[Resolved] MISERABLE build experience. Did I destroy my CPU?

[WinnieThePujols]

I was putting together my seventh or so computer today. I liberally applied some thermal paste, installed Windows.... every time I would install VGA drivers (I'm using integrated graphics and an i3, by the way), my computer would crash. It would just reboot. I looked online, and determined that my processor was probably getting too hot.

I take off the heatsink, and the thermal paste is EVERYWHERE. Trickling down the sides of the processor. I pop the processor off of the motherboard, and a stream of it had dripped alongside the BOTTOM of the processor! Wow, no wonder why my computer was crashing.

Obviously that thermal paste had to be removed, so I gently dabbed the bottom of the processor with a paper towel. I'm guessing that was a stupid idea, but I was furious / frustrated at the time and acted in haste. After doing so, parts of the bottom of the processor were then slightly discolored.

The computer still boots up, but it still crashes like before. Now I maybe get an extra few seconds out of it before it restarts.

I'm obviously a complete idiot. I'm assuming this is probably completely hopeless and that I essentially just flushed $125 down the toilet? Or...?

 

[pm]

I would think that it's unlikely that your CPU is destroyed. You could probably put a CPU through the wash and it would be ok (but not the dryer - the static would nuke it).

You probably already know this, but you want the absolute thinnest possible layer of thermal paste. So little that it's almost like it's not there.

Thermal paste is normally not electrical conductive - in fact, I don't know of any that are. Usually your best bet for removing it is high % isopropal alcohol (rubbing alcohol). I usually go with the 90% stuff.

It probably also goes without saying but be very careful removing it around the pins of the socket. They bend ridiculously easily and it's very hard to fix them when they are messed up - I know this from personal experience. It can be done but it takes steady hands, patience and either extremely good eyesight or a microscope. So, be super, super careful around the socket pins, but make sure that they aren't covered with paste to the point that they can't make good contact with the CPU.

The problem might have started off as overheating (although I personally doubt it) but the result is likely that the paste is causing something to not have contact with the CPU very well. I don't have great advice except to clean it up as best you can, and to say that as a CPU designer for Intel, I would bet you $5 that your CPU is fine and it's your socket not making good contact with the CPU or drivers, or memory, or power delivery or any other thing that can go wrong that doesn't have anything to do with the CPU. Not that CPU's are perfect or awesome, but they are monolithic things and they tend to either work, or not.

I'd also test out the system using something like Ubuntu that's less flaky than Windows with regards to drivers (I love Windows but when you get crashes, you never be sure if it's drivers or hardware). Fire up Ubuntu on a USB and boot up in that, if you crash or have problems, it's easier to troubleshoot because it's almost certainly hardware - even if you've never used Ubuntu before, it takes Windows drivers out of the equation and it's pretty user friendly. Also fire up Memtest86+ (or your favorite memory checker) and check the memory - memory timings and SPD can be totally unrelated to the CPU thermal paste thing and cause instability that seems heat related. Oh, and see if you can borrow a different power supply - flakiness in booting can be linked to power delivery issues.

 

[AdamK47]

I doubt you did any damage to the CPU by rubbing a paper towel along the back of it. What kind of temps are you getting?

 

[WinnieThePujols]

I doubt you did any damage to the CPU by rubbing a paper towel along the back of it. What kind of temps are you getting?

Well, what do you mean by the "back" of the processor? I'm talking about the bottom of the processor -- there was thermal paste on the pins on the bottom of it. I had to rub a few of them to get it off.

Also, I can't even really look at temps. It crashes like immediately going into Windows.

 

[bigpimpatl]

I have no experience with i3, but perhaps there is a conflicting BIOS setting in regards to video?

 

[notposting]

Try cleaning with 99% IPA.

edit: wheeeee I love tabs, bit late to this party

 

[pm]

You can usually check the temps in the BIOS - they usually have a hardware monitoring tab thing. Check there. I still don't think it's temperature related though.

In the lab we occassionally get temporary technicians that don't know what they are doing who douse the CPU's in AS5 (it's been a while but it happens). They'll cover a CPU like it's BBQ sauce or something. I've never seen damage to the CPU from this sort of stuff. Not that this couldn't be a first time, but I doubt it. If your thermal paste is electrically conductive then maybe... but that would be weird. About the only thing to be super careful of is static.

 

[WinnieThePujols]

You can usually check the temps in the BIOS - they usually have a hardware monitoring tab thing. Check there. I still don't think it's temperature related though.

One of the motherboard's CPU pins had trace amounts of paste near the top; I gently wiped this off and was able to get it off. There is also a tiny bit on the motherboard itself, near the "stem" of a couple of pins. I think the stuff I wiped off could possibly interfere with operations, but probably not the stuff at the bottom. I'm gonna boot up and see if it has any affect...

 

[Drsignguy]

Just recently, I had the same thing happen ( other than the thermal paste thing) But the Heat sink wasnt seated on the cpu correctly and it shut down immediately. I knew, once I fired it up that was the case so I removed it and reseated it and all is just peachy!

Also, you don't need to put a liberal amount of paste on the cpu, A very small amount would be sufficient. Rub it completely over the cpu using plastic wrap on your finger and all will be good. The reason for the plastic wrap can be many reasons but personally it makes cleaning up much easier. There are other reasons why too but no need to explain.

 

[WinnieThePujols]

OK, looking in the BIOS, the CPU temperature is running at about 33 degrees celsius.

 

[Zap]

I'm with the other guys in that I seriously doubt the CPU was damaged. I've seen both a Pentium 4 and a Phenom survive being installed rotated wrong in the socket. After straightening the couple of bent pins, both CPUs worked fine. The P4 motherboard died from the experience but the AM2 motherboard survived.

Even if you can't get into Windows, you can check temperatures in BIOS.

WinnieThePujols, if you have difficulty judging how much thermal compound to use, I would suggest one of two things to try in the future. First thing would be to use a heatsink that has thermal compound pre-applied. This includes the stock Intel heatsink. The second thing would be to use a thicker thermal compound, such as Shin Etsu. It tends to stay in a thicker layer so it won't squeeze out and drip as much. Any excess usually just stays as a lump right where it squeezes out.

 

[WinnieThePujols]

Here's an update on the situation:

After playing around with the computer for a while, I was really convinced that my issues were actually on the software side; in particular, I thought it was a driver issue. Here's why: when I would initially install Windows, everything was completely stable. I could surf the 'net, install drivers -- do anything I wanted, and it would be fine. It seemed like I didn't have problems until after I installed a graphics driver (which required a restart). After installing that, I could log on to Windows and last 30 seconds if I were lucky. I figured it was a faulty driver or something.

However, I once again wiped Windows clean, and this time didn't install any drivers at all. I hypothesized that the computer would run all night. It lasted about a half hour, and then restart -- exactly the same as the problem I was having before.

I then went into the BIOS to look for settings to play around with; while in here, I also checked the temperature of my CPU, which was one again in the 30's (celsius). Interestingly, I had a full-on lockup in the BIOS as well. Now I'm back to thinking that this is definitely a hardware issue.

I'm not really sure what to think anymore. It doesn't seem like an overheating issue... it does seem like I cleaned everything off really well too. I'm getting really perplexed here...

 

[Dadofamunky]

Not an immediate hardware issue. If you're getting into the BIOS and such the CPU is fine and the mobo socket is fine. Which mobo is it? Is it a cheap Taiwanese offbrand? It's possible the BIOS is old and incompatible with something that you're doing. It all sounds like a compatibility issue to me.

 

[pm]

Well, if it hangs/reboots in BIOS that makes life less confusing - you can be sure it's hardware which removes one big variable (windows and drivers).
Remove any and all cards, except video (which sounds like it's integrated).
Clear the CMOS on the motherboard with the jumper.
Strip down to one DIMM, and run memtest86+. If that fails swap memory and try again. If you can, test the memory in another machine with memtest86+.
Try swapping power-supplies - unless you are sure the one you have is good.
Flash the BIOS.
And my last idea when I'm all out of ideas is that I remove the motherboard from the case and power it on on a piece of non-conducting cardboard - every once in a while I mess up and screw down the board too hard or something...

and if not any of these, I'm out of ideas. At that point, I try for an RMA of whatever I think is responsible - usually the motherboard.

 

[VirtualLarry]

I recently purchased a mobo from a forum member here, and when I recieved it, I noticed that it had a tiny, tiny, bit of what appeared to be AS5, lodged into two of the mobo socket pins. I put in a CPU anyway, and some RAM and a video card and tested it. Other than the mobo didn't like my unbuffered ECC memory initially, it worked fine (well, as far as booting into the BIOS for a little while).

So I don't think getting AS5 onto the mobo pins will necessarily kill the CPU or the mobo. In my case, it seemed unaffected. Though I didn't install Windows and test extensively.

 

[Campy]

Since the pc simply reboots and it's not heat related i'd say its most likely the motherboard. It could possibly be the cpu or perhaps even RAM but it's unlikely

 

[Vdubchaos]

Why would you put thermal grease on a new CPU? They usually come pre greased....

 

[degibson]

Well, if it hangs/reboots in BIOS that makes life less confusing - you can be sure it's hardware which removes one big variable (windows and drivers).
Remove any and all cards, except video (which sounds like it's integrated).
Clear the CMOS on the motherboard with the jumper.
Strip down to one DIMM, and run memtest86+. If that fails swap memory and try again. If you can, test the memory in another machine with memtest86+.
Try swapping power-supplies - unless you are sure the one you have is good.
Flash the BIOS.
And my last idea when I'm all out of ideas is that I remove the motherboard from the case and power it on on a piece of non-conducting cardboard - every once in a while I mess up and screw down the board too hard or something...

and if not any of these, I'm out of ideas. At that point, I try for an RMA of whatever I think is responsible - usually the motherboard.

This.
Check the RAM.
Try all the DIMM/slot combinations (could be faulty DIMM, faulty slot, or both).

 

[Numenorean]

Why would you put thermal grease on a new CPU? They usually come pre greased....

Uhhh, no they don't.

 

[Axon]

You can (and I have) cleaned the bottom of a CPU and even the CPU socket with a soft bristle tooth brush and some rubbing alcohol. The CPU is a lot more durable; it's the socket you really need a gentle hand with, because pending a pin can spell disaster. As long as you're careful and gentle, you'll be fine. Once you clean the paste off, just let both air dry and you'll be good to go.

 

[HumblePie]

I've dropped a CPU in a cup full of alcohol and then pulled it out to let towel off and let air dry and it was fine afterward. You could do the same thing with water, so long as there was no electricity around it and it was given sufficient time to dry thoroughly before use.

You can do the same thing with the motherboard and it would be fine as well. Just don't bend those pins. Make sure to use a magnifying glass to check if the pins are out of alignment or bent slightly. Use a needle or safety pin to bend pins back into place if need be if they aren't really far bent out of whack.

Also, there may be a plethora of reasons why you are having issues. It could be a bad harddrive, bad memory, bad BIOS settings, or a bad power supply as well. I suggest checking all those for starters.

 

[WinnieThePujols]

OK, thanks for the recommendations guys, I'm going to take a look into some of those suggestions.


For what it's worth, let me elaborate a little bit on the stability issues. I reinstalled Windows something like four times, thinking that this was a driver issue.

Attempt #1: hardware was fine during installation (no power failure). Windows booted up perfectly. Ran for probably 20 minutes (installed drivers, etc) without any problems. After restarting (having installed 4-5 drivers), it would immediately crash upon signing on to Windows. I restarted 3-4 times and it would crash within 15 seconds of showing the desktop.

I figured this was a driver issue, so I reinstalled Windows (Windows 7 only takes like 20 minutes to install so it wasn't that big of a hassle). I couldn't get into safe mode to roll back either.

Best-guess: driver issue.

---

Attempt #2: hardware was again fine during installation. Windows booted up perfectly. Ran Windows for a good 15 minutes; this time, I only installed the video card driver. I decided I'd install drivers one-by-one to see if there was a faulty one I was working with. I installed the driver, restarted, and was again greeted by instantaneous crashing upon going to the desktop.

At this point, I figured it was either the CPU overheating, or that the video card driver was bad. This time, I took the computer apart and noticed the thermal paste fiasco. I was convinced cleaning it off would solve my problem.

Best-guess: overheating or driver issue.

---

Attempt #3: hardware was, again, fine during installation. Windows booted up perfectly. Played around on Windows for a few minutes without a hitch; installed the video card driver, restarted, and BOOM... it once again crashed right as I logged in and was shown the desktop. I also checked my temperature in the BIOS -- the CPU was running at a normal temperature.

I was now certain I was dealing with a bad video card driver. I reinstalled Windows (it would also crash going into safe mode, by the way) one last time...

Best-guess: driver issue
Ruled out: unlikely to be overheating

---

Attempt #4: no hardware failure during Windows installation. Windows booted up fine. I decided to see how long I could let the computer run with stock drivers. I was on it for about a half hour, at which point I received another unexpected restart. During this I attempt I didn't install any drivers at all, so that obviously rules out a driver issue.

I then went into the BIOS, was cycling through some menus, and everything locked up. It didn't reboot like it usually did, it just outright locked up.

Best-guess: [the suggestions given in recent posts -- RAM, hard drive, BIOS, motherboard, etc]
Ruled out: unlikely to be drivers, unlikely to be overheating

---

Could it be merely a coincidence that the crashing seems to be more frequent after installing the VGA driver? In any event, there is obviously something beyond just a driver issue since it locked up in the BIOS too.

 

[Seero]

Note that video processing is done on CPU before the proper driver is installed, and IGP does nothing but forwarding data towards the display. Afterwards, the IGP starts to work up, and in your case, failed. Assuming you have not play with bios settings, reset bios settings to factory default and see if it helps.

Get another PSU and see if it helps.

Disable IGP and use a discrete video card and see if it works. If it does, then the problem is either on the IGP, meaning bad motherboard, or RAM. Download memtest and see if it is RAM related.

 

[Insomniator]

Why would you put thermal grease on a new CPU? They usually come pre greased....

Where the hell do you get your cpu's?

Also how can you build ~7 computers and not know to use a very small amount of thermal paste?

 

[nyker96]

just a few short advices:
1. don't use too much TIM, a tiny bead is all your need. anymore you actually separate the heatsink from the cpu.
2. run something like prime95 to see if your cpu temp is a problem, if it is then you need to recheck your cpu installation.
3. Is your power supply big enough to handle your rig?