HP 8590A Anomaly in trace/scan

[n6oli147]

Hello all,

New to the forum here. My 8590A has developed a problem that I'm having a hard time troubleshooting. When I power it up, everything looks fine, but after a few seconds I get traces that scroll across the screen in the noise floor.
Could it possibly be in the video filter on A16 processor board. Kind of hard for me to explain, so I posted a video on Youtube.
Any help would be greatly appreciated.
Thank you.

[Testtech]

I would look in the IF and variable resolution chain. Do you have another spectrum analyzer available? A scope may be enough until you get into the IF system. Something is "leaking" into the signal path I think.

[MarkL]

Try to characterize the problem a little more.  Does the jump in dB appear the same with other Amplitude settings?  If you give it a signal, does the glitch add to the signal or does it just raise the noise floor around the signal?

It could also be an ADC issue.

Before digging into anything, take a look at the Aux Sweep Output to see if the problem is there too.  If it is, it's an RF/IF problem.  If not, start looking at the processor/ADC section.

[n6oli147]

Yes, the jump in dB does appear with other amplitude settings. The glitches do not add to the signal, only the noise floor. I was thinking the same thing about the ADC issue. When I do a Conf test it does fail with errors Vid-BW fail in 1000, 300, 100, but I think
that's due to the fact that it has the glitches. I'll try your suggestion on the Aux Sweep Output. Also, when I disconnect the 2ND converter cable going to the 2ND IF, the glitches are still there. So, no RF in the signal path I believe.

[n6oli147]

I just checked the Sweep Output, and it looks good to me.

[MarkL]

Ooops, very sorry.  I meant the "Aux *Video* Output".  This should be a copy of the analog amplitude before it gets to the ADC.

[n6oli147]

The AUX Video Output signal looked okay on the Oscilloscope, and when the glitch went across the screen on the analyzer, the Oscilloscope pattern didn't change. I'm thinking it's the A/D Micro Interface that's controlling the Video Filter on the A16 Processor board.
Could be why I'm getting the VID-BW FAIL on the confidence test.

[MarkL]

I agree it seems likely that's why the video confidence test is failing.

I don't have the full 8590A service manual.  Mine is missing the processor, A/D and other sections, so what I'm about to say here is somewhat blind.  I do have a 859xE manual, which is similar, with many of the same parts with the same Uxx designations.  If you know of a good copy somewhere, please point me to it.

So, caveat, caveat...

In the 859xE, the video filter is before the AUX VIDEO output.  So, if it's the same for the 8590A (which it should be), I don't think it wouldn't be the video filter.

The 859xE A/D has a MUX (U64) which selects the AUX_VIDEO signal on pin 4.  I would start there, and verify the you can see the AUX_VIDEO signal.  As a sanity check, provide an input signal to the analyzer, and you should see the curve on that pin, without the glitches.

After the MUX it becomes more complicated, since you need to synchronize your measurements with whatever is going on with the MUX input selectors on pins 1, 15, and 16.  With any luck, the system is not doing anything with the MUX while sweeping and it's always set to input #0 (pin 4).  But you can check the MUX selector inputs for activity.

I would also take a look at the ADC itself (U47) and make sure the REF OUT (pin 9) and REF IN (pin 10) are stable.  Check other pins to see if there's any activity that correlate with the glitches: +15V (pin 7), -15V (pin 11), +5V (pin 1), 10V SPAN (pin 13), BIP OFFSET (pin 12).

You also have a gain stage (U45), and a sample and hold (U46) which deserve some scrutiny.

There may also be some diagnostics which could help verify the A/D operation, but again I don't have that section.  There is brief mention of a +10V and -10V reference check that is apparently described in the section I'm missing.

EDIT: "Wouldn't" --> "would", but hopefully the meaning was obvious anyway.

[n6oli147]

Well, I found the problem. It was the fan introducing noise on the +15 volt rail. Incredible! It passes all confidence tests, and no more anomalies in the trace. Maybe they should have included that in the service manual.
Thanks for all your help MarkL, and Testtech.

[MarkL]

Great - glad you found the problem!