
18. ged as soon as you spot the problem and all of the aluminum can electrolytics should be replaced 50 volt capacitors aren t needed in most cases 35V is fine for supply buses up to 15 volt and their exact type tantalum or aluminum isn t important either Just get those old caps off the board Keep in mind that affected traces can be several inches away from the nearest capacitor A careful inspection of the entire PCB is necessary when this failure occurs Front end mixer damage causes and cures Spectrum analyzer front ends are winerable to damage due toa variety of causes RF overload Keying a transmitter into the input jack Use of an efficient antenna near a high powered transmitter Connecting the analyzer to a oscillator or other source gt gt 100 mw Application of DC to an unprotected input Failure to use a feedthrough capacitor or DC block Failure of a feedthrough capacitor or DC block at the component level Use of a feedthrough capacitor large enough to dump a substantial charge into the front end when power is applied to the unit under test Failure to verify that no DC bias is present on an unfamiliar signal connection ESD Setting RF ATTEN to 0 then touching the center conductor of the input jack Failure to use an external attenuator and or preamp when working with an antenna then touching the antenna or exposing it to high potential sferics Use of a long piece of coax that carries a leftov
19. htness and renew the heat sink compound while you re at it Avoid losing track of the mica washers and aluminum spacers that may fall out when you remove the PCB Use 105 degree C rated electrolytics where possible e g 3 3 uF 350V 105C Digi Key part 493 2046 ND These can replace all of the smaller high voltage capacitors in the power supply 6x 2 2uF 200V 1x 1uF 350V The two large filter electrolytics on the AC line side seldom cause trouble They can be replaced if necessary with Panasonic EETED2E821DA Digi Key P11644 ND or Farnell 1198619 According to a Usenet post sci electronics equipment 2 Sep 01 the large stud mounted transistors with Tektronix part number 151 0703 00 cross reference to the industry standard part number 2N6586 10 450 12 5 MHz fT TO61 package With the appropriate mechanical modification 25 4237 transistors have also been successfully used Symptom Display noise jitter The 49x 275x CRT display subsystem is one of its weaker areas You will not see HP like levels of display stability and sharpness in these instruments but there s often a lot of room for improvement When servicing instruments with the pre 1988 deflection amplifiers it s a good idea to replace the 620K and 300K carbon composition resistors the negative feedback paths Many of these resistors have drifted upward over time and or become noisy Typical replacement parts would be P620KBBCT ND two each and BC300KW 2CT ND
20. ith calibration of Tektronix 49xP printf spectrum analyzers using routines fromvolume 1 of the Te printf service manual 070 3783 01 FEB 1984 edition n if p 2 printf nUsage cal tek490 lt address gt n exit 1 addr quM atexi t shutdown for rintf n 1 Digital storage cali ion pa 3 70 n 2 1st 0 ga War teet Cl L ver 031 CF control 3 0 MHz n 4 4 2 GHz step 2c 3 52 5 4 278 GHzV n 6 5 5 GHz step 2c 3 52 7 lst LO tune sensiti vity CF E R1028 page 3 55 8 2nd LO tune range CF control R4040 page 3 5 J 9 2nd LO tune sensitivity CF control R3040 page 3 56 n ESC Exit n n Choice C8 ch getch if 27 ee exi t 0 printf bc n ch connect ee ch case 1 ds_cal break case 2 cf cal break case 3 GPIB write FREQ 0M break case 4 GPIB write FREQ 4200M break case 5 GPIB write FREQ 4278M break case 6 GPIB write FREQ 5500M break case 7 lol sens reak case 102 Longe cal break case 9 page 3 51 n 02 break ge If you are following the calibration procedures in the 492P manual 070 3783 01 note that the roles of the 1st LO Offset R1032 and 1st LO Sense R1031 controls are reversed in steps 2 a and 2 b Use R1031 to bring the two signals to the same horizontal position
21. itude control is not needed for preselector calibration Maintenance Replacing the Crt p 6 29 Like the logamp alignment instructions this is an area of the manual where the engineers and technical writers at Tektronix never quite found themselves on the same page Personal injury or equipment damage can result from following Tektronix s CRT installation procedure The key point is simple before reinstalling the clear plastic implosion shield and metal bezel frame you should loosen the four plastic mounting blocks around the perimeter of the new CRTs face This will allow the CRT to rest as far back in the instrument as possible while the bezel is being tightened The goal is to be able totighten the bezel s four Allen screws in a cross pattern without putting any pressure on the CRT face You should be able to slide the implosion shield around freely with one hand while you tighten each bezel screw If the plastic shield binds up as you tighten one of the bezel screws the mounting block screw at that corner is still too tight In particular don t follow step 6 on page 6 30 of the 494P manual 070 4416 00 remove the bezel and tighten the mounting block screws or any similar instructions that may be in your edition of the service manual Instead once the bezel screws are tight and you ve verified that the plastic shield is still free to slide across the CRT face tighten the mounting block screws evenly in a cross pattern to approximat
22. not be said of their service manuals Calibration procedure change notes were never issued in many cases so it s best to follow the instructions in later 49x 275x instrument manuals whenever they apply to the modules in your instrument When searching for the correct schematics and service information for your instrument date codes on components are often better clue than the model nomenclature on the front panel For example you may not need to install your logamp on extenders and remove its covers as recommended in step 8b on page 5 14 of the 494 P manual if your logamp s control adjustments are easily accessible through the slots in the housing as they were in all later instruments The first step in logamp calibration is to identify which of the three possible Log and Video Amplifier test point and adjustment locations figures 5 17 in 070 4416 00 or 5 13 in 070 5560 00 with two possible control layouts applies to the module in your instrument An analyzer with the older style logamp uses figure 5 17 in the 494 P manual while logamps in newer 492A 492B 494 495 497 analyzers and their 275x series counterparts use the physical layout in figure 5 13 of the 494A P manual The catchis that at least two different layouts were used for the controls in figure 5 13 At some point probably around the time Tek redesigned the board using surface mount components in the late 1980s early 1990s the trimmer functions were reordered while their physic
23. nt and other applications LEVEL FREQUENCY SPAN DIU REF amp GDBM CEN 1 000 6 GHZ MKRx 93 9DBH HZ MKR 999 99 MHZ TEK 494AP 18087 4eDB 1 8 EXT 3Hz 2 VERTICAL RF FREQ REF VIDEO RESOLUTION DISPLAY ATTENUATION RANGE OSC FILTER BANDHIDTH In short GPIB support is a very useful feature one that often goes unappreciated by hobbyists If you have a programmable spectrum analyzer that isn t connected to your PC you re missing half the fun Recommended books and app notes Tek Scope volume 12 issue 1 1980 available in PDF format at the http www ko4bb com manuals page introduces the 492 with descriptions of various production advances and key design concepts The Manuals Tektronix 49x directories contain other useful application notes as well including the well regarded Fundamentals of Spectrum Analysis As the directory layout on this site may be subject to change over time it s not practical to include direct links It s very worthwhile to explore the Manuals Tektronix pages on your own Spectrum and Network Measurements by Bob Witte is a great tutorial on all facets of spectrum analysis A more technically oriented and much harder to find volume is Modern Spectrum Analyzer Theory and Applications Artech House 2nd edition 1984 by Tektronix s Morris Engelson Neither of these books is oriented towards cutting edge signal analysis t
24. oldering iron It may be difficult to preserve the original gold ribbons during dissembly be prepared to replace them with new ribbon stock Amateur wire bonding and gap welding work is discussed on Luis s chip hybrids Yahoo group When replacing front end components use extreme care to avoid overtorquing the 0 141 hardline SMA interconnects Don t use needle nose or serrated pliers to work with gold plated connectors use a 5 16 or 8mm wrench ideally an SMA torque wrench Inside the first mixer Photos courtesy Luis Cupido CT1DMK of the chip hybrids group showing the diode carrier from the 21 GHz first mixer with its substrate removed undergoing diode replacement The mixer diodes themselves are almost too small to see even in the magnified view at bottom Actual width of the entire substrate is approximately 15 mm Parts accessory sources The TekScopes Yahoo group is a great resource for repair information for all Tektronix instruments new and old Watch for contributions from Tektronix veterans such as Stan Griffiths whose work at the company included technical marketing for the spectrum analyzer product line PC board extenders are needed for many 49x 275x service and calibration tasks High quality reproductions are frequently available from Norway Labs ask for Matt North when inquiring about them Extenders for the keyboard and power supply assemblies were also issued by Tektronix but they are not included in N
25. or similar model up to 494AP era standards remember to install the 3 MHz filter FL36 on the RF deck as well Firmware Firmware images and notes on DIP switch settings invaluable for upgrading as well as troubleshooting can be found in the Manuals ROM Images section at http www ko4bb com Each set of 49x series firmware images is shared across several models in the product line with the DIP switches determining the instrument s actual personality For example the v9 0 firmware stored at the Tek 495p 9 0 link is applicable to the 494A P 492A P 495 P 2753 P 2754 P 2755 P and 2756 P models while the v9 7 firmware stored under Tek 494AP 497P 9 7 added support for the newer 492B P 2755A P and 497 P models I m not aware of any benefits to upgrading a 9 0 instrument to 9 7 sol don t recommend doing so in most circumstances There is also an earlier version Tek 492ap 6 0 which is suitable for use with the original 495 P and 492A P ROM boards that have only eight EPROM sockets rather than the ten that are present on the later instruments Search the various notes and spreadsheets in the ROM Images section carefully for the latest information applicable to your model Users of the original 492 should look for the Tek 492 1 6 link which contains the original 1 6 firmware for the 492 plus a consolidated file suitable for writing to a single EPROM with notes by Luis Cupido I m unsure if this file can be used
26. orway Labs s package and appear only rarely on the surplus market While nice to have they are not as important as a full set of PCB extenders Quickar Electronics and Sphere Electronics are good sources for Tektronix specific semiconductors and other components You can also search for Tek part numbers at these sites http www american milspec com http www talonix com http qservice tw Greece Don t neglect eBay when searching for parts and documentation it should be your first resort If you can afford a parts mule a junked instrument whose modules are compatible with yours by all means grab one Module swaps and upgrade possibilities Modules can typically be exchanged between instruments of similar vintage 492 496 494 492A 494A 492B and so on Early model 495s are most compatible with the 494 and 492A models while late model 495s with the 10 Hz RBW feature post B030000 are essentially 494As without the 829 MHz IF strip That being said Brian Henry has noted a few differences between the VR modules from his 494AP and a late model 495P after observing that he needed to maximize the 110 MHz IF gain to calibrate the latter properly I checked the input level to the 495P s VR module and it was at 25 dBm In the 494AP manual it is supposed to be 35 dBm at that point In light of that I decided to back down the IF gain to 35 dBm After doing that the output level from the VR module dropped by 10 dB
27. out 300 Hz At offsets of 10 KHz and beyond all 49x 275x analyzers should generally perform within 5 dB of these traces Allthree analyzers meet Tektronix s rather loose PN specifications easily Appendix Calibration program for 494P and earlier analyzers For analyzers that don t have firmware procedures to assist with digital storage and 1st 2nd LO calibration here s a C version of the BASIC code from the 492P service manual 070 3783 01 A GPIB adapter from Prologix or National Instruments will be required This code can be compiled with the MS Visual Studio Express package and linked with the gpiblib lib library provided with the KE5FX GPIB Toolkit Specifically you can save the file in the Toolkit s installation directory as cal tek490 cpp and compile it with cl cal tek490 cpp gpiblib lib then run cal_tek490 lt addr gt from the command line to display the required calibration pattern Replace lt addr gt with the analyzer s actual GPIB address A ready to run copy of this program is included in the GPIB Toolkit under the name cal tek490 exe incl ude st di incl ude con i ncl ude assert incl ude lt stdlib define W N32 LEAN AND MEAN incl ude wi ndows h include gpiblib h 532 addr WNAPI GPIB error C8 msg S32 ibsta S32 iberr 32 i bent 0 gt 0 gt rt h gt h gt printf s msg exit 1 void connect void GPIB connect addr GPI B error void disconnect voi d
28. r Again this operation is not necessary and will not work with the base model 492 496 ROM images end
29. s rubber silicone grommets and or gasket compound Do not attempt to bolt the fan to the power supply housing without a vbration dampening mount unless you enjoy looking at noise sidebands Similarly careful electrical isolation is needed to maintain the instrument s original performance Twist the fan leads together and keep them as short as possible Bypass the fan with a parallel combination of a large 470 to 1000 uF electrolytic capacitor and smaller tantalum capacitor 10 uF and use a series choke with a value of 100 uH or more that s rated for the fan s supply current A more subtle consequence of using a replacement fan is extended warm up time for the entire analyzer assuming the fan is no longer temperature controlled Another take on fan repair is excerpted here by permission from its originator on the Tektronix customer forum Noisy Fan Repair My fan was making a loud buzzing noise When turned on the hub is pushed back into the fan assembly say 3 16th of an inch by air pressure When off the fan spins freely in its forward position but not so freely when it is pushed back 3 16th of an inch If I place a small 1 4 1 16 Neodymium disk magnet to within 1 Sth of an inch of the center hub of the fan then turned the fan on the hub stays in its forward position and no more buzz I took 2 inch by I 8th inch diameter wire and bent a ring on one end and slipped two of these magnets into the ring and glued in place two m
30. two each Calibrate the deflection amplifier and digital storage subassemblies after replacing these resistors This particular board already had 300K metal film resistors 3 4 in place so they were left alone The 620K 596 carbon composition resistors 1 2 were measured at 580K and 980K respectively It may be necessary to adjust the frequency compensation trimmer caps for stable readout characters as well Misalignment here causes instability as well as tearing Ensure that the ribbon cables that go to the CRTs X and Y deflection plates are dressed away from the deflection amplifier PCB If these cables are routed too close to the deflection amplifier input components they can cause feedback problems Check the 100V and 300V supplies under the aluminum cover on the Z axis RF interface board Low voltage or excessive ripple indicates failure of the miniature high voltage electrolytic filter capacitors in the main power supply See notes on PSU servce Symptom Baseline lift or instabilty when servicing VR 1 and VR 2 assemblies The short wire pigtails from the various SMB connectors should be routed close to the assembly walls away from the PCB and other components Some possible layout related issues near the 20 dB Gain Steps subassembly have been observed to cause baseline lift at least in later 494A P era instruments and particularly with the VR assembly covers removed for servicing Final calibration of the VR assemblies
31. ure them independently Check for dirty intermittent operation of the loop gain trimmer on the PLL assembly and for noise in the 10 volt reference on the 1st LO Driver board All electrolytics on these assemblies should be checked for ESR as well It is less common to see excess noise at narrowband less than 10 kHz offsets See the graphs at the end of this document for an idea of what to expect from your analyzer Helpful hint listen to the power supplies and voltage references with a crystal earphone Use back to back 1N4148s across the earphone to awid damaging the barium titanate crystal or your eardrums with a 1K series resistor to the supply being monitored Symptom Low or unstable amplitude response A wide variety of problems can cause this behavior You ll need to troubleshoot the analyzer just as if it were any other multi conversion superhet receiver One common issue is overtorqued SMA connectors in the front end If someone has been too heavy handed with a wrench the ferrules at the hardline terminations have probably been crushed and will no longer make good contact regardless of the amount of torque applied Nothing can be done here except to search eBay for microwave quality SMA jumpers This is especially likely tobe the problem if you re seeing a lot of FM TV or cellular pager leakage into the analyzer with nothing connected to the input jack From what I ve seen this was a bigger issue with earlier 49x analyzers th
32. usly don t need to bother with the step attenuators Users of HP 8640s and other generators with vernier level controls will probably find it easier to use the outboard attenuators as specified When servicing your logamp module watch out for corrosion from any surface mount electrolytic Capacitors that may be present 494AP era logamps were among the earlier surface mount assemblies produced by Tektronix in the late 1980s The quality of the surface mount capacitors used at the time was terrible as owners of other Tek instruments such as the early TDS series DSOs can often attest I ve had tojunk atleast one logamp module due to instability and intermittents caused by this corrosion See Symptom Unstable amplitude response for more on this 8 Adjust Resolution Bandwidth and Shape Factor p 5 14 of 494A P manual 070 5560 00 Step i At the conclusion of this part of the procedure reconnect J693 13 Digital Storage Calibration p 5 22 The layout shown for the trimpots on the horizontal digital storage board in figure 5 24 is wrong The correct order for these controls from left to right is OUTPUT OFFSET R1039 OUTPUT GAIN R1041 INPUT OFFSET R1046 INPUT GAIN R1048 Unlike the logamp controls above don t believe this control layout varies with the board revision As noted elsewhere the digital storage controls should be aligned much earlier in the calibration process than the 494P manual recommends Perform this procedure immediatel
33. ventional single phase DC fan is installed the multiphase driver board will be unused Ideally the new fan will use some combination of the supply voltages available to the existing driver board 49V 5V 17V and ground A 12 volt fan can be used on the 17V bus with a suitable Zener diode for instance The original Buehler tubeaxial fan was rated for 31 CFM of airflow at 12 VDC consuming 2 4 watts Superficially the EBM Papst 622HH model is a close substitute 33 CFM 12 VDC 3 5W but its EMI characteristics are much worse than the original part and also noticeably worse than the Papst 622H model 27 1 CFM 12 VDC 2 3 watts This is most likely due to its higher current consumption For most environments the 622H will be good enough although neither of these fans can be used without additional EMI filtering Some EMI advantages may be gained through the use of a higher voltage fan such as the Papst 624H 18 28 volts 27 1 CFM at 24 volts nominal or 624HH 18 28 volts 33 CFM at 24 volts nominal A 624HH fan on the 17V bus is likely to work well It should cause less EMI than the 12 volt 622HH especially when run at a lower than specified supply voltage Various resellers for the EBM Papst fans can be located by entering part numbers such as 624HH at http www octopart com or http www findchips com Regardless of which replacement fan you select it will be necessary to fabricate a mechanical shock mount with materials such a
34. y after servicing or recalibrating the deflection amp 15 Preselector Driver Calibration p 5 24 Follow the procedure 13 Adjust Preselector Driver p 5 21 in the 494A P service manual when aligning the preselector It s normal for the 19 and 21 GHz response peaks to be rather broad or to occur at the far extent of their respective shaper adjustments In most cases it will not be possible to achieve perfect preselector tracking in any one band much less across all bands and in any event thermal factors will limit repeatability You can spend hours tweaking the preselector response if you like but as soon as you put the cover back on the analyzer the effort will be wasted Ultimately the response with the PEAKING control centered may be 3 4 dB down at various frequencies in the different bands but with correct alignment it should be possible to peak the response at any given frequency between 1 7 and 21 GHz without leaving the shaded area If your preselector response already meets this criterion there is little benefit in adjusting the preselector driver assembly If the TM500 series time mark generator and comb generator specified by Tek are not available an HP 33002A SRD comb generator module driven by a 23 dBm 100 MHz signal source makes a good substitute You can also perform the procedure with a CW capable sweep generator e g HP 86290A H08 or synthesizer HP 8673B or equivalent but it s more time consuming Absolute ampl
Download Pdf Manuals
Related Search