HP Programmable RF Attenuator Repair

[rubidium]

Just a couple of initial questions for now: Can these things be repaired? What internally is it that actually fails. I have an 8664A signal generator (and the aching back to prove it!) that has "gaps" in the output level as it is varied from the front panel. The transitions from/to sensible output to/from just noise coincide with audible relay clocks in one of the 2 attenuators. With the top cover off, and if my 65yo ears aren't deceiving me, it's the 33321-60030.  Before rolling up my sleeves and doing any further disassembly to remove/replace/inspect the attenuator, I'm just wondering if they can actually be repaired with a sensible level of effort. Replacements can be had on eBay, but I suspect it's just a crap shoot as to whether any of those would be in better working order than my own. The sig gen otherwise works fine, and I can get by output-level-wise with the presently functioning range of the internal attenuators combined with fixed external attenuators. But if the internal one can be fixed, I'd like to invest the time and do so.
Thanks,
Jim

[George Edmonds]

Hi

May I suggest that you ask the question at

https://groups.io/g/HP-Agilent-Keysight-equipment

I believe that the normal problem if it has not been damaged by excessive reverse RF Power is some small O rings.

George G6HIG in the UK

[rubidium]

Yes, thank you for that tip. I also discovered several videos tied to o-ring replacement (e.g. ). It appears that this is such a common problem that it's worth getting some o-rings (about $0.10 each it turns out), pulling the attenuator, and taking a look. As you say, there's always the possibility that one of the pads is cracked or fried, but worth the time to hopefully discover it's merely an o-ring.
Thanks,
Jim

[rubidium]

Follow up: I did the "dirty deed" of removing the necessary hard and flex coax cables, ribbon cables, and about a dozen screws to extract 2 step attenuators from my Agilent 8664A. After removing the attenuator covers, a quick inspection of the plungers that the relays push and pull upon to engage spring contacts with the actual attenuating elements revealed that a few of the o-rings were missing, while many others were well on there way to disintegrating and departing the plungers. Disassembling  the attenuators themselves involves the removal of several dozen more screws (whew!), the plate that secures the attenuating elements, magnets, and finally the o-ringed (or bare as the case may be) plungers themselves. When I inspected the attenuating elements under the microscope I found them all in pristine condition - and breathed a big sigh of relief when I realized that I could save this by just replacing the o-rings. I used 0.5mm cross section, 1mm, Duro-70 o-rings that I bought on line, and they fit pretty well. Getting it all back together is a juggling act that leaves one wishing one had a couple more hands, but it all went pretty straightforwardly. When I finally fired the 8664A up, I had working step attenuation throughout the entire amplitude range. This was about a 3-hour job from end-to-end, with a lot of manipulation of tiny parts, but well worth it. Funny, but about the time you complete 2 attenuators you finally feel that you've got the hang of it and are ready to take on a few more - but then there's no more to do.
Jim

Original worn o-rings.

Replacement o-rings.

[Rax]

Hi rubidium,
I think I'm just about to embark on this project (I have an 8657B with gaps in output every 5dB).
Could you possibly report the image showing the specifics on the o-rings you used? For some reason, the full size picture does not correspond with the preview.
Or maybe just post here that tag again?...
Thank you!