Gaming I probably just fried my CPU

[RupeeClock]

Alrighty then, my CPU is probably fried now, after everything that's happened and me not knowing better.

Last night I basically backed up all my important data, took all day but it was done.

So first thing this morning, before booting my computer for the day I thought I'd try and push my CPU a little farther with an overclock.
It's a Core 2 Duo E8600 @ 3.33ghz, it was set to 3.6ghz with 1.5v, and I upped it slightly to 3.75ghz with 1.625v
Something went terribly wrong there, probably the voltage tweak, because after that the display wouldn't come on. Windows was definitely booting as normal, but no signal.

I had to open up my PC guts to reset the CMOS, which also involved taking out the GPU for a moment to access the jumper pins.
Whilst I was in there I probably made a huge mistake by removing the CPU cooler fan to inspect the thermal paste. It looked a bit thin, but I had no way to remove the old paste in order to apply new paste, so I just put the cooling fan back on.

When I booted the system after that, the signal returned and I was able to get into the BIOS again.
Windows 7 tried to boot the first time, and failed spectacularly. It froze on the booting screen with only a red spec of the windows logo having animated.

When I got into the BIOS, it became pretty clear I just messed up big time.
The CPU was reading 2.00ghz, and wouldn't adjust along with the FSB.
Then I examined the H/W health and was shocked to see the CPU temp had reached 105c.

I immediately powered off the system when I saw that, and now I'm afraid to even go in there again for threat of causing further damage.

So long story short, one thing lead to another, now my CPU may be burnt out.
I have no idea if the system is salvageable at this point, I may need to take it to some nearby computer specialists and have them sort things out.
Worst comes to worst, I'll need to buy a new computer. I had been wanting to buy a new computer in the near future anyway, but not by losing my current system. It may be nearly four years old, but I was hoping to give it to my brother who's using a 7-8 year old system.
For the time being I'm stuck on my laptop.

Anyhow, do you guys have any idea about the state of my system?
From what I've described, is there any hope for the CPU? Could it be the 2ghz reading was only because of the critical heating?
The system didn't shut off or make any usual noises, so I don't know.

 

[Sicklyboy]

105c? Holy shit, man o.o


Your pc will still boot, at least, right? You can get into BIOS? You wouldn't be able to if the processor was completely fried, so you have some bit of hope there. Restore it to factory defaults in there, then see what happens. If it still reads at 2.0ghz, try to set it to it's default clock speed and everything.

 

[RupeeClock]

Yes, thankfully it still boots. But right now there's no way I'm going to try re-booting it without taking it to someone or reapplying the thermal paste correctly.
I don't want to dig a deeper hole at this point.

That's the part about the 2ghz I don't understand, by resetting the CMOS I'm pretty sure everything is back to factory defaults.
The CPU should read 3.33ghz, as is the speed of a E8600. The 105c was probably screwing that up.

 

[Originality]

As mentioned, since you got into BIOS, the CPU is NOT fried. The speed can be easily explained - after the temperature goes above a certain threshold (90C on mine, 100C on my brother's), it will throttle the CPU. This means drastically reducing the clock speed (and voltage) in order to try and get the temperature back down and prevent the CPU from frying (funnily enough).

So, chances are the thermal paste is ruined (not completely covering the CPU, probably full of bubbles, and definitely not transferring heat like it's supposed to). I'd say that's 90% likely the cause. Now, go online and buy yourself a cleaning agent for removing the thermal paste (either citrus based or alcohol based) and a new thermal paste (either ceramic based or silver based) and find a guide to removing and re-applying thermal paste. I'm not the most knowledgable of thermal pastes, but if you do a lil research online you'll learn all you need.

Quick and dirty guide to re-applying thermal paste: Apply the solvent to the CPU and back of the HSF, let it soak for a couple minutes, then wipe it off with a lintless cloth (the type used for glasses, although a kitchen cloth will do). You should be able to read the small print on the CPU and see a shiny surface on the HSF. If not, repeat as necessary. Next, apply a small blob of thermal paste on the CPU and spread it with a Q-tip card. Ideally you'll want it to have the thickness and consistency of nail varnish (i.e. very thin but even). When you put the HSF on, it'll get squished out anyway but it's OK so long as you didn't put too much on.

Remember to secure the HSF from opposite corners (top left then bottom right, etc) to keep even pressure on the thermal paste/CPU. This reduces the chance of bubbles forming (slightly).

 

[RupeeClock]

Thanks Originality, you've been a big help.
There's a nearby computer specialists shop called Syn-star, they will probably have the cleaning solution I need, I already have thermal paste.
I'll have to visit them tomorrow now since it's late Sunday.

 

[Zetta_x]

Go to the nearest health store and pick up some isopropyl alcohol. Either that or buy 99% isopropyl alcohol online.

 

[p1ngpong]

Thanks Originality, you've been a big help.
There's a nearby computer specialists shop called Syn-star, they will probably have the cleaning solution I need, I already have thermal paste.
I'll have to visit them tomorrow now since it's late Sunday.

You really don't need any specialist cleaning fluids to remove thermal paste. I've removed paste a hundred time from various machines and only ever use lint free (disposable) cloth to do it. Once you have taken off the majority of the paste from the cpu and the heatsink just use a clean lint free cloth to buff them until they looked polished, job done.

And yeah it sounds like your cpu simply throttled itself to avoid being damaged, 100c is usually the threshold for throttling. Its never a good idea to remove the heatsink to "check" on the thermal paste, only ever remove it if you intend to replace the paste immediately.

 

[RupeeClock]

Well, I don't have any lint free cloth anyway so I'll still need to go out tomorrow.
Can't do anything now, it's too late in the day, and we're about to have barbecue.

 

[meornot0]

This may seen like a stupid question, but did you remember to plug the cpu fan power back in?

 

[RupeeClock]

This may seen like a stupid question, but did you remember to plug the cpu fan power back in?

I didn't unplug it to begin with, I only unlocked the four screws to look at the CPU, and locked it back on.
The problem was the thermal paste being voided by that little action, I hopefully just need to re-apply it.
Can't do that till tomorrow.

 

[Originality]

If the HSF isn't plugged in, it would tell you in POST and it wouldn't go any further.

 

[RupeeClock]

Okay, an update.
I got around and mustered the courage to reapply the thermal compound, after some struggling to clamp in the CPU fan, the system is booting normally now.
Temperature is nominal, safe at 39c and the CPU is A-okay, 3.33ghz as expected.

But I have a problem with the BIOS now, after configuring it there's a chance that the signal stops going out entirely.
I have to keep resetting the CMOS because I can't see anything on screen.

I think I get what the problem is though, for whatever reason the BIOS configuration, if I just look at a certain option and save it, begins to ignore that I have a graphics card plugged in.
The DVI cable is connected to the GPU obviously, not the embedded VGA cable.

Does anyone have any input on this? I'd really appreciate this, because I don't want to struggle with the possibility that I'll have to be resetting my CMOS all the damn time. I have to remove the GPU just to reach the jumper pins.

 

[moose3]

Do you by chance have a VGA cable you can use to hook up to the onboard graphics to the monitor and switch the default settings to use the pci card from within the graphics control panel from within windows, not just BIOS?

I had a computer with onboard graphics that even with the BIOS set to recognize pci over onboard graphics, it would still default to onboard. I had the onboard connected to the monitor, then as I mentioned, switched the pci card set to the primary connection. Followed by switching the connector over to the pci card in the 15s as it waits and asks if you want to keep these settings before resetting. You would just need to hit the input button to change from vga to dvi on the monitor. I also disabled the onboard graphics in windows just for good measure.

Good luck.

 

[RupeeClock]

That's what I'm going to have to do I reckon, I just can't be bothered to right now.
I've been beginning to appreciate using my laptop just for beating the summer heat, using my deskop is like having two radiators in my bedroom.

 

[RupeeClock]

Here's an update.
I hooked up the VGA cable, and like I suspected, BIOS was sending video out the integrated graphics.
It was easily resolved once I could see what I was doing, and made sure that the video was being sent from my GPU, had to set it to PCI-E instead.

PC health seems pretty much fine now, no overheating, and only the usual sluggishness that follows windows not shutting down properly.
Thanks for helping me out guys.

 

[alex_0706]

well it is imposseble to fry your cpu with overclocking
its just an fable

that says computertotaal (ducht pc magazine)

greetings alex_0706

 

[Originality]

It's not impossible, it's just that motherboards tend to have a series of failsafes incase you push it too far or something isn't right.

 

[RupeeClock]

well it is imposseble to fry your cpu with overclocking
its just an fable

that says computertotaal (ducht pc magazine)

greetings alex_0706

That wasn't the problem.
I realise what actually happened though, my display wasn't getting a signal, not because I had tweaked the CPU clock/voltage, but because I had never properly configured the video output (defaulted to internal)

After that, I opened the guts to reset CMOS, when I only needed to grab a VGA cable.
Whilst doing so, I stupidly took the fan off the CPU without reapplying thermal paste, and in hindsight I may not have even properly reclamped the HSF.
THAT'S how my CPU hit 105c

Anyhow it's all fine now, thanks for your concern. I might even try overclocking it again, see how 3.75ghz fares, maybe with the voltage tweak after all.

 

[alex_0706]

if you want do overclocking do it with watercooling (much better)

1 thing it could be eexpensif

 

[FireGrey]

Interesting story..
Good that everything turned out ok.
And, Holy Crap 105 Celsius

So are you going to have another shot at over-clocking?