Windows 7 Ultimate 64-bit crash/BIOS crash

[password]

i have been having increasingly frequent bsods/ complete bios hangs"- ie the keyboard numlock does not even respond. Today i needed to restart enough times for my bios to check its NVRAM.

when the bsod started i didn't install any new software and last week i reinstalled windows 7 .after the re-install my pc ran fine for 4 days and i was back to square one.

i believe i have found a clue. I disabled cores 2,3,4 and my system now seems stable. I doubt its the cpu itself because the problem was postponed with the reinstalling of my windows

oddly enough my minidump folder is completely empty

Specs
ram : 2x2GB , DDR3 , dual channel
Cpu:AMD phenom II x4 965
Power supple: 550w - LPG-43
OS : windows 7 ultimate 64 bit
GPU:nvidia 9800 GT (problem existed also with a 6600le)
Motherboard:ASUS

PS: i have updated my bios to check if it solves anything and there was no change

 

[Mark56]

Check these settings to see if it will then save future minidumps. If it then starts to save them, put them in a zip file and add as an attachment to future posts.

As the mindump file is empty your system is not set to save memory dumps. Follow this guide to set it up correctly.

I would also suggest you check that the voltage settings are correct for your RAM and CPU. You will need to find the recommended voltage from the manufacturer's site.

Click on Start > Control Panel > System > Advanced System Settings (left pane) > Advanced tab > Settings button for Startup and Recovery. In the System Failure box make these settings.

Check > Write an event to the system log.

Check > Automatically restart.

Select > 'Complete or Kernel Memory Dump' in the Write debugging information box.

Small Dump Directory should read %SystemRoot%\Minidump.

At the bottom of the box make sure "Overwrite any existing file" is not checked.

Then go back to the Write debugging information box and set it to "Small memory dump".

Hit OK, OK again and close the System box.

 

[password]

ok i did what you said i should do and i have also reanbled my other cores. PRoblem seems gone now as my pc is now running fine for more that 2 days without any glitches. i dont know what the problem was but it gone now.

 

[Mark56]

You are now set to save any minidumps, if or when you have any more BSOD's, it will just be a waiting game to see if any more happen.

 

[password]

And the waiting game is over. I have had a few completly hangs and bsod between today and yesterday. attached is the three mini dumps that was created.

 

[Mark56]

You have three different error codes from these minidumps. One states hardware, one states memory corruption and the other is inconclusive. The most likely cause is that you have faulty RAM.

Run Memtest86+ overnight and allow it to complete at least 8 passses or until it shows errors. Remove one of the sticks before running the test and then swap the sticks and repeat the test.

Download Memtest86+ from here.

You need to scroll down the page until you see the three blue screenshots. Just below is the list of downloads for Memtest86+ V4.20. Click on Download - Pre-Compiled Bootable ISO (.zip) and save to your desktop.

When the download is complete right click the file and select Extract Here and burn the image to a CD.

In windows 7 right click the extracted file, select Open With, then select Windows Disc Image Burning Tool then follow the prompts. For all other versions of windows (if you do not have an ISO burner) download this free software. isoimageburner

Boot the PC into the Bios setup and set the CD/DVD drive to 1st in the boot sequence. Insert the disk in the drive then reboot and the disc will load into dos. Leave the test to run through at least 8 cycles or until it is showing some errors. If errors show in the test, remove all but one of your RAM sticks and repeat the test on each stick until you find the one that is faulty. This is a long slow test and should ideally be run overnight.

The memtest will not be 100% accurate but should easily detect any major faults.

 

[kristain]

Hi,

Do you remember making any recent changes to the computer prior to this issue?

I would suggest you to follow the steps in the link and check if the issue persists:

http://windows.microsoft.com/en-us/windows7/Resolving-stop-blue-screen-errors-in-Windows-7

Also check event viewer for error follow the steps in the link:

http://windows.microsoft.com/en-US/windows7/What-information-appears-in-event-logs-Event-Viewer

 

[password]

@kristian nope I did not make any recent changes

@Mark56 i ran memtest86+ for 4 runs and it found errors. I removed one stick and ran memtest86 again. It ran for 22 runs without errors. So i rebooted into windows and 5 minutes later my pc restarted.it didn't generate a new minidump though.

 

[password]

i have tested my other ram module with 10 passes and memtest did not pick up any errors at all, and 5 minutes after booting with both sticks in my pc i get a restart

 

[password]

And here is *three more minidumps if that would help

 

[password]

Here is some more hardware information from AMDoverdrive that may be helpfull in some way

 

[Mark56]

@ i ran memtest86+ for 4 runs and it found errors. I removed one stick and ran memtest86 again. It ran for 22 runs without errors. So i rebooted into windows and 5 minutes later my pc restarted.it didn't generate a new minidump though.

I just want to confirm what you have done. After finding errors with both sticks inserted you removed one stick and continued the test for 22 passes and it found no errors. Yes?

When you rebooted had you put the other stick back in? If not try running the PC with just the single stick that memtest found no fault with.

The last three minidumps all suggest a memory fault but includes one that could be Malware (amongst other things)

-----------------------------------------------------------

i have tested my other ram module with 10 passes and memtest did not pick up any errors at all, and 5 minutes after booting with both sticks in my pc i get a restart

If either stick is faulty this would be expected, you need to run the PC on the stick that memtest passed and leave the other one out. If that still produces crashes then we need to dig deeper as it is very unlikely that both sticks have errors.

From the memtest results it is clear that one of your sticks does have errors and should be replaced.

Another check you should make is that the RAM voltage is set in the Bios correctly. You will need to look up the recommended voltage on the Kingston site for your model of stick.

 

[password]

I just want to confirm what you have done. After finding errors with both sticks inserted you removed one stick and continued the test for 22 passes and it found no errors. Yes?

When you rebooted had you put the other stick back in? If not try running the PC with just the single stick that memtest found no fault with.

The last three minidumps all suggest a memory fault but includes one that could be Malware (amongst other things)

After scanning one stick for 22 runs and finding no errors i rebooted with only that stick and less than 5 minutes later my pc rebooted(was still loading apps and desktop stuff).
I even at swapped the sticks and did 10 runs on that stick and no errors were found.

i'll check the required voltage of the sticks and adjust them if necessary.

One weird quirk i forgot to mention is that memtest86+ detected my one stick as dddr1333 and the other as ddr400, which is weird since are a matched pair and have identical specs.

Hope that clarifies it

 

[password]

i justed checked and for my memory "KHX1600CD93/2G" the recommended voltage is 1.7v-1.9V and it was set to 1.590v. I incremented it to 1.605V. Will check stability and increment again if necessary.It will now be just a waiting game,again.

 

[Mark56]

This incorrect voltage setting may explain some of the strange behaviour and misleading Memtest results. Can you set it higher so that it is within the recommended voltage as 1.605V is still below what it should be.

 

[password]

sure i can set it higher and i will in the future. I just have some weird fear of adjusting voltages, so i will slowly day by day increase it bit by bit.

 

[Mark56]

There is nothing to fear from setting the RAM to the manufacturers recommended setting. Anything below what they recommend is underclocking the RAM and prone to cause errors.

As long as you do not go above the maximum recommended voltage you will be perfectly safe.

 

[password]

ok i tried setting it to 1.7v but it kept jumping back to 1.695V will try again tomorrow after my semester test by actually setting it in bios(currently adjusting via AMDoverdrive).

thanks again for everyone helping me out with my pc. I really appreciate it and i'll keep you posted if anything happens.

 

[Mark56]

I would set it mid way using the Bios to 1.8V and see how it performs.

Only happy to try and help but we are not there yet.

 

[password]

And once again I spoke too soon , attached is the newest minidump

 

[password]

i am now working on a hunch that my pc bsods less if i do a proper restart after a bsod

 

[Mark56]

All the last five minidumps have produced different error codes which, as I have said before, is typical of a RAM fault. One of the last two states 'memory corruption'.

Have you now got the RAM voltage set between the recommended settings?

It's clear from what you said earlier that you have run the PC on one of the sticks without the other fitted, have you tried running the PC on the other stick (on it's own)?

If the voltage is set correctly and you have tried to run the PC on each individual stick then there are a few possibilities.

First, there is a slim chance that both of your RAM sticks are faulty. If you can borrow a RAM stick from another PC to try in yours that would be a quick way to see if we are dealing with some other fault, (make sure you adjust the voltage to match the sticks recommended settings).

There could be a bad driver but I would expect to see more consistency in the minidump error codes if this was the case.

There could also be an infection in the system but I think a hardware fault is more likely.

I think it is important to prove, before moving on to other tests, that your RAM is 100%. Try to do as I suggested above and borrow a RAM stick to test the PC without your own pair of sticks fitted. If errors still occur then we can be sure these BSOD's are being caused by something else. The other option would be to fit your pair of sticks into another PC and see if the errors continue (again make sure you adjust the voltage to the correct setting and remove any other sticks).

I would also suggest that you make a close inspection of the RAM slots to make sure there is nothing stuck in them and give them a blast with a can of compressed air. Also, clean the stick contacts with a soft pencil eraser to remove any oxidisation.

The fact that Memtest is misreading your RAM's frequency may suggest a faulty slot.

 

[Mark56]

i am now working on a hunch that my pc bsods less if i do a proper restart after a bsod

You posted this while I was writing the above post, try what I have suggested with a borrowed stick and see how it goes.

 

[password]

ram slots is clean , i cleaned them a few days ago.
I have tried each stick on its own.
I have set the voltage correctly.
I will see if i can borrow some ram somewhere.

 

[Mark56]

Ok, let us know how it goes.