HP 54542x PSU question

[wje]

The PSU in my HP 54542C just failed, but only partially. Just started diagnosing, digital supply is fine, scope runs but no A/D or some other analog bits, so seems like one of the analog supplies. Given that the LV side is pretty crude, half-wave with filter caps, I'm assuming it will be relatively easy to fix, bad cap, bad diode, or both. Waiting for a 2ohm load resistor. There is some burnt smell, but I can't visually identify the fault. Caps aren't leaking.

I did get running again by swapping a psu from my 54542A, but I want to repair it.

Anyway, this brings up a question, is this PSU used in any other HP instruments of a similar era? It's a 0950-2369.

[wn1fju]

These basic power supplies were used in many HP pieces, for example 54502A scope, 53310A modulation analyzer, 165x Logic Analyzers, etc.  They may have different part numbers and this is usually due to a slightly different pinout or placement of the output connector.  But the basic Boschert supply is the same.

If you search through the eevBlog forums, you will see quite a few threads on repairing these supplies (there is a current one on the 1653B logic analyzer).  There is also a hand-drawn schematic from a French website that has been widely circulated - you'll come across it as you read the various forum entries.

The usual fault with these supplies is leaking electrolytic capacitors.  Even though you may not see a lot of visible leakage, there might still be some puddles underneath the caps that you can't see.  Most everyone ends up re-capping the whole thing (after a careful cleaning and inspection).

[valley001]

I have the 1653B in the other thread.  While you're in there you may consider replacing all the 1000uf 35v caps.  12 out of 14 were completely fubar on mine, and several of them leaked all over the board. 

[max-bit]

As long as they have the same voltage and current efficiency, you can apply.
It is worse if the power supply is also controlled by the device.

And you can't buy an original power supply, for example on ebay?

[oz2cpu]

please post a few pictures of your exact power supply type and model,
for better more correct help,
open the lit and take pictures,
you will see its design is very compact, and full of caps,
start counting and you be all dizzy and tired :-)
see also how many are directly glued to hot heatsinks with white silicone.
this is ofcourse why they fail.
see this video
https://youtu.be/c3eyTzcAnXE
start from 32:00
double check the AC input area for RIFA caps, REMOVE them all, and put in proper caps.
some powersupplies this model, do have rifa !!

[wje]

I've narrowed it down to the -12v analog supply, and I'm totally confused. I's a very simple half-wave supply with an LM350 regulator. I've pulled all the caps associated with it, I've disconnected a few other bits, etc. The raw DC is fine, verified by breaking the circuit to the LM350. So, this should be simple, no? But what's happening is that somewhere there is a low-ohms path from the -12v line to ground (ps not connected to the instrument). The output voltage is good for a couple of seconds, then the LM350 gets too hot to touch and goes into thermal limiting. I've traced everything, can't find any likely short I haven't already isolated by pulling caps.. This is driving me up the wall. Any suggestions welcome. Really seems like a cap issue, but again, I've pulled all of them.

As to just buying a replacement supply on EPray, absurdly expensive, can buy a complete scope for the price, and no reason to think such supply would be in any great condition. It was pulled for a reason, I assume.

[wje]

Ok, fixed, but still a mystery I'm not going to take the time to solve. I decided to use a delicate analysis technique, back-feed the -12v with a hefty power supply, wait to see what would catch on fire. It was, of course, a cap. But one on the +3.5v output. There should be no connection between the two supplies from the schematics I could find. Anyway, everything working again. BTW, a full set of new caps runs $45 from Mouser using Nichon, Panny, and Vishay caps. All 25 of them. Really? Who designed this thing? (Oh, 25 electrolytics, not counting 1 tantalum and 5 ceramics, or even more on the little daughter board)

Bill