HP 8648C Signal Generator with No output

[jiarui]

I just purchased a HP 8648C signal generator on eBay.  The unit passes self test, and the display of frequency and output level responds to the front panel input. When I turn on the RF output, I can hear a relay click. But no signal comes out of the RF output. I checked it with a scope, as well as with a spectrum analyzer. 

I'm not sure what exactly the self test checks.  Would it report an error, if no RF signal is generated by the internal unit at all? Is the problem likely at the mechanical attenuator? Any suggestion is appreciated.

Thanks,

Jerry

[Tony_G]

Somebody might have blown the attenuator by feeding too much power back into it. The Service Manual is on the Keysight website with block diagrams that should show you where to test in the RF output chain.

There might be a CLIP as well that you could search for.

TonyG

[hscade]

Easiest thing is to check the mech. attenuator. The ALC is all devices before the mech. attenuator so the device can't check the signal directly at the output N-Type connector.
There should be a coax cable from a module to the mech. attenuator, just disconnect this and connect it to your SA.

Sharia mentioned that in one of it repair videos. (all credits to signalpath)

[jiarui]

Thanks for the replies. 

I opened it up and checked for signals on the input of the attenuator and on the output.  There is signal on the input, but not on the output of the attenuator.  The problem is apparently with the attenuator.  I took the attenuator out and opened it up.  It turned out that, for some reason, all of the contactors slipped of the plunger.  So when the solenoid is dis-engaged, the plunger is pulling the contactor back. Thus, no bypass contact.  I pushed all of them back onto the plunger.  I'm now getting the default bypass connection.  I put everything back and I'm now getting signals out.  It seems that there is still some problems with the attenuator. Maybe one of the solenoid is not working.  The output level changes in weird ways.  It jumps by quite a bit when the attenuator is switched.  I'll try to open it up again and do some more checking.

[Oldtestgear]

The solenoids in the attenuator can stick. In extreme cases the result is no output & the attenuator appears to be dead. I took the attenuator apart & then gently worked the sticking solenoid until it was free  again.  I used IPA to help clean the shaft being careful not to flood the working parts. I got a couple of these working perfectly again after a couple of attempts.  There is nothing to lose in carefully taking it apart as it is faulty anyway.

Just my opinion though.....

[Tony_G]

There was another thread on repairing an 8648C attenuator:

After more research I found that 33322-60011 is the only attenuator that attenuates up to 130dB using 10, 20, 30, and 60 dB sections. So using any other attenuator would result in discontinuities when sweeping the signal amplitude since the attenuation that the 8648C tries to select wouldn't match what the mismatched attenuator gives.

With that in mind I successfully repaired my 33322-60011 thanks to a series of small miracles. For those Googling, the problems with the attenuator were:

Contact fingers not mating well with the bypass pads / attenuator pads

   Solution: clean pad mating surfaces with a q-tip and rubbing alcohol until a good connection is observed with an ohmmeter. [1]

Plunger o-rings missing / deteriorated

   Solution: ideally buy new o-rings, but being impatient I used a small amount of hot glue to secure the plungers with bad o-rings to their respective metal plates. Totally amateur but it worked. [2]

Solenoid wire was severed

   Solution: re-solder the very fine green wire back to its respective post

[1] Take a look at page 9 of this PDF showing attenuator internals. Note the gold flaps between each bypass surface, and how the ends of the gold flaps make contact with the bypass pad:

   http://ve2azx.net/technical/HP_StepAttenuatorRepair.pdf

[2] This video goes into detail regarding the o-rings:

   

Hope this helps.

TonyG

[jiarui]

I have watched this Youtube video as well as a couple of others about attenuator repairs.  My problem now seems to be that one, probably the 10db one is not engaging.  It does not click at all.  It's possible that the driver chip may be dead. I'll try to take a look when I have a change.  This Youtube video talked about the exact same attenuator.  It should help me with this.

<iframe width="560" height="315" src="https://www.youtube.com/embed/0dAv6Z1CCwQ" frameborder="0" allow="autoplay; encrypted-media" allowfullscreen></iframe>

[jiarui]

I opened up the 8648C again. It's a little pain to take a attenuator out. You'll have to remove the motherboard, which means removing almost everything.  After I took out the attenuator, I did a little probing to figure out the controls. I hooked up a 24V power supply to it and tested all of the controls.  It turned out that all 4 controls responds to the input control.  I hooked up the attenuator to another signal generator and a spectrum analyzer so that I can check the attenuation.  Everything checked out.  I put everything back and now the output level matches with the display on all ranges.  I'm not sure what happened before. For some reason, the -10DB control was not responding before.

I'm pretty happy with this.

[Tony_G]

Congrats - Maybe the actuator was stuck and needed more current than the control circuit could provide but your 24V supply gave the push it needed to free itself.

Anyway glad to see it working for you.

TonyG