• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Memory keeps failing

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

rjohannesson

New Member
Joined
Apr 8, 2007
Hello all,
I am having some issues with my memory sticks failing after about a month of use. I am using the normal testing methods to isolate the problem (memtest, prime95, isolating stick, swapping slots, etc) and have isolated the problem to one stick. The thing is, this is the third stick that has failed.

I have two questions:
First, how common is it for memory to fail after a month (three times)?
Second, could it be the mobo causing a memory stick to fail?

I would appreciate you feedback.

Lemme list some details about my system:
mobo: ASUS p5b Deluxe/Wi-fi
memory: Crucial Ballistix PC2-6400 (i've used both SPD and 4-4-4-12 timings)
video card: Sapphire/ATI X1950XTX
PSU: Hyper 550W
Overclocked?: NO - I'm afraid to do this much until I get this memory thing fixed.
System Temps: 34 to 44 C, depending on room temp
CPU temp: 39 to 56, full load
 
how much voltage are you giving them ?? i think with some of those asus boards they undervolt the dimms allot at stock bios settings.. and if its happened more than twice.. i would be very wary of that motherboard, or id at least use dif ram slots..
 
Thanks guys,
It is definitely heat related - the stick passes memtest - it runs for hours. The only way I narrowed it down to one of the memory sticks is that Prime 95 fails within 15 minutes (Large FFTs) for the bad one, the good stick will run forever. I'm gonna go ahead and replace the stick if Crucial will do it a third time. When I get the new one I'll test it/them both thoroughly - If one of the fails again then I'm going to try to get Asus to take my board back. BTW where can I get a RAM fan?
 
jordon said:
how much voltage are you giving them ?? i think with some of those asus boards they undervolt the dimms allot at stock bios settings.. and if its happened more than twice.. i would be very wary of that motherboard, or id at least use dif ram slots..


Can you give me the name of good (free) utility to tell me what my RAM voltage is? I have CPU-z, but don't see it on there. My BIOS won't tell me unless I enable AI.NOS and set it manually.

Thanks
 
rjohannesson said:
Can you give me the name of good (free) utility to tell me what my RAM voltage is? I have CPU-z, but don't see it on there. My BIOS won't tell me unless I enable AI.NOS and set it manually.

Thanks

You need to find the spot on your board and check your vdimm with a DMM. Neither your BIOS or CPU-Z will give you an accurate reading.
 
It's not the memory - I'm pretty sure

So I got the new sticks from Crucial in like two days; they issued an RMA no questions asked (Crucial rocks) . I stuck them in and did a stress test and it failed just like last time. after the system warmed up. I did Memtest overnight without a failure. Since I've replaced a total of 4 memory sticks without a change in the stress test results I'm now convinced that it is the motherboard.

Now the question I have is: Do I need to do any more diagnostics or should I just go ahead and get a new mobo?
 
What test are you using? Run orthos small FFT to test cpu, run hci memtest for mem. I would make sure you know its the board before getting a new one.
 
Eldonko said:
What test are you using? Run orthos small FFT to test cpu, run hci memtest for mem. I would make sure you know its the board before getting a new one.


I used MEMTEST98 for testing memory (ran all night without failures) and small FFTs for testing the CPU. The CPU test ran without failure overnight. What led me to believe that the memory was bad the last time was that when I ran Prime95/Orthos Large FFTs (which tests memory and a little CPU) I get failures.

Do you know of any other tests I should run?

Thanks
 
CPU temp is about 12-15C above system temp during the small FFT tests I think. My system temp ranges from 36 to 45, depending on how hot the room is.
 
Similar discussion here as well

http://www.xtremesystems.org/forums/showthread.php?t=139544

Eldonko said:
run hci memtest for mem. I would make sure you know its the board before getting a new one.

Ive run two instances of this program for 30minutes on a pair where I'm really really sure that one is bad ...they both passed...


The fact that rjohanssen and dumpalump also have a P5B is a bit disheartening but I dont feel like it's totally time to blame the board...
The bad stick clearly had a warped heatspreader with no contact with a ram chip.

I just ordered a new case and bought a cheap 1 gig stick as well... I will give the Ram one more RMA shot before i switch boards... I'm also getting a DMM
 
Last edited:
Wow, It's interesting to know that some others are having the same problems. I read the other threads and it looks like some people had bad heat spreaders on the RAM not touching one of the chips. I noticed that my latest set of Ballstix now has a new type of heat spreader - wonder if that 's the reason?

Others also have had problems with the RAM after two or three months - this happened to me on the first two sets, but the set I have in there now seems stable enough to run very intensive games, but fails orthos large ffts within 20 minutes. The difference this time is that it is happening right out of the box, without waiting 2 months.

Anyhoo, I have really done nothing to the vdimm settings. I have the vdimm on "Auto" in bios. I will now try setting the Vdimm to 2.2 and see if that makes a difference.
 
Well,
Since I changed my Vdimm setting from AUTO to Manual and set it at 2.2V, my system has been rock solid. Orthos ran for 24 hours on Large FFTs.

I had set it on Auto when debugging my last set of ram and never set it back.

So I guess now I'll wait another couple of months to see if these sticks go bad....
 
Back