Memory issue (?) Very mind-numbing...

Ok so I have to give the full story on this otherwise it just leads to an easy answer... :pfff:

I've had my comp stable for a pretty long while, started getting some BSDs.
Reformatted, was ok for a while then started getting more BSDs.
Reformatted again to no avail, actually got a BSD DURING the install period and had to try it again. Got it to install after the second try.
Did memtests and everything, nothing checked out to be wrong.
Reset my CMOS. RAM errors like crazy from memtest 86+. Noticed my motherboard autoclocked my RAM to 800 and 4-4-4-something timings. Put it back up to 1066 @ 5-5-5-15 and reran memtest. No errors.
Ran a western digital scan to see if there was something wrong with the HD and got error 0266, Gutted the drive and got a new one
Still having BSDs (wut).
Ran the WD scan again, no errors on the new drive.
Started to put the things that I use normally back onto my new HD overnight (WoW and Aion at this particular time) - froze overnight with my monitor being asleep or whatever you want to call it. Wiggled the mouse, HD lit for a second or so, then stopped dead. Screen wouldn't display any image, no HD activity.
Hoped it was just some sort of random mishap, the next night I attempted the same thing and woke up to the exact same issue.
Button-booted (due to the step above) into a BSD, button-booted again into a C1 error on my 750i FTW board. Rebooted several times, it got to different parts of the rebooting process but never could load windows. Sometimes getting to the Microsoft bar of death and then freezing. Sometimes only getting to the first line of BIOS text.
Took out two memory sticks (out of 4) after trying to reboot several times. Booted fine, re-inserted the possible bad sticks, managed to boot after a few tries.
Ran memtest and OCCT. OCCT overnight -> woke up to a frozen screen @ pass #7 @50%
Took out the "bad" sticks.
Ran memtest and OCCT on the current "good" sticks, they checked out.
Put the "bad" sticks in... booted fine (???)
Ran memtest and OCCT on the "bad" sticks... they checked out as well... (???)
Now sitting here stable for at least 3.5 hours playing Aion with maxed settings breezing through everything (lol...)

@ the last step of this... This basically happened every time I got bluescreens. It would be very very persistent one day, maybe the next but then be absolutely fine for the next week..

So, either I have gotten two bad drives in a row from WD (unlikely, especially since WD tests wouldn't detect any errors on the new drive)
Or my motherboard is bad (I really hope not...)
Or my RAM is acting up (Easiest to fix. PLEASE be the RAM)
Or my voltages suddenly decide to not let 4x1 GB of RAM work @ their current settings (auto everything except the RAM settings which I've set to factory defaults)

Voltages
CPU VID....1.21V
CPU FSB....1.10V
Memory.....2.11V
3.3V..........3.26V
3.3V Dual..3.26V
12V...........11.98V
5V.............5.06
Vbat..........3.10V

nForce SPP (Auto)...................1.40V
NF200 Voltage Level (Auto)....1.20V

I've never ever had the issue of my computer freezing into a blank screen like that. Once I tried to install the programs overnight again I was furious because I thought I had a bad drive. I just finished installing WoW with the 2 sticks overnight and I did not have that issue. I am going to let my computer run overnight again to see if I can recreate that HD light going on and stopping cold again with 4 sticks of RAM in, if not this computer is confusing the living HELL out of me as to why the errors are so random and hard to recreate!
^ Woke up with a C1 error after trying this.

I've correctly installed my OS/drivers etc. There is no way there is anything wrong with the software... Also heating is not an issue either, checked the temps, and I have a very fair amount of cooling fans/good ones.

750i FTW board by EVGA
GTX 285 GPU by EVGA
e8400 Wolfdale @ stock settings
2x1GB Corsair Dominator RAM (Two sets of these for a total of 4GB) http://www.newegg.com/Product/Prod [...] 6820145043
VelociRaptor Hard Drive by WD http://www.newegg.com/Product/Prod [...] 6822136296
Antec TPQ 850

Message edited by c1o5ry1991 on 10-09-2009 at 07:19:13 PM

Add a reply

Register or log in to remove.

OK, my mind is officially numb now.

Sounds like you've done great diagnostics on your system - covered everything I can think of to narrow down the problem. But you've never mentioned one item I'd suspect right away!

First of all - the HD light when screen power off, don't worry about it. The computer is already locked up and wants to show you the BSOD when you wake it, but it can't. The HD led is just part of its feeble attempt to bring the screen back up.

You don't have bad RAM. The memory test didn't fail, didn't find any errors - it just locked up in the middle of the program just like everything else does. In addition, if this happened with a test run from diskette/CD - like memtest86, then it also clears Windows as a cause of the problem.

Hard drives don't give BSODs that I'm aware of. The OS knows how to handle a bad drive and will tell you it can't access it.

What are we left with? Video card, motherboard... and power supply.

I would swap out the power supply. It's the easiest item to test and replace and can cause the problems you mention.

So could the motherboard. The video card - seems unlikely.

Two choices on replacing the PSU. One is to get a cheap tester model from a source you know there's no problem returning just for testing. Other is to order a good replacement with the confidence it's going to fix the problem!

------------------------------ Gigabyte MA785GM-US2H, AMD Phenom II x2 550 BE (4 cores@3.6GHz), G.Skill 2x2GB DDR2 1066 (820@CL4), Scythe Katana3, Hec 585W, Samsung 2232BW+, WD Caviar Black 1TB SATA, Sea 500GB USB, 12 IDE HDs w/Masscool IDE-SATA Convertors, Ugly Old Case, Win XP 32bit
Reply to mongox

LOL, let me make it worse for you. I just put in all the sticks again overnight to finish a WoW patch. Woke up to a C1 error. No bluescreen. No freeze, Just the error and a blank monitor.

Do you think I should up voltages somewhere because I'm running two sets of 2GB RAM kits? Right now I'm going to run the "good" sticks and just see what I get with those. Maybe my comp has an issue with running 4x1GB RAM of two kits? The only really stupid thing is why does it decide to act up now... Notice how OCCT only froze with 4 sticks in but not just the "bad" ones in. So maybe it's my motherboard or what i said before?

Message edited by c1o5ry1991 on 10-09-2009 at 07:20:19 PM

Reply to c1o5ry1991

No, I'd quit worrying about testing anything until I swapped out the PSU.

------------------------------ Gigabyte MA785GM-US2H, AMD Phenom II x2 550 BE (4 cores@3.6GHz), G.Skill 2x2GB DDR2 1066 (820@CL4), Scythe Katana3, Hec 585W, Samsung 2232BW+, WD Caviar Black 1TB SATA, Sea 500GB USB, 12 IDE HDs w/Masscool IDE-SATA Convertors, Ugly Old Case, Win XP 32bit
Reply to mongox

The only problem with that is ....
Return for refund within: 30 days

What if my computer decides to behaive for those 30 days. (trust me, it has done the impossible before). What other options can you give other than purchasing to guess-and-check hardware. Also that's quite possibly the worst part I could change with all the cable routing I have done in my case :[. I'd rather replace the motherboard if I have to.

Message edited by c1o5ry1991 on 10-09-2009 at 07:30:52 PM

Reply to c1o5ry1991

Swapping out hardware is, in fact, the most reliable and only way to diagnose and resolve many problems with computer systems. If you have a system that demonstrates repeatable errors and you replace one item and the errors stop, then you have solved the problem.

My impression of your post here was to obtain the best advice from those here. This is my best advice - replace the most obvious - and cheapest - item that is likely to be causing the problems.

If the PSU doesn't seem to affect the errors, then you have been able to cross it off your list and move on to the motherboard. And it's easy to return and even if you pay a restocking fee, you've spent the same as less then 10 mins of a technicians hourly rate. And what is the value of your own time?

As someone that has built perhaps 1000 systems, my advice is to replace the power supply and test it.

------------------------------ Gigabyte MA785GM-US2H, AMD Phenom II x2 550 BE (4 cores@3.6GHz), G.Skill 2x2GB DDR2 1066 (820@CL4), Scythe Katana3, Hec 585W, Samsung 2232BW+, WD Caviar Black 1TB SATA, Sea 500GB USB, 12 IDE HDs w/Masscool IDE-SATA Convertors, Ugly Old Case, Win XP 32bit
Reply to mongox

Check the stickers on the RAM. See if all 4 modules are the same revision.

------------------------------ Official Corsair Tech Support. If You Need Help, Just Ask.
Reply to XMSYellowbeard

^^ Doubt it, but a possibility. PSU test seems to be the next opiton at this point, just to be safe.

------------------------------

http://valid.canardpc.com/cache/banner/622598.png

Reply to gamerk316

No, it is definately a memorry issue. I think one or two of my sticks are shot. Although tests arent catching it I seem to only bluescreen and C1 error on the one pair of sticks. I'm going to let this run straight for 2 days with the "good" sticks in. If they hold up that has to be the issue.

Reply to c1o5ry1991

Well, confidence can be a useful character trait. I hope it works out.

------------------------------ Gigabyte MA785GM-US2H, AMD Phenom II x2 550 BE (4 cores@3.6GHz), G.Skill 2x2GB DDR2 1066 (820@CL4), Scythe Katana3, Hec 585W, Samsung 2232BW+, WD Caviar Black 1TB SATA, Sea 500GB USB, 12 IDE HDs w/Masscool IDE-SATA Convertors, Ugly Old Case, Win XP 32bit
Reply to mongox

What does loosening the memory timing and increasing the memory voltage do?

Message quoted 2 times

Reply to jsc

jsc wrote :

What does loosening the memory timing and increasing the memory voltage do?

+1 run it at 800mhz at 5-5-5-15 and see if it bombs..

also what is the recommended voltage for the ram.

If you do get new memory make sure you get 2x2gb chips

running 4 chips isnt ideal

Reply to rand_79

jsc wrote :

What does loosening the memory timing and increasing the memory voltage do?

LOL, is this a pop quiz??

I would say, makes the memory more stable and reduce chance of errors! But in this case, the memory hasn't ever failed a test.

------------------------------ Gigabyte MA785GM-US2H, AMD Phenom II x2 550 BE (4 cores@3.6GHz), G.Skill 2x2GB DDR2 1066 (820@CL4), Scythe Katana3, Hec 585W, Samsung 2232BW+, WD Caviar Black 1TB SATA, Sea 500GB USB, 12 IDE HDs w/Masscool IDE-SATA Convertors, Ugly Old Case, Win XP 32bit
Reply to mongox

Did you ever confirm that all 4 modules are the same revision? If they are not, they CAN cause the issues you are seeing.

------------------------------ Official Corsair Tech Support. If You Need Help, Just Ask.
Reply to XMSYellowbeard

Focus On

Forum Motherboards & Memory : Memory - Memory issue (?) Very mind-numbing...