RAM Error in Memtest86+

[alderosa]

Hi, I am running Memtest86+ after some BSOD crashes on my system (I was told by someone else on this forum that the crashes were due to graphics drivers which I have now changed). I know that this sort of problem is also commonly RAM related so have been testing the sticks. I got a single error in Memtest with all 4 sticks of RAM on the board and now I am currently running the test with individual sticks to try and find the source.

Given me previous BSOD problems I should probably RMA any defective RAM but I was wondering if the RAM voltage and settings could be the source of the single error? Also, my BIOS is not the latest version. I have a Gigabyte GA-880GA-UD3H Mobo and the BIOS version is F4 while the latest version is F7. Could upgrading to the new version get rid of the error?

Alex

P.S. I will post back if I get more any Memtest86+ errors but I guess this will mean I need to RMA the sticks. The link to the previous post is;

http://www.sevenforums.com/crashes-debugging/129954-bsod-again.html

 

[SledgeDG]

You are supposed to test single RAMs with that test (it also says so in the readme) and let it run through 4, 5 times not just once
If the refresh rate is set too low you usually get more than 1 error (because the ram "forgot" the information already when it's due to a refresh. with the wrong voltage it's being unstable which also shows you more than just 1 error. Just set it back to AUTO or check the manufacturer's page for the appr. settings

-DG

 

[alderosa]

Yeah, I'm running it with each stick at the manufacturer settings atm. I have 4x2Gb ADATA Gamins Series DDR3 1333G running at approx. 1.7 volts and 8-8-8-24 timings. I thought you could test multiple sticks with memtest86+ though first to see if there was a problem at all with any of them then try to narrow it down? I was following this guide...

Memtest86+ Guide/How To's - Overclockers Forums

 

[alderosa]

I've also been running 5 passes on each.

 

[SledgeDG]

Sure you can run it with multiple sticks but the result is not as reliable as if you check one stick at a time.
That was the guideline when memtest was published first. But since you already had an error there it's moot to worry about that.
BTW not sure if your walkthrough mentioned that anywhere) if all sticks come up faulty test them in different slots to make sure, it's the RAM that's faulty and not the slot

Well 5 passes should give you an overall idea.
If one fails just test that one longer/harder

-DG

 

[alderosa]

Yeah man, thanks. Currently looking good for 2 sticks in slot 1. I'll let you know how I get on...

 

[SledgeDG]

Please do
g'night

-DG

 

[alderosa]

We have a winner. Defective RAM. Looks like I'll have to RMA. I'm still gonna check the other slots on the board though. I guess this was also probably the cause of the BSOD in my other post (link at top)?

 

[alderosa]

So I changed the RAM over to using 2 sticks which passed 6 Memtest86+ runs each. I also tested each Mobo slot and all ok. System crashed to BSOD on first reboot. Then when restarted, it went to BSOD when I tried to shut down. I attached the dump files for this recent problem here...

http://www.sevenforums.com/crashes-debugging/129954-bsod-again.html#post1120814

 

[SledgeDG]

So it's not because of faulty RAM...dang that would be it
Please post that in the BSOD section. There are the experts for those errors.
I'm not sure if they see that here and I'm not trained to deal with those issues.
I wouldn't want your post "go under" here.

Oh I see you already did
-DG

 

[alderosa]

Could it be RAM timings and voltage? I've read a few things here and there about this RAM I have (ADATA Gaming Series 1333G 2Gb) being pretty voltage hungry and fussy about stability. I've just set the timings from 8-8-8-24 to 9-9-9-24 and upped the voltage to 1.8v. Hopefully this helps... would I need to change the northbridge voltage or VTT too?

 

[SledgeDG]

If I'm not mistaken those RAMs are really power hungry...they should have the range printed on the side...up to 1.85V. I wouldn't touch the north bridge tho (unless it's really underpowered) or you risk frying your cpu and/or board.
Overclocking shouldn't be an issue until the BSODs are solved. Until then everything should run with default/clock settings (except your RAM of course
-DG

 

[alderosa]

I haven't tried putting the voltage up this high yet. The range is up to 1.75 but I've heard similar so I'll probably go higher now. I've updated the BIOS too. Is it just the direct RAM voltage I need to change then? Should I leave the timings at mobo recommded (9-9-9-24) or manufacturer recommended (8-8-8-24)?.

 

[alderosa]

My RAM is now at DRAM voltage 1.8v and timings 9-9-9-24, I haven't been able to alter the VTT voltage though (it's either DRAM voltage or VTT, not both) so my VTT voltage is stuck at 0.75v. Is this going to be a problem? I read somewhere that they should be within 0.5v of each other but I can't change it.

 

[SledgeDG]

As for the timing go with the mobo..the manufacturer says you could go faster but that's no good if the mobo doesn't support it
Well if you have to alter the RAM voltage then do that. I trust you tried with default voltage and ran into an error, right? (just so I understood that right) If not leave it at stock voltage until BSOD is solved. It's generally a bad thing, to alter too many variables at once while on the hunt for errors.
So change just one thing, test if the error is still present and go from there.

-DG

 

[alderosa]

Yeah, it crashed on the default settings. I've been talking to a couple of people on here about the BSODs I had and examining the dump files. I'm currently waiting for a reply on the latest ones. At the moment I'm on mobo recommended settings and increased voltage (manufacturer says 1.55 to 1.75V & I'm using 1.8V). Like you said, it seems to need more voltage. We'll see how it fares over the weekend. I'm just off to UPS to RMA my defective RAM sticks now too. Might get new stuff end of next week... maybe...

 

[SledgeDG]

Those guys sure are capable so you can expect the issues being resolved before your RAM comes back
I'll keep my fingers crossed.

-DG

 

[alderosa]

Just had another BSOD on first startup this morning. This is really starting to be a pain...

 

[SledgeDG]

Have you heard back from the BSOD experts yet? I fully agree with you it really starts to be a PITA.
Not to mention all the frustration and wasted time. I just hate when that happens
-DG

 

[alderosa]

Not yet, I know it was something to do with ntoskrnl.exe from BlueScreenView though. Is it possible this is still RAM related or should I start to look elsewhere?