Repair of an HP8904A signal generator

[douardda]

Hi,

I have a couple of HP 3904A signal generator. One of them had a failing PSU and I recently took the time to work on it.

I've made a description of the problem and the repair on a small blog I'm beginning there https://whatever.sdfa3.org/hp8904a-multifunction-synthetizer-restoration.html

Long story short, the culprit was a resistor on the small "secondary" board of the PSU (labeled R7, a 20k 1% resistor involved in the current limitation portion of the voltage regulator

I've made a (very) partial schematic of the +16.3V rail of the PSU, if it may help someone:

 

Now, I have another (small) problem? I've discovered that the 10MHz ref output is flat on the other HP8904A unit. I've begun to probe a bit around the 10MHz output (which is near the 10MHz xtal), and it looks to me the transistor Q106 is dead.



But I cannot find this reference on the net. It looks it's a Motorola, but the numbers (30659) does help so much to find same specs.  Having a couple of other same transistors on the board, I think it's a simple PNP bipolar transistor. I wonder if I can use a basic ref like a 2N3905 instead there.
Anyone have an idea where I can find some specs of this transistor?

BTW, does anyone have the schematic of the A2 (logic) board? It's not in the service manual...

David

Last edit: PNP, not NPN

[SAUL BRITTO]

Ins't 2n3905 a PNP?

[rf+tech]

douardda,

The complete HP part number is 1853-0659, one has to know the proper prefix to prepend to these mysterious part numbers. 

The HP-JEDEC cross reference lists this as a 2N3906(SEL), which indicates it is selected for a specific parameter. We have no way of knowing what the selection criteria may have been but I suspect the single digit 3 on the top line might be a clue. Seeing how this transistor sits in a schematic might provide another hint.

For now, pop in whatever 2N3906 happens to be lying about. 

RF+ Tech

[brob]

It looks like the transistor has a broken wire on the one leg. might be able to patch it back together...

[douardda]

Ins't 2n3905 a PNP?

Indeed. I've corrected my post. Thanks

[douardda]

douardda,

The complete HP part number is 1853-0659, one has to know the proper prefix to prepend to these mysterious part numbers. [emoji14]
The HP-JEDEC cross reference lists this as a 2N3906(SEL), which indicates it is selected for a specific parameter. We have no way of knowing what the selection criteria may have been but I suspect the single digit 3 on the top line might be a clue. Seeing how this transistor sits in a schematic might provide another hint.

For now, pop in whatever 2N3906 happens to be lying about. 

RF+ Tech

Thanks a lot for pointing me to the  exact reference number.

These  custom part references  are really a pain in the ...

David

[douardda]

It looks like the transistor has a broken wire on the one leg. might be able to patch it back together...

Amazingly you are right! I did not notice this (impossible to see  it by direct sight, it's obvious on the  macro picture.) I've made a quick fix  by putting a solder drop on the  wire  and my 10MHz signal is back on the  BNC socket.

Thanks

David