Solved BSOD after computer has been off for 12+ hours

[mjm64]

This has been a long time problem that I'd like to figure out once and for all.

I purchased an AMD AM3 Phenom X4 processor & a Gigabyte MA785GMT motherboard with 2x2gb Gskill DDR3 ram, Corair CX430 power & Seagate 500gb HD. I purchased an OEM Win7 Home Prem 64.

It worked for a bit then started to generate random BSOD on startup. If I wiped the drive, and re-loaded windows, everything worked until the computer was turned off for more that 12 hours. If you put the computer into sleep mode, it was fine. You could reboot, just not turn the machine off, or if you unplugged it, it would BSOD.

One of the consistant errors was with the Catalyst video drivers. As the chips on the MB are from ATI/AMD, I turned off the onboard video card and tried buying a new ATI video card thinking it would have the least issues with the MB support chips. (ie ATI chips should work with ATI video cards). This didn't fix it. So I replaced the memory cards. I tried a different motherboard from ASUS. I tried a new case/power supply. I ran memtest and no errors in a 38hour run. I ran Prime95 and it was fine without errors for 18hours.

Eventually I gave up on trying to fix it and lived with just not turning the computer off.

Then I moved to a new house and the computer got shut down for several months. When I got arround to re-starting it, I figured I'll just try replacing the one part that hadn't.. I got a new Asus P8B75M motherboard and i5-3570 CPU. I replaced the Corsair power supply with a TX650. I added an OCZ Agility 3 SSD.

And now I am getting the same BSOD if you turn off over night, and added that after a round of BSOD the computer now then looses your win7 install. The data is on the drive but you can't do anything.

Attached is the last round of minidumps from the computer minidump.zip. If anyone can point a direction to look for what is going wrong I would love it. I have spent too much money on making a computer, and just should have bought an overpriced Dell and got some extended warranty as it would have cost me less that what I have paid so far. HELP!!!

 

[mjm64]

In order to get the rest of the information that is requested in the BSOD how to instructions, I have wiped my SSD drive and re-loaded windows 7. I have installed all the Drivers that came on the CD that ships with the MB, and run the SF Diagnostic tool to get all the info for the install before it crashes. I have turned off the computer and will give it until tomorrow and see if it will repeat itself and BSOD on startup so I get some minidump errors.

 

[mjm64]

I have taken some time to read through some of the other threads here in hopes to find something that might explain my issues. I have come up with a question however with running memtest+

In a thread about the STOP 0x50 error (that I have gotten quite a few of) one of the posters wrote:

No. Sometimes testing all give false results. The test must be done the way that I said. I can, also tell you, I have never seen consistent bad memory results that were not correct. I can also tell you to press down hard on each stick, sometimes they are not set correctly.

This brings to question. If your motherboard has Dual Channel memory, would you get problems if you put in a single Dimm to test it? Or would you need to put in 2 dimm's into either the A1/B1 or A2/B2 slots to properly test?

 

[mjm64]

Just for my own sake, I ran memtest+ with the full 8gb of ram installed for 5 hours (for complete tests) and didn't get any errors. I know that this memory is OK as I have run it perfectly fine in my other work computer without trouble for two years.

So I turned off the clean install off overnight and rebooted this morning. Window came up OK, but as soon as I ran IE, it locked up, and eventually generated a STOP 0x3b (I believe) but it was so fast on screen I didn't really get a chance to see it. The system rebooted, and when into a chkdsk and recovered several bad blocks from the SSD.

Then windows restarted, and is seemed pretty stable. IE worked, and I ran CPU-z and looked at the processor, and I was trying to verify if I got conned with my i5 that was supposed to come with the faster Intel HD4000 video onboard according to the sales person who sold it to my. It turned out only the i5-3750k does.

Then when I shut the computer down, it had another BSOD with STOP 0x50

Attached are the reports that you fellows ask for. What floors me is this is doing this exact same problem now with two totally different everything (cpu, mb, psu, ram, hdd) the only thing that hasn't changed is my Windows 7 install DVD and it's install key. It works fine until you power it down for a while. I'm guessing 8-12 hours. Then it starts to generate BSOD when you start to use it.

HELP please. I desperately want to get to the bottom of what is going on.

Matt.

 

[Layback Bear]

For us to have a chance to help could you fill out your system specs?
((My System Specs)) This will help you.

Speccy - System Information - Free Download

 

[mjm64]

I thought that I had. My machine is not a pre-built one (like a Dell or HP) if it was, then I would be talking with them to get the problem fixed under their warranty.

As listed on my CP page my current machine is:

ASUS P8B75M-CMP motherboard with bios 602 (the latest version from ASUS)
Intel i5-3750 cpu
8gb GSkill F3-12800CL9D-4GBNQ ram
Corsair TX650v2 PSU
Nec EA231WMi riunning at 1920 x 1080 if that really matters
OCZ Agility 3 128gb SSD

The case is black and made of aluminum. I don't know nor really care who made it. (does that really matter??)
I have a Saphire HD5670 ATI video card, but I have pulled that out and running with the onboard video to eliminate that as another possible cause of issue.


I had a Corsair CX430 psu prior.
The ram, I have been through 3 sets just to make sure it isn't memory. I can run them on my other i7-2600K computer without any issue. I have memtest+ them for over 40 hours now without any error.
I have tried working with Seagate HDD as well as Western Digital HDD.

Every combination for has the same issue. It works fine from original build until you turn the unit off for 12 hours. Then you start to get BSOD.

You can even do a build, and pull the power cord from the PSU and leave it for 12 hours with NO power plug, and on reboot it will BSOD.

 

[Layback Bear]

When I clicked on you specs before all I got was your operating system. Now I can see all your specs. Thank you.

 

[mjm64]

Some more errors from the latest reboot after being turned off for the day.

Again sadly it rushes by too fast to see what the actual error was. It booted up into win7 initally, then when opening up IE it generated the first BSOD. The machine rebooted and went into a chkdsk and found a slew of orphaned files on the SSD, then it would not reboot into windows 7. It will work with safe mode.

 

[Layback Bear]

When you boot and not use I.E. will you get a BSOD? I'am unable to read dumps sorry. Hopefully some one will come along for that. I have just read over your post again and it just confuses me. One post your computer has this,that and ding dong parts and the next it has something else. I hope you My System Specs is correct. Correct information helps us help you.
1. Because memory sticks work in one computer doesn't mean it will work in this problem computer. Your ram speed and voltages should match your specs of the motherboard and cpu.
2. As you posted; you run Prime 95 for 18 hours. That sure is a long time. What temp were you getting?
3. Running memory test 38 hours. I have never heard of such a long test.
4. Where did you buy Windows 7 from?
5. If you are overclock in anyway stop and use default settings.
6. What anti virus and firewall are you using.
-------------------------------------
I will see if I can find someone to analize your zip files.

 

[mjm64]

1. Because memory sticks work in one computer doesn't mean it will work in this problem computer. Your ram speed and voltages should match your specs of the motherboard and cpu.

The ram was bought because it was listed on the original motherboard tested list. The voltages and speed have been checked with the RAM manufacturer and correctly entered into the bios setup.


2. As you posted; you run Prime 95 for 18 hours. That sure is a long time. What temp were you getting?

When you have tried just about anything else, you go to extreme measures to try and find the problem. The MB got up to 58c. The CPU ran at 67, 69, 71 & 74c. That was the maximum it has ever gotten to. And zero errors, no Blue Screens.


3. Running memory test 38 hours. I have never heard of such a long test.

I was told to test the ram, and so I turned it on and left it for the weekend. Started up on Friday night, turned off on Sunday morning after I got up. Again no errors. But it was all the ram installed. I am trying each Dimm individually right now. At 9min per test cycle, I figure if I give is Dimm 2 hours, if it's flaky, it should fail by then.


4. Where did you buy Windows 7 from?

It was bought from my local computer store (Canada Computers) when I bought the first iteration of this computer. I know my post is confusing. I bought the components for a computer to do Photoshop. It worked for a month then started to have errors. After about 6 months errors, I gave up and replaced the machine with an i7-2600k. Not wanting to throw in the towel, I played to get the machine stable. From that I tested the RAM in the new i7 computer as I bought the exact same ram for the new machine. I bought a new case/psu to put the MB into. Also a new DVD drive, video card, mouse.... Then I figured out that if you don't power down the computer it would work mostly stable. So I lived with it until I moved 6 months ago. I decided to just replace the CPU as I thought that was what was likely defective and reuse the rest of the machine. It tuned out that it wasn't the CPU as the problems continue with the new i5-3750 and new motherboard that was bought two weeks ago.

5. If you are overclock in anyway stop and use default settings.

I don't see a point in Overclocking. Never have, never will.

6. What anti virus and firewall are you using.

I use Microsofts.. But it crashes reguardless of if there is any AV installed or not. If I do a clean install from the MS Win7 dvd and after it is finished, I turn the computer off for 12 hours. I don't install any software, or even hook up to any network, it will crash just the same.


I will let you know how the individual dimm memtest+ goes.

 

[Layback Bear]

I pat you on the back. You have doing a lot of testing. I think I understand what you have done and in what order. I will wait for your test results.

 

[mjm64]

I think I have found a defective Dimm. It fails if you have it cold and put it into the computer by itself on memtest+. if you put it in as dual channel, it tests fine. if you let it warm up and test it even by itself, it tests fine. Guess it's getting shipped back to GSkill as defective and get them to replace it. I will test the rest of the memory and then give it a try rebuilding tomorrow. Will know if it's solved next week I guess to give me time to re-build and then wait my 12 hours to see if I get the errors on a cold startup.

Thanks for the help.

 

[mjm64]

Not so fast... the plot actually thickens.

So I tested my Dimm's last night, and the first one had an error on it. So I flagged that one with an 'X' so that I could keep check on it. The other 3 dimm's all checked out perfectly fine. I went to bed.

Now it's morning, and I powered up the computer with the 4th Dimm still in it that checked out OK last night. I had my memtest+ CD in the drive, and it boots up into testing. The Dimm that checked out perfectly fine last night, Fails horribly this morning. Like in the first test it gets 100+ errors.

So I turn off the computer before the first check is even completed, and pull the 4th Dimm out and put back in the one that I marked with the 'X'. It checks out perfectly fine for 4 tests. So I put back in the 4th Dimm that worked fine last night, but failed first thing this morning, and re-checked it, and now it's perfectly fine.

The variable now is the temperature of the motherboard, as when you do memory tests it would generate heat. A Dimm sitting on a table would be at room temp, so putting that Dimm into a computer, it would be at or close to room temp for the first pass and heat up from there. If it passes when the CPU/MB is hot, but fail badly when the CPU/MB is cold (like after 8 hours turned off) then the issue probably is not in the memory.

This has got me thinking. I had the exact same problem with my last CPU/MB in that it would fail when cold but work fine when HOT. It would BSOD on startup that could be dumping corruption onto the hard drive that would get it to BSOD from that point onwards. When you install, you are heating up the computer, so it would install fine and stay working as long as you don't power down and let the computer cool down.

Logically, if two different motherboards do the same thing, then it's very unlikely that it's the motherboard at fault. Perhaps its the case it's mounted in that is not flat, and the distortion of the mounted motherboard is enough to cause errors when it's cold, but as the motherboard heats up this fixes itself??

Does that sound like a plausible problem? I guess I could try doing memory tests from cold without the case. ie have the motherboard sitting on a antistatic bag on a desktop and see if I can get my memory to fail from cold.

How is that for a way out there problem???

 

[Layback Bear]

I like the idea of testing with motherboard out of case. Get a window fan handy so you can do test with different amount of cooling. Things to think on.
1. A computer will go back to ambient temp in less that a hour so I don't get the 12 hour thing.
2. Quality of case maters; the way the motherboard mounts is critical. One stand off in the wrong place can distroy computers or do crazy things. A motherboard that is not tightened down evenly can also cause crazy problems.
The I/O Panel where it goes through the case in the back of the case can cause a twist. If that hole in the back of the case is off a little can cause a twist of the motherboard. Cheep cases are cheep.
3. We cant't see your computer and instulation properly even with pictures. Your Specs, test and reports, observation, is what we got to work with.
4. Never assume their is only one problem.

 

[mjm64]

I thought I'd give a different thing a try.

1. I removed the CPU cooler from my new CPU/MB and moved it over to the CPU/MB.
2. I removed all but the 4 corners of the offset posts out of the case that hold the motherboard to the case.
3. I cut up and made some rubber washers from an old bicycle innertube for the 4 remaining posts
4. Put the case together and gave it a test with one of my original Seagate harddrives that I had installed windows 7 for this CPU/MB combination

Powered it up and it worked without a single error. I am stress testing it right now to see if I can get it to BSOD.. but the ultimate test is turning it off tonight and seeing if it will work tomorrow morning.

I think it was that the case is twisted slightly and/or one of the offset posts is either too hight/low and is deforming the motherboard just enough to cause memory read errors on startup from cold. This is trashing the boot drive with errors and giving me the BSOD. With just the outside edges supported, the motherboard can lay totally flat now and the errors seem to have gone away.

Now I have an extra i5, Asus P8B75M, Saphire HD5670, SSD and Corsair TX650. The only thing I am missing is a case and some RAM to build a 3rd computer.

Will post a follow up next week if it's all stable.

Matt.

 

[Layback Bear]

Be sure to get back to us on this crazy problem.

 

[mjm64]

Nope.. Not fixed.

On the rebuilt AMD FX 955 machine, it works fine unit you turn it off for at least 8 hours.

I rebuilt Win7 from my install DVD, and got it to install all the updates, and loaded the lastest motherboard support drivers for the HD4000 onboard chipset. If I turn the computer off and back on again, FINE. If I turn it off for 2-3 hours.. FINE. I turned it off and went out with the family to the beach and came back 8 1/2hours later. BSOD.

To add even more fun. I get a STOP 0x19 in safe mode when i try and run the SFDiagnoticTool program to get the MSInfo.

I would really appreciate if someone can look at these dump files and give me an idea as to what is crashing. After installing the motherboard with minimal offset mounting bolts, I can't get the memory to fail in memtest+

I don't know what to test now. I need a starting point.

 

[kegobeer]

I'm assuming that you visited the motherboard manufacturer's website and downloaded the correct drivers, and you also checked for BIOS updates and flashed to the newest revision, and you loaded the default BIOS settings, followed by the optimal settings, and ensured the correct settings were chosen for your hard drive (AHCI/SATA/etc). I also assume you did the same for your other hardware. If you did all of that, then...

Since this started immediately after you built the system, I believe you have a hardware issue. Start by removing the motherboard from the case and putting in on something that electrically isolates it (like a piece of plywood, etc). Also remove the power supply, then connect it to the motherboard. Connect only the power button from the case to the motherboard. Connect the video card, keyboard, and mouse, and the hard drive. Boot the system and see if you have BSODs.

 

[mjm64]

I ended up taking the motherboard out of the case, and it still game me the error after being off for the night. So I put in a single Dimm and left it off for 5 hours then did a memtest+ and found that in actual fact, there were two dimm's defective from cold.

I took the memory that I knew worked from out of my work PC and put it into this computer, and it checked fine. Put everything back into the case, and it boots fine now.

I have opened a case in the update forum as now I can't get the windows update to run.. but that's another story.

 

[Layback Bear]

Two bad ram sticks. You are so lucky. Thanks for letting us know. Happy computing.