I have a PSU tester in my bag for just such events. This is $9 on Amazon; https://www.amazon.com/Sutinna-Suppl...s%2C164&sr=8-5

More money gets a fancier tester.

I rigged up a heatgun as a load. It presented a 10Ω resistance. The power supply drew 1.23 amps @ 12V, .5 @ 5.08, .31 @ -12 (which made the power supply's fan come on).

The -12 is rated at .2A. Thus I overloaded it at 10Ω. So I rigged up a slow cooker as a load; it's 120Ω. The -12 delivered .28A into a 120Ω load and turned the fan on. At 10Ω the drop was 3.34 across the external load; at 120Ω the drop was 3.59.

At 200Ω the fan didn't come on, -12 delivered .04A

What say you?

You have more experience than I. I replaced a bulging capacitor in my furnace last winter. All those in this computer look good.

That reports only voltages, doesn't load them; I do that with my meter. And I have a 'funny' PS : 1 14-pin connector, another 4-pin connector, not the standard ATX. If I had a larger soldering iron - or had bought that 12V one, I could use that as a larger load. Perhaps a toaster would do the job.

Is that 14 pin actually a typo and you meant 24 pin? That would actually be a standard ATX2.x motherboard. If the motherboard is just a generic ATX and not a specific manufacture like Dell that would be indeed funny.

The -12 VDC was/is basically used only for RS-232 and I would not necessarily dwell on just that to determine if the PSU is bad or good IMO. What you posted for the other voltages basically matches ohms law but your not really taxing them to their limits.

No. It is actually 14 pin. The CPU has separate power supply wires, 4, at 12V.

It's a Lenovo ThinkCentre M93P 10A8


It has a serial port. I don't use it. It has the chip for a parallel port even, though not the cable and jack.
Isn't it curious that using it makes the PS fan come on? I'd have thought it would always be on or thermostatically controlled.

I noticed. It shows that it works at least that well. The question is how PSes fail: so badly they can't even deliver 15 watts, or they can't deliver 100. The PS testers that people have recommended work if they deliver enough current to light an LED.

Yes. The +12 provides 12A to the motherboard; 16A to the CPU; the 5.08 2.5A to whatever uses that.
I'd need a 1Ω 150 watt resistor to test it - or an equivalent load. My heat gun and slow cooker were the highest loads I have around the house. If I had an inverter to turn 12DC into 120AC - that'd do the job. I have a vague memory of having owned one. I wonder where it would be.

My pickup's headlights! They draw 65 watts in high beam @ 12V! I'll get on it tomorrow.

Beware that the bulbs initially have much less resistance than Ohm's Law would predict. 65W/12V - 5.4166A or just under. 12V/5.4166 = 2.215 ohms. But initial current might be 5x that, or up to 10x in freezing conditions. Once the bulbs go yellow the resistance drops. Maybe connect them one at a time?

I planned to connect only 1, figuring that if it worked at 5A that's probably good enough. I asked on a car forum, found a guy who uses old headlights to test power supplies he salvages from discarded computers.

I decided it wasn't worth the trouble, bought a barebones used.

The power supply is good.