PC screen goes blank... overheating?

esuhl

My mate's given me his PC to look at because the screen goes blank after it's been powered up for a few minutes (if you're lucky), and sometimes nothing appears on screen at all.

It's a refurbished P4 from PC World with no case fans at all (!) so I thought it might be overheating. It wasn't that dusty, but I've given it a good vacuum just to be sure but that didn't make much difference. I tried taking the side panel off and pointing a big fan heater (with the heat turned off!) at it, and it seemed to be working okay, so I thought I'd cracked it - I just need to get him a decent case with 3 or 4 case fans...

To rule out any video-card problems I removed the add-on card and tried using the on-board graphics, but exactly the same thing happened.

The only thing is... I tried running SpeedFan to check the internal temperatures, and they all seem to be around 30 to 35 degrees (except the graphics card which is maybe 38 to 42-ish).

Is it safe to assume that the PC is over-heating? Is there anything else I can check to confirm this?

And if it is overheating, would it be a good idea to try re-seating the CPU or just get a decent case & fans?

Incidentally, I can't do anything more than log in to Vista today even with the big fan pointing at it's innards... but it is particularly hot today...

Thanks for any suggestions! I'm really puzzled by the low temperatures reported by SpeedFan...

joeypesci

Check the RAM as well? That can do funny things when it's faulty.

You could try reseating the CPU but use some decent cpu paste for the cooler fan that will sit on it. That might be old or PC World happy shopper techs haven't bothered to use any.

GeoffX

Have you tried monitoring the temperatures in the machine's BIOS? They should be fairly accurate.

ddoris

esuhl wrote: »

The only thing is... I tried running SpeedFan to check the internal temperatures, and they all seem to be around 30 to 35 degrees (except the graphics card which is maybe 38 to 42-ish).

...

You only need to really know cpu temp ... which is ????? Some progs don't measure older cpu temp - does speedfan?
Has one of the cpu cooler fixing clips come lose - they can!

Rich99

Check the temps in the BIOS too - sometime speedfan can get things wrong.

alexanderea

It its the problem, try to use cooling pad in future to avoid it.

esuhl

Thanks for the replies everyone :-)

GeoffX wrote: »

Have you tried monitoring the temperatures in the machine's BIOS? They should be fairly accurate.

Rich99 wrote: »

Check the temps in the BIOS too - sometime speedfan can get things wrong.

I had a quick look in the BIOS, and the CPU temperature started at 30C and gradually rose to 36C after ~90 seconds(during which time the system temperature went from 28C to 30C). I repeated this a few times, and it seems to be game over at 32C now (although the system temperature is about the same - surely it wouldn't shut down because the system temperature is too high... at just 30C?!

ddoris wrote: »

Has one of the cpu cooler fixing clips come lose - they can!

No - the (bog standard) heatsink/fan seem to be firmly attached...

I just found out that my friend has had the PC for a year and has been working fine till now...

I'm really confused... it *appears* to be overheating, but 32C (or even 36C) seems really low for the thermal cut-off to kick in. I've just tried running on each stick of RAM individually to rule out any memory problems but the same thing happens.

I can get some chip cleaner and thermal paste to try re-seating the heatsink/fan for about £10 or get a new case and a few fans for about £50 to £60... but my friend is pretty skint so I don't want to spend his money without being certain that it's gonna fix things...

EDIT: Actually, I've got an unopened tube of Arctic Silver thermal paste from about 4 years ago. Does anyone know if thermal paste goes off or dries out after a certain amount of time?

totalsolutions

Open it and see.
Can you set the shutdown temp higher in the BIOS

esuhl

totalsolutions wrote: »

Open it and see.
Can you set the shutdown temp higher in the BIOS

Unfortunately the BIOS is pretty basic and doesn't seem to have that option...

Lil306

esuhl wrote: »

Unfortunately the BIOS is pretty basic and doesn't seem to have that option...

What can you see ?

I'm inclined to agree, 36c is completely safe for a CPU mine reaches 50c without a sweat although it does have crappy cooling and ambient temp is like 24c

It's possible the thermal shutdown temp is active but that wouldn't cause the monitor to lose it's signal that's more the chipset.

Try taking the motherboard out the case and rest on a bench with onboard graphics, yank all components and reseat leaving only cpu, ram and heatsink, connect onboard vga and load BIOS.

Additionally, I'd do a BIOS reset on it to restore factory defaults, could be some settings have been messed on with (AGP Modes or Display modes and it's causing a conflict)

I know my screen went blank when my graphics card was damaged and hitting 80c (seized fan lol) and it was fixed with a new card, but if it's shutting down at 36c and no thermal shutoffs are activating you could be looking at a motherboard short or damaged motherboard.

esuhl

Lil306 wrote: »

What can you see ?

I can't get it to stay on long enough to have a good look now, but I did look through a few days ago. The only relevant section is "PC Health Status" which just shows the core & system temperature, CPU fan speed and PSU voltages. There's definitely no user-changable temperature setting.

Lil306 wrote: »

It's possible the thermal shutdown temp is active but that wouldn't cause the monitor to lose it's signal that's more the chipset.

Try taking the motherboard out the case and rest on a bench with onboard graphics, yank all components and reseat leaving only cpu, ram and heatsink, connect onboard vga and load BIOS.

Additionally, I'd do a BIOS reset on it to restore factory defaults, could be some settings have been messed on with (AGP Modes or Display modes and it's causing a conflict)

I know my screen went blank when my graphics card was damaged and hitting 80c (seized fan lol) and it was fixed with a new card, but if it's shutting down at 36c and no thermal shutoffs are activating you could be looking at a motherboard short or damaged motherboard.

Ahh... I did wonder whether the video signal being dropped was what normally appears to happens when the thermal cut-out triggers. So it's more likely to be a motherboard/chipset fault, you reckon...?

I'll set it up outside the case this weekend, unplug everthing non-essential, reset the BIOS and see what happens.

I really appreciate your reply - a mobo fault/short is a more logical explanation than a thermal cutout at 32C!

Cheers! :beer: