[justin89h]

I've been having 'Memory Management' BSOD showing up since I built my PC around 2 months ago. As well as occasionally 'Page Fault' BSOD and others related to memory corruption.

 

After receiving these errors I've performed tests using the Microsoft Memory Diagnostic tool, and MemTest86 and they've come back with hardware issues within the first few seconds.

 

However, after having the computer turned off for a while, to try the RAM in a different slot configuration, or just to slam my head repeatedly against the wall, a second memory test reports no errors with the RAM whatsoever. I have run tests over 16 hours with absolutely no errors shown.

 

I have since reformatted my system due to corrupted system files, and have yet to receive a BSOD. I am however expecting one as I have performed a diagnostic test and received a hardware issue. After further messing around I have thankfully found out how to recreate this issue, and clear it, but not the root cause.

 

 

This is how I can cause a memory test to show a hardware issue:

 

1) Have Windows running, all startup software running etc. 
2) Shut down the system through the start menu option or the power button on the front of the PC. 
3) After the system has shut down immediately start it back up again and allow windows to boot. 
4) Start Microsoft's Memory Diagnostic test, allowing it to restart the computer.

 

 

The following is how I can clear this issue resulting in a memory test completing several passes with no errors:

 

1) After the previous diagnostic test completes with errors, shut down the computer. 
2) Leave it off for several minutes. 
3) Boot Windows and start Microsoft's Memory Diagnostic test, allowing it to restart the computer.

 

 

After significant searching on tech forums I have failed to find anyone that has a similar issue. Does anyone know what the root cause of this may be? Could the memory be set-up incorrectly or could it be due to the power supply? My system details are as follows:

 

CPU: Intel Core i5-4590 
Motherboard: ASRock B85M-Pro4 
Memory: 8G Kit 1600 Patriot Signature Line CAS 11 
GPU: Gigabyte R9 290 4GB 
PSU: Fractal Design Integra M 550W 
OS: Windows 8.1 fully updated

[Advertisements]

I've done some further testing which I should probably have done as soon as I figured out how to recreate the problem.

 

I managed to recreate the problem with just a single 4GB stick in slot one. After recreating it, I swapped over the stick to the other one and managed to recreate the problem on that one as well.

 

After that I booted the computer with just a single 4GB stick in slot 3 and manged to recreate the problem with that one as well.

 

Every time I tested I verified the ram passed (a single pass) with no errors, before a quick shutdown and bootup, followed by a second test with errors after a few seconds

 

As it would be unlikely for both sticks of RAM to have the same issue I believe this discounts the RAM as being the root cause? Same with the slots on the motherboard.

 

Does this mean the CPU's memory controller is the most likely suspect? If this is the case it would be somewhat difficult to explain and receive an RMA so would someone be able to confirm this with a better explanation? or offer an alternative?

 

I'll add as well that following a memory test with errors shown on just a single 4GB stick, Windows had significant trouble booting, flashing a BSOD, offering system repair options, or running for just a few seconds. One time it appeared the computer failed to even POST.

 

So once this hardware issue comes about it appears to render the RAM close to completely unusable.

[justin89h]

I've done some further testing which I should probably have done as soon as I figured out how to recreate the problem.

 

I managed to recreate the problem with just a single 4GB stick in slot one. After recreating it, I swapped over the stick to the other one and managed to recreate the problem on that one as well.

 

After that I booted the computer with just a single 4GB stick in slot 3 and manged to recreate the problem with that one as well.

 

Every time I tested I verified the ram passed (a single pass) with no errors, before a quick shutdown and bootup, followed by a second test with errors after a few seconds

 

As it would be unlikely for both sticks of RAM to have the same issue I believe this discounts the RAM as being the root cause? Same with the slots on the motherboard.

 

Does this mean the CPU's memory controller is the most likely suspect? If this is the case it would be somewhat difficult to explain and receive an RMA so would someone be able to confirm this with a better explanation? or offer an alternative?

 

I'll add as well that following a memory test with errors shown on just a single 4GB stick, Windows had significant trouble booting, flashing a BSOD, offering system repair options, or running for just a few seconds. One time it appeared the computer failed to even POST.

 

So once this hardware issue comes about it appears to render the RAM close to completely unusable.

[iammykyl]

 justin89h.

A few steps to take before making any claim.

Clear the CMOS to  set the BIOS to defaults, > restart and enter BIOS, > check the RAM is set to Defaults, > exit menu. F10, > Save settings.

Run The Intel Processor Diagnostic Tool, HWMonitor, PC Wizard 2010,   Downloads, > http://www.intel.com...b/CS-032037.htm

 

When testing, disconnect all external devices except, keyboard, mouse, monitor.

If possible.   Try swapping in a know working PSU,, if you remove the video card, a 350W would suffice.   Test the RAM in another computer.

 

If the issue persists.

Remove the MB, > Remove the heatsink/fan and CPU.  Does the TIM look OK.

Inspect for damage to the socket, bent pins, burn marks, cracks on the CPU.

[justin89h]

BIOS is all set to default. 

 

I have completed two full tests with the Intel Processor Diagnostic Tool, they both completely passed everything. I noted that this test stressed tested the memory controller as well. It didn't test the integrated graphics card. I don't think that's too related to what we need to test but if you think it's worth it I'll disconnect my dedicated graphics card so it can test that as well.

When running PC Wizard Windows reported compatibility issues, but then it appeared to run without any issues. i did a global memory benchmark and everything appears to be fine, but to be honest I'm not sure what values are considered acceptable. I also did a system stability test for a few minutes, but that doesn't appear to have any kind of a report. The processor went up to 100% use regularly without any issues. Are there any specific tests that would be a good idea with this software?

 

HWMonitor doesn't appear to show anything abnormal but gave the option to export results so I've uploaded them here in-case it may help. https://www.sendspace.com/file/32ow9g

My PSU in my old (unused for a few years) desktop is a 650W Extreme Coolermaster. Bought it in 2007. My old PC still boots ok with this but I feel its a little risky putting it in my new system as it's so old. I'll give it a shot tomorrow if we can't get any results any other way.

Thanks for your help!

[iammykyl]

will a stick of RAM from your old rig run in your present? and visa versa?

For PC Wizard, from the left window, click the volt icon, how do the readings compare to HWMonitor?

Sorry, but I was unable to interpretate  the HWMonitor infor, please run again, drag the bottom of the open window down to see all info, take a screenshot and include in your next reply. 

Preferred method  to upload a file or image to G2G. 

 

First, > click the More Reply Options, (bottom right next to the blue reply button.)  > Type in your text.

If the file or image already exists, > Click Choose file... (bottom left) or see the next part to capture an image.

In the window that opens, > Browse to the file or image you want to upload, > Click the file/image you want, wait for confirmation at bottom left, 

Repeat if you have more files/images to upload, > When finished, > click Reply button.

 

To capture and post a screenshot; Instruction are for anyone who is unsure of the steps.. 

 

With the image you want displayed on the screen.
> Hold down the ALT key and Press the PRT SCR key..its on the top row..right hand side of the keyboard.   (nothing appears to happen, but the image is copied to the Clipboard.)

> Click on start...all programs...accessories, > click paint.

In the window that opens, >  left click in the white area, then > Hold down CTRL and press V. 

On the top menu, > click File > click Save as, Be sure to save as jpeg,...otherwise it may be to big to upload,

save it to your desktop...name it something related to the screen your capturing.

 

See if you can spot a recurring event using the  > http://www.makeuseof...bility-monitor/

[justin89h]

My RAM from my old system is DDR2 and isn't compatible with my new systems DDR3 motherboard unfortunately.

I pulled out my old power supply and gave that a go on the new system. Worked out fine once I did some research and realized the 4 pin 12v ATX plug works fine with the 8 pin motherboard socket. (I uninstalled the graphics card first of course)

With the different power supply I was able to recreate the failed hardware test on the RAM exactly as I detailed in my first post.

The voltage readings between PC Wizard and HW Monitor appear to only have a little variance. I've included screenshots as requested.

 

I do find it a bit odd that the '+12V Voltage' on PC Wizard is showing 6.41V. Is there a reason for this or does that indicate a problem? I would be expecting a lot more severe issues if 12V components were running on half voltage.

Ill take a look at that windows 7 reliability monitor. (Your link appears to be broken)

Thanks again for your help. From the tests you've suggested as well as the ones I've done before I'm suspecting the motherboard to be at fault, but anything else you could suggest to look into it further would be much appreciated. I don't particularly wish to be up for costs and wasted time if it turns out the motherboards fine.

Attached Thumbnails

• 
• 

Edited by justin89h, 20 April 2015 - 03:23 AM.

[iammykyl]

Sorry for the late reply.

Fixed link > http://www.makeuseof...bility-monitor/

Unfortunately non of the software for reading sensors is foolproof.   Sometimes they can't see the sensor or mix up which is which.   HWMonitor does not see the 12v and PC wizard reports it as low, think that's wrong as the computer would not start.    All rails specs allow +/- 5% of the value.   Initial voltage control is via the MB, secondary is via the CPU.   Without proper instruments, not possible for a user to determine which is faulty.

 

I don't think there is much more you can do for troubleshooting.   I would attempt a RMA of the CPU first with the retailer and if successful, try a test.

If the issue persists, then try a MB RMA, hoping it is still under warranty. 

[justin89h]

Was searching around a bit more on the net and found this:

 

http://superuser.com...old-boot-attack

 

The research mentioned seems to indicate that old types of RAM stored data for up to five minutes after system shutdown. But new types of RAM are meant to lose voltage after only a couple seconds as a security measure.

 

This might be related as if there is a fault in whatever component is responsible for controlling voltage to the ram, which causes data to stay in the memory for a few minutes after shutdown, this could get corrupted on a second boot-up.

 

Thoughts? Am I on completely the wrong track? What is responsible for controlling voltage to the RAM?


Just thought I'd mention this just in case. If I don't figure out anything new over the weekend I'll be attempting to RMA the CPU or motherboard next week. I'm hoping the reseller will organise it for me as I only bought the parts a couple months ago. Would save a fair bit on postage.

 

Thanks heaps for your assistance.

[iammykyl]

Best Answer

My understanding, settings for the RAM, voltage/speed/timing, are controlled on the MB, the CPU receives Data from the RAM, processes it, then passes it back to the RAM, so I don't think the fault wold be the CPU, rather think if there was a fault, it would be the MB..

This may not do any good, but please try.

Run your test but instead of just a restart, start in Safe Mode.   If the error does not happen, could be a non Microsoft driver or software issue.

Try running the test with a different mouse/keyboard in case there is a hardware fault there. 

 

Your most welcome

[justin89h]

I booted the system in safe mode (using msconfig) but it started with RAM issues, same as usual,

 

I've decided to call either ASRock or Patriot Tech Support, go through the issue with them in case there's something I've missed. After that I'll probably end up sending back the motherboard under warranty as it seems to be the most likely suspect.

 

Will probably do this in a month or so after I've finished playing GTA V though, because as long as I start it up in the correct way the system runs with no issues.

 

Thanks for all your advice, it helped a lot

[Plastic Nev]

I think you hit the nail on the head with the data retention you mentioned in post #8

That could be due to a faulty voltage levelling resistor within the power to the RAM section of the motherboard, thus allowing a smoothing capacitor to hold a charge for too long. That in turn will be allowing the RAM to retain partial and therefore corrupted data for a time, hence why the fault showing on immediate restarts, but not if given a reasonable time before a restart.

That of course means a RMA of the motherboard.

 

Nev.

[iammykyl]

We would very much appreciate an update on the results, not only helpful to us for troubleshooting, but for any member who have similar issues..

 

Thanks.