Solved Random, None-recoverable freezes. No BSOD, No Memory dump, No eventlog

[Antarez]

Recently assembled a new build, everything seemed to be running fine, however after a few days of normal use, the build now freezes randomly. This has persisted across multiple clean installs, driver updates, bios updates and even different drive configurations, a new psu and a new cpu. I'm at my wits end! I've tried stress testing with prime, and I can run both small and large tests seemingly indefinitely, but blend test crashed within a few seconds. Ive ran memtest86 overnight with no errors and ive tried all manner of suggestions ive found around the web from people with similar problems, to no avail, I am, quite frankly stumped, any help would be greatly appreciated! As the title says, I havent had any dumps or event log entries when i freeze and it seems to happen almost entirely at random, however running games through steam seems to increase the likelihood of it happening.

 

[HarriePateman]

Try to scan your HDD for errors.

also try to repaip windows through the disk.

 

[Antarez]

Try to scan your HDD for errors.

also try to repaip windows through the disk.

I will try that later on as I'm currently running memtest again, do you really think it could be the HDDs at fault? All parts are brand new and less than 6 weeks old.

 

[HarriePateman]

Its possible.

It could just be a bad install of windows.

 

[Antarez]

Its possible.

It could just be a bad install of windows.

I'm doubtful of it being a bad install as the issue has persisted across 6 seperate clean installs and multiple hard drive reformats, I initially thought my SSD to be at fault as I was using an SSD for windows and my RAID array for everything else but after removing the SSD and doing a clean install to the HDDs the problem persists, its very annoying. I'm grateful for the help though, thank you!

 

[HarriePateman]

Is the firmware for the SSD and hardware upto date?

 

[Antarez]

Is the firmware for the SSD and hardware upto date?

Yep, that was the first thing I looked into when the problem made itself apparent, it really feels like a memory issue but no testing has proved my suspicions so far.

 

[HarriePateman]

hmmmm, i take it the RAM is at stock setting and speed? Voltage?

 

[Antarez]

hmmmm, i take it the RAM is at stock setting and speed? Voltage?

I used XMP to configure the RAM but I checked over it quickly and it seemed to match the specification set by corsair. Could there be an issue there?

 

[HarriePateman]

Disable the XMP, and run at Stock for a little while and see if this improves.

 

[Antarez]

Disable the XMP, and run at Stock for a little while and see if this improves.

Ill try that, but the stock values specified by corsair seem to match what XMP set the values at, but I will try that, and report back with my findings later, thank you!

 

[HarriePateman]

No worries.

This may not be the Problem, but if we troubleshoot as we are we cant miss anything!

 

[Antarez]

Somethings just struck me, would memtest86 pick up on an error if the XMP is the issue or not? Or could there be a chance of false positives or negatives etc, with incorrectly configured ram? Because I'm running memtest right now but I havent disabled XMP yet and I'm unsure of how it may affect the test, if it would at all.

 

[HarriePateman]

im not to sure on that one, Try to run it at stock then with it on. What profile are you using with XMP?

 

[Antarez]

im not to sure on that one, Try to run it at stock then with it on. What profile are you using with XMP?

My BIOS only has an option to enable or disable XMP, theres no sort of profile selection, but, incase ive misunderstood, my ram is configured thus; 9-9-9-24 timings, 1600MHz and 1.5V If that is helpful in anyway haha

 

[Antarez]

Well, as it turns out, I disabled XMP and my memory returned to its stock clock of 1333MHz (Despite the manufacturer saying that the ram was rated for 1600MHz) and, I decided as I was restarting anyway, it was worth a punt at trying Prime95 again before I ran a new memtest, I've been running a blend test for 10 minutes now, and normally it would have frozen the system within seconds, now I'm not saying its fixed, I know this is very much one of those things that you can't know if its fixed straight away, you have to play the waiting game and see it if the problem rears its head again, but right now I'm hopeful. Assuming that the XMP is the problem, I'm assuming its something to do with undervoltage for the higher clock rate, what would be your opinion?

 

[HarriePateman]

okay thats fine, lets see if memtest picks anything up.

 

[Antarez]

I'll run another memtest but honestly I believe the problem to be fixed, only question now is how to get my system stable with the RAM running at 1600MHz, like Corsair claim it should be.

 

[HarriePateman]

When using XMP there shouldn't be a problem. If there is i would take the RAM Back. As "XMP" is a Pre-safe-overclock mode !

 

[Antarez]

Well thats rather worrying, only thing is, I'm unsure how to isolate which stick it is that can't handle running at that clock, because sometimes a freeze wont happen for days and memtest isnt picking anything up... It could take me weeks to find the faulty stick haha!