HP 8566A large amplitude error in the 2-22GHz range

[matthuszagh]

I recently purchased an HP 8566A on ebay that was listed as functional. I tested this using an HP 8341B synthesized sweeper. Although the range up to 2.5GHz seems to be fine (I haven't tested it much, but the few frequencies I've checked seem to be ok), the higher frequency range shows large, progressively increasing amplitude errors. The amplitude error (using a 0dBm input) at 2.2GHz is about -4dB. At 20GHz the error is around -37dB. Although the 8341B sweeper hasn't been calibrated, I've tested the output power with a calibrated power meter, so the error definitely appears to be with the spectrum analyzer. All cables etc. are of high quality. I've performed the error correction routine using the calibration output, which appears to work fine. I also frequently see YTO unlock messages, even though the instrument has been given adequate warm-up time.

The ebay seller insists that the instrument worked when it was sent. I'm skeptical, since the instrument did not have any visual signs of damage when unpacked and I've been extremely careful with it. I've only taken off the cover once to replace a stripped screw and I did this at an ESD-safe workstation. I haven't yet checked and replaced old caps, so I guess it's possible this could be a source of failure that didn't manifest when the seller tested it.

Before I do a deep dive into the service manual and start pulling the device apart, replacing old caps and doing other checks, has anyone experienced this behavior or something similar with an 8566A or B? Any advice on how to proceed? Also, I'm happy to post more information and pictures/videos if those are helpful. Just let me know!

[cncjerry]

Lots of help available on the HP groups.io forum.  Many, many 8566Bs on there. YTO unlock shouldn't be happening at all, except maybe when just plugged in and not warm.  There are some easy fixes for that problem and some that aren't easy at all.  The amplitude errors could be the attenuator so you might want to try bouncing that up and down a bunch of times and trying again.  Also try the amplitude at a couple different attenuator settings.  You should run the calibration and error routines (page 104) and see if they fail because at that level of variance they will fail if there is a problem the calibration can't adjust out.  Make sure the calibration settings are set on, can't remember how, like shift-x or shift-y.  You should also look at the calibration constants and make note of them now highlighting those that are out of spec, if any.  It is nice to have a baseline that can help diagnose future problems.

I wouldn't wholesale replace caps but check voltages to see if they are within spec.  Across the three I have i've had only one capacitor fail in 5yrs.

The 8566 and 8568 are just great analyzers but have been bounced around a lot between people over the years.  I have an 8566b and two 8568s for some reason. 

Jerry

[miken]

On my 8566B there was an obviously bad cap on the A21 board, which lives under the swinging plastic door. A21, A20 would be where to look first I think.

[matthuszagh]

Thanks for the recommendation cncjerry, I've been getting a lot of help here (https://groups.io/g/HP-Agilent-Keysight-equipment/topic/hp_8566a_large_amplitude/78770386?p=Created,,,20,1,0,0). Miken, I'll take a look there.

[0culus]

Before you do anything crazy make sure the preselector peaking is correct (read the fine manual for more details). I have found that this not being set correctly will cause odd amplitude problems in the upper bands. Basic thing to try is to put a tone in the area that you suspect and manually hit "PRESEL PEAK" and let the automatic routine run.