What happened to that trace - did the copper evaporate?

Hi

actually I'm not sure.
When inspecting the pcb I did checkfor continuity and everything was fine. To be sure, I added a wire between the solderjoints.
On the other side of the board are I beleive an inductor, -12V regulator and an potentiometer.

greetings
dura-zell

dura-zell wrote on 2024-02-24, 18:27:

actually I'm not sure.
When inspecting the pcb I did checkfor continuity and everything was fine. To be sure, I added a wire between the solderjoints.
On the other side of the board are I beleive an inductor, -12V regulator and an potentiometer.

Desolder all nearby electrolytic capacitors, this is usually a sign that one of them leaked it's fluid and corroded the trace. This would not be obvious from the top side if the leak was slow, along the lead, as this is single-sided PCB and the holes are not plated so any fluids tend to easily wick into the holes. If in doubt replece the capacitor. The secondary side of SMPS is usually where caps would age quickly and once the ERS goes up it's a runaway situation - temp. rises and the cap will either dry out or leak, or sometimes even vent. If this was 85C then think about 105C replacement. Preferably low-ESR but it's not critical (as on mobo CPU rails), any decent modern electrolytic will work fine.

EDIT: If you already replaced that one, ignore my ramblings. As for the PSU, make sure you attach some load to it, I like using 12V automotive light bulbs - you can use it for 5V as well, or get another 6V bulb. Do note a 12V bulb at 5V will draw more current since it'll be colder than it is supposed to be.
Do not connect your scope to the PSU primary side unless you have a differential probe rated at 400V at least. Yes, there are tricks, like isolating the scope ground lead, but none of that is safe - for you or the scope, or the PSU. It's just better not to do it then regret it. If the PSU doesn't start even with the load, perhaps the PWM chip is dead? If it's cheap to buy then get a replacement before investing a lot of time into circuit analysis.

Hi

I did all tests with 4 sacrificial harddrives connected to the PSU. These are rated with 0.6 amps wich exceedes the minimal load of 2 Amps a noted on the PSU itself. Thought about lightbulbs too but didn't have any around so the harddisks were the way to go.

In the direct vicintiy there is one electrolytic. Did not look bad, neither on the top nor on the bottom after desoldering it. I did not replace it as I do not have this specific one in stock, but I did measure it. According to the device it's fine. (One of these LCR-T4 things you get for a few bucks from amazon, doublechecked with my dmm.). But now when checking a bit closer, the -12V regulator doesn't look that great. Maybe check this one a bit more and order the cap nonetheless. (Any idea how to test a -12V regulator?)

As for the specs of recapping in general:
I tend to use caps which are rated for a higher rated voltage and temperature if possible. I do not change electrolytic for ceramic ones. I did not pay special attention for low-esr and had no issues so far but YMMW - I know about it, just didn't care for PSUs.

For the scope:
I do have a probe rated for 450V but this is only half of my concern. The other half is that while probing I'll short something out or connect something like 385DC to the probe without intending it. As said: its pretty cramped there.
And no - I won't do any tricks here but instead try without the scope.

The PWM chip seems to do "something" but I'm not sure if it's doing the correct thing. I think its not that expensive from what I found with google. The main issue would be to find a shop which has it in stock.

greetings,
dura-zell

dura-zell wrote on 2024-02-24, 19:36:

Any idea how to test a -12V regulator?

The only true test for any voltage regulator is to load it and measure the output (voltage, ripple). That being said it's rare for these things to go bad. I think it would be safe to remove it from PSU entirely if you suspect it's shorting the rail and tripping it into shutdown. The main rail is either 12V or 5V and that must be loaded, other ones not so much (though without nominal load the voltages might not be in spec).

dura-zell wrote on 2024-02-24, 19:36:

I do have a probe rated for 450V but this is only half of my concern.

Differential probe. Unless you do have one? The problem is shorting, yes, but not with the probe but the ground clip. It's tied to earthing via the scope body and everything on the primary side tends to float at half the mains voltage for 230V regions. Some PSUs can even be at line voltage for the primary "ground", depending how the L/N lines go into it.

dura-zell wrote on 2024-02-24, 19:36:

The PWM chip seems to do "something" but I'm not sure if it's doing the correct thing.

This is already secondary side, you can double-check that but it's typical configuration. So the chip can be probed with the scope safely, and the scope ground clip can be then attached to the ouput GND. If you worry about space inside the PSU then just solder a short wire to the point of interest and attache the probe to that. Just make sure the insulation on the wire is rated at 300V AC / 500V DC and not pinched or pricked by any sharp metal part. You can solder more then one wire but take care not to short them (including to each other).
If you ever have doubts about safety please measure the voltage (both AC and DC!) from the point you want to measure to the earthing pin/contact in the wall socket. Same with grounding point for the scope probe - this must result in 0V measurement. Then it's safe to attach the scope there. With differential probe both the signal and "ground" lead have high impendace to scope body so those can be attached anywhere as long as you don't exceeed its rated voltage.

Is that same PSU as used in the Tower AT model?

I'm asking, because I've essentially rebuilt that thing a few years ago. 😁

It took some tome, but it was worth it. Replacing the diodes by bigger ones (4007?) helped.
My MOSFET was broken, too.

Other than that, my memory is a bit foggy.

I vaguely recall that the negative voltage was being created in a very interesting way.
Essentially by using a trick circuit, hah. 🙂

The little daughter card had some "power good" circuit that remote-controlled the PSU, essentially.

Oh well, that's all that comes to mind rught now.

Good Morning

Unfortunately it is not the same. I could only find one picture on a german site (http://retro-net.de/?id=73 / Click on the pictureof the system and then "right" two times). It looks a bit similar with the daughterboard but still quite different.

The negative Voltage is created with a regular ST L7912CV and the rest of the circuit looks more or less "run of the mill" - seems that they did nothing "naughty" here. I think I'll remove the IC later today and also check a bit closr on the PWM chip. Obviously I can replace most parts "just like that" (except for the transformers) but if possible I would like to find the cause of the issue.

greetings,
dura-zell

Update: removed the Cap and -12V regulator and had no change in behavior. Waiting now for the new Cap before reassembling and testing again.

Hi

Short Update:
Replaced the Cap und -12V regulator with no difference.

Currently I'm now concentrating on the SMPS-Controller daughterboard. I'm powering the board with a regular 12V power supply. This gives me the possibility to measure with the scope.
I also dug up an application note: https://archive.org/details/bitsavers_sgsdata … ual2Ed_71700791
I'm linking it as the complete file is about 70MB. The TEA2260 is explained in great detail from page 579 upwards.

At least I can say that the internal oscillator of the TEA2260 ist running (20KHz) and all "outputs" are nominal. For some reason it still refuses to pulse the transformer. Not sure why.
Could be that it expects some kind of input that I'm not "emulating" correctly. I'm still evaluating this.

keep you guys posted...

greetings,
dura-zell