Repair of Agilent E3620A(now working 100% again)

[TheSteve]

I bought an E3620A power supply off ebay which was supposed to be working but alas one output is damaged. The seller offered a full refund and to pay return shipping but what fun would that be? I thought it would be fun if I posted some pictures to see if anyone was interested in "playing along at home".

For reference the power supply has two isolated outputs that are 0-25 volts each at up to 1 amp. There is no adjustable current limit on this model.

The failure is on CH1, instead of getting 0-25 volts I get 41 to 42 volts when I turn the pot. This is what is displayed on the power supplies internal meter and I have confirmed that the same voltage is present on the output terminals.

CH2 works as intended and provides the proper 0-25 volts.

I opened the supply up and found what appears to be two damaged resistors. It kind of looks like only one may have failed from a fault of some sort the other may simply have been damaged by the extreme heat next to it.
So with that said I have attached a few pictures of the PCB.
The service manual can be downloaded from Keysight but it is easier to grab it off this site:
http://user.engineering.uiowa.edu/~expeng/equipment/HP_E3620A_manual.pdf

So if people are interested in playing along at home I'd be happy to take resistance/voltage etc measurements anywhere needed to determine if anything beyond the two resistors are damaged. It might also be interesting to discuss what might have caused such a failure.
If there is no interest I will still post my findings after it is fixed of course.

[PA0PBZ]

Looks like Q2 is shorted, creating a path from the output via A1 and CR27 to R77 & R76 to the +5 rail. That's about .5W, maybe enough to cook the resistors if you give it some time. Not sure if U19 is still feeling healthy, and check CR27 while you are there.

[TheSteve]

Looks like Q2 is shorted, creating a path from the output via A1 and CR27 to R77 & R76 to the +5 rail. That's about .5W, maybe enough to cook the resistors if you give it some time. Not sure if U19 is still feeling healthy, and check CR27 while you are there.

Hah, well that didn't take long now did it? R76 is actually still showing the correct value. R77 would cook if the voltage adjustment pot is turned down and Q2 is shorted collector to emitter. The input resistance on U19 is the same as on the good channel so it may still be alive. Q2 isn't actually a dead short but is it under 200 ohms which is still enough to do the damage.

So far this is a nice simple repair so hopefully those new to repairing stuff can follow along. It also shows how nice it is to purchase gear that is easily serviced and has published schematics, parts lists etc.

For me this is a nice distraction from the real project on the bench - fixing an HP8921A service monitor.

[TheSteve]

Received the replacement pass transistor today(ordered off ebay, came from Germany). In the end I replaced R76, R77 and Q2. I also found that it must have been dropped during shipping by Fedex as there was an LED floating around in the box. I decided to inspect further and removed the display board. The overload LED's for both outputs were snapped off. One of them was in the packaging, the other was wedged inside the front panel. Neither had any lead left on them so I used a sharp knife to scrape epoxy away to expose enough lead so they could be resoldered to the PCB. This was a bit of the pain as the display board is two boards in a sandwich. To solder the LED's I had to remove the boards which was 32 connections to desolder. It all went great though, the LED's are solidly mounted and work just fine - as does the entire power supply now. The PCB was burnt pretty bad near R76/77, all traces were intact but one was bad enough looking I ran the resistor lead directly to where the trace originated.
All in all a very simple repair that even a newbie could have handled I think. Hopefully this helps someone else down the road or gives them the confidence to have a go at repairing an Agilent supply. If anyone has any questions feel free to post or PM.

[HighVoltage]

These are some great little power supplies and usually easy to repair and well worth it. I have few of them in use all the time and just love them.

Congratulations on your successful repair.
Why did you have to order the parts in Germany and where are you located?

[TheSteve]

The transistor was cheapest with shipping from Germany. I am located in Canada and as usual shipping cost more then the part itself.

[JohnPi]

How do you get the board out of the case ? After I remove the top, and then the two screws at the side near the rear, the PCB can slide back a little and I can lift it up over the rear two keyhole clamps -- but I can't free the ones near the front. Any hints ?

Never mind -- you also remove the screws on the side of the plastic front; remove the front and that gives clearance for the PCB to be removed from the frame.

I also have 40+ volts out and have the same burned R76, R77; not sure about Q2 yet; it's not a dead short.