Tektronix 492BP Spectrum Analyser Repair: Only fundamental 0-1.8GHz band works

[adilmalik]

Hi Everyone,

I have got my hands on a 21 GHz Tek 492BP and was hoping it would be my first SA for university RF work . The instrument seemed to be in good nick, with original cal stickers from 2005 still sealed. I tested the instrument and apart from a dim CRT it seemed operational. However, I noticed that the instrument only displays signals in the fundamental band of 0-1.8GHz. When I switch to other bands, the noise floor jumps up but I cannot see any input signals displayed, it is as if the input is disconnected.
I believe the relevant manuals are: http://www.ko4bb.com/manuals/86.21.69.68/Tek_492_Microwave_Spectrum_Analyzer_Service_Manual.pdf , http://www.ko4bb.com/manuals/86.21.69.68/Tek_492_All_Schematics.pdf

Any ideas about how to proceed debugging, i believe the issue here might be something to do with the harmonic mixing circuit inside.?

Thanks

[pbarton]

I know absolutly nothing about the Tektronix 492BP Spectrum Analyser, however if its working on Band 1, then that proves that the majority of the instrument is working (from input to the display). There is very little extra functionality/circuitry employed in the addition of the other bands.
I did look at the manual...
http://www.ko4bb.com/manuals/109.145.77.118/Tek_492_Microwave_Spectrum_Analyzer_Service_Manual.pdf
I did notice on .pdf page 7 of 96 (the frequency control section block diagram <7>) that there is a Span Attenuator <35> box in the lower part of the schematic.
This box shows a relay which is closed for Band 1 and which feeds the FM coil+ signal.
If you jump to .pdf page 69 of 96 (the span attenuator <35> schematic) at the RHS you will see relay K4062, the output of which feeds the FM coil+ signal.
Its probabally worth checking that this relay is doing what its supposed to be doing. Well, its a start.

Two more relays to check, in the 1st Converter pdf p13 of 96 <12>. In the top left corner S11 and S12 within OPTION 1.

There are a limited number of baseband checks that you can use, to diagnose the problem.
Ultimately you probably need access to some specialised equipment to investigate the presence/absence of Band 2 signals.
For example, it would be nice to know if you have a LO signal coming from the Output of the 1st and 2nd LO Connectors, just below the CRT (across multiple bands). However, you may not have anything that could indicate the presence of the LO signals?
Further investigation into the Frequency Control Section (p7 of 96 <7>) depends upon what equipment you have access to?

Band   Freq Range        LO Freq         1st IF
1         0-4.2GHz         (2072-6272)   2072MHz
2         1.7-5.5GHz      (2529-6329)   829MHz
3         3.0-7.1GHz      (2171-6271)   829MHz
4         5.4-18.0GHz    (2072-6276)   829MHz
5        15.0-21.0GHz   (4309-6309)   2072MHz

Above table extracted from .pdf p14 of 81,   http://www.ko4bb.com/manuals/109.145.77.118/Tek_492_Instruction_Manual-1979.pdf

[adilmalik]

Thankyou so much for taking the time to go through the manuals and helping me!

I have managed to borrow some equipment to test this. As per the LO ranges you have mentioned I checked the output of the LO connectors on the front and I am happy to report that the 1st LO used for the first and last band remains always on. The LO used for the other bands is off in band 1 and band 5 but turns on in bands 2-4 as it should.

 Now I don’t have any equipment to test most of the bands but I hooked up one of those cheap 4.4Ghz ADI synth boards hoping some of the harmonics would be detectable up to 18/20GHz. My thinking is to try and eliminate if the band 2-4 IF strip is bad or is the problem further up front the signal chain. Assuming the IF for bands 2-4 is bad I should get signals through in band 1 and band 5 as they use the other IF which I know works for band 1. But this is not the case and bands 2-5 seem equally dead. I feel this means the IF for bands 2-4 is most likely functioning and the problem is in the YIG preselector path....?

PS I have temporary access to an older 1.8GHz SA, 1.3GHz generator and a dodgy ADI 4.4GHz synth PCB to debug this.

[pbarton]

According to the functional block diagram, pdf p1 of 96 <1>, in the top left corner, a pair of relays are used when bands 2 to 5 are selected.
If either relay S11 or S12 is defective, then this could account for Band 1 working and the other Bands not working.
Is it possible for you to eliminate this possibility?

[Testtech]

As mentioned,  the RF relays are a source of problem on this model. Kind of difficult to get to, but the entire relay assembly can be removed complete. Spend a little time looking at it, and the method should be somewhat apparent.

[adilmalik]

Ok, I have opened it up and everything looks connected, plugged in etc. The relays seem intact, with the control line connectors firmly inserted. I suppose I should first try and bypass the assembly and see the result?

BTW, shall I get myself a SMA torque wrench before doing this or is over/under tightening not soo critical?

[pbarton]

You mention that the fault symtoms include a raised noise floor on Bands 2 to 5.
You could use your "older 1.8GHz SA" as a diagnostic tool, just by monitoring the noise floor.
For example, look at the output of the 1st Converter while on Band 1, observe the noise floor.
Then switch to Bands 2 to 5 and observe the noise floor.
If the noise floor is raised, then the fault is within the 1st Converter.

[adilmalik]

I think increased noise floor on bands 2-5 is normal? The higher bands use harmonic mixing where the conversion loss is higher so the dynamic range takes a hit?

[tggzzz]

BTW, shall I get myself a SMA torque wrench before doing this or is over/under tightening not soo critical?

Getting it wrong can mean subtle measurement errors that are not only difficult to track down but also difficult to fix. It would be a shame to foul up an expensive precision instrument.

You will be horrified at the cost of the torque wrenches. You can fake them with a spanner and a spring balance; it is easy to measure the spanner's length and then to calculate the required force.

Note that some of the connectors are pretty inaccessible. The blue widget can help.

[pbarton]

Clearly while on maximum span the noise floor tends to rise with frequency.
However, you mention "When I switch to other bands, the noise floor jumps up"
This sounds like a significant noise floor jump, how many dBs?
I would have thought that you could utalise this noise step indication, to locate the problem using the other SA.

[adilmalik]

Ok, I will have a look today and get back to you.

Also, it seems the manuals are not totally accurate for the 492BP. In particular, in addition to the 1st Mixer and 2nd Mixers I also have a "Harmonic Mixer Assembly" board: https://www.sglabs.it/en/product.php?s=tektronix-a25-harmonic-mixer-119-1640-01&id=1804. Not sure how this is different to the 1st Mixer? I ask because Im getting a good deal on a used "Harmonic Mixer Assembly" board, so if that is critical I can buy this and test if it helps.