New build unstable at stock, please help to troubleshoot

Got home from work today to find 2 errors in Memtest after running for 21 hours, test 7, pass 13 and 25 if that means anything to anyone.

And what does this mean, that my RAM is bad and the cause, or that its ok and errors happen every now and then. Not sure what to make of it?
 
How would OCUK qualify the RMA? I dont want to be told theres no problem with the ram as it was after 21 hours that I found an error?
 
MrM said:
Got home from work today to find 2 errors in Memtest after running for 21 hours, test 7, pass 13 and 25 if that means anything to anyone.
Click to expand...

MrM said:
How would OCUK qualify the RMA? I dont want to be told theres no problem with the ram as it was after 21 hours that I found an error?
Click to expand...

It wasn't in the 21st hour that it developed the fault. The first one occured in pass 13 (I'm guessing here but approx ~4hrs) so Overclockers would pick up on the fault 'if' it replicated the same error at the same stage - but that would only be 'if' it is the memory at fault. If you decide to rma the memory you could request that the memory be tested for x amount of hrs...

And no, it shouldn't give up any errors - but unfortunatley it doesn't conclude that the memory is at fault - it could equally be one of the DIMM slots or the memory controller - or other. Yu will need to further testing by alternating the memory in different slots.

What cooler have you got stuck on top of the CPU?
 
Last edited:
Cooler is Scythe Shuriken big quiet low cooler

I set memtest going at around 8pm, and at 6am there were no errors, so definately after 10 hours. Will test each stick in different slots. 2 reiterate, errors only on test 7 and only 2 errors. Windows diagnostics memory test finds no errors. Timings set manually as per spec and voltage at highest recommended value.

Cooler makes it difficult to get access to nearest slot to cooler, and heatsink itself was a pain to fit.

Will test hdd drive when external drive arrives
 
Until you get the new HDD testing each stick individually is the logical next step - although it may be long drawn out process if it replicates the same time pattern :/

The idea behind the next question is a long shot - when you fitted the cooler did you notice any considerable bowing in the MB?
 
Last edited:
It was a pain to fit, what is your thinking?


So 9 hours in memtest and no problems on the one stick so far...
 
MrM said:
It was a pain to fit, what is your thinking?
Click to expand...

It doesn't sound relevent as you didn't mention any bowing - and as mentioned it was a long shot.

Sometimes, when fitting a cooler (usually heavy duty or push pin), the resulting bowing in the MB can cause the DIMM slots to become unstable - usually the closest ones to the socket. However, usually, the problem presents itself far more brutally - i.e. system wont post, or you have errors almost instantly in memtest etc... So it's looking very unlikely that this could be a factor in your case.

But, if you did have considerable warping after fitting and you continue to have inconsistent errors in memtest then it may be worth keeping in mind for future testing...

MrM said:
So 9 hours in memtest and no problems on the one stick so far...
Click to expand...

I'm hoping that the next stick starts giving up errors - as it would be an easy fix...

Keep on plodding on - it's a PITA i know.
 
If the current stick on test produces no errors, I'll assume that its good, and that the DIMM slot is good. I will then use this stick in the other DIMM slot to check that.

Assuming all good, I can assume the other stick is bad, but will look to verify that on test.

External HD should come tomorrow or Monday, so will test that then.

With regard to bow, no, I do not have a permanent deformation. As with my experience in building PCs, there always seems to be one corner on a mobo that doesnt have a chasis screw under it, and inevitably when installing the cooler you get flex down in that corner, but its not lasting.

I also switched over the SATA lead that I had to the SSD. Im stiull puzzled as to why my mobo cannot remember that when in AHCI mode, the SSD is the primary boot device. Every time from a cold start it seems to forget this.

Also, without AHCI mode enable, the machine will ignore the memtest usb stick... very bizarre.
 
MrM said:
Assuming all good, I can assume the other stick is bad, but will look to verify that on test.
Click to expand...

Only until the other stick has tested as such (as in fails memtest) - but sensibly you're going to verify this.

I never assume anything with computers - even if the error initially seems obvious, sometimes it can be something seemingly unrelated...

MrM said:
Im stiull puzzled as to why my mobo cannot remember that when in AHCI mode, the SSD is the primary boot device. Every time from a cold start it seems to forget this.
Click to expand...

Do you lose any other setting that you have changed manually - if you do try changing the battery.

Additional: This has probably been mentioned before but is the BIOS upto date?

I'm hoping the 2TB HDD connected up on it's own will help clarify a few things.
 
Bios is up to date and was from factory. No other settings affected. Have changed cable so will see if that's the fault causing the boot order issue.
 
MrM said:
Ok, so a quick update for you. Cant remember what happened last night, if I recall the machine crashed for some reason and I opted to let Windows attempt a repair when given the option on boot. Wnd forward a few mins and Windows tells me it is unable to do a repair, and then wont boot and comes up with anothr BSOD.
Click to expand...

Hi mate, i dont want to off your original topic too much, but i think i have been suffering a similar problem to yourself.

I built a system last year with the following spec:

I7 950
Gigabyte ga-x58a-ud3r
6 GB Corsair XMS 3
(This was one of the pre-overclocked bundles from Overclockers)
Gigabyte Nvidia GTX 580
WD Raptor 74 GB O/S Drive
Seagate Barracuda 1.5 TB Storage drive
XFX 800 Watt PSU
Windows 7 64 Professional


Whilst playing Arma 2 last week, the system suddenly rebooted after a BSOD, then entered startup repair, and did this 3 times before finally saying this computer coud not be repaired automatically, so i was basically left with a dead machine at that point.
After a complete format and re-install of Windows 7, the system ran ok for 2 days, then one morning when i went to turn the system on after using it and shutting down normally the previous night, it entered the repair loop again.

I then reinstalled Windows 7 again, but this time i replaced my 6 year old WD Raptor 74 operating system drive with a new Hitachi deskstar as i was suspecting the old drive was possibly nearing the end of its useable life, and causing errors and data corruption.


After 2 days using the system with the new Hitachi unit, the pc suffered anothe BSOD crash, and yet again entered the startup repair loop. I have however found a solution to fixing the system and allowing me to get Windows working again without having to carry out a ful re-install, but for the love of me cannot figure out what is causing the crash and boot up issue in the first place.

I may create a thread on this of my own, but if you like i can give u the fix for the startup repair loop to save u having to re-format again, as in y case it does not solve the problem. :(
 
Last edited:
Ok, so another update for you.

Lets call it stick 1, survived a 24 hour memtest in each DIMM slot, so assming this stick is good and the mobo is ok. Put stick 2 in lets call it slot 1 for arguments sake, after 10 mins I had lots of errors on test 7, so I thought Id nailed it. Took stick 2 and put it in slot 3 and no errors after a couple of cycles, which dismayed me a little. So I put it back in slot 1 and has been going for 4 hours with no errors, so have to admit, Im a little confused.

The fact it was test 7 that failed seems too coincidental, but Im wondering if I hadnt seated the ram correctly as its not been instantly repeatable. So the memtest test gooes on.

Just to point out, on the original memtest error, I only got 2 error messages after 24 hours. When stick 2 was in slot 1 earlier I had about 80 in a matter of mins.....?

External HDD not arrived yet either.
 
I'm getting a feeling of de-ja-vu - see link - and in that epic thread the problem 'eventually' turned out to be the CPU.

I'm not saying that that is your problem - but it is becoming increasingly as disparate.

I'm inclined to blame stick number 2 as i've had memory problems in the past that were inconsistent with testing. It was only when i tried a different set of memory that i was able to prove it beyond all reasonable doubt. Problem for you is that you don't have another set to test... (You could always order another set with the intention of keeping them if it cured your problem - but if they didn't return them using DSR.)

Alternatively, perhaps you could run the system with stick 1 for a few days and then stick 2. If you get BSODs only with stick ,2 in the same DIMM slot, then that coupled with the 'ambiguous' memtest results maybe about as conclusive as you'll get.

That is if the HDD doesn't solve things...
 
Plec, you took the words out of my mouth. Proving without doubt that the issue is the mobo or cpu is going to be difficult I think. In thr back of my mind, I am wondering if the memtest errors were due to the ram not quite being seated correctly. I stopped stick 2 in slot 1 after 18 hours, as it had done 50 test loops and no errors. Now on stick 2 slot 3 and been going for 2 hours no problems. Assuming it makes 50 loops I have to assume memory and DIMM slots are fine.

I will be able to rule out the HDD easy enough when the backup drive comes, and Plec, I thank you for keeping an interest in this, somewhat keeping me sane in this tedious process. Machine was prime stable for as long as I tested it, so other than that, I am not sure how I can isolate the cpu. One thing lingering in my mind is this whole bios forgetting the boot order when AHCI is selected. I cant work out why...
 
Yet another update. Stick 2 in slot 3 survived memtest for a day, so, aside from running memtest 24/7 ad infinitum, I have to draw a line somewhere. I am now assuming the ram and mobo DIMM slots are good. The HDD test will commence next week. Have been running Prime 95 now for 5 hours and no errors......
 
MrM said:
Plec, I thank you for keeping an interest in this, somewhat keeping me sane in this tedious process.
Click to expand...
Not a problem – i just wish there was a faster process for you but without spare components it can feel like you're treading water in one of Dantes circles...

MrM said:
Yet another update. Stick 2 in slot 3 survived memtest for a day, so, aside from running memtest 24/7 ad infinitum, I have to draw a line somewhere. I am now assuming the ram and mobo DIMM slots are good.
Click to expand...
Seems the most logical conclusion with the facts you have to hand. You’ve been more thorough than most in your troubleshooting efforts. I too think the memory has at least proved itself, for now - but i’ve always found memory to be very duplicitous, in the past, so i only ever completely rule it out when another component has been identified as the cause.

MrM said:
Have been running Prime 95 now for 5 hours and no errors......
Click to expand...

Stress testing apps (prime 95) are great for general stability testing and are ideal for identifying most aggressive faults quickly.

But, like a 'presumed' stable overclock which can prime for 48 hrs straight yet will still unexplainably BSOD in light apps use – you can, likewise get a rogue new build that will punish all stress test apps yet will buckle at the thought of browsing the internet while playing a tune through media player.

The only analogy i can think of is it's a bit like an F1 driver who's done 65 laps at the limit of the car's capabilities and slows to save tyres and then inexplicably throws the car into a tyre wall... Unfortunately, you may have a Kimi.

MrM said:
The HDD test will commence next week.
Click to expand...

The HDD will hopefully help clarify things either: identifying your SSD and 160GB as faulty/compatibility issues - or you’ll be restraining yourself from hurling the rig against the nearest supporting wall as it may replicate the same problem :/

In the mean time, if you don’t have a known working psu to test in the rig, you could try testing the PSU. It’s possible that this could be a contributing factor – but wouldn’t explain everything. And unfortunately you wouldn’t be the first person to have received two faulty components in a new build (myself included.)

Below are links to Huddy's site which has a step by step guide on how to test your PSU using a volt/multimeter. You seem very dedicated in trying to find the problem and although this isn’t a definitive PSU test it would at least rule out fundamental issues.

'Use a multimeter to check your PSU' and one to 'How to check if your PSU is faulty'

The process is fairly easy and certainly less time consuming than your other tests.

Good luck!

Additional: Hope the above is intelligible - couldn't sleep so decided to post - i'm not prime stable myslef at this time...
 
Last edited:
Ok, so yet another update for Plec et al. Given the Windows installation issues I have had, with Windows kicking out mid way through unpacking and expanding files, I downloaded aj identical copy of my oem disc, to see if that made any difference. Installing Windows on the SSD when I can get it to work is not a long task, so worth a shot I thought. Same problem.....

So with files backed up onto the external drive I have installed Windows on the larger, slower HDD. Everything seems fine, so problem on install, no random crashes since, no complaints from Windows, nothing, so a huge improvement, now again its early days, but I need to make a conclusion as some point as this was not what I was aiming for in my build.

Other points to note. I have downloaded Win 7 SP1 onto dvd and I cannot get it to install, it comes up with an error, and does this regardless of which drive I had Windows installed on. That said, when the OS was on the SSD, it would frequently have issues downloading Windows updates. On the HDD I have not had a single issue.

I have this morning been able to update the firmware on the SSD now it is not the drive with the OS on, but to be frank, not being able to update the firmware of the drive if it us your primary is far from ideal.

Now here is a general question, on the OCZ drive there is a folder called boot, and a load of language subfolders in it, which amongst other things includes a file called memtest. WhenI have tried to install Windows in the past, I have been unable to format this small partition, its always been protected in some way. Do I need it, and what does it do? If I can delete it via Windows should I?

At this stage it certainly seems like a SSD issue, but an issue that may be extremely hard to identify in an RMA. Perhaps the new firmware will have solved the issue, any chance you think?

With regard to the psu, I suppose I could test it, but in fairness I went a little OTT with a Corsair premium modular premium unit, forget exact model, 850W if I recall, which should be easily able to cope with my punitive system.
 
You must log in or register to reply here.
Back
Top Bottom