HP 3325B repair

[vindoline]

I recently won a "for parts or repair" HP 3325B function generator for a great price from our favorite auction site. The auction description was actually pretty honest in it's description. I was obviously hoping for a functional unit or a simple repair. This time Murphy got me!  Upon opening it up and taking a quick visual inspection some severe damage was immediately observed on the output attenuator board and the output amplifier board. As you can see, the resistors in the /10 attenuator were literally fried. The resistors in the /3 attenuator also appeared to have suffered - perhaps collateral damage. At this point I had no idea if the Aromat RF relays (made of unobtainium) have suffered any damage. 



Not surprisingly, a quick power-up resulted in multiple fail codes on the display and no signal was observed at the output.

The first order of business is to repair the attenuator board. Although free scans of the 3325B service manual are readily available online, the ones that I found had poor quality schematics. I decided to buy a commercially scanned copy of the manuals from artekmanuals.com For $7.50 I was able to get excellent quality scans of the users manual, installation manual, and the service manual. IMO well worth it. 

The burnt out resistors were cut out and desoldered and the board carefully cleaned with IPA. You can see that the damage burned down into the weave of the fiberglass on the board!



Next replacement resistors were ordered from Mouser. The original resistors were all 0.1% low TC precision resistors. They appear to have been custom values for the 50R Pi attenuation network. I decided to use parallel combinations of available E96 value resistors to rebuild the board. This is based on the excellent information at www.pulseresearchlab.com/download/attenuatorappnote.pdf

The repaired attenuator board is shown below.



To be continued...

[vindoline]

The output amplifier board also showed significant heat damage.

It's a little easier to spot looking at the back of the board.

Both transistors in the push-pull output stage were shorts on all terminals. I decided to simply replace all the transistors in the output stage and any passives that were obviously damaged or located over the heat stressed area of the board.
The 2N3866 is still available from Mouser, but the 2N3866A (a higher bandwith spec) is not. Also, I had to track down the PNP transistor. The markings on the part itself 3-440 402 were not helpful to me. The service manual gives an HP part number 1853-0440 wich I was able to cross-reference to an MM4018 with Google's help. I was able to get NOS 2N3866A and MM4018 transistors from US suppliers on eBay
Again, the damaged components were cut out, leads desoldered and the board cleaned with IPA. This time one trace had lifted and a repair bodge would be necessary.


After reinstalling the board and powering up the 3325B didn't throw a "Fail 023" which it had been  However, hooking up the scope showed no output 
At this point I powered it off and was going to put it aside for a while, but I decided to track down the part I had heard rattling around inside the case first. After tilting the rather large instrument this way and that and gently shaking, out popped a loose scorched resistor. WTF? The markings were unreadable, but a quick check with the DMM showed it to be 50.0 . No, I thought to myself, it couldn't be the output resistor. A quick check showed that sure enough it was. I had overlooked the filled solder pads assuming they were filled vias  After replacing with a new 50 resistor the generator has an output signal  It's clearly a little out of spec, but I can confirm that at least all 3 attenuator relays are working!
Next I need to tackle the odd HPIB problem...

[amspire]

I used to use 3325As back in the early '80s. They were pretty good, but the most common fault was output attenuator/amplifier issues. Not sure how the 3325B is different but I seem to remember the 3325A had a more expensive and delicate custom HP-magic attenuator.

If you ever want to use the generator as a stable AC voltage source for things like calibrating meters, you might want to check the output AC voltage drift. For most purposes, the type of fixes you have done should be fine.

Richard.