It’s not my primary PC, just a spare but would like to fix if possible.

It’s a Dell Optiplex 990 so about 8 years old-ish.

The problem is that it won’t stay powered on so I suspected a faulty PSU and that has been replaced but the fault persists.

I’ve managed to test the memory and that seems fine. I’ve removed any peripherals but that hasn’t helped.

A bit of googling suggests the motherboard – is this likely? There’s not much else it could be now!

Any advice appreciated!

Thanks in advance,
GQ

How long does it stay on for? Long enough to inspect log files to see if any error messages logged?

Sounds like you’ve already taken it back to bare bones (unplugged hard drive, optical drive etc).
If you’ve got more than one memory dimm in there, take one out. Power it up and go into the BIOS and leave it there to see if you get the same fault. Try different slots for the memory / swap the dimms over. If the problem persists then it sounds like you’re correct and it’s the system board.
I assume the on board diags haven’t thrown up anything helpful?

Was the PSU brand new or second hand? If not brand new it may have the same issue as your original.

Not much help but it really sounds like you’ve done most of the troubleshooting already!

Are all the fans working? Are they showing as working?

If it’s recent I’ll give very good odds it’s temperature related.

I’ve had similar with a machine where the cpu fan wouldn’t register properly and it shutdown to prevent over heating (fan was fine but an old one with no speed sensor).

I’ve also had this when faffing with clock speed and the like, again over heating related but that tends to be very obvious as you’re doing something you expect might go wrong at the time

I had a PC that would power off like that and it needed a new motherboard.

I agree with dangeourbrain, probably temperature related.
Try running it with the case off or open if possible. If it still does it could be the processor fan/heatsink could try reseating with some fresh thermal paste.

Temp related can also be due to a dry/bad solder joint.

There’s a good chance that it uses a proprietary format case and motherboard so you can’t just swap in a new MB. If it does use a standard format motherboard, swapping that for a secondhand one would probably be pretty cheap. Obviously need one compatible with your CPU and memory.

Thanks for the replies all.

It stays powered on a varying amount of time – few seconds to a few minutes. I’ve tried removing the memory (4 dimms in total) and have tried one by one.

The PSU was brand new albeit not an original Dell. Fans all look okay, BIOS reports the correct fan speed and I’ve cleared out the dust from the heatsink (suspected temp problem also).

The diagnostic lights on the front report either possible motherboard or CPU fault.

Cheers,
GQ

Friend of mine had similar problem.
Turned out the fans were clogged with dust and the CPU was overheating.
He blew out the dust using with compressed air at work and it was fine after that.

Use something like HWmonitor to check the actual temperatures..

It’s very rare for a CPU to fail but it will cause shutdowns if it gets too hot. Maybe the thermal paste between the CPU and the heat sink has gone bad which won’t be obvious other than it shutting down shortly after booting up depending on how fast it hits it’s thermal limit.

If your temperatures check out then my best guess is a motherboard fault.

I had an issue where I upgraded my nans duel core to a quad core CPU and re used the stock fan, the CPU could easily hit over 90c and would shut down just from installing an application.

Problem is keeping it powered up long enough to run HWmonitor!

I think I do have some thermal paste so maybe the next step is to try that.

Cheers all 🙂

Might be worth going into the Bios as scc99 mentioned and see if there’s any temperature reporting in there, but with it being a dell rather than a super turbo gaming motherboard, the Bios might be very limited in terms of what you can check.

But it would also be a good test as there will basically be zero load on the system so if it still crashes after a few mins/an hour of just being in BIOS that could indicate a board failure.

Also take the side of the case off if possible and finger the CPU heatsink.. If your CPU is over heating the fins on the heatsink will be very warm maybe almost too hot to touch for more than a second or two.

If it stays at just above room temperature I.e. Just slightly warm you could maybe rule out a CPU overheating issue without going to the faf of replacing the thermal paste.

Even getting into the BIOS is tricky – very hit and miss how long it will stay powered on! I can’t see anything in the logs other than lots of “side panel has been removed”.

The heatsink doesn’t seem to be particularly warm.

Motherboard then? There are a few on eBay so might be worth a punt.

Sorry, I wrote this 3 hours ago and failed to hit Post.

If it stays at just above room temperature I.e. Just slightly warm you could maybe rule out a CPU overheating issue without going to the faf of replacing the thermal paste.

A cool HSF assembly could mean that the heat sink isn’t sinking heat.

Thermal shutdown was my first thought. Typically you’ll see a pattern though – it’ll run for the longest time from cold, them once it shuts down it’ll fail again much faster until it’s had time to cool off. Also can be load-related, dues it occur when the PC is working hard?

First thing I’d do is take a can of compressed air to all the fans, fins and vents and then replace the thermal paste on the CPU/HSF. It’s a cheap fix so it owes you nowt if it doesn’t work. Pro tip, get a bottle of TIM cleaner to shift the old stuff. It cuts through dried-solid thermal paste like nothing else, leaving it lemony fresh.

I had a pc while at uni trying to write my dissertation that kept doing this. Someone suggested overheating so I bought a cheap desktop fan, opened the case and fired the fan into the opening when the pc was on. Once my dissertation was written it went back for a warranty repair.

This would be an easy way to diagnose an overheating issue…

bitcoin mining virus?

Thermal shutdown was my first thought. Typically you’ll see a pattern though – it’ll run for the longest time from cold, them once it shuts down it’ll fail again much faster until it’s had time to cool off. Also can be load-related, dues it occur when the PC is working hard?

First thing I’d do is take a can of compressed air to all the fans, fins and vents and then replace the thermal paste on the CPU/HSF. It’s a cheap fix so it owes you nowt if it doesn’t work. Pro tip, get a bottle of TIM cleaner to shift the old stuff. It cuts through dried-solid thermal paste like nothing else, leaving it lemony fresh.

It tends to run longest when cold (but we’re not talking long here – maybe a few mins tops) – and then does get shorter down to a few seconds. And can be zero load i.e. just sat in the BIOS screen.

I’ve cleaned all the crap out – I think the next step is to re-apply some heat sink and see if that helps then.

Cheers.

Yeah, if the heatsink isn’t properly taking away heat then even being in BIOS will overheat the cpu despite no load.

Mine just started cutting out in this warm weather. I know it is temperature related as I have an external temp reader on the front panel and when it gets to 32 degrees it shuts down. I have 6 fans with full speed control and have had to run them flat out to maintain 29-30 degrees. But if I start to do anything with graphics like editing pics or, god forbid, play a game, it is instant shut down. My place is very dusty (old stone walled barn conversion) so I have got the air line on it in the garage and cleaned it right out. I also removed the CPU heat sink and replaced the thermal paste. It now runs 6 degrees cooler with the fans turned down to a more acceptable speed (hence volume).