Maximum PC

I have a Intel i7-920 based PC running W7 Pro 64. Home-built the rig a couple of years ago and it has been running fine. Tonight I shut it off and when I attempted to turn it back on, the computer suddenly shuts itself off. The first time it got into POST, but died before trying to start Windows. Afterwards, it would only last 2-5 seconds before shutting down, not even bringing up the BIOS.

I opened it up and re-seated all the connectors. I thought that had fixed the problem as Windows was starting and applying some updates. Then it just quit. After that, it was back to it's 2-5 seconds then dying behavior.

Then, after waiting a while, I tried again. In POST, it gives me this weird message (in red text no less) about how overclocking has failed and its resetting to defaults (even though I have not overclocked anything on this system). It then dumps me into BIOS setup. After about 30 seconds, off it goes again.

Seems to be the pattern. If I wait a while (~20 mins), it will get into POST, and even may try to start Windows, but 30-45 seconds or so into the boot, it just shuts down.

Here is some more info
Dual Radeon 5870's
PSU: Corsair 750
Boot Drive: Intel M25 SSD.
MB: MSI X58-ProE

Any suggestions? I wanted to see if there was anything else I could try or if anyone had experienced this bizarre behavior. I am leaning towards the MB being the problem, but I just don't know.

TIA!

I have had a similar problem in the past, and it turned out to be a PSU with a defective +5V rail (it was only putting out +4.05V according to the motherboard's "PC Health Status" readings in the BIOS). Amazingly enough, it was a PSU manufactured by Corsair as well. After leaving the computer off for a period of time (30 minutes to an hour, maybe!) turn it on and try to get to the area in the BIOS settings where you can see the voltage readings for each of the rails on your PSU. Note what they are and contact Corsair support. Before you do that, however, try swapping out your Corsair PSU with a known good PSU that will power up your system (and take note of the voltage readings on each of its rails as well). When I called to tell Corsair about their PSU with the defective +5V rail, the support technician's first reply was that the motherboard's voltage readings aren't necessarily accurate and should be taken with a grain of salt. When I told him that I had swapped the PSU out with a competitor's PSU (I did not tell him that it was a PC Power & Cooling 750W PSU) and found the system to be perfectly stable (meaning no random and unexpected shutdowns of the computer) with the motherboard's health status reading of +4.89V on the aforementioned competitor's PSU, he immediately went about processing an RMA number for me. Total turnaround time from the time I sent the defective PSU in to Corsair to the time I received an exact replacement: 15 days (including weekends and holidays)! Just FYI.

OK, thanks, I'll look into that and try to get those readings. Unfortunately, I don't have a spare PSU to try. I'll have to think on it.

That's fine. Check the voltage readings on each of the rails of the PSU. If any of them seem out of range or abnormally low, give Corsair a call and give them the readings. If they try to stonewall you by telling you that the motherboard's voltage readings aren't accurate, bluff them by telling them that you have swapped out the PSU with a known good one and tell them that the machine works fine with it, but not with the Corsair PSU. It's definitely worth a shot, and Corsair will replace the PSU with a brand new one (if it's still under warranty, that is!).

Sorry for the noobish question, but where do I look for those voltage readings? I only had a chance to try it once before I had to rush off to work. I looked in the "H/W Monitor" section. I saw some "5V" lines that read 5.0something V. However, there was no mention of PSU or rails or anything. There was some mention of CPU, but not on the lines I was looking at.

So given the limited amount of time I have to search the BIOS settings, any pointers for me in my search for these readings?

Thanks!

I would definitely try another power supply before you damage the motherboard. If it is not a high quality supply it is suspect. Buy one at BB and take it back after testing.

I believe you were in the right place in the H/W Monitoring section of your BIOS. What you will see there are readings for +12V rail, +5V rail, and +3.3V rail (you will not see anything that refers to your PSU specifically as you are checking the voltages of the PSU's power rails in this section of the BIOS. These readings need to be within +5% of their purported voltages on each of the rails of the PSU. If they fall outside these voltages, then you have a defective PSU. However, if all of the voltage readings are within specified tolerances, then I would say you have a problem with the mobo.

as peeps have stated, its either an electrical problem or an overheat issue, if your able to get into the monitoring tab in the bios, it normally gives all that information including CPU temp, its thats about 70C(which would be under load temps normally) i would look at reseating your heatsink as well.... If all voltages check out it still may be a PSU problem as well as alot of problems dont actually show until power is being pulled(such as like using the cpu when loading up windows)

Let me know how the cpu temp and volts are looking and we can go from there

OK, looks like the voltages are good, but the temperature may be off.

Here is what I get after booting a cold (left off overnight) machine:

CPU Temp 85C
IOH Temp 26C
SYS Temp 48C

Vcore 0.952V
3.3V - 3.344V
5V - 5.087V
12V - 12.232V

I'll try reseating the sink when I get home from work this evening.

And the problem was the temperatures.

Taking a closer look, the CPU heatsink was loose. After tightening back down, all is good!

Thanks to all who helped! Much appreciated!