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

[Zap]

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

I was wondering the same thing, but didn't want to be rude. :sneaky:

 

[brencat]

Reading through this thread, my guess is one or both memory sticks are bad.

OP, can you please list your full system components for us? Thx

 

[clok1966]

You sound like you have a ticky one for sure. I would suggest a memtest (if you have access to another PC) make it a boot cd and run it. I would also wonder about the PSU with the random thing, but it could really be any part of your hardware. The fact that it 'sorta" seems to happen when you where installing vid card drivers makes me lean towards possilby power issues.
You might want to list your parts, what type of MB, PSU, RAM etc are you using? Even good ram can be finicky in some MB's.

 

[slag]

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.

You really need to be careful. Liquid can work its way under the cap an cause issues with the chip under the metal cover.

 

[WinnieThePujols]

Reading through this thread, my guess is one or both memory sticks are bad.

OP, can you please list your full system components for us? Thx

Yes, this is what I have:

- Processor: i3-540 Clarkdale (link)
- Motherboard: ASRock H55M (link)
- RAM: G. SKIL Ripjaws (link)

Now that you mention the memory stick thing... my motherboard said it supported DDR3 1600 RAM, but after you said that, I checked a compatibility chart on the ASRock website and this RAM is NOT listed. Am I really that stupid?! It's miraculous I've been able to build one computer. . .

I ordered a set of RAM that's listed on that compatibility chart... I'll post if I have better luck with that. -_-

 

[Zap]

Usually the "compatibility list" is just RAM that the motherboard manufacturer lists as having been tested by them and found to work. Has anyone ever seen an "incompatibility list?"

What you may want to try is with your existing RAM just lower the settings to 1333 or so and see if it improves stability.

 

[brencat]

What you may want to try is with your existing RAM just lower the settings to 1333 or so and see if it improves stability.

I don't think he was overclocking so if the Bclk base is 133, even with the system set to Auto, it would use the max memory multiplier of 10 putting the speed at DDR3-1333 with no overclocking.

OP, just in case, do what Zap suggests... manually set the main 4 timings of that RAM to its rated spec of 9-9-9-24 with 2T command rate, and manually set memory voltage to 1.50v. Make sure your Bclk x mem multiplier results in DDR3 speed of < or = 1600.

Then run memtest86 (if you can). Let's see if it really is the RAM as the problem.

 

[pm]

Right. Usually I take the compatibility as a reference but I pretty much have never bought memory that's actually on the list.

Test out the DIMMs individually with memtest86+

 

[SickBeast]

If the thermal paste actually seeped into the socket, you should probably clean it all out of there using some alcohol as others have mentioned. My guess is that it's interfering with the electrical connections. Even though I'm recommending this, you're doing it at your own risk.

At least it's only an i3 BTW.

 

[WinnieThePujols]

Well, I got off work tonight and played around with this some more. Mixed results.

The first thing I tried was removing the RAM and putting it back in, to make sure that it was seated properly. I don't think I've tried this -- I'd just been focusing on the processor / heat sink until now. No luck. Same results as before.

I then tried troubleshooting the RAM. I tried only using one stick, and had encouraging results: the system worked for 20 minutes. I decided to go forward with the installation of the VGA driver, which in the past had been extremely troublesome. For the first time since installing a VGA driver, I ran for more than 30 seconds. In fact, I had the computer running for about 20 minutes; at that point, I opened Windows Media Player and played one of the sample songs. It crashed almost immediately.

Attempt #1:
Used... “first” stick of RAM
Ran for... ~20 minutes
Crashed when... I played a song

---

I shut down the computer and tried swapping out to the other stick of RAM; the system crashed before I had a chance to do anything after the desktop loaded.

Attempt #2:
Used... “second” stick of RAM
Ran for... ~10 seconds
Crashed when.... the desktop was displayed

----

Next, I reverted back to the first stick of RAM. I again had great success, with the system working for over 30 minutes. I tried to play a YouTube video, and it crashed about a minute in.

Attempt #3:
Used... “first” stick of RAM
Ran for... ~30 minutes
Crashed when... I played YouTube

---

At this point, I was thinking my issue was maybe related to a sound issue -- possibly the driver I was using? I didn’t have any luck installing the audio driver from the ASRock website, so I proceeded to play around with the computer some more in the interim. I wanted something that was sorta-GPU intensive, so I fired up a N64 emulator and let it run (I disabled the sound from within the program before trying this). I played for about a minute -- no problems at all. After this, I was just browsing through some random folders with Windows (not Internet) Explorer. The computer shut down again. This was the first time it had crashed when I wasn’t playing audio of some sort.

Attempt #4:
Used... “first” stick of RAM
Ran for... ~25 minutes
Crashed when... I was aimlessly browsing Windows

=====

I think the results are at least somewhat promising. Although still very unstable, it's dramatically more stable than it was yesterday. It seemed that using only one stick of RAM really helped. It’s getting late, so I don’t want to run memtest64 tonight; reason being is that I had some more RAM overnighted to me, so that will be here tomorrow afternoon. I think that testing another set of RAM is just as good as running memtest, so I’ll see what kind of results I have with that.


Anyway, do any of these developments spark any thoughts? Being that the system is more stable now, is there any good logging software that I could run that could help pinpoint where the problem is coming from?

 

[PlasmaBomb]

Feel free to post pics of the CPU/mobo socket if you are still concerned.

Fire up memtest86+

If that doesn't give you anything follow pm's advice.

Originally posted by 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...

 

[PlasmaBomb]

*snip*

Anyway, do any of these developments spark any thoughts? Being that the system is more stable now, is there any good logging software that I could run that could help pinpoint where the problem is coming from?

The fact that you are getting more stability with one stick than the other suggests that one stick is possibly bad... I say possibly as it maybe that there are compatibility issues with both of them...

Good luck with the new sticks.

Still probably a good idea to memtest86+ the new sticks before anything else though, that way you know what your starting point is (that the new memory is good, and fully functional!). Again, start with one stick and test, then install windows and see how you get on... then you can start adding components/drivers in - then retest...

 

[brencat]

I don't think he was overclocking so if the Bclk base is 133, even with the system set to Auto, it would use the max memory multiplier of 10 putting the speed at DDR3-1333 with no overclocking.

OP, just in case, do what Zap suggests... manually set the main 4 timings of that RAM to its rated spec of 9-9-9-24 with 2T command rate, and manually set memory voltage to 1.50v. Make sure your Bclk x mem multiplier results in DDR3 speed of < or = 1600.

Then run memtest86 (if you can). Let's see if it really is the RAM as the problem.

OP, did you try manually setting the timings, command rate, and ram voltage like I suggested ? Using 1 stick or 2 sticks is irrelevant if they are not running at their proper spec. Lots of motherboards set the timings improperly if left on Auto since many sticks are different obviously.

 

[dbcooper1]

After you're sure you have good RAM, I'd swap out the power supply and/or boot a linux live CD and see if it runs; that would give you confidence in your hardware.

 

[blckgrffn]

hah - I am getting a kick out of how much time you are wasting playing around with Windows.

RUN MEMTEST ALREADY.

If you are unable/too lazy to get that downloaded and burned, at least run the microsoft memory diagnostic that comes with the Vista/7 installation media.

 

[WinnieThePujols]

OP, did you try manually setting the timings, command rate, and ram voltage like I suggested ? Using 1 stick or 2 sticks is irrelevant if they are not running at their proper spec. Lots of motherboards set the timings improperly if left on Auto since many sticks are different obviously.

Yes, I did try doing this. I forgot to mention that.

The settings for the RAM were set to auto; however, it displayed what the automatic settings were, and they were set exactly as you had listed.

Just to be sure, I set it to manual anyway, configured it as described. It was still extremely unstable with both sticks. After that I went back in the BIOS, changed it back to auto, and then started messing around with the stick combinations as I described earlier.


My only question with memtest -- how long do I have to let it run for? I was looking into it yesterday, and previously conversations I read said that it needs to run for hours. So realistically, I'll have to test one stick tonight to let it run over night, and then swap the stick and let the other one run before I go into work or something, right?

After you're sure you have good RAM, I'd swap out the power supply and/or boot a linux live CD and see if it runs; that would give you confidence in your hardware.

I'm confused about the Live CD -- what do you mean by that? Are you saying that would help pinpoint a compatibility/driver issue with Windows if the Live CD worked properly?

 

[blckgrffn]

Yes, I did try doing this. I forgot to mention that.

The settings for the RAM were set to auto; however, it displayed what the automatic settings were, and they were set exactly as you had listed.

Just to be sure, I set it to manual anyway, configured it as described. It was still extremely unstable with both sticks. After that I went back in the BIOS, changed it back to auto, and then started messing around with the stick combinations as I described earlier.


My only question with memtest -- how long do I have to let it run for? I was looking into it yesterday, and previously conversations I read said that it needs to run for hours. So realistically, I'll have to test one stick tonight to let it run over night, and then swap the stick and let the other one run before I go into work or something, right?




I'm confused about the Live CD -- what do you mean by that? Are you saying that would help pinpoint a compatibility/driver issue with Windows if the Live CD worked properly?

If you've got nasty bad ram (sounds like a good bet) it will probably manifest itself in the first pass, which could take up to half an hour or more. Since you are seeing unstable behavior so quickly my guess is you'll get it right away as one of the sub tests will pound the problem area.

If you don't see anything right away, go ahead and run it over night...

 

[dbcooper1]

Yes, I did try doing this. I forgot to mention that.

The settings for the RAM were set to auto; however, it displayed what the automatic settings were, and they were set exactly as you had listed.

Just to be sure, I set it to manual anyway, configured it as described. It was still extremely unstable with both sticks. After that I went back in the BIOS, changed it back to auto, and then started messing around with the stick combinations as I described earlier.


My only question with memtest -- how long do I have to let it run for? I was looking into it yesterday, and previously conversations I read said that it needs to run for hours. So realistically, I'll have to test one stick tonight to let it run over night, and then swap the stick and let the other one run before I go into work or something, right?




I'm confused about the Live CD -- what do you mean by that? Are you saying that would help pinpoint a compatibility/driver issue with Windows if the Live CD worked properly?

It's a quick test to run the hardware in a different environment and doing so from the live CD doesn't put anything on your Windows install. I'd run the memtest first and probably switch out the power supply if you have one handy.

 

[WinnieThePujols]

I don't have any blank CD-R's, so I couldn't use memtest. I had to settle for the Windows Memory Diagnostic Tool. I'm sure that thing sucks by comparison, but it's all I had to work with.

I popped the new RAM in, used the Windows tool to do two passes, new RAM checked out OK. I booted into Windows, played around for about 20 minutes with no problems. Note that prior to this, I think I've lasted all of one minute with two sticks of RAM with the original set.

I shut down and installed the old RAM again. I did two quick passes with the diagnostic tool. This RAM checked out fine too. I did just do the "basic" test -- it only took like five minutes. Anyway, I booted into Windows, and the system reset itself within five seconds.

I'm currently running an "extended" test on the old RAM to see if it finds any errors this time. In any event, it really looks like the RAM is the issue. After going back to the old stuff I immediately experienced the problem again.

I was also able to borrow a power supply from work. Regardless of the results of this test I'm running right now, I'm going to use the new set of RAM again and see if I experience the issue again. If I run into the issue again, I'll swap to this new power supply and see if that's what was causing the problem.

EDIT: I didn't find any errors on the second round of testing on the original RAM. I put the new stuff back in, fired up and.... yep, it crashed almost immediately. I think we can rule out RAM itself now. Currently have the system hooked up with the power supply from work. So far, so good...

 

[WinnieThePujols]

Well, both sticks of RAM ended up passing the memory tests from the Windows utility.

I tried swapping to the new power supply, and it ran continuously for four hours last night without a problem. I'm going to let it run all day while I'm at work to make sure that it can stay powered for a long period of time. I have it playing / decoding some high-def movies, so hopefully that'll make it work a little bit instead of just sitting there idle doing nothing.

Anyway, I think it looks like the problem was the power supply. Interesting that both the power supply that came with the original case I bought and the backup I had in my storage room for were both bad.

 

[HumblePie]

Well, both sticks of RAM ended up passing the memory tests from the Windows utility.

I tried swapping to the new power supply, and it ran continuously for four hours last night without a problem. I'm going to let it run all day while I'm at work to make sure that it can stay powered for a long period of time. I have it playing / decoding some high-def movies, so hopefully that'll make it work a little bit instead of just sitting there idle doing nothing.

Anyway, I think it looks like the problem was the power supply. Interesting that both the power supply that came with the original case I bought and the backup I had in my storage room for were both bad.

No, not interesting at all. PSU's that typically come shipped with cases are usually generic cheap pieces. I never factor in the PSU when I am deciding on what case to purchase unless I know the PSU to be a quality one. A bad PSU is what causes hardware instability 9 times out of 10 in my experience, baring user imposed damage on a component such as spilling water or something while the system is running.

Also, if you check back to my original post in this thread, I was the first to suggest checking the power supply

 

[WinnieThePujols]

No, not interesting at all. PSU's that typically come shipped with cases are usually generic cheap pieces. I never factor in the PSU when I am deciding on what case to purchase unless I know the PSU to be a quality one. A bad PSU is what causes hardware instability 9 times out of 10 in my experience, baring user imposed damage on a component such as spilling water or something while the system is running.

I was fully-expecting the one that came with my case to die pretty quickly. It was remarkable how many reviews on NewEgg spoke negatively about the bundled power supply ('d say at least half of them). In that respect, no it was not at all surprising or interesting. I just think it's sort of humorous that my replacement power supply was also bad. That's just flat-out bad luck!


On a side note, thanks to everyone who contributed suggestions or information. I appreciate the help on getting this resolved.

 

[pm]

Glad you got it working.