Vintage 130W IBM5160 Schrack Elektronik EG07168-E PSU Repair

[hozone]

Hello,

I'm trying to repair an old PSU of an IBM 5160.
It's a Schrack Elektronik EG07168-E, 130W supply.

The problem is that on the 12V and 5V line I've low voltage output.

Of course I can not find any schematic for this or similar models.
I'm not an expert in switching supply, expecitally old design one. So I'm here asking your help.

Those kind of supply needs some load to work, but even with load I read low voltage.
With PSU disconnected I can read
+9.7V almost (on the +12V line)
+4.7V almost (on the +5V line)
With all connected (and PC powered on - note: there's no problem on motherboard and devices cause even with an external "working" load, like 2 or 3 HDD same low voltage behaviour)
+10.5V almost (on the +12V line)
+4.4V almost (on the +5V line)
With just one "modern" PATA HDD connected
+10.5V almost (on the +12V line)
+4.4V almost (on the +5V line)

What I've done so far:
Resoldered all the joint.
Replaced the 3 main switching transistors (BUT11A) they are working but I have 3 new one so I replace them: ok.
Checked output capacitors: ok. (at present I've connected spare one)
Checked the others "big" capacitor: ok.
Checked the two "big" diode: ok.
Checked the only 7815 rugulator: 15.10V out. Injected 15V here, request is 150mA, and makes no difference if is 15V bang on down to 10V.

I can't find a PWM switching regulator chip, just some comparators UPC339G and one HEF4013P (for protection circuit).
BUT I've notice that If i warm up (with a hot air soldering station) the secondary PCB mounted, the only with some SMD component, main voltage output goes down. It has something that seems Carbon Printed Resistors. Seems to be this board to control the output.

Tracing the board is somehow difficult cause it goes into "unknown" or custom model coils.

I don't know if it's something related to the feedback flayback regulator or other. Do the comparators works as flyback regulator for the switching?

I attach here some picture of the PSU I've found on web. Mine is like this. My pictures between those one are the one of the SMD board.

Any suggestion appreciated.

[Tim@psutest.com]

Check out this link:

http://asknotes.com/2021/09/16/ibm-5150-psu-repair/

[fmashockie]

You really need to provide more info to get more help on this issue.  You have to make an effort to trace out this circuit a bit.  Especially if it is not clear to you from looking at it how it works.  For example, you have two quad op-amps here.  One of those could be used as a PWM for switching.  From the photos you shared, the board is one sided, which means it should not be that difficult to trace some of it out. 

If you are saying the 7815 is outputing 15V, then I would suspect your issue downstream of that.  Check all those electrolytic caps on the secondary side first. 

[hozone]

Thanks,

@Tim@psutest.com
the asknotes.com post unluckily does not help me, it's a different damage.

@fmashockie
All caps tested. No problem at all.
At visual inspection nothing seems wrong.
I have the hope it was something "simpler", as I said tracing is a bit of a mess. Anyway, find attached the main output trace, and also what I think is the trace of the right zone (the one with MC3403) of the SMD board. With picture. This is the zone that warmed up change the voltage output. Feedback from +5 and +12 seems to go in the middle C339G.

[xavier60]

The MC3403 being opamps likely for error amp, the blue diode could be the reference Zener.
The 339s can be used for various functions like, oscillator, PWM comparator, steering logic, rail monitor. The 4013 would be the flip-flop for the steering logic.
Faulty Zeners are usually temperature sensitive.

[xavier60]

With that large, gapped inductor, could this be a resonant converter?
The output inductor doesn't seem large enough to suit a half-bridge PWM type converter.
Id like to see the coil drawn between the main transformer and diodes. Transductor maybe?

[fmashockie]

Nice job tracing out the circuit a bit!

I've definitely seen this kind of issue due to an intermittent/bad zener diode part of the feedback circuit for other SMPS I've worked on. 

That being said, if the feedback circuit was the issue, I'm having trouble seeing how that wouldn't affect the 15V rail.  OP says the 15V is measuring fine correct? 

[hozone]

@fmashockie
The only diode i can see in the SMD additional board is the blue one... and maybe the other thing on top right? Never seen a component like that.
On the main PCB there are a lot of diodes.

The 15V rail comes from a 7815, and it's fine at 15.1V almost. I've also removed the 7815 and inject external 15V from my bench supply. All works fine (well with lower output error but works). Down to 10V, then it stop working.

@xavier60
I'm not an expert in SMPS, expecially old one design. Both the diodes for the 12V and 5V rail comes from the large yellow transformer. Then on both the lines there are two smaller inductor, I think for filtering signal.
 
I think the 15V rail supply the power to this board (and the other component next to this board... R,C,D one UPC339C and one HEF4013). My suspicion goes to this board cause if I warm this up, voltage goes down. But maybe that's something "normal".
Maybe Ive to trace the feedback too, but this is starting to be an hell

Do you think measuring the PWM on the base pin of BUT11A could help?

[xavier60]

Looking at the PWM might tell us something. It will be safest to probe it at one of the transformer outputs. You could also check the 5V and 12V outputs for excessive ripple.
That BUT11A at the corner of the PCB near the large black reservoir capacitor. If it has a direct connection to the large, gapped inductor, means the PSU has a PFC stage which would be unusual for such an old PSU.
Can you direct heat directly at the blue diode on the small board while to see what effect it has on the output?
Or heat it directly while off with a soldering iron, then switch it on while the diode is still warm.

[xavier60]

If it does have a PFC stage, a fault in the PFC could result in low output voltages.

[hozone]

Finally I had time to make a couple of tests. I'm so sorry for my delay.
I'm using a portable DSO150 oscilloscope to prevent my main scope from damage, don't want to risk with high voltage.
I've tested both three BUT11A. The two next to the "big" transformer seems to have a 21.5k base square both. If i warm the "smd" circuit the freq rise up to 21.7k slowly and the voltage goes does. But the 21.5k and square shape makes me think there's nothing wrong there.
Then I've tested the first BUT11A, the one on the high voltage side. It has an "ECG" like shape. See attached picture (probe at 10x) connected from Base to Emitter.

[xavier60]

Show the waveform at one of the output stud diodes. If it's full duty cycle, it could mean that the PFC stage isn't working properly.

[aeg]

I don't know the situation in whatever market this was made for, but in the US, PC power supplies certainly did not have PFC in 1983.

[xavier60]

I don't know the situation in whatever market this was made for, but in the US, PC power supplies certainly did not have PFC in 1983.

There needs to be a purpose for the large, gapped inductor on the right and the 3rd BUT11A. Buck regulator maybe.

[xavier60]

A Buck regulator feeding the half-bridge stage wouldn't need an output capacitor. The power path needs to be traced out.

[p.larner]

what scope is that?,it looks handy for low freq stuff.

[hozone]

Thanks all. Now I'm going to office, this evening I'll do the tests.
DSO150 has 200kHz bandwith, not that much, but with high voltage and possibly mains connection my main bench scope may risk.

[hozone]

Attached the scope output from one diode, both looks the same. Probe at 10x. Ive take the signal from the output of the diode (anode) and the catode, but I think I've made a mistake, which should be the ground for this measurement?
Also the circuit traced from output to the big yellow transformer. The two big diodes (the one I've take measurement) are the one on top. I think there's a sensing wires on the inductors you find in picture.
It's difficult to trace it out more due to the lack of my knowldege on this task, and due to the fact that I read 0.15 to 0.20 ohm on trace there, but that may be winding of the transformer.

[xavier60]

That output stage is unusual with the 5V being choke filtered and the 12V being capacitor filtered. The reason could be to achieve independent 5V and 12V regulation. You could check the waveforms at the 5V and 12V diode cathodes for now with respect to ground.
I would like to know how the large, gapped inductor is being used just to get a better idea of how the PSU operates.

[hozone]

Thanks!
I'll check waveforms tomorrom morning before office.
As for inductors, do you mean number 1, 2, 3, 4 or other?
3 and 4 are the output filtering inductors.
1 and 2 are a bit difficult to trace, It seems to me (looking at the traces) they are "sensing" output and input like in the image... but without disconnecting wires it's difficult, cause it's a mess of sub ohm values reading in that zone.

[xavier60]

This one. Also, what is your mains voltage?

[hozone]

Attached the two wave at the output of the diodes, not much ripples and voltage almost the same read from the multimeter.
Forget to mention that there are a couple of trace, from +12 and +5 that goes in the SMD circuit, ?probably for feedback?

The 300V input positive goes in the first BUT11A (the one with the "ECG" form), then goes to this inductor you are asking me, then to the yellow inductor.

[xavier60]

The single BUT11A and the large inductor is a Buck regulator feeding a push-pull stage consisting of the BUT11A pair and the yellow transformer with center tapped primary. The blue capacitor near the large inductor which I thought was the coupling cap for a half-bridge stage now appears to be a snubber for the BUT11A pair.
 Except maybe for the messy Buck drive waveform, I think that there are no faults in the power paths.
You could scope the high voltage waveforms for curiosity if can be done safely.

That Zener diode on the SMD board is now suspect.

[hozone]

Thanks!

You could scope the high voltage waveforms for curiosity if can be done safely.

Where do you want me to test it?

That Zener diode on the SMD board is now suspect.

You mean the blue one (number 2 in image)? number 1 I don't even know what It can be.

Back from office I'll test.

[xavier60]

Id like to see the waveform at the Emitter of the single BUT11A and the Collector of one of the BUT11A pair.
Both with respect to live side ground, the negative of the large black 385V capacitor.
This is the live side of the PSU, referenced to the mains supply. What safety precautions with you be taking?

The blue diode in the photos in post #3. You could try to read its part number or carefully measure the voltage across it.
I sometimes tack in some wires from an old IDE cable and run them to a DMM. Avoids probe slip accidents.

Reference Zeners are almost always 5.6V.