I got home last night and my computer was off. I tried to turn it on, but to no avail. I think the power supply is burned out (cause it happened a couple of years ago) but I am not sure. This might be a dumb question, but there aren't any fuses in computers are there?

This might be a dumb question, but there aren't any fuses in computers are there?

Yes, there are fuses. In your power supply, there's likely a large and easily visible one.

There's often a fuse in the PSU, although it's usually a pain to get to, and potentially dangerous if your not electronically savvy.

I'd question the merit of doing this anyway - unless the fuse blows when you are there because of some stupid mistake, it's probably not the fuse. From your post I infer that you leave your PC on a lot (or 24x7). Eventually the PSU fan is going to fail which will cause heat build up, and can lead to the or other part of the PC (especially if the PSU fan was the only case cooling) failing.
But I'd guess the PSU has failed.

But I'd guess the PSU has failed.

Eh? You mean you could diagnose that from his posting?

I am obviously not worthy, Oh Great One!!!

It's a guess. Last time he saw his PC it was on and working. Now it is off and not working.

Horowitz and Hill postulated (*1) that the most unreliable components in an electronic system would be (in order):
1. Cables and connectors.
2. Switches.
3. Potentiometers and trimmers.

I think that the empeg has borne out their postulation rather well, assuming that we class the rotary encoder in the same class as a pot or trimmer, which I believe is reasonable.

But I also think that H&H didn't really go far enough in that chapter. Admittedly at the times that AofE was written and updated, the technology was transitioning from a mix of wire-wrap and PTH pcbs to a mix of PTH and SMT pcbs, and the resulting power densities of the era were far lower than those seen today, but they did actually discuss cooling and the use of fans just prior to that postulation. They didn't go the extra step though - they didn't say anything about _electromechanical_ reliability, eg fans, (or generators and other motors). Ironically they did state that, "in any cooling situation it is a good idea to design conservatively; failure rates for electronic components rise dramatically when equipment is operated hot." (*2)

If you asked Slashdot readers (!) what failed in a routine manner in their PCs most often, I am confident that you would get 50%+ of the answers involving electromechanical devices; Hard Disks, CD/CDROM/DVD drives and fans, and another 30% saying PSU (without knowing why).
That would also sit with Enterprise Server product lines - why do Enterprise servers have hot-plug disks and hot-plug redundant PSUs? Why not hot-plug network cards?
It would also match my experiences pretty well too - my linux server has run 24x7 for over 3.5 years now with only the occasional reboot. In that time the CPU fan has been replaced twice IIRC (maybe 3 times..), and the PSU has been replaced once. It has also had 2 new drives (RAID1, constantly logging so never spun down), but the old drives hadn't failed - I simply ran out of space and it was as cheap & easy to replace them completely than to augment them (which would have required more IDE channels). The only other death in the family has been a mouse, very recently after 40+ months of dedicated service.

I don't believe that the CPU is to blame (other prime heat candidate). Most PCs will crash long before the CPU dies even if the fan dies, and a crashed CPU doing nothing doesn't usually consume or dissipate much energy. Most modern CPUs have thermal protection built in anyway. (This is assuming of course that the heatsink has not become dislodged somehow which could possibly cause the temperature gradient to exceed the protection's ability to deal with it.)

I'm going to discount the PCs power switch as a cause - ATX only uses a momentary switch to signal a power state change, and if it failed it would not shut the machine down. (I suppose that if a lightening strike caused a power outage then a failed switch would prevent the PC being turned on again, but most people would mention that they suspected a power outage may have occured. Besides, I just don't see it being the most likely cause.)

Furthermore, he stated that he had tried to turn it on "to no avail". I would guess from this and the topic title that the machine is _completely_ dead, ie no LEDs, no fans etc as most people would state if it were otherwise. That would suggest that the PSU is not turning on at all.
ATX1.1 specs dictate that the PSU turns on when signal PS_ON is pulled to TTL low, ie 0V, and turns off when PS_ON is pulled high or allowed to float. (Pulled up in PSU). The circuitry required to do this isn't exactly rocket science, even allowing for adding WOL, WOM, Keyboard Wake functionality. Since this signal basically carries little current, this circuitry shouldn't really be prone to overheating either.

I suppose that H&H would suggest checking cables and connectors, ie power lead (it didn't get unplugged by the cat by chance?), ATX motherboard connector and the ATX power switch connector. But connectors are far more likely to fail when they are prone to connect/disconnect cycling, or noticeable physical vibrations. I don't see that being the case here.

Finally I suppose that the PSU could be locked in a shutdown mode due to a short circuit on one of the power lines (+12V, -12V, +5V, +3.3V or +5VSB). But again, this isn't the kind of failure I'd expect to see hit suddenly. This shutdown mode can only be cleared by cycling AC power - I'd expect to see it kick in again as soon as the power switch is pressed, probably with a gentle click or whir from the PSU.

But yes, it is just a guess.

*1 Art of Electronics, 2nd Ed, Chapter 12, Pg858.
*2 Also Pg858.

I am obviously not worthy, Oh Great One!!!

I save those kind of accolades for people who do crazy stuff like designing a radical in-car entertainment device 5 years before all the sheep do, or putting binocular remote vision on an RC car. But thanks anyway.

Yeah it was the PSU. This is the 2nd one now. It probably keeps frying because the computer is in a cubby hole in a computer desk and it gets too hot.

He needs a new computer anyway it's only a 1 gig athalon

no i don't need a new computer...just a small upgrade....

It probably keeps frying because the computer is in a cubby hole in a computer desk and it gets too hot.

It might be that. On the other hand, it might be bad power. Either way, I'd be wary of this. I've had PSUs that have failed on, rather than off, meaning that they dump a metric shitload of juice through the mainboard and fry everything. At that point, you've got no choice but to upgrade.

metric shitload

Oh, you silly Europeans, with your strange units of measurement.

strange units of measurement

I believe that they keep the international standard metric shitload in an outside lavatory in a vault under Paris, and get it out once a year to clean it.

This is where the phrase "polishing turds" comes from.

Where do they keep the chocolate teapot? Or is that just another name for the same standard?

They keep that in a refrigerated display case. Note that they don't actually use it for anything, not even as a measurement standard; a chocolate teapot is completely useless. That's why you only hear people say, "It's about as much use as a chocolate teapot.", and not, "It's less useful than a chocolate teapot." The second phrase (aside from being impossible) implies that that someone has somehow measured the usefulness of a chocolate teapot and obtained quantative results, which is, again, impossible.

(I knew that. I was trying to be funny. Serves me right, I guess.)

(Well, I knew it apart from the refrigerated case. )

I prefer "chocolate fireguard" myself...

in an outside lavatory in a vault under Paris

Surrounded by the 64 lasers of the awesome GEMMD*

*Le Grandiose Ensemble de Meurtriers de Mouche Domestique

Jim, you're weird.

The words "pot", "kettle" and "black" spring to mind

For the record, I've used "metric shitload" here on this side of the pond, too.

I prefer the ``metric assload'' myself. Or is that ``assloadde''?

Jim, you're weird.

Why, thanks.

You are all excused. (I except the complement.)