Hardware failure

[JustinAWR]

When I set up my computer initially, it was running fine for a long while (probably 3-4 hours), and I had run performance tests on my GPU to test framerates, as well as a test to see how hot my system would get while it was under maximum load (100% usage on all 4 cores of my cpu for 70 minutes - it got to 57 degrees max). It was when I was attempting to watch a video on firefox and update my Windows Live account through a videogame simultaneously that I got my first blue screen of death. After that I got the BSOD, or a system freeze or some sort of crash every time I loaded up Windows, though the amount of time and BSOD error aren't necessarily consistent. I have since run Memtest on my RAM and it has passed all of the tests given it. My system temperatures seem fine. I have tried to repair the problem through Windows but it never succeeds, and system restore doesn't do anything either. I have tried to boot from the Windows CD in case it was a hard drive issue, but once Windows loads up it always gets the BSOD at the same point before I even am able to attempt to repair or reformat or anything. My system has never crashed in BIOS though.

I cannot figure out what might be the problem. I initially though hard drive, but the fact that the system crashed when booting from the Windows CD makes that somewhat illogical. I thought maybe the motherboard, but since BIOS has never crashed that doesn't seem like the problem. The RAM tests fine. The Power Supply provides more than enough power (600W), though it could still be the issue. The GPU worked fine for the benchmark tests, but may have malfunctioned since then. The CPU might be at fault, even though it worked without issue when it was running at 100% on all four cores for 70 minutes. I really am at a loss for which component might be the issue. Do you have any ideas on what the culprit(s) might be?

 

[aford10]

First, pull the CMOS battery for ~5 minutes.

Then try running with 1 RAM chip at a time.

What are your hardware specs?

 

[Doom3klr]

test your system with hot cpu tester pro for 6 hours and make sure you have plenty of airflow neat cables inside and plenty of case fans as well as no internal dust before testing.

 

[cpatel1987]

When I set up my computer initially, it was running fine for a long while (probably 3-4 hours), and I had run performance tests on my GPU to test framerates, as well as a test to see how hot my system would get while it was under maximum load (100% usage on all 4 cores of my cpu for 70 minutes - it got to 57 degrees max). It was when I was attempting to watch a video on firefox and update my Windows Live account through a videogame simultaneously that I got my first blue screen of death. After that I got the BSOD, or a system freeze or some sort of crash every time I loaded up Windows, though the amount of time and BSOD error aren't necessarily consistent. I have since run Memtest on my RAM and it has passed all of the tests given it. My system temperatures seem fine. I have tried to repair the problem through Windows but it never succeeds, and system restore doesn't do anything either. I have tried to boot from the Windows CD in case it was a hard drive issue, but once Windows loads up it always gets the BSOD at the same point before I even am able to attempt to repair or reformat or anything. My system has never crashed in BIOS though.

I cannot figure out what might be the problem. I initially though hard drive, but the fact that the system crashed when booting from the Windows CD makes that somewhat illogical. I thought maybe the motherboard, but since BIOS has never crashed that doesn't seem like the problem. The RAM tests fine. The Power Supply provides more than enough power (600W), though it could still be the issue. The GPU worked fine for the benchmark tests, but may have malfunctioned since then. The CPU might be at fault, even though it worked without issue when it was running at 100% on all four cores for 70 minutes. I really am at a loss for which component might be the issue. Do you have any ideas on what the culprit(s) might be?

If your mobo has an integrated card, take out your graphics card and use the integrated card and see what happens. If it works fine, you most likely made a defect of the graphic card worse by running a stress test on it. Although not a bad thing, at least you know your card's bad now. If you return it though, don't mention you did a stress test they might not give you your money back depending on who you bought it from they might consider that user damage.

Also, hit f8 to get the "disable automatic restart on system failure" option. That way, you can read the BSOD and post it here to see if it says something relevant.

I use to work as a representative for computer repair techs. One big rule about diagnostic software is its not 100% guarantee it will find the problem, good old fashion "take this out and see what happens" is more definite.

Definitely post your BSOD error if the card is not the issue for further help.

 

[JustinAWR]

Ok, I tried your battery removal method and left in only one stick of RAM but the system isn't even starting up any more. This is not a new issue btw, it has happened a couple of times where it doesn't want to start up.
I cannot run any program for 6 hours, as I cannot really get into Windows for more than 5 minutes.
I have had 3 different BSOD errors throughout this process.

What is new is that I just checked and according to the RAMs website, it is not compatible with my motherboard. I bought them as a combo through New Egg, so I just assumed they would be. My RAM is G.Skill F3-10600CL9D-4GBNT and my motherboard is an AMD 870A-UD3. It is strange that New Egg would have combo'd that if it was incompatible, though they do claim they do not assure compatibility with combo deals. If this is such an issue, would it make sense? AKA, if they are indeed incompatible, would I have been able to run my system for so long and do so many things before it finally crashed?

 

[aford10]

The RAM (DDR3 1333) is compatible with the motherboard (DDR3 1333). If it wasn't, it wouldn't have booted before.

When you say 'it's not starting up.' Are there any signs of life? Fans? Beeps?

 

[JustinAWR]

Well that is unfortunate - I was hoping that might have been the issue.

The PC Speaker beeps 3 long beeps, pauses, then repeats - basically endlessly. The fans and lights and all are working, but nothing other seems to happen, aka the monitor shows nothing coming from the computer.

EDIT: as far as my components. I have a Gigabyte HD 4850 1GB GPU, a WD 500Gb hdd, athlon 2 x4 2.9ghz cpu, OCZ 600W psu, and the ram and mobo listed above. also a dvd drive and wireless card.

 

[aford10]

3 long beeps doesn't match any error codes in your manual.

Did you try each stick of RAM?

Try removing and reseating the video card and the RAM.

 

[JustinAWR]

I have looked around for what the beeps could mean but as far as I can tell they don't mean anything so who knows.

I have tried removing the RAM and running it one at a time and such to no avail. I have reseated the video card many times as well trying to figure this out.

I am thinking about just packing up most of my computer and sending it back for replacements. I am considering sending back the motherboard, cpu, gpu, psu, hdd, and RAM for replacements. Essentially I have no clue what the issue is but I am assuming it is with one of those components. Can I safely remove any components from that list as being the issue?

 

[aford10]

I doubt it's the hard drive. That's not needed to POST.

CPUs don't die real easy, so it's not likely, but not impossible. Make sure the CPU/HSF feels secure, and isn't loose.

Have you tried breadboarding? If not, remove the motherboard and place it on a piece of cardboard next to the case. Connect the CPU/HSF, PSU, 1 RAM chip, video card, and power switch. Everything else can be disconnected.

Same beep code?

If yes, disconnect the PCIe power to the video card.

Same beep code?

If yes, my $$ is that it's your video card.

 

[Doom3klr]

Make sure cpu was well seated on the board with the heatsink and proper thermal paste. To get a screen to come up again reset the cmos there should be a jumper on the motherboard. Check pdf documentation for your motherboard manufacturer online. Make sure all conectors are pluged in well. Sometimes ram will give all sorts of issues. Like shutdown errors blue screens etc. Check ram and reseat. Good luck

 

[cpatel1987]

I doubt it's the hard drive. That's not needed to POST.

CPUs don't die real easy, so it's not likely, but not impossible. Make sure the CPU/HSF feels secure, and isn't loose.

Have you tried breadboarding? If not, remove the motherboard and place it on a piece of cardboard next to the case. Connect the CPU/HSF, PSU, 1 RAM chip, video card, and power switch. Everything else can be disconnected.

Same beep code?

If yes, disconnect the PCIe power to the video card.

Same beep code?

If yes, my $$ is that it's your video card.

I second that, I didn't know of another way to test it without an integrated video card, but that works just the same to determine if its the video card.

 

[JustinAWR]

Ok, thanks for the help guys! I took your advice and looked more into the RAM. I ran Memtest again and got millions of errors so I sent the RAM back to new egg. They did not have any more of that RAM in stock so I had to order more. I ordered the same brand (G.Skill) but different RAM - I chose one of the customer favorites this time. I got the new RAM in today and ran Memtest on it again - I was going to run it for 7 passes like they suggested. I DID NOT BOOT THE COMPUTER. I went straight to Memtest without doing anything else just to make sure the new RAM was ok. I went four passes without errors. On the fifth pass I got two errors. I am not on the sixth pass as I am writing this and have a million errors so far. My question is this:
Is this bad luck having my RAM fail again, or could it be something else causing this to happen, aka a bad motherboard or something? Thanks for your help so far!

 

[aford10]

It sounds like some bad luck.

What RAM did you get?

You can try running memtest on 1 RAM chip at a time to narrow down the bad egg(s).

 

[JustinAWR]

The new RAM I got was the same brand - G.Skill. It is F3-12800CL9D-4GBRL - it was a customer choice award winner on New Egg and had hundreds of positive reviews so I assumed it would be a safe bet. I will try it one stick at a time tonight just for the heck of it, and mail them both back tomorrow (thats not New Egg does RMAs). Is there any chance it could be something somehow corrupting the RAM? Maybe the port?

 

[aford10]

Make sure they are installed firmly, and in the same colored slots. You can also move them to the other colored slots. This will help rule out a slot issue. If you're still getting errors, you should send them back.

 

[JustinAWR]

Thanks for the advice; will do. I'll update this thread if anything new happens... which knowing this build, it will.

 

[JustinAWR]

Ok, an update already.. I read through the Memtest manual more, and it sounds like it might be a motherboard issue. I am trying each stick individually now - one ran overnight with 16 complete passes without error. The other made three complete passes, but stalled on the fourth pass (no errors, but would not progress Memtest past 50%). I am retesting that stick now. If I don't find errors with that stick, or if it freezes again as it did, it has to be the motherboard right?

If it is the motherboard - I have an issue. In order to send it back, I have to remove the CPU. The CPU is secured onto the motherboard naturally, and there is a heatsink on top of the CPU which is attached via thermal paste. How do I separate the heatsink from the CPU without damaging it or the motherboard?

 

[aford10]

Once you find out the RAM that had errors, try the RAM chip that tested good, in that same slot that the other RAM tested bad in. If the RAM that tested good, then tests bad in that slot, the slot is the problem. If the good RAM tests good again, the other RAM is bad.

If you need to remove the CPU, you can disconnect the HSF. The CPU will likely be stuck to it because of the thermal paste. They should come apart easily if you pull on the CPU.

 

[JustinAWR]

I have completed five passes so far on my other stick with no errors, which leads me to ask; if neither stick shows errors after 7 passes, what gives? I got errors with both sticks in, but would those have appeared solely because of the faultiness of the second slot, or would one stick of RAM have to be faulty as well as the slot for errors to show up?

As far as removing the heatsink fan from the CPU, my concern is the termal paste being stuck on - is there a risk of damaging the CPU by pulling on the HSF too much?

 

[aford10]

Are they still in the same slots they were in before, when they showed errors? If no, and memtest comes up clean, the first slots are flakin out. If you had thoughts of adding more RAM in the future, and would need those slots, it may be best to send the motherboard back. Even if the other slots/RAM test good.

The CPU shouldn't be stuck on that hard. The thermal paste should still be in a paste form, and not crusted together. If they get stuck together, you can start up the PC, and let it run for a few minutes. That will heat up the CPU and paste. They will be easier to seperate then.

 

[JustinAWR]

So I just did seven passes on the second stick, and completed them all without error. I then booted up my system for the first time since getting this new RAM. Windows said it hadn't shut down properly last time and wanted to run startup repair, which I did. It then tried to do a system restore, at which point I got a BSOD. What could that mean? I wasn't using the second memory slot at all.
It just doesn't make any sense to me, and right now I am tempted to just sent back everything I can for a refund (less the 15% they charge for restocking and the shipping cost - aka losing a lot of money). Does that make sense to do?

 

[cpatel1987]

Replace the motherboard, your prediction of the memory slots themselves being bad is most likely correct. Your concerned about taking off the heatsink with thermalpaste, that isn't a big concern at all it shouldn't be stuck on that much. Use water or IPA to carefully remove thermal paste from the cpu after pulling of the heatsink. If you didn't buy thermalpaste and use the slight amount that came with the stock fan, your gonna need to purchase some once the new board comes in.

No, does not make sense to take back everything. Figure out whats defective, in addition to learning something new for all of us (since it seems nobody on these boards including myself know whats 100% wrong here) its less of a hassle to ship things one at a time back and get another.

 

[aford10]

Do you remember what the BSOD message was?

Can you boot into windows now? If no, how far does it get?

 

[JustinAWR]

If I don't send it all back for a huge loss and do end up trying another round of replacements (my 30 days for nearly everything is up next wednesday), I would need to send back a bunch of stuff now for replacements instead of testing one at a time. I am thinking I would need to send motherboard, gpu, and cpu back now for replacements (should I send back the RAM as well?). The only trick will be getting microsoft to let me activate my system builder copy of Windows again I guess.

I don't recall what the first BSOD I got was, but the second time I tried I got a Quota_Underflow error which is weird because that seems to have something to do with drivers (I got it after I had updated Windows and restarted strangely - could it be an HDD issue?). The one after that was a Sytem_Service_Exception. So yes, I can boot to Windows now, and it loads up my profile and I can tinker around a little bit, but no more than a couple minutes at most before it crashes, just as it was doing before.