PC upgrade going bad

[Brianmil]

I've just bought myself basically a new PC by upgrading all the internals as follows:

ASUS P5N-E SLI NF650i 775 mobo
Core 2 Duo E6300 CPU retail
Patriot 2GB kit PC6400 800MHz 4-4-4-12
nVidia 7600GT 256MB GFX card
Seagate 320GB SATA II HDD
Seasonic 430W S12 PSU

Also replaced the standard thermal paste with some Artic 5 silver paste. And kept my IDE DVD burner and case.

All went well at first installing XP and the various drivers for the hardware, then I tried running a utility called Heavyload to test the system stability.

Here's where things went bad...

I was monitoring the system temp using the bundled ASUS probe software and I kept the case sides off to start with just to be sure. Idle temp for the CPU was around 30 deg and the mobo about 28 deg.

At first Heavyload would run for up to half an hour then crash and/or the system reboots other times it would freeze after running for a few minutes then crash but XP was fine, but after restarting Heavyload it would eventually reboot the system and constantly reboot just as it's booting back into XP.

I then thought it might be the utility which wasn't stable so I tried installing Quake 4, this went well, the game installed and was okay until I tried moving the camera around quickly, this caused an endless series of reboots too. Tried another game which installed and started okay but also caused it to reboot shortly into the game.

The two things that first came to mind were the PSU wasn't enough for the system, but the mobo manual says at least 400W for a 'fully configured system' whatever that means. The other thought was maybe the memory wasn't seated properly. So I decided to swap the two modules around, this was when I noticed that one of the memory slots was right next to the mobo chipset's heatsink, which gets very hot, maybe the memory is overheating, I thought. At first I still used this slot but it made no difference then I tried swapping the memory to the other two slots, away from the chipset heatsink but again no difference. i.e. system starts rebooting eventually once it's under load. Although if just left idle it is fine, no reboot's, but also no good if I can't use the damn thing.

The temp when it reboot's is also varied, with the case sides off it gets up to about 34-36 deg under load before rebooting, with the case sides on idle is around 32 deg and it reboots between 38 and 42 deg under load. The mobo temp never goes above 34 deg under load.

I called a friend who's also upgraded his PC to a dual core, and he said his system did this when the CPU heatsink wasn't in place properly. I checked mine and it didn't seem to be, but I made sure anyway and tried again, but still no difference so I looked closer at the heatsink.

The heatsink retention clips are two piece push in types with a split pin that grips the underside of the mobo, and has small grooves in the top part which holds the bottom part in place. Looking at the the clips I can't see how they could be any tighter, but there is still a gap of around 1mm between the mobo and clip on the top side, and another 1-2mm between the clip and the metal leg of the heatsink. Shining a torch between the heatsink and motherboard I can see that the some of the artic paste has been squeezed out from between heatsink and CPU so I assume it's a tight fit.

So in conclusion, Is this heatsink any good, is the PSU powerful enough, or is there some other explanation?

Any help would be greatly appreciated.

[Anonymous]

It doesn't sound like dodgy temps to me, maybe dodgy ram? Use memtest and find out if you have fully functioning ram.

[mac124]

I would agree with Blaine its probably memory though i would say its more likely settings, particularly voltage, rather than a faulty stick (or sticks) i have personally seen similar things happen and know of numerous people who have had stability issues and it turned out to be the memory voltage was set too low.

I suggest checking the voltage in the bios as according to patriots website the memory is rated at 2.2v and i wager the bios setting for the memory voltage has defaulted to 1.8 or simiar, try setting it to 2.1v and recheck the stability.

Before you go changing stuff in the bios though please check CLICKY>> this the memory you have <<CLICKY (http://www.patriotmem.com/products/detailp.jsp?prodline=5&catid=2&prodgroupid=38&id=317&type=1) and i haven't looked at the wrong stuff.

[Brianmil]

Your both correct, it was memory, Memtest found 50+ errors within a few minutes of running. Mac124 I didn't see your reply until just now as I only have internet access as work while the PC's been misbehaving. I'll look at the voltages tonight and seem if that's the cause. Thanks both.

[mac124]

Nice one, looks like your on the right track anyway, deffo check the voltages as it could cure the memtest errors by setting the voltage towards the top of the specified range.

No guarentees though :wink:

[Mattbot2]

it could just be a genuinely broken one. But most of those come with a liftime guarantee. but if its dead, its not in its lifetime :P

[cabbages]

psu?

[Doctor 8Ball]

as, cabbage questions above., the PSU spec looks way low and if it is generic then I would suspect this is the issue.

I had this problem last year, constant reboots on various applications, even though windows xp would be fine on a cold start from no overnight usage (PSU nice and cold) the damn thing would crash by mid-afternoon. After intense process of elimination it was found the PSU was faulty (cheaply made with crappy voltages when the system reached standard tempertures)

When i read your problem it instantly reminds me of that time.

So if you have a spare PSU (even test one from a dif case) note the wattage and go for it, DO A PSU TEST!

[mac124]

Unless it is faulty the Seasonic 430watt psu (no its not a generic one) is more than man enough to run the op's above spec, i am running my media pc which is a similar spec, check sig for actual specs, off of a Silverstone 400watt psu just fine and it is overclocked to 3200mhz.

It wouldn't hurt to try a different psu to prove a point, assuming the memory suggestions didn't resolve the issue.

[t31os]

Seasonic S12 is a decent PSU, i very much doubt the PSU is at fault, especially with the memory faults now found.

Might be helpful to test each stick individually to.. ;)