Computer randomly rebooting for no apparent reason

Bit of an old computer, but it has been completely stable for the previous owner. Computer (motherboard/cpu) was transferred into a new case. Computer was given to me after a power surge took out my old build (apparently protector had developed a fault).

Build:
ASUS M3N78 Pro
Phenom 9500/stock AMD cooler
2x Crucial PC2-6400 2GB 6-6-6-18
2x Corsair PC2-6400 1GB 5-5-5-18
(memory is running @ 667 MHz due to AMD limitation - only 2 sticks can run @ full 800 MHz)
Geforce GTX 650
Crucial M500
Seagate SSHD
Ultra LS-500W power supply <- I am skeptical of this one
Windows 7 64-bit

Usually while Internet browsing or Youtubing (Google Chrome), computer will just randomly shut off for no apparent reason. No blue screen, no restart - just off. It turns right back on as if nothing happened. When I have had overheating problems with another computer, it would refuse to turn back on unless the power supply was turned off first, and then would display a warning in the BIOS when booting up again. This does NOT happen as any regular occurence and I can not induce it if I want to - completely of its own accord. Sometimes fine for a whole week, sometimes multiple times a night.

CPU heat sink re-pasted - idles @ 90F; never seen it go over 130F flat out @ 100% usage.
Northbridge re-pasted - this motherboard has just a giant heat sink. I briefly tried a northbridge cooler + paste; it crashed anyway.

Memtest x86 was run through 4 passes; did not see any errors.

I have stressed the machine for hours at a time with video conversion and playing games on top of that video conversion (100% CPU usage and heavy graphics card/northbridge activity) and it has been 100% stable.

I do not see any leaking capacitors on the motherboard. Power supply voltage looks stable as far as I can tell - system has been running for 4 hours with HW Monitor in the background and voltages are within .1V of 5 and 12V on each side.

Aside from a problem with the power supply, or possibly a RAM issue that's not showing up, what am I missing?

The whole build pulls under 200W full-tilt; most of the time around 100W. Could the power supply still be flaking out even though I am not seeing it in any monitoring programs?
 


Got another hard drive used as a backup with what is pretty much a clean install of Windows 7 and the latest version of drivers for this computer in a cabinet. Unhooked the other 2 hard drives and just ran off this one for a while. Still shut itself off.
 

Vic 40

Titan
Ambassador

Think it's the psu too.Where are you looking? From these choices the evga.

Which monitoring programs used? Tried this?

Download hwinfo32,

http://www.hwinfo.com/download32.html

You can open hwinfo32,after that open the sensor window by marking sensors=V,after that click run,,then click at the bottom "Logging start" ,after that choose a name and place for the log. like Hwinfolog at your desktop.Leave it open and do game/stresstest or whatever you do until a crash or shutdown.The log will end at that point but will be saved.
You could look at the log yourself with either excel or libre office calc.
 

spp85

Distinguished
Aug 11, 2011
699
0
19,360


Any working PSU that's available with you, friends etc to rule out a possible PSU failure. If the system works after changing the PSU, then you can buy a good quality PSU later....

 


Yeah, I have been monitoring the hardware voltages in HWMonitor. Over 4 hours last night (where the system did NOT shut off), I didn't see any voltage fluctuations between .1V in either direction of 5V and 12V.

Tried to use a friends power supply from an old computer, but it was from a Dell, and didn't have the small 12V connector to the motherboard that I needed. All my other friends have laptops, unfortunately...

I'm going to pick that EVGA up from MicroCenter this morning and give that a shot. It may be several days before (or if) it happens again, but I'm hopeful that is is the PSU.
 


Yes, latest version - released in 2010/2011 or something. I tried re-flashing the same version BIOS back to it just in case it was corrupt.

I have removed the 4 gig sticks for now, and am just running on the old 1 gig sticks that I know are good. It has bumped itself back to DDR2/800 running like this.

I also have a cheap USB3 card I added for an external hard drive. I'll remove that as well, just to test - I should think that would blue screen too, and it wasn't accessing the external at the time it crashed.
 
Nope, did it again. Expansion cards replaced; 2x1 gig sticks removed; 2x2GB sticks installed.

Here goes...

At this point, just wondering if the motherboard is bad. Can't really be much else.

I've seen other complaints of the M3N78 series motherboards doing this, but never a follow-up or actual solution to what it was.
 

Vic 40

Titan
Ambassador
Maybe the motherboard get's a bit "worn out",you could try to add a bit voltage top the ram,like 0.5V on top (for ddr2 about 2.1 Max.) and see if that helps.Maybe add some to the northbridge too if this isn't enough.

You said that you changed back to the old psu,this also happens with the new one?
 
F8CFD461-D28B-4460-9565-22F9349892AB_zpsawxaaxp0.jpg


111A4D59-7329-403F-BA90-214674569994_zpsya62vuad.jpg


Prior to this EVERYTHING was in auto. I have no idea what I'm doing with voltages on AMD machines.
 

Vic 40

Titan
Ambassador
2.080V for the memory should be enough,at 2.155 will it probably turn red meaning it being a bit too much.
When setting the NB Chipvoltage to manual is it set at a certain value,bump it one notch up if upping the ram voltage won't be enough.

See first how it does with only the ram voltage at 2.080V,if it doesn't help try the other voltage with one/two notches up.

You installed everything alright when changing the pc to another case? Like standoffs installed (correct) and such.

Maybe you can make such a log with hwinfo32 and upload it to dropbox,or similar (without to much crap),and post a link here.See if it gives something.
 
Yeah, everything installed fine - I have actually taken the whole board out to install a Northbridge cooler, which didn't really seem to help matters at all - I just thought that the NB chip was getting pretty warm with only a heat sink on there. Nothing should be shorting it from underneath.

I have woken it up and done lots of web browsing off-and-on today, and it seems to be fine with only 2 sticks of RAM in there. Really wondering if this has done it or not.