momaka wrote on 2025-07-18, 07:35:

But this has nothing to do with the industrial PSUs I'm talking about. Most of them come with 5 screw terminals: AC input (L & N), ground, and output (+) and (-). The ground terminal is highly RECOMMENDED, but OPTIONAL. It is isolated from the PSU output terminals, but you could reference one of these to it, if desired.

The B2B sector (industry) implies the presence of skilled employees, due to which it consciously goes for reducing electrical safety and cost.
And then once a year the housewife gets an electric death in the shower because she bought a charger without galvanic isolation on AliExpress.
It's better not to do that.

OK, I desoldered what I think is the 12V diodes (MUR1510CT, in TO-220 package), and both diodes in the package test fine on my Fluke multimeter -- OL one way, and 400-something mV of drop the other way. So no luck there.

Also, I did replace all the electrolytic caps -- the Jackcon, etc. were what was available through Jameco, and while they may not last as long moving forward, they all tested OK on my LCR meter before I installed them.

Any ideas what to try next?

Thanks!

Also, I did double-check the 115V/230V switch, and it is in the correct position (115V for United States) and shows proper continuity -- I worked it back and forth a few times just to make sure.

momaka wrote on 2025-07-18, 04:17:

I've something similar too, except instead of light bulb sockets, I have 2 wall outlets (and a switch) wired in series. One outlet is connecting my DUT (device under test), while the other is for connecting my "current limiting" device (e.g. desk lamp with incandescent bulb or floor lamp with a high-power halogen bulb... or when heavier currents are expected/allowed, I use a 2 kW space heater or 700W/1.4 kW mini oven.) Yes, such large "current limiting" appliances can indeed take too much space on the bench (and I don't keep them there for that reason)... but sometimes, there's just no other alternative to such high power "resistors" - namely when testing ATX PSUs with APFC circuits, as those APFC circuit absolutely do NOT like low-power current limiting devices (e.g. 100W or lower light bulbs.)

I also have such a "series plug" device - but almost never use it. The problem with big heaters etc.
is that they take longer to get hot - so higher that wanted current occurs longer, and more often than
not it's long enough to cause damage.

What I do use more often, is a setup with 1A, 2A, 3A and 5A circuit breakers.... (I only use one at a time, otherwise
trip characteristics aren't terrible predictable).

I've also got one with Bakelite screw terminals where I can fasten a piece of wire to be a "homemade fuse".. I spent
a few days testing(*) various "thin" wire types to be able to create some very specific current limiters. Also not one I
use a lot, but very handy when I need it.

(*) Fairly easy to test current trip with a 5A ammeter, some high power resistors and a variac.

biggieshellz wrote on 2025-07-18, 22:15:

Any ideas what to try next?

Thanks!

with minimal PSU load ( waste HDD, which has not yet been disassembled into magnets)
- c5/6 each must be charged to 110V AC (for US) *1.4 = 150Volt DC
- C9 - measure for capacity and resistance. If it is bad, then this capacitor will limit the transmitted power.
- R/C/D around power BJT - incorrect driving of BJT.
- Power BJT - Starting from the fact that one of them may be broken and ending with the loss of gain. It is better to check by replacing them with similar ones.
- driving circuit- may disrupt self-oscillations during startup.

^ Agreed with all of the advice given.

Although in regards to the BJTs failing with loss of HFE - IME that is unlikely. If HFE falls too low, the BJTs won't be able to fully turn on, and thus overheat and blow up (either short-circuit or open-circuit.) When that happens, one or both may either short-circuit or open-circuit (I've seen both... and all I can say is, crappy BJTs tend to open more often than go short-circuit.) The fact that the PSU can regulate without a load suggests that both BJTs are OK. Otherwise, the PSU won't be able to turn ON at all (asymmetric driving resulting in magnetized transformer core and not possible to transfer energy anymore.) So since the PSU voltages seem to fail only with a load, then the issue is more likely somewhere on the voltage feedback circuit. While I may have said that PWM ICs don't fail often, I think in rare cases it can happen... and this might be one of those cases - well, it might be either the PWM IC or the LM339 quad comparator... or small resistors and diodes around them.

That said, BEFORE you try any of the above further troubleshooting, I STILL think you should try the PSU with a PROPER load on the 5V rail - i.e. 20-25W 12V auto bulb. When troubleshooting PSUs, I always like to do as much testing first BEFORE I start removing components from the PSU. Otherwise, you could end up chasing a ghost problem, and that's just a waste of time. I can't tell you how many times I've seen this issue before, where someone on a forum tried to repair some PC PSU only to find out the real "problem" was that he/she was trying to test the PSU without a proper load and thus getting bogus results.

So again, test the PSU with a proper load first and then proceed to dig in further.
And by the way, could you also take note of the -5V and -12V voltages when the PSU is tested with and without a load?
Looking at the picture of your PSU, I can see the common-mode output toroid has two different windings made from thin wire, suggesting one is for -5V and the other for the -12V rail. This is important to note, because it means the -5V and -12V are generated separately from each other - same way as the 5V and 12V rails (on some PSUs, -5V is generated from an LM7905 regulator, that's why I'm noting this.) Thus, if the 5V rail is going low and loosing regulation, you should see the -5V rail voltage also go similarly low. Likewise with the 12V and -12V rails - they should follow each other to within a few volts.

Other components I would suggest to check are:

1) the 1 uF 250V polypropylene cap on the primary side. These -rarely- fail (I don't actually recall ever seeing one do), though it is theoretically possible. If this cap looses capacitance, it will limit how much power can be transferred on the primary side to the main transformer, and that itself can directly affect the output voltages.

2) bad component in the OPP (over-power protection) circuit. The circuit is very crude on these old half-bridge AT/ATX PSUs. It's where you see that small toroid with the red wire going through it in the middle of the PSU. The large red wire going through the inductor generates a current pulse every time a BJT on the primary side turns ON. This pulse is relayed to the secondary side via the small red windings on the toroid. This is how the PWM controller can "tell" if an OPP condition is present and start to limit PWM values. The inductor itself on this circuit is extremely -unlikely- to go bad. However, the components that follow afterwards on the secondary side that relay the pulse to the PWM controller could go bad. This would correspond to D13, R31, R32, R33, R34, C15, and R19 in the schematic posted on the previous page. If any of those components is not in spec, it can cause the PWM controller to mis-interpret an OPP condition.

3) bad/poor solder joints on the common-mode output inductor. Had one like that. PSU would sometimes regulate OK and sometimes not and shut down. This would be an extremely unlikely failure for your PSU, but still worth mentioning, just to make sure everything gets checked.

4) the NTC thermistor on the input going high-resistance. This one is even more unlikely to be the problem... but I've had a case of that too (on an ATX PSU, but that's besides the point.) It would allow enough current for the PSU's 5VSB to run, but not for the main PS to turn ON... basically letting in a little bit of power only. But that one became a bit more obvious, because I could see some heat marks around the NTC, since its dissipation was very high when the PSU was trying to run.

On that note, you can check #4 above by measuring the voltage on the primary caps as was suggested shevalier while the PSU is running (JUST BE VERY CAREFUL, as that's the PRIMARY SIDE and has relatively HIGH DC VOLTAGES.) Generally speaking, the voltage on these caps should be close to 160V DC... but of course may fall slightly lower under load (so 150V DC is not impossible either.) If the voltage goes below 140-145V DC, though, that would suggest voltage is getting dropped somewhere along the way on the primary. But I don't see any burn or overheat marks, so I think this and #4 above are unlikely to be the case... and overall why I suggest to test the PSU with a proper load on the 5V rail once more.