System Rebooting without BSOD regardless of temperature or load

[Loder]

I've recently been getting some intermittent, but persistent problems with my PC rebooting with no warning, and no BSOD. Crash logging is enabled, but there is no blue screen and no dumps are being created. The only insight offered by Event Viewer seems to be critical warnings about the computer having lost power unexpectedly. My PSU is a brand-new RMA replacement unit, but I will still check it tomorrow when I can get my hands on a multimeter.

My specs are:

Core i7-930 @ 4.0GHz
ASUS Sabertooth X58 Motherboard
12GB G.Skill Triple Channel RAM @ 1600MHz
Corsair TX850W PSU
EVGA GTX 480 GPU
128GB Samsung 840 Pro SSD (Boot)
128GB Corsair Force SSD (Games)
2x1TB Seagate Barracuda HDD (Data, not in RAID)

The problem does not seem to be related to load at all, nor to Windows or any other software. I am almost certain that this is hardware-related, because I have had the problem even while sitting in the BIOS or while trying to run MemTest86+. Typically I will go anywhere from a day to several days without seeing the problem, but when I do, the system seems to be very difficult to get running again. Leaving the computer powered off for a while seems to be a temporary fix.

Every time this happens, the machine tries to POST and immediately tells me: "Overclocking Failed; press F1 to enter setup and change values or Press F2 to load default values and continue." My processor has been overclocked to 4.0GHz without issue, but I have also tried using the PC with default settings recently, without seeing the problem change. After the overclocking failed message, I can go into the BIOS and do whatever the *censored* I want—I can even set defaults from within the BIOS and reboot; no matter what, I get the same message again: "Overclocking Failed." The only way to get around this seems to be to accept the F2 option at the prompt and continue into Windows with the default values. Otherwise, powering off the machine for a while seems to be the only fix.

I have tried the following:

- Defaulting BIOS settings
- Reseating RAM
- Removing and replacing CMOS Battery
- Checking all power connections
- Checking temperatures (Everything is running at more than acceptable temperatures, even under load)
- Reseating GPU
- Testing RAM with MemTest86+

I'm having trouble getting information on my specific situation by searching around, so I'm hoping that someone might be able to offer some insight. I currently suspect the motherboard or CPU, but I am not sure if I am correct in doing so. I would love to hear what anyone's thoughts are. The problem seems to be getting worse and worse, so I don't think I'm going to be able to ignore this for much longer. What does it sound like to you? Thanks for your time!

[Calum]

Have you updated the BIOS?
Is your RAM 6 x 2GB or 3 x 4GB?
Have you tried a different slot and different power connectors for your graphics card?

[patio]

The problem with testing with a multi-meter is it's not under load...replace the PSU.
Also spend a bit more on the next one and use the RMA credit towards it so you've lost nothing...

[patio]

PS...i'd also throw a brand new CMOS battery in there as it seems to not be holding settings...
Although that could also be the flakey PSU causing that issue...but it's 4 Bucks so no biggie.

[Loder]

Thanks for the responses, guys. Calum, the BIOS is up-to-date, my RAM is 6x2GB, and I'm currently trying different power connectors. It's a little tricky to try a different PCI-e slot at the moment because of the positioning of some things in my system.

I tried testing the PSU, and all the 3.35v pins were at 3.37v, the 5.0v pins were all at 5.06v, and the 12.0v pins were both at 12.13v. The -12.0v pin was at -11.92v.

Patio, this PSU is a brand-new unit that was shipped to me from Corsair just a few weeks ago as a replacement for one that died on me previously. Also, you say spend a bit more next time, but I was under the impression that Corsair makes quality power supplies. Is that not the case? It wasn't cheap when I bought it, that's for sure. I know that not being able to test it under load leaves some questions unanswered, but do you think the fact that I've had this issue even while sitting in the BIOS might indicate that load doesn't seem to have a particularly clear effect on the issue? Furthermore, I've gamed for hours without it happening, and the resetting doesn't seem to correlate with gaming sessions. With this information, do you still think it's the PSU?

Also, the CMOS battery holds the settings just fine, it just says that my overclock has failed when it tries to boot up. I've noticed that after the reset, the system tries to boot up several times before it manages to get to that message. I'm thinking that maybe the system is basically reporting that it couldn't boot up properly, and assumes that the issue is related to poor overclock settings, but I could be wrong.

Another thing to note is that I took a look at all the capacitors that I could see without taking too much out of my system and I didn't see any that seem to be bulging or leaking. I know this really sounds like a PSU issue, but since I just received this one and the tests seem fine (at least while it's not under load), and because load doesn't seem to have any effect on the problem, I'm just having trouble accepting that it could be the PSU. Let me know what you think, and if you still think it's the PSU I'll contact Corsair, but I'm not sure how they'll feel about giving me a replacement unit when all of the tests that they suggest are working fine and they just sent me this one recently. Maybe I'll just keep this one as a spare and buy a different one entirely if it comes to that.

Anyway, I have a habit of typing particularly long replies, so I'll leave it there and see what you think. Thanks again for your time guys, I really appreciate it.

[patio]

Corsair does in fact make some good PSU's...
They also depend on which grade it is...they also make budget units...
As to the boot error message are you in fact OC'ing ? ? If not that message can happen with a weak cmos battery which is why i suggested replacement...

At one point i had 2 RMA PSU's in a row shipped to me that were both bad for a total of 3 units...not sayin it's the case here but i have had it happen...ask for a different batch # if you decide to go with the same unit.

[Loder]

This PSU was considered quite good quality when I bought it; it was a little bit before the modular thing started to catch on so that's one thing that might make it seem like a low-end unit. However, it was definitely not a budget PSU at the time.

I am OCing, yeah. I've seen this message before when the PC had trouble booting up properly, which previously was related to not having the OC settings set properly. They're set just fine now and the system has been running stable for a long time, and I even tried tweaking the settings and I'm now running the system at all stock settings just to see if there's any difference. (It doesn't appear so.)

As for testing the PSU under load, I did try some software-based measurements while under load just now, just to see if I saw anything drastic, and everything seemed to be in order. I checked the reported voltages before and during stressing with prime95, and the voltages didn't drop very much at all, and were still well within spec. I know that software is not as accurate as a multimeter, but I can't help but feel that I might notice something awry even with software measurements, despite them not being as accurate as measurements done with a multimeter. I'm not sure if prime95 stress testing is the type of thing that would reveal any issues in this situation, but I tried the test either way. Maybe you have an opinion on it.

Thanks again.

[Calum]

The TX850 is a solid, excellent quality unit, no need to worry about its quality.
Software based voltage measurements are unfortunately woefully inaccurate.
Have you tried your RAM one module at a time?  It's possible for RAM to pass Memtest yet still be faulty.
Certainly worth trying a different CPI-E slot if you can, or even a different graphics card if possible.

[Loder]

Thanks for the additional suggestions and info, Calem.

I didn't test one RAM module at a time; I figured that I would try a test with all 6 for a start and start narrowing it down from there if I saw any issues. My computer generally is usable in its current state and thankfully I haven't had a restart since changing the PCI-e connectors (maybe coincidence, maybe not)—so perhaps over the next few days I will try testing each module individually with several passes whenever I get a bit of spare time (I only did 1 pass with all the modules in.)

As for trying a different PCI-e slot or video card, I might be able to kill two birds with one stone. I know someone who has a spare PSU lying around that's around 500w, so while that isn't enough to power my PC properly with my current GPU in, I'm thinking I could swing by my parents' place and snag the cheap GPU that I put in their PC when I built it for them. They're out of town for a few days so they won't miss it. I could try the spare PSU and the cheap GPU for a while and see if the issue shows up again.

Honestly I'm not particularly suspicious of the GPU itself, because in my experience there's usually some more obvious indicators of a bad GPU that would present themselves, like visual artifacts for example. I've never seen anything like that. But either way, at least this will give me the opportunity to potentially find out if my PSU is the problem.

Thanks again, and I'll let you and anyone who checks this thread out with a similar issue know what the results of my next tests are.

[Calum]

Sorry, to be clear I was suggesting running the system with individual modules rather than testing them each with Memtest - if they all pass together, they'll almost certainly pass individually.  Running the system with individual modules to see if you still have the problem would be a good test.

The 500W unit would power your system under lighter loads, assuming it's a good unit.

I have seen graphics cards cause this problem, with the error in the event log that you mentioned - assuming it's even ID 41, kernel power?  For some reason the GT430 seemed to cause this issue a lot, we had a ton of those cards we had to replace - although they would pass every other test they would cause the system to reboot randomly with no other indication of anything being wrong.

[Loder]

Okay, sounds good. At this time, I still haven't encountered the issue since switching to different PCI-e cables, so it's possible that that was the problem, but I can't be sure quite yet since the problem is so intermittent. My GPU takes one 6-pin and one 8-pin PCI-e connector, so maybe the small 2 pin piece was a bit loose?

As for the spare PSU, it turns out that it was an older standard, so it's not going to be of any use to me. However, I actually ordered a new graphics card recently and it should be here on Tuesday or Wednesday, so if I do experience the issue again before then, I can see if I have the same issue with the new graphics card installed when it gets here.

Running my system with one RAM module at a time will be the last of these that I try if necessary, because it would make it difficult to use my PC the way that I want to. I'll try it if I have to, though. Thanks again!

[Calum]

It's quite possible, although the additional 2 pins on the 8 pin PCIe are just additional grounds.

Let us know how it goes over the next few days.

[Loder]

Will do. Still going strong, so I might switch back to the other PCI-e connectors soon and see if anything changes.

If those two pins were loose, do you think that it could be possible, being that they are ground pins, that my power-hungry graphics card draws more power than the 3 grounds in the 6-pin connector alone can handle, and that excess power might have been building up over time, eventually causing the system to power down? Maybe the trouble with getting the computer up and running again right away, but working after a while, might have been to do with excess power draining over time, allowing me to eventually boot without issues?

I don't know, I'm not really sure about any of that even making sense, but maybe you know what I mean.

[Calum]

The GTX 480 has a max power draw of 250W.  PCI-E spec allows the slot to provide 75W, a 6 pin connector to provide 75W, and an 8 pin (with just two additional grounds and no additional 12V pins) to provide 150W.  So, allowing for effectively two 6 pin connectors, that's only 25W off it's maximum draw.  It's definitely possible that one of the grounds is defective and causing the problem, though, faulty PSU connectors can cause all sorts of weirdness.

[Lisa_maree]

Hi

Just something I would check is the 5v SB the green wire on the motherboard plug. It should read 5v with the meter when plugged into the motherboard if it is low or noisy it will cause the symptoms you are having.

Lisamaree