New build problems - faulty GFX and RAM?

Soldato
Joined
12 Jun 2004
Posts
5,475
Location
Exeter
I build a new PC about 5 days ago with these components;

Intel C2D E8400
Asus P5Q Pro
OCZ 4GB (2X2GB) PC2-8500C5 Platinum RAM
EVGA GTX260-216 SSC
WD Blue Caviar 640GB HDD
Enermax Modu82+ 625W

Basically sometimes when playing games such as CoD4 or Crysis my PC would crash or the screen would become distorted then restart.
It first happened in CoD4 when the screen went a solid grey colour and I had to restart my PC twice before I could see the screen and boot into Windows.
The exact same thing happened in Crysis, so I know it's a hardware issue.

I have reseated everything except the CPU, checked cable connections, temperatures, etc.
I ran memtest86 and my RAM always gets errors. Here are some pics;

Both sticks;

71pl5y.jpg


One stick;

kbziv4.jpg



And this is what just happened 15 mins ago when I was playing CoD4;

vctkba.jpg


The screen went all distorted then the PC restarted itself.

Does it looks like both the RAM and GPU are faulty? The RAM settings in the BIOS are all as they should be (DRAM freq; 1066MHz, Timings; 5-5-5-18, DRAM Voltage; 2.1V) so it shouldn't be failing memtest.

I suspect the EVGA graphics card can't handle it's pre-overclock and is overheated or something. The annoying thing is that it doesn't happen very often, so if I RMA it I expect the shop will test it for ~ 1 hour then say it's fine and charge me then send it back. :(

I'm not sure what to do, any help would be appreciated.
 
First of all remove any overclock on the cpu...

Then i'd reduce the ram speed down to 800 or even lower, that way your ram wont be working so hard... It might be giving errors due to dropping packets when running too fast...

Your gfx may well be fine, i'd suspect ram for the moment.
 
First of all remove any overclock on the cpu...

Then i'd reduce the ram speed down to 800 or even lower, that way your ram wont be working so hard... It might be giving errors due to dropping packets when running too fast...

Your gfx may well be fine, i'd suspect ram for the moment.

Thanks for the reply.

The CPU and RAM are both running at stock speeds.
 
I posted a pic of the memtest errors on the OCZ forums and they think the RAM is not defective, but I need to change some BIOS settings to get it to work. I'll try it soon.

It just happened again in CoD4;

2ez0hee.jpg
 
I ran memtest all night with the changes in the BIOS that the guy on the OCZ forums suggested. Memtest didn't show errors until ~6 hours into the test. When I woke up it had 50 errors. The guy on the OCZ forums says it is not defective.

I don't know what to do now... I need to rule out faulty RAM but this support guy from OCZ seems certain it is not defective.
I know it's either my RAM or GFX but I don't know which. :mad:
 
Imo, if the RAM is running at stock speeds with the correct basic settings (e.g. if it's set to default to the SPD values) and you're getting errors, the only thing you can check is the RAM voltage. You could try increasing this by 0.1v and see if that helps - but make sure you keep the voltage within the manufacturers specified range.

If you're still getting errors then, whilst you may be able to get it stable by fettling some more, I would personally RMA it as faulty.

GPU is most likely fine imo, it's clear there is a problem with the RAM so until that's sorted I wouldn't go fishing for other problems that may or may not exist.
 
Imo, if the RAM is running at stock speeds with the correct basic settings (e.g. if it's set to default to the SPD values) and you're getting errors, the only thing you can check is the RAM voltage. You could try increasing this by 0.1v and see if that helps - but make sure you keep the voltage within the manufacturers specified range.

If you're still getting errors then, whilst you may be able to get it stable by fettling some more, I would personally RMA it as faulty.

GPU is most likely fine imo, it's clear there is a problem with the RAM so until that's sorted I wouldn't go fishing for other problems that may or may not exist.

The RAM is definitely running at stock speeds with the correct settings.

DRAM freq; 1066MHz
DRAM Voltage; 2.1V
Timings; 5-5-5-18

I will try increasing the voltage a bit more as it can go up to 2.2V. If it fails again tonight I think I will just RMA it.
 
Yea I would, it seems you know exactly what it should be running at and you've set it accordingly. If it doesn't work with the specified settings, then it's likely that you may have a compatability issue or a defective product (despite what you've been told on the OCZ forum!). Be safe and RMA it :)
 
Have you tried running it with one stick of ram at a time?

I've tried one stick on it's own and weirdly enough it got even more errors (over 2000) than with both sticks. I haven't tested the other stick yet.
I am now pretty certain that the crashes are caused by the RAM because I just put everything in the BIOS back to AUTO and it hasn't crashed once. It's only running at 800MHz in AUTO so I guess it's unstable when at 1066MHz.
 
It's common to find only one faulty stick - make sure you check the other one in the same RAM slot, and if there's any doubt check each stick in a different RAM slot also. It may become very clear what the problem is if you check the other stick and get no errors :)
 
I will check the other stick tonight with the same settings.

Last night I ran memtest with everything in the BIOS at AUTO (causing the RAM to run at 800MHz) and it passed with no errors... It also hasn't crashed once since changing everything to AUTO.
So it seems this RAM just doesn't like running at 1066MHz which is what it should be running at.
 
Sometimes the Auto settings don't provide for the rated latencies etc - the SPD sets them to slightly slower settings to provide more chance of getting a flakey system to boot. This allows the user to fine tune the settings in BIOS afterwards.

I would go back and change nothing apart from RAM settings, perhaps even one at at time and then memtest it (tedious job, I realise!). try to figure out what causes the crashes. If you're confident that it won't run at the rated speeds (with any combination of sticks installed) at the rated voltage then I'd RMA it ASAP.
 
ive been getting the same problem with a new build.
everytime i run 3d applications in fullscreen i get a distorted graphics lockup, happens with games, direct x test and the run automated clock configuration utility on the ati drivers.

funny thing is everything was working okay at first for a few days then i started getting the problem. another thing sometimes a game does by chance actually work in full 3d then maybe 3hours later the same error will occur. games run okay in windowed mode so generally ive not been running into the problem or trying to sort it out as im playing games in windowed.

things ive tried
-updated mobo bios
-different gfx card drivers (cats/omega/saphire cats on cd)
-different psu
-low underclocked settings
-mem passed memory test

things to note
-i bought a psu from mm but it got returned to sender, sender forgot to put his address on it so its still in the offices where its gonna be for a while id expect. bought some cables from ocuk to run the new setup on this old 400wt psu. i was having problems getting the system up and running prior to these gfx problems, bought a 500wt psu from pcworld which i took back at a later date as this wasnt the problem. problem was couldnt have bios set at auto for the asus p5n-e sli, had to configure manually.

-gfx errors on both psus, so i took 500wt back and am now running on this 400wt. i was using ati tool and it strangely started to report the voltage regulator of the gfx card was reaching 105 degrees (was reporting normally for first few days). uninstalled and started using riva tuner which reported temps correctly.

i aint got a clue whats the problem. gfx card? psu? memory? mobo? combination of more than 1.
i dont know whether ive stressed something i.e. the gfx card, from using this lower than recommended psu, or if the ati tool temp readings messed something up.

i can still get the error in windowed mode, but only if i whack gfx up to full, if i use lower res and medium gfx for example games will run fine for as long as i play for.

maybe the psu isnt giving sufficient power to run the games at full settings. but maybe the gfx card or something else does actually require replacing.

SDC10045.JPG
 
Last edited:
Sometimes the Auto settings don't provide for the rated latencies etc - the SPD sets them to slightly slower settings to provide more chance of getting a flakey system to boot. This allows the user to fine tune the settings in BIOS afterwards.

I would go back and change nothing apart from RAM settings, perhaps even one at at time and then memtest it (tedious job, I realise!). try to figure out what causes the crashes. If you're confident that it won't run at the rated speeds (with any combination of sticks installed) at the rated voltage then I'd RMA it ASAP.

I am certain whenever I run the RAM at the speeds advertised it causes the PC to become very unstable, so for that reason I think I'll just RMA.

Another support guy on the OCZ forum (the support manager) has actually given me RMA instructions now and he said it should be replaced as something is not right. I think I'll do it straight after Christmas.

Does anyone know how long it usually takes to RMA to OcUK?
 
I am certain whenever I run the RAM at the speeds advertised it causes the PC to become very unstable, so for that reason I think I'll just RMA.

Another support guy on the OCZ forum (the support manager) has actually given me RMA instructions now and he said it should be replaced as something is not right. I think I'll do it straight after Christmas.

Does anyone know how long it usually takes to RMA to OcUK?

Yeh defo RMA that badboy. Mine was pretty quick.. like a week or something.
 
Back
Top Bottom