Got a real odd ball of a issue on a pc I have just built last week, advice suggestion

[grart]

Hi,

I hope someone can may be point me in the right direction as I am at my wits end with my machine.

I have had nothing but problems when ever I attempt to play games, I get into to a game, be it, BF2, BF2142 and the new Quake wars beta and it plays fine for about 23-30mins some times only a few minutes then the screen locks up and the only way to get out it is to restart the machine.

But when I run sis Sandra on burn in test, 10 times all passed, even left it running for 4hrs no issues found. only issue that was spotted once was a memory test which failed on my memory but was never seen again.

I built a new pc last week, spec included.

I have updated all the nvidia drives that I can get hold of, even trying the beta versions, made sure every windows driver has been updated sill the same issue. Windows is not giving me a blue screen or a error code after restart. I have also updated the mother board bios to the latest release. I have reinstalled the os twice.

I have changed the following parts using a friends machine as a donor and some old bits I had all bits I believe are working fine, memory swapped out, graphics card swapped over, still the same issue.

I thought that it was the graphics card after I had read a few articles on here, but when it was taken out the pc has the same issue, but sadly no one I know has a psu that will handle the card so I cant test it any where else.

Going off my kit can anyone suggest anything else before I format my secondary drive and try loading the os again on that incase its the hd, i have also swapped over the sata cables incase one of the was duff.

[GoNz0]

check the event viewer for starters, but a lock up normally leaves no clue.

system temps after 30 mins would be the 1st thing to look at, if your thrashing the PC your creating a hell of a lot of heat, straight after a crash, whats the system temps in bios, or get an app called everest ultimate edition as that will monitor all system temps.

[grart]

Sorry GoNzO,,

Forgot to post the temps, CPU 27, MB 42, GPU 64.

Thanks, I will get hold of everest when I get home from work, as the temps that I seen were after a reboot.

[GoNz0]

yeh may well have all dropped 10 degrees during the power cycle.

i find it hard to believe CPU temps as low as ambient (on air anyway) with mobo at 42, a motherboard hitting 50 degrees C under load could cause a crash, whats the case ventilation like?

hopefully everest will give a more realistic set of temps

[grart]

I have a 2x 120mm AK-183-L2B Akasa fans one on the front of the case taking air in and one on the rear venting. There is also on the case a direct draw shroud that goes over the cpu fan and the msi motherboard fan is on too.

I dont think I can put any more cooling on this particular case, other then something more exotic, water cooling.

[abeacock]

To rule out over-heating, why not try taking the side off the case and positioning a large 14" desk fan straight into it.

As for other ideas, do you know what windows is doing behind the scenes when you are playing your games? Is drive indexing turned on? Any other background processes, virus/malware/spyware scanners kicking in?

Just a few suggestions, I hope you get to the bottom of it soon, it must be driving you mad!

[MD]

I don't see anyway that CPU temp is right....I mean its a good 10 degrees less than mine was over the weekend and thats a water cooled Q6600...not that much thermal diff between chips...especially not on air.

[grart]

I will put everest on tonight as GoNzO suggested, see if that gives me a more accurate temperature of the cpu. I have little faith with what is coming up in vista 32 temp wise, alot of the reporting from ntune and the like all seemed a little off. Still half tempted to reinstall XP on the machine.

[grart]

I don't see anyway that CPU temp is right....I mean its a good 10 degrees less than mine was over the weekend and thats a water cooled Q6600...not that much thermal diff between chips...especially not on air.

MD, I agree the temp for the CPU is well off for a stock cooled chip. I ran everest and it said that the cpu temperature when under the stress test maxed at 32c (cpu only). It did not go higher then that after it was ran for 25minutes, mid late forties would have sounded about right.

I then ran the other tests that are on Everest, CPU, FPU, Memory and disk, within 1.5mins the test failed, I ran the test 5 times and each time it failed from 10seconds up to 4mins. I turned the test parts on and off in turn until it was found that that the memory test was causing the fails.

I luckily had two friends pcs to hand, to try parts from.

I ran the everest tests on both their machines and it passed on all accounts, so I swapped my memory over for the exact same memory and his memory failed in my machine and my memory passed in his.
So, i swapped my CPU out for a Pentium D (3400 not sure of the exact model number), I ran the tests again on my machine still with the diff memory in, my pc now passed on all tests was fine running for 30mins, sadly I could not test my cpu in his machine. The cpu temp on the Pentium D went up to 67c and stayed there (bit more realistic temp from the cpu). Swapped my memory back still with the Pentium D in and passed again.

So that leaves me, unless someone can suggest something else with a possible motherboard issue, incompatibility or other wise or cpu issue. ( the motherboard has the latest bios update 1.2 which only came out a last week)

If any one can shed any further light on the problem, which I would be most great full for I think I will have to RMA both the motherboard and cpu for testing.

Sorry for the long post, just tried not to miss anything out.

[GoNz0]

stick your ram back in, up the ram voltage a point and see if that makes it run stable?

[grart]

stick your ram back in, up the ram voltage a point and see if that makes it run stable?

GoNzO, sorry , just to check I updated my post around the time that you so kindly replied. I tried my complete pc with just the processor being swapped over for the Pentium D cpu and all worked fine on the everest tests, I am not sure I follow what you mean about putting the ram back in and raising the ram voltage to see if that makes it stable.

Was up till 4am removing and refitting parts so bit slow this morning, sorry.

[DratUK]

often the ram will be set to a mobo default voltage ie 1.8. 6400 Ram often requires higher voltage to run stable in demanding process's ie games etc.
The ram you have should be set to 2.1v, please check with Corsair website, it may vary as there are 2 flavours one with 4-4-4-12 the other is 5-5-5-12, I am not sure whether the ram voltage required is different.
The other is the timings, ie 4-4-4-12 etc. I had a similar problem with my corsair Dominator, marketing speak for lets have more of your money!, my game would crash for no apparent reason.
The reason was another setting in the bios for memory this was the Trfc by default it was set to 30, my abit board didnt like that. I was advised to change it to 42, problem went away.

Rather than fiddle with the bios there is a program called memset 3.0 this allows you to set the memory timings in windows. You could give that a go.

Two I would then eliminate any dodgy memory by running memtest. You can get an bootable iso to burn and run that from boot. Run it overnight and see what you get. It will indicate any memory faults.

Also perhaps run Orthos to stress test the cpu.

Not saying it will solve the issue just different options to help pinpoint the problem.
Good luck.

[DratUK]

I don't see anyway that CPU temp is right....I mean its a good 10 degrees less than mine was over the weekend and thats a water cooled Q6600...not that much thermal diff between chips...especially not on air.

Thats probably the bios temp recording, my Abit Uguru shows 29 degs on a Q6600 I wish it were. Coretemp shows 38 average across the 4 cores. So add 10 or 12 degs to his temps that would be more realistic.

[grart]

Well thanks for all the replies and the suggestions that were given, but after last nights testing the setup i am calling it quits.

With the Pentium D in the machine all appears to run stable, i can play games, run everest tests, memory tests, cpu tests all pass.
Put the E6850 in the machine and a whole host of issues come up, system hangs, driver issues.

I found the timing settings for my memory entered them into the motherboard bios as suggested, making sure the voltages were running as they should be ie 2.1 volts I even tweaked the settings to see if the machine would run stable as suggested but still to no avail.

Graphics drivers were saying they were falling on both a 7800GTX and a 8800ultra they were not with the Pentium D in, cpu clock speeds were as they should be at all times with the Pentium D, but not with the E6850. Loading drivers or loading anything onto the machine, would cause lock ups.

Since I have changed prity much every component on the machine other then the Motherboard and testing the chip else were i can only assume the motherboard does not like the chip or possibly the mother board is faulty too, on the back of that I am calling it quits.

Again thanks guys for all the suggestions and advice, its been very much appreciated .

[abeacock]

Have you tried updating the BOIS of the motherboard?