Sign in with
Sign up | Sign in
Your question

3rd homebuild unlucky!

Last response: in Systems
Share
February 10, 2008 11:36:58 AM

Hi All,

I hope you don't mind my first post here being a request for help but you guys seem to know your stuff.

I just built myself a new system yesterday but I am getting frequent BSOD's pretty much always when gaming.

My Spec:

Intel C2D E6750
Gigabyte GA-P35
Inno 3D 8800GT
4 x 1GB OCZ PC6400
500GB Seagate Sata II
XP SP2 Home

And some other parts I salvaged from old PC:
570W Trust PSU
M-Audio Delta 2496 Soundcard

Everything seems peachy until I try out a game (Graw2 or COD4). I will go for between 10 - 30 minutes before I get a BSOD and restart.

I have tried swapping the RAM around and am currently using just 2 sticks which have passed a lengthy memtest with no errors. I have also tried several different sets of nvidia drivers (am currently trying the omega drivers).

After having this problem all day yesterday I reformatted my drive again with a clean install of XP but to no avail.

Looking at all my minidumps (with my very limited knowledge) I am seeing most point to nv4_mini.sys but several others point to ntoskernel.

Given my PSU was fine and powerful enough and I am getting the same problems after a fresh install I am starting to suspect my new hardware.

Does anyone have any tips on how to narrow down my problem?

I don't have a spare GFX card to try as this is my first PCI-E system, *sob*.

More about : 3rd homebuild unlucky

February 10, 2008 12:38:34 PM

Ditch the PSU. An old, no name psu is likely the problem. Get a quality psu. Also, what are the temperatures of the cpu and gpu?
February 10, 2008 12:56:47 PM

mmmmmm...sounds like something with yr mobo,..or something screwy with your memory!
Related resources
February 10, 2008 1:20:58 PM

Thx for replies.

I think the PSU looks like the first place to try before I try and RMA the gfx card or mobo.

Temps: CPU is around 32 - 35, GPU I haven't seen go over 53.

Are there any little apps that will help me determine it is the PSU, like a voltage monitor or something?
February 10, 2008 1:25:50 PM

how old IS the PSU? that's another question
February 10, 2008 1:37:18 PM

PSU is just under two years old i think.
February 10, 2008 1:39:44 PM

I say PSU also! Try Corsair,Silverstone,PC power & cooling...Go with a quality unite.
February 10, 2008 1:43:42 PM

Tried to edit last post, it wouldn't let me?

Anyway I have been running Sisoft Sandra to monitor the voltages and they all look pretty stable, although I haven't done anything very demanding yet. But the thing I have noticed is the power being drawn swings from 40 to nearly 60 watts seemingly randomly. Does that sound normal?

I think I'm gonna replace the PSU and see if that helps. It is the only bit of crap in my machine now (hopefully ;) 
February 10, 2008 2:20:58 PM

Try removing your Sound Card and it's Drivers.
Use OnBoard Sound as a test.

Alternately, Did you disable the on-board Sound in the Bios?
Did you uninstall those Drivers or was the card disabled in the BIOS when you install Windows?
February 10, 2008 3:16:38 PM

PSU maybe. Will it game ok with just 2 gigs? If it does then you probably need more voltage on your ram. Make sure you ram is running at manufacturer specs. Your mobo might of defaulted it to 1.8v and it might need 2.1-2.2v. Check it out.
February 10, 2008 4:56:32 PM

instead of stabbing blindly into the air based on hardware, how about you write down the BSOD's hex code and/or corrupted file/driver message and post it... then we have a better understanding of what's going on... if your BSOD doesn't stay on screen before the system reboots, you can disable automatic system restart...

I'd say it's a mix between bad PSU and bad RAM... you have 4 x 1GB sticks, and as roadrunner and I have discussed in a previous thread, the chance that one of them could fail is 4 times greater than the chance that 1 could fail...

I suggest running memtest 86+

keep us updated
February 10, 2008 5:12:43 PM

Thx for suggestions, it's really quite frustrating having just built what I was hoping would be a kick ass rig.
I tried upping the RAM voltages but with no change.

@ truromeo.

I am only running 2GB at the moment and have run memtest for 2 hours with no errors.
I will disable automatic system restart but from my noobish examinations of the dump files it is always 'IRQ not less or equal to" at memory address 0x0000000. The windows debugger thing points to different sys files each time. Sometimes ntoskrnl but my last one was to "kmixer.sys ( kmixer!MxWriteComplete+1ea )" which is an audio thing?

I don't really want to disable my onboard sound as I use it for headphones, but will disable it in bios for purposes of testing.

Thanks again for help.

February 10, 2008 10:56:04 PM

Well I accidentally switched my cpu to 1600mHz instead of 2600 in my bios without noticing while looking for an option to switch off onboard sound (which doesn't exist incidentally).
I then proceeded to play over 2 hours of crysis without a crash, although it was damn laggy.
When I realised what I'd done I put it back to 2600 and managed about 10 minutes before another BSOD which was a rather unhelpful DFS_FILE_SYSTEM.

Dunno man, any other suggestions before I stump up for a new PSU?
February 10, 2008 11:52:12 PM

In that case, it sounds like you may not have installed the CPU properly. check to make sure that you have the heatsink/ thermal grease applied properly and maybe even reseat it for good measure (you never know). Then, my next bet would still be the PSU.
February 11, 2008 12:57:07 AM

I would think that the way it acted with the lower CPU speed could also indicate the PSU is the problem. Higher speeds = more energy needs.
February 11, 2008 1:18:33 AM

Since it worked at lower speeds, it could either be the PSU not keeping up or a problem with the cpu. Run Prime95 while keeping the voltage monitoring program open, that way you can watch what happens to your voltage output. If prime95 errors and the voltage doesn't drop, you may have a bad cpu.
February 11, 2008 3:05:27 AM

now hold on guys, overclocking his CPU can be done by increasing the speed of his RAM... this points back to a bad stick, or a bad slot... or bad PSU or even mobo... could be that your chipset is overheating and causing the problem (especially if it's pointing at an onboard audio driver)...

I would remove your M-Audio card and try again
February 11, 2008 9:26:39 PM

Appreciate all the replies chaps.

The M-Audio card doesn't make any difference, still BSOD's regularly without it.

I tried Prime 95 (cool little prog!) for a bit over an hour @ 100%. No crash and CPU temp didn't go over 47 which I think is OK for an C2D?

I bit the bullet and bought a new PSU (s'OK I don't need food for the next couple of weeks LOL). Gonna fit it tomorrow and will let you know how it goes. Thanks all.
February 11, 2008 9:49:34 PM

Run everything on 1 stick of Ram...

Just because you pass Memtest doesn’t mean anything. Eliminate things 1 by 1.. do all this testing by removing what you don’t need and see if it runs proper, once it runs proper add stuff back in.

Run everything on 1 stick of Ram... I bet that’ll fix this.

**It is almost always the Ram**
February 11, 2008 9:54:04 PM

How about Bios? which one are you running?

Is this your MB? GA-P35-DQ6 (rev. 1.1)
If it is… here is a link to newest Bios http://www.gigabyte.com.tw/Support/Motherboard/BIOS_Mod...

I bet ya it is the Ram.

***Also since you just built this system, that suggests computability issues.. Bios update/ram ??
February 11, 2008 11:36:27 PM

Cheers Grieve. I have 4 sticks of 1GB OCZ 6400 that I have swapped continuosly with little to no difference in lack of BSOD's.

This is my mobo : Gigabyte GA-P35-DS3 iP35.

I updated the bios from f10 to f12 within an hour of finally booting my system into windows, again with no change.

I pray it is the PSU, otherwise it is between my CPU (which doesn't overheat and performs well), my GPU (which hasn't failed under stress tests) or my mobo (and fk knows how you test that???).

Thanks again for ideas all.
February 14, 2008 12:08:25 PM

Well, I finally replaced my PSU but am still getting the BSOD's. ARrrrrgggggh!!

Well, its gotta be either a gad Mobo, GFX or CPU. I have tested the ram and swapped my sticks around so I'm pretty sure it's not that.

My most recent BSOD's have said "MEMORY_MANAGEMENT" which given the fact my RAM tests OK leads me to believe my mobo is duff?

Are there any tests I can run to be sure it's not the cpu or gfx?
February 14, 2008 12:59:30 PM

To test the cpu, run prime95 stress test. You can try atitool to test the gpu. Run it and have it check for artifacts. But now it sounds like a bad mobo. You did already run memtest86, right?
February 14, 2008 2:03:38 PM

Doesn't Prime95 only test one core? Should he be running multiple instances of prime95 to make sure he is fully stressing the chip?
February 14, 2008 5:23:08 PM

Aragorn said:
Doesn't Prime95 only test one core? Should he be running multiple instances of prime95 to make sure he is fully stressing the chip?


The new version, 25.6 supports multiple cores.
February 19, 2008 7:25:41 PM

Well, I finally have what I hope is a resolution. I haven't had much time to spend with my pc but finally got time to mess around with it today.

It was running OK with 1 stick of RAM and having tested all 4 sticks thoroughly I was just about ready to pull the thing apart and RMA my mobo, but as a total stab I upped the RAM voltage to 2.1v and tried with 2 sticks: Viola no more restarts.
Thanks Roadrunner for the idea. I did try it up to 2.0V but the increases in the bios settings after that were in red, which scared me from doing it!!
My ram is supposed to be able to run on 1.8 so I still think my mobo isn't 100%. I haven't added the other two sticks yet so I may have to up it some more.

Anyway, Thanks for all help and suggestions.
February 19, 2008 10:42:31 PM

Which exact model of OCZ ram do you have? Every model except one on Newegg is listed at 2.1V or higher for 2x1gb PC2 6400. And that 1 is the Platinum rev 2 listed as 1.9-2.1V. What frequency and CL is your ram running at? Use CPUz to check.
February 19, 2008 11:41:24 PM

Yea, I too have OCZ ram and I had to up the voltage when I first put my rig together. My ram basically requires something above 2.0 v. I found that out when I tried to install windows and my machine kept freezing.
Another thing you could do, although I don't know why you would, is decrease the ram speed. For some reason I set my ram down to 667mhz instead of the stock 800mhz. That fixed it too, but adding more voltage doesn't hurt.

Congrats on finding the problem!
February 19, 2008 11:56:25 PM

markbowyer8 said:
Well, I finally have what I hope is a resolution. I haven't had much time to spend with my pc but finally got time to mess around with it today.

It was running OK with 1 stick of RAM and having tested all 4 sticks thoroughly I was just about ready to pull the thing apart and RMA my mobo, but as a total stab I upped the RAM voltage to 2.1v and tried with 2 sticks: Viola no more restarts.
Thanks Roadrunner for the idea. I did try it up to 2.0V but the increases in the bios settings after that were in red, which scared me from doing it!!
My ram is supposed to be able to run on 1.8 so I still think my mobo isn't 100%. I haven't added the other two sticks yet so I may have to up it some more.

Anyway, Thanks for all help and suggestions.


Whenever you run four sticks of RAM you almost always have to supply them with more voltage, most mobo's think that anything higher than 1.8 is either in the yellow or red. I wouldn't worry about it because you can't expect your RAM to recieve 1.8 volts x 4 sticks set at 1.8 volts. Resistance is the problem here.
!