"Jeff Keyzer" <jeff@...> writes:
I have an HP 8753D with the internal 50 ohm test set (6GHz) and what I
suspect is an intermittent step attenuator.
If I connect a short cable between ports 1 + 2, I will occasionally see a
steep rolloff in the uncorrected S21 below about 300kHz. Cycling the power
ranges manually a few times usually fixes it. Sometimes leaving the analyzer
alone for a few minutes or giving it some "percussive maintenance" will fix
it as well, so I believe that the mechanical step attenuator may be dirty
and is not always making contact. Certain ranges are more affected than
others, so I suspect that one of the stages is going bad.
I can't find information about the step attenuator in the service manuals,
where can I find more information? I'd like to know the part #, how
difficult it is to extract from the VNA, and if anyone has had any luck
cleaning or repairing these.
I've read that HP considered these non-serviceable items so you're
unlikely to find service manual information other than electrical
connections to the attenuator. However, in many cases they are
repairable and definitely worth the attempt.
I don't have any experience with the HP 8753D, but I would probably just
open the unit up and take a look if you can't find much in the
manual. I'd expect it to be right next the the RF connector, so probably
at the front. If you take a picture of the label on the attenuator
people on this list can probably help out. I "repaired" an HP 33321
attenuator in my HP 8663A a few months back. Unfortunately, I neglected
to take notes on the process but I did take pretty detailed pictures and
I think I could remember from those if you need some assistance.
You can sometimes implicate the attenuators without opening the
unit. For instance, in my unit it would jump up by 20dB or something
when I would decrease the power 1dB for some levels. The attenuators
work (at a very high level) by switching in attenuation pads. These are
often 1, 5, 10, 20, 40 dB (probably others too, but you get the
idea). They achieve the desired attenuation through combinations of
these.
The most common fault (judging by number of posts) seems to be faulty
o-rings. Do a search of those here and maybe on eevblog too. You should
get a bunch of results. There are also youtube videos last I checked
that could prove useful.
In my case the o-rings were totally fine and the switches were just a
bit sticky. Attempting to actuate it electrically didn't do the trick,
but opening it up and actuating it mechanically (many times) worked. I
haven't had an issue with them since.
With anything RF precision like this you do want to be careful when you
open them up. But, with some care it's not difficult to avoid any
damage.
I've tried cycling the attenuator manually a bunch of times and it's still
exhibiting the problem. I was hoping that by exercising it a lot, the
problem would go away. Wishful thinking, I guess.
Thanks in advance for any replies.
Best of luck and hope we can help.
Matt
