I kind of need to test my pc power outlet while the air conditioning is running on the same circuit to see if that's maybe been causing some of my problems, but I probably need a nortamericano to explain how to do that

Never mind all that. What you really have to beware of is finding a fault with your kit, replacing some parts and then being left with some perfectly good bits and pieces. If that happens, a little voice starts saying 'you could build a server with that. Go on, you know you want to' Look, you've got the CPU and memory - and that old case in the attic - all you need is a cheap motherboard and...' and before you know it you're £lots ($lots) out of pocket and somehow you have more bits left over than you started with.

As for the air conditioning thing: I imagine a multimeter with its probes stuck in a molex would show any spike or drop out associated with a switching event. You'd have to sit and watch it as the timer or thermostat or whatever is tripped.
 

And I've got the dusty old boxes of kit stacked around my midden to prove it!

At least I tossed a couple of old pc's a few months ago. Pretty sure I've still got a fucking huge old 8-bay server tower case stored in the back room. I'm actually using MrsD.'s old sff pc as a development server, also driving a scanner from an xp virtualbox on it.

Ok. Paperclip and multimeter on the go. It seems to only be giving 3.3V out of the 12V line. That's broke then is it? Only thing that makes me doubt it is that maybe it has some clever "keep volts low if there is no load" thing going on? It's quite a fancy PSU (BeQuiet P8 1200W that I got second hand and had running for a couple of years now) so I wouldn't be that surprised. It's just the rather precise 3.3 that gives me that thought. The 5V line is still 5V.
I also tried pulling the RAM and the graphics card out, that didn't help. Still no life, and none of the board's diagnostic LEDs are coming up. Reset the CMOS too, no help.

The psu sends a power_good signal to the motherboard once voltages are all in range, without that signal the motherboard won't boot up. If you really are getting only 3.3v on the 12v line then that would be an issue and prevent the pc from booting.

Are you *definitely * checking the 12v pin? What about the disk drive power connectors, is it giving 3.3v there too? As you say, 3.3v is a very specificly wrong figure to get.

My PSU is fully modular apart from the motherboard cable. So that's in there with the paperclip, and then I've plugged in a HDD cable to get the molex power socket. It's got yellow, 2x black, and red cables. With black from the multimeter into black on the molex, and red into red I get 5v as (I think) I should. Red into yellow I get 3.3v. There's no other cables (other than the one going to the wall!) connected at all.

Is any of that me being a doofus? I am prepared to accept the possibility as I aren't very experienced with electrickery.

Oh hello. Different cable (same type) into different socket on the PSU and it reads 12.5V. Hffffffff.

I read a few things yesterday saying that some PSUs do need a load, might be worth plugging in an optical drive and testing some more. Check what's coming out of the ATX connector since that's what is actually powering the motherboard.

Yeah, ATX is showing 3/5/12 as it should. Well, 12.5. Maybe one duff HDD molex cable was enough to cause this, plugging the other into the same bit works as well for the case fan. Weird. Time to put it all back together.

And... Still nothing ffs. Going to just smash it all with a hammer. Think Truffy had something about that in his advice.

Have you got an old/another PC tucked away somewhere, as in try powering it with your BeQuiet PSU or try powering your current PC from the other PSU? The BeQuiet does look as though it has all sorts of fancy dohickeys and magic electric juju inside.

Do you have a hard drive (not solid state) you can use as a load i.e. plug in directly? I think these need 12V to spin up which would show whether it's supplying enough.

Edit: OK, I see you've moved on beyond this. Serves me right for taking a break to watch the telly before clicking on post.

Ha. Yes. I am a bit bereft of spare things (small flat plus baby means clear outs were enforced) but I'm back to being flummoxed now anyway.

Rrrgh. So I started again thusly: PSU connected to nothing, ATX shorted with paperclip. Powers on, volts read as they should. Plugged in the 'other' motherboard power (that fires up the CPU I think) and powered up, volts read as they should. Plugged in PCIe for the graphics, volts all good. Plugged in AIO CPU cooler power, volts all good. Plugged in hard drives power, NOTHING. Hello. Went back a step, nothing. Mixed and matched a few steps, nothing. Went back to just PSU With ATX cabLe shorted, nothing.

*flings hands in the air*

*goes to bed*

When it does nothing, the PSU doesn't power up?

Sounding like a PSU on its way out.

Broken and/or stuck power switch? You can bypass the switch by stabbing wildly at the motherboard shorting the two pins that the power lead from the switch is plugged onto, with a screwdriver. If the PC turns on, the switch is faulty.

The relevant wire will be a very fine twisted pair and probably have PWR written on the plastic plug bit in itty bitty writing. It may also be silk-screened on the mobo near the pins.

I've had bad hdds totally nix a pc posting (not stopped power on though). 

Nah, forget the pins thing, I didn't spot that you left the paper clip in. Although shorting the pins will work.

So all the power stuff worked, fans on etc. until you plugged a hard drive in?

Yes... But once I unplugged it again (and I'd had the same HDD on moments earlier no problem) it still wouldn't work. So unless it outright killed the PSU, I dunno.

My mobo has a power button on the board itself so I'm reasonably satisfied the case isn't a problem.

This morning, I tried PSU with paperclip plugged into nothing and it switched on at correct voltages again. Then I went to work. 

When I found myself with a similar dilemma a few weeks ago, I was happy that the motherboard was faulty as the multimeter readings for the PSU were fine and it powered a hard drive, fans etc perfectly well when not connected to the board. It was also a reasonably expensive item even if not quite up there with yours.

However, I decided that I couldn't be sure, and as I hadn't done any PC building for ages I bought a new motherboard/CPU/memory and a new PSU. The old CPU was an early i5 and as compatible second hand motherboards are stupidly expensive, I couldn't reuse it.

I was then left with the old PSU which I used to build a second PC/Server using an old Athlon 64 and motherboard from the I-can't-throw-that-away pile. I learned a lot about the fibs told by Nvidia and several mobo manufacturers about how much memory the early Nforce chipset/bios combo could support (up to 4GB but only if you could find memory modules that nobody actually made) and had fun playing with it until it started making crackly noises and finally the PSU burst into flame with a loud bang.

So my present theory is that the PSU may well have killed the original motherboard by serving up generous portions of voltage at inappropriate moments.

Of course, I then decided to buy an ultra cheap PSU and discovered that I could make use of my old i5 and memory with a new motherboard from Gearbest who manufacture a whole range of boards based on old chipsets for a far more reasonable price than second hand ones fetch on ebay. Expensive things computers.