System Reboots

[TopChip7]

Hi.

When playing internet games ONLY (usually TrackMania) my system periodically shuts down and reboots.
It is definitely NOT a temperature problem. Sometimes I can play for 10mins, other times 3 or 4 hours.

Do you think it's possible it could be a PSU issue? 500Watt Antec SmartPowerII running 2 x 7900GT SLI, X-Fi, and 2 x Hard Drive Raid0.

If not, does anyone have any other suggestions.

As always, thanks for your help ;o

[Whyzman]

It appears that you haven't turned off the XP default "auto-reboot" feature. Disabling will then produce an error message which might be helpful in diagnosing...or point us in the right direction.

Also, these folks have done an excellent job outlining and troubleshooting the usual random reboot culprits. Here's (http://www.hardwareanalysis.com/content/topic/57410/) the link.

[TopChip7]

Thanks Wyzman. Have now disabled 'auto-restart' (for others wishing to do so goto CONTROL PANEL - SYSTEM - ADVANCED - START UP & RECOVERY Settings - Uncheck Automatically Restart)

Will now wait for the system to crash or give me an error message (you watch, it wont happen for ages now lol) and will post error message.

Cheers

[Whyzman]

Oops, sorry I forget to post the info on how to disable. :confused: I'm glad you found the path... ;)

[TopChip7]

Ah atlast - after about an hour or so the game crashed out. Returned straight to desktop - no error message - GPU temps 53 + 48c, CPU temps 43c, mobo temp 38c, ddr2 temps 40.1 + 38.5c max average 34c, HDD 28c.


No error No error No error ?????????


So what next?

[TopChip7]

Strange - My DVD-Rom and hard drive just started spinning up on its own after submitting the above thread, yet windows task manager (processes) shows nothing? Could that have something to do with it - perhaps a power boost???

[Whyzman]

I would follow through with my linked troubleshooting procedure. Being as you didn't get an error message, I would suspect something more hardware than software...

Do you have a PSU you can swap?

[TopChip7]

i dont have another PSU of similar power (dont think 187Watt PSU will cut it lol).

Memtest86 showed errors on RAM but unable to RMA as no more in stock (Corsair Xpert Modules) and would have to put £100 to the refund to get from elsewhere. - Anyway, although Memtest showed errors I have never had a problem with them and do NOT believe memtest is that accurate on EPP modules anyway as I know they are completely test prior to packaging and come with 10year lifetime warranty against defects.

GPU drivers are not WHQL however they are what came with the GPUs (Leadtek) think drivers are v83.91 which are used by thousands of people.

Could it be something to do with the firewall settings, router or internet speed? Or perhaps AVG Free Anti Virus, Windows Auto-Update or other software that runs in the background?

[TopChip7]

Will try dropping the overclock on the GPUs, then if that dont work will try dropping the Cas of the RAM to C5.5.5.15.2T.

Anyone got a 700Watt PSU I can borrow and live locally (I'll get the beer in whilst we test it playing games lol)

[Whyzman]

You might want to crosscheck the RAM with this:

http://www.simmtester.com/PAGE/products/doc/docinfo.asp

[jlreich]

I would test the PSU with either a PSU tester or a multimeter (preferred).

[TopChip7]

Well Whyzmann, I ran the memory tester and that gave even worse results than memtest86. After 201 errors the whole PC locked up and took 2 reboots to get windows back up - All the errors seem to be at address 2046.9M - which makes me wonder whether there are any errors or not or whether it is just trying to test the address where the EPP/SLI settings are stored on the RAM? ANY OPINIONS PLEASE???

As for the PSU testing jlreich, I dont have a PSU tester however ASUS Probe reports around 11.73v maximum coming off the 12v rail - is it possible that 0.27v drop although in +/- 5% specification is not enough?

Really appreciate all the help guys.

[jlreich]

That's well within the acceptable range. But I wouldn't trust the software, I would test it. You can get a PSU tester for around $15 at your local comp store.

You can also try the new Windows memory diagnostics (http://oca.microsoft.com/en/windiag.asp). Burn it as an image and boot from it, it will start automatically. Hit the "t" key for to run the extended tests.

I really have no idea, but it is possible Memtest isn't compatible with the new EPP memory.

And try the ram one stick at a time if you are still getting errors. It's unlikely that both sticks are bad.

[mjc]

The Asus Probe reading, was that under load or while idle (or in BIOS)?

If that is an idle reading, while it is within the range it may drop even more under load.

[TopChip7]

checked out "theramguy.com", corsairs own forum - they told me that none of the memory testing software would not work on my modules without a slight change to the program.

however - they also told me that my ram should be set to 2.1volts - in auto my mobo was only pushing 1.98v - 2.01v.

set voltage manually now to 2.05v which is actually now pushing through 2.06v and 2.09v - tried 2.1v but was pushing through 2.15v then??????????????
arrrrggghhhh!!!!!!!! any explanations please????

i also dropped off the GPUs overclock and reset them back to manufacturers spec.

anyway the result: just played 3hours non stop internet gaming with the ram at just 38.4c maximum and 37.2c average without any crash whatsoever!!!!!



Thanks guys.

[Whyzman]

We like success stories...

I failed to notice that you had the GPUs overclocked...that just added another variable to the mix...at this point it's hard to say what the actual solution was...

But hey, no crashes...sounds good! ;)

[TopChip7]

I doubt the o/c GPUs had much to do with it to be honest - they have never crashed on me yet as the samsung ram is good to 714 (1430mhz) - never understood why nVidia underclocked their cards to 660 (1320mhz)????

Thanks very much for all your help (as per usual).

[TopChip7]

Arrrggghhh!!!!!

can anyone decipher the following error (given on complete system shutdown)
***STOP: 0x0000008E (0xC0000005, 0xBFB334FQ, 0x91c3DB4c, 0x0000000)
***nv4_disp.dll_ Address at BF9D4000, Date Stamp 43f16f88

If it helps find problem, the 2nd 7900gt seems to be 7c lower than it usually runs at.
Is it possible maybe it got warm, the fan kicked in on the card and drained power to it?
Or could it be an internet problem as when rebooted, opened internet and got error need 2 close???

thanks

[jlreich]

The first error (http://support.microsoft.com/kb/315335) shows faulty ram according to MS KB. Note that it can also be ram that is not compatible with XP. So it's possible the EPP ram is not playing nice with XP. It would be nice if you could put some regular ram in there to see if the errors go way. If not I would RMA the ram.

The second one no doubt has to do with the video cards. I would look at trying the latest drivers, especially since you are using SLI. Perhaps see if disabling SLI gets rid of the errors, just to troubleshoot. But I wouldn't be surprised if you replaced the ram and those errors went away.

[saphalline]

The latest BIOS version for your mobo is 0401, dated December 22, 2006. If you don't have the latest BIOS version (or worse - if you still have the first BIOS version) I would try that next. If the latest BIOS and some more RAM tweaks don't work, I would get the RAM replaced.

Also, did you use the latest v 1.65 of MemTest86+? There's no reason to think it wouldn't work with your RAM - all the EPP stuff is handled at the BIOS level anyway.