Tektronix 495P Fix and Alignment

2021/4/20

Before writing this article, I already worked on this beast for a rather long time. And have try many times to bring it back to work.

It comes to my workbench with a seriously low amplitude response, about ~50db low than the calibration signal source.  Self auto-alignment could not to be completed, said no calibration signal detected.  

The scan base line is very funny, amplitude is so low, even the zero span.  The higher frequency even worse.

Reading the Service Mannual

A marked photo is the most powerful Service manual.  This is an exciting step to get familiar with the device.  It may benefit lots of hobbyists, and make my restoration life happier, enjoy it.

It's worth bring up, all 492/494/495/496/497 modules, almost exchangeable. Some awesome guy said there are exceptions,  but I change 492's log amplifier to the 495, it works, even though lacks one SMB output.

Fix Log Amplifier

By my instinct, it must the log be wrong. I need SMB connector, i feel needs all kinds of connector converter, -:) 

Sadly, there is no 495 schematic at all.  Lucky one is that  492/494/496, they all basically share the same module but different version. Refer the 496/492 log schematic also very helpful.  I just blindly checked all components on the board, bingo! one of the inductors open!

Well, looks like it works again.  It worth to note , there is one spur response near the third vertical line(200Mhz).  And signal amplitude still missing  30dB !  For no reason, I believe the culprits must still be the log amplifier.  

Alignment IF module

I even did not check the schematic carefully, I assume it was the log amplifier is out of alignment for some reason, may the previous owner did terrible thing to it.

It's worth trying alignment.   The most sad thing is that alignment the Log amplifier help nothing at all,  always  30 db less than normal, if you trying hard to tune the Output Ref and Input Ref, the dynamic range gonna reduce to 30 db or even less,  this is another symptom to say it's lack of gain.

It's strange I so obsessed to fix log amplifier and video amplifier while it's not very clear if it is wrong or not.

Reverse Engineer Log Amplifier PCB

Log amplifier installed in TEK495  is new, there is no layout and schematics match it. The 496 Service manual is the nearest version I could find.

It's better isolating problem before fixing,  I totally forgot divider and concoque principle months ago while facing this unfamiliar beast.  But becoming familiar with the details of log amplifier and video amplifier,  help to find which part broken.

VR2's 10Mhz IF input is splitted. One of the paths buffered and output to rear panel, another path , the main path, go into log amplifier.  the last stage is detector amplifier and the detector.  Log Gain trimmer control the couple between log amplifiers and the detector, then could setup a correct millivolts per dB (or per 10dB normally)

video amplifier Note:(while i start fixing,  i know  nothing about this)

• convert millivolts/dB to millivolts/Div by program video amplifier gain per db/Div button config

• switch modes: linear or log etc...

• ref level shift:  U5041 given a DC voltage according Ref level (50mV/ref-level-1-dB-change), shifting the response waveform up and down.

• -20 dBm is special.  On -20dBm ref level, U5041 should output 0V , at -20dBm , minius 96 dB dynamic range(12dB/div)  ~= -117dBm minimal reference level.  -30 dBm U5041 will shift the windows up by add a 1 Div voltage, -10dB gonna switch in 10dB attenuator instead.

• Video 1 come from video processor, compensate the RF front un-flatness

Back to month ago, when I start to investigate how to fix it, I barely know how the video works and  what exactly function it designed for.  It's strange to determine the Log amplifier is responsible for all the low amplitude response.  I try to fix it nothing ...   but all my attention is focused on lacking of a PCB layout.  I decided to reverse engineering the layout ! Starting from a nearest service manual , the journey begin....

And one month late, while I'm very proud of what i'm done,  guess what I found, look!

Evaluation Alignment Procedure

It's hopeless while do Rounds of rounds alignment,  I did know better what the several trimmers for, though.

• INPUT REF-level:  alignment 2dB/Div and  10db/Div response, while switching 2 mode alternatively, you should got same reading. Refer to the video amplifier schematic,  in 2dB/Div mode U5041 will set a DC voltage according Reference level, this voltage cut-off  most of voltage corresponding to reference level (RefLevelDBm - 2db*8Div). 

• OUT REF-level  :  the trimmer add a voltage to final video output, the overall effect is shift the whole response window up/down a little bit to align the Full Scale output to Ref-level, that's is to say, on top of the screen. 

• Linear Balance:  did similar thing alike INPUT-ref-level for 2dB mode.

Service manual steps seem a little bit confusing and not convenient for evaluation,  following  steps present a  quick roughly alignment for evaluation only,  this gonna make the log amplifier works almost ok, before you do a full serious alignment:

• set reference level to -20dBm and reduce reference level  to -70 dBm,  adjust the LOG CAL on front panel, make sure the line step 5 divs.  note:  use zero span and fine video filter mode.

• switch between 2 dB/div mode and 10 dB/div mode, adjust INPUT Ref Level Trimmer, let the 2 modes  all reach top graticule.  you may need to tune the OUTPUT Ref Level otherwise, it's gonna more than full screen.

• switch between linear mode and 10 dB/div mode, adjust Linear balance  Trimmer, let the 2 modes  all reach top graticule.  

• Connect signal generator, set to -20 dbm, and set Reference level to -20 dbm too, adjust OUTPUT Ref Level ,  align trace to top graticule,  switch in 50 dB attenuator,  adjust  LOG Gain make sure signal change just 5 divs. repeated until get satisfied or boring.

• Connect Calibrige signal, and set Reference level to -20 dbm, switch in 40 dB attenuation,  now it's almost on -40 dBm graticule line,  adjust LOG linear timmer, let it on -40 dbm graticule line.  if need a big change on this trimmer, you' d re-do last step.

Auto Calibrate

why auto calibrate process say ' no calibridge signal present', even you got a -20dBm signal perfectly aligned to top graticule?  In endless LOG amplifier alignment,   I got the answer: the 2dB mode is not work!  follow the alignment process for  INPUT Ref Level,  it would work!  Off cause, the signal still  missing  30 dB , by abuse OUTPUT Ref Level,  you could bring the response almost close to FS, then the auto calibrate process proceed smoothly,  seems promising,  but it's not working like this way, though.

Reference Level and dynamic window operation Theory

Learning stuff by fixing things is very interesting and fun.  How the Spectrum Analyzer working?  Understanding  dynamic windows setup is key. 

Overall, Spectrum works in this way: 

• Assume RF stage provide to IF stages with a fixed gain ( It's not flat, though), but equipment a step attenuator.

• IF stages provide configurable gain

• Log amplifier convert the linear signal into logarithm output ( millivolts/db)

• video amplifier convert millivolts/db to  millivolts/div

• deflection amplifier provides a fix millivolts/div interface to display signal trace.

Then an import configuration set the signal amplitude alignment: 

• On -20dbm Reference level,  feed into a -20dbm signal ,  while no RF attenuator switch in,  IF gain is set to MAX value.  IF signal feed into Log amplifier is round  -6 dBm (for Tek 495) . 

• LOG amplifier output feed into Video amplifier: convert millivolts/dB to  millivolts/div,  shift response waveform and up/down, align response waveform to correct level.

• While Reduce reference level 10 db  (to -30dbm),  IF gain keep MAX, shift window up by adding 1 div  voltage into the video amplifier.

• Increase 10 dB reference level (to -10dbm), IF gain keep MAX , switch in 10dB attenuator,  the signal trace lower  1 div.

• Switch in 10dB attenuator, ref-level will keep unchanged,  response trace  1 Div shorter， at same time the whole dynamic window(ie, the response waveform) gonna  shift up 1 div at same time, ie, the signal amplitude not change but bottom line raise 1 div.

Knowing the reference level/ attenuator and If gain changing protocol  help a lot. It should have study that all at very begin of fixing -:)

40 dB dynamic range, really?

After several rounds of alignment,  I noticed I got a response seems only 40 db above noise floor?  looke the HP8566's response, what happens? lack of gain?  bad video amplifier?  mis-alignment LOG amplifier?

I tried one 'backup' log amplifier came from Tektronix 492 spectrum,  it's better, 10dB better than the 495 log amplifier  (following photo).  Seems 495 log amplifier did have some defect.  

Calibrate the log amplifier by inject a 10Mhz signal from a generator,  the 495 log amplifier perform just well,  it could output 500mV/10dB in at least 70 dB range!  Things became very confuse,  

What's wong?

NOTE: 

The logamplifer is not alignment correctly in the above screen capture.

First thing i trying is fixing the inductor, use a TV balun core, 30 Turns , inductance just about 30 uH, its works well as RFC.

This RFC is works better,  the noise floor reduced several dBs!  I wonder the TV core might works better than original open inductor,  but sadly it's not.  2x TV core just change nothing.

Since we targeting to reduce the log amplifier noise,  replacing caps is obvious choice. But that board's capacitor is replaced already,  ESR is well below 1 ohm, that's seem OK.  IF signal is 10 Mhz RF, I wonder parallel ceramic capacitors gonna help.

So let go...

Bingo, we fix it!  Look, works as well as backup log amplifer.

We Fix it, are We?

We are here now, the log server should be good,  now it's time to calibrate it. The mysterious  problem happened again,  carefully, i mean correctly alignment Log amplifier,  I find the response waveform could not reach to -20 dBm (Full Scale). Checking the Log amplifier with external signal generator,  it works well!  

I summarize the symptom:

• Inject 10 Mhz IF signal by external signal generator,  Log amplifier works well.

• On machine, response could not reach top graticule  ( -20 dBm CAL signal,  -20 dBm Reference Level), 30dB gain missing

• I remember VR2 output could reach -6 dBm, I remember!

• There is a false signal response!  nere the 200Mhz response line!

Such chaos ! It's time to sort everything's straight.  Finally, I realize,  I had to checking the signal level to isolating the problem. Thanks OH5IY,  There is a wonder table, it make world beautiful.

Soon, I found VR2 output (to the Log amp input),  not -6 dBm, instead , it's almost only -50 dBm, 30db missing!  Wait, I remember IF signal could reach  -6 dBm, I did check that !

Ha, I recall something else, I had touched the RF path to checking things after fixing log amplifier ! Especially the second Mixer to third mixer SMA cable,  so, rechecking that, and aid by a torque control screwdriver,  fixing is easy. 

Sometime, CPU report 2nd LO tuned failed while use narrow scan. That's might be another tough fixing...  But it report failure occasionally , I suspect some where had touching problem. Look, I pulled the YIG control line and put it back, and so did to another control line, problem solved! lucky! 

This Might be OK

I occasionally read this on the TEK 495 operators manual: while MIN Noise button lighted, the RF attenuator reduce 10dB ( on Ref level >= -20 dBm), so did the noise!  I push that to light it up, magically, the noise floor reduced 10dB!  TEK 495 operators manual said MIN noise model is default, but it's not on my TEK 495 (configurable?).

By active the MIN Noise model,  it seems regular as my Anritsu 610B now!   Signal to noise level: 60dB @1Mhz RBW. Another significant factor for Noise is  peak/avg mode, and video filter. Avg/peak mode and Low noise mode is new to me, Ha, It could be relax if i had known all of these stuff at begin.  But anyway, this is learning fun!

the Average mode is more like a digital Video filter, it's more faster than video filter, and almost same effect somehow.  and, for 1Mhz 3Mhz RBW, high noise floor is expected,  almost every SA achieved  60dB down from Ref-Level, it's a normal thing, I know that now.

Alignment

Yeah, we now finally be here, now we follow Tektronix 495(or 496)  alignment process, at least following modules: 

The Deflation amplifier,  Log amplifier, and digital store (I noticed the saved waveform is low than pure analog trace).

Deflection Amplifier Alignments 

This alignments ensure the vertical deflection amplifier drive the CRT at a certain millivolts per DIV(500mV/DIV). Method is use a signal generator and a oscilloscope,  inject 500 Hz signal to the Mark/Video rear BNC.  to enable Video mode, you need short Pin1 with Pin5. 

external oscilloscope is used to monitor the signal coming from generator, ensure it's exactly 4Vpp, the SA CRT should display a Full Screen Sine Wave as this screen shot. The signal generator muse could tuned the DC offset to shift the waveform  on SA CRT. 

1-4: safe turn on machine!

5. Select the following 495P Spectrum Analyzer settings.

TIME/DIV ................................ ................................ ............. MS

TRIGGERING................................ ................................ ........INT

VIEW A AND VIEW B ................................ .......................... OFF

6. Set VERTICAL POSITION fully counterclockwise.

7. Select the following signal generator settings.

SIGNAL TYPE ................................ ......................... SINEWAVE

FREQUENCY................................ ................................ ...500HZ

8. Select the following oscilloscope settings.

VOLTS/DIV ................................ ................................ ........... 1V

TIME/DIV ................................ ................................ ............ 1MS

9. Adjust signal generator level and offset until the oscilloscope and CRT displays 4 Vp-p sine wave.

10. On J104 ACCESSORY connector, connect pin 1 to pin 5.

11. Adjust A1A64A1 R1066 (fig FO-14) and signal generator offset control for eight division vertical display on CRT

•· Positive sine wave peaks should touch the top graticule line.

•· Negative sine wave peaks should touch the bottom graticule line.

My 495 is on it's perfect state on this alignment item, Good. 

Log amplifier alignment

1.-2.  setup and power on

3. On 495P Spectrum Analyzer, press Blue-SHIFT and PULSE STRETCHER keys.

4. Use DATA ENTRY to select 5=DISABLE/ENABLE USE OF CAL FACTORS.

5. Use DATA ENTRY to select 0=DO NOT USE RESULTS.

6. Select the following settings.

CENTER/MARKER FREQUENCY ................................ 100MHZ

SPAN/DIV ................................ ................................ ........2MHZ

REFERENCE LEVEL ................................ ....................-60DBM

PEAK/AVERAGE................................ ....... FULLY CLOCKWISE

7. Select the following signal generator settings.

FREQUENCY................................ ................................ ..10MHZ

AMPLITUDE................................ ................................ ....+6DBM

8. Set signal’s variable attenuator to 50 dB.

NOTE Steps 9-14 adjust LOG CAL control  to ensure REF LEVEL Shift will move waveform correctly: 50dB for 5 DIV.

9. On 495P Spectrum Analyzer, adjust LOG CAL to center of range.

10. Adjust A1A62A1 R1030 to place trace five divisions below the top graticule line. 

11. Set REFERENCE LEVEL to -110 dBm.

12. Adjust LOG CAL for a five division (50 dB) change on CRT.

13. Select -60 dBm REFERENCE LEVEL.

• Check that trace moves five divisions (50 dB) on CRT.

14. Repeat Steps 10-13 until trace moves exactly five divisions between -60 dBm and -110 dBm REFERENCE LEVEL settings.

NOTE

•· Steps 15-21 adjust A1A62A1 R1025 (LOG Gain): ensure Log amplifier output 5 Div while signal strength change 50dB use external generator.

•· When A1A62A1 R1025 is adjusted correctly, the trace moves one division on CRT each time the variable attenuator setting changes by 10 dB.

15. Set REFERENCE LEVEL to -20 dBm.

16. Set external signal’s variable attenuator to 0 dB.

17. On 495P Spectrum Analyzer, adjust A1A62A1 R1025 to align trace with top graticule line.

18. Set external signal’s variable attenuator to 50 dB.

19. Adjust A1A62A1 R1030  for a five division (50 dB) change on CRT. 

20. Set external signal’s variable attenuator to 0 dB.

• Check that trace moves five divisions (50 dB) on CRT.

21. Repeat Steps 17 through 20 until trace moves exactly five divisions between 0 dB and 50 dB variable attenuator settings.

NOTE

           Steps 22-26 adjust A1A62A1 R1037(INPUT-REF), A1A62A1 R1030(OUT-REF), and VERTICAL POSITION: 2dB/mode and REF-LEVEL accuracy alignment.

22. Set external variable attenuator to 0 dB. (signal +6 dBm)

23. On 495P Spectrum Analyzer, press 2dB/DIV key and note trace location on CRT.

24. Press 10dB/DIV key and note trace location on CRT.

25. Adjust A1A62A1 R1037(INPUT-REF) to match level of traces for 10DB/ and 2DB/ settings.

26. Repeat Steps 23-25 until traces match for 10DB/ and 2DB/ settings.

27. Disconnect cable assembly between A1A62 J621 and variable attenuator.

28. Adjust VERTICAL POSITION to align trace with bottom graticule line.

29. Connect cable assembly between variable attenuator and A1A62(LOG AMP) J621.

30. Adjust A1A62A1 R1030(OUT-REF) to align trace with top graticule line.

NOTE

          Steps 31-35 adjust A1A62A1  R1060 (LOG Amplifier Bottom 40dB change accuracy)

31. Set the external variable attenuator to 40 dB.

32. Adjust A1A62A1 R1060 to position trace four divisions (-40 dB graticule) below top graticule line.

33. Set the external  variable attenuator to 30 dB.

• Check trace position. Trace should be three divisions (30 dB) below top graticule line.

34. Set variable attenuator to 50 dB.

• Check trace position. Trace should be five divisions (50 dB) below top graticule line.

35. Adjust A1A62A1 R1060 for best display at 30 dB, 40 dB, and 50 dB variable attenuator settings.

36. Repeat Steps 15-35( restart from LOG GAIN) if a large change to A1 A62A1 R1060 setting is required.

37. Set variable attenuator to O dB.

38. On 495P Spectrum Analyzer, press SAVE A key.

39. Set variable attenuator to 10 dB.

• Check that trace is between 0.9 and 1.1 divisions below SAVE A trace.

40. On 495P Spectrum Analyzer, press SAVE A key to turn SAVE A mode off.

41. Repeat Steps 38-40 to check remaining seven divisions. Use following table as a guide for variable attenuator settings and test limits.

NOTE

• Steps 56-58 adjust A1A62A1 R1012 (Linear Mode accuracy)

• Steps 56-62 verify LIN mode accuracy.

• After A1A62A1 R1012 is adjusted, trace should align with top graticule line when LIN, 2 dB per

division, and 10 dB per division modes are selected.

56. Set variable attenuator to 0 dB.

57. On 495P Spectrum Analyzer, press LIN key.

58. Adjust A1 A62A1 R1012 to align trace with top graticule line.

59. Press SAVE A key.

60. Set variable attenuator to 6 dB.

• Check that trace is between 3.6 and 4.4 divisions below SAVE A trace.

61. Set variable attenuator to 12 dB.

• Check that trace is between 5.6 and 6.4 divisions below SAVE A trace.

62. Set variable attenuator to 18 dB.

• Check that trace is between 6.6 and 7.4 divisions below SAVE A trace.

63. Disconnect equipment from 495P Spectrum Analyzer.

64. Connect A1W682 to A1A62 J621 (fig FO-13).

65. On 495P Spectrum Analyzer, connect CAL OUT signal to RF INPUT.

66. Press Blue-SHIFT and CAL keys to begin CAL function.

67. Use FINE key and follow on-screen prompts to complete CAL function.

68. Disconnect equipment from AN/USM-620.

Digital Storage Alignment

1. Turn on SA

2. Select the following 495P Spectrum Analyzer settings.

SPAN/DIV ................................ ................................ ....100MHZ

VERTICAL DISPLAY ................................ ........................ 2DB/

3. Adjust VERTICAL POSITION to align trace with bottom graticule line.

4. Adjust HORIZONTAL POSITION to align frequency dot to center of CRT.

5. Press Blue-SHIFT and PULSE STRETCHER keys.

6. Use DATA ENTRY to select 2=DIGITAL STORAGE CAL.

7. Press SAVE A key to display the test pattern.

8. Adjust  A1A61A1(vertical) R1033, A1A61A1 R1045 (figFO-13), A1A60(Horizon) R1040, and A1A60 R1050 (figFO-12) to match the test pattern with example waveform.

• Use A1A60 R1040 and A1A60 R1050 to align patterns with vertical graticule lines.

• Use A1A61A1 R1033 and A1A61AI R1045 to align patterns with horizontal graticule lines. 

9. Press SAVE A key to continue.

10. Connect CAL OUT signal to RF INPUT.

11. Press SAVE A key.

12. Adjust A1A60 R1055 and A1A60 R1060 to match horizontal position of alternate sweeps. See example waveform. 

• Verify that traces match at right and left edges of CRT.

• A1A60 R1055 and A1A60 R1060 interact.

Repeat adjustments until alternate sweeps match.

13. Adjust A1A61A1 R1034 and A1A61A1 R1046 to match vertical position of alternate sweeps. See example waveform. (left amplitude response math to left saved waveform, adjust for minimal changing while it sweep alternatively)

Note for step 12:

use the saved A trace is hard to tune, may because the alternative sweep rate is annoying.  I prefer to toggle View-A , you could directly refer to pure analog result, work well.

All Done, 495 restored, accuracy is satisfied.