Maximum PC

I've been getting BSODs, seemingly without connection to any piece of software running. It might be driver conflicts. It might be the page file. Here's the small dump: https://docs.google.com/file/d/0B5zIz1Y ... WpScVpya3c

Here's my PC's stats: https://docs.google.com/file/d/0B5zIz1Y ... EtWGc/edit

I have my page file set to 1024 initial, 6142 max for the primary drive only, no page file for the secondary HDD.

It's doing it like every ten minutes now. Here's another dump: https://docs.google.com/file/d/0B5zIz1Y ... k5VTA/edit

Have you cleaned the machine out recently? Tried running memtest on the ram?

I'm also working on this PC, and I've ruled out any software issues, as it reboots at completely random times (during boot, installing windows, in a command line boot). I can also rule out video card, hard drive, and RAM (I've tried different sticks in other slots and all that), so that leaves CPU, mobo, and power supply as the remaining culprits.

I ran Memtest and came up with 83 errors. Is this good/bad/normal?

I'm curious as to why the power supply would be giving BSODs though. The failure message is different every time it comes up, so that seems a bit odd.

83 errors is not good, do you have some ram laying around you could swap in? Sounds like bad ram or maybe the board is failing.

Ideally Memtest should not come back with any errors. I might be able to overlook a handful over a 24 hour period though. However, 83 errors doesn't sound like RAM just yet. 100% "yes it's RAM" is when it flatout barfs errors.

In regards to the power supply, power supply issues mask as hardware issues to no end. I've had my hard drive supposedly give me the click of death one time, but a PSU change fixed that and the hard drive went on to live for many years later. If there's anything that's a hardware issue and I've ruled out RAM and my HDD/SSD connections, I blame the PSU. Especially if the PSU is over two years old.

If this is an old rig, I would be looking for bad caps, if you have any at all, you will need to replace the caps or the motherboard, examples below:

http://www.google.com/search?num=10&hl= ... UkTPaFsweY

Nasty

Ah... Well if you were a hardcore DIY guy, sure replacing the caps would be a viable solution, except... Even if you can find the exact cap used in the motherboard or an equivalent replacement, you either have to spend time and energy replacing caps deemed bad. Or you can replace all the caps. Oh, and how do you know it's not just the through-hole caps? What about the surface mount ones?

Both options are not worth the time and effort for something non-critical. If this were the Space Shuttle, I could understand.

LatiosXT, this sort of thing happens with old motherboards as they had a real bad run of caps about ten years ago. If the caps go bad you will get all kinds of blue screens or the system may not boot at all because of this problem, I know, because I had an Abit board go bad on me and it was the caps. They are used for timing and if the caps are bad, blue screens will happen because of corrupt data. This is not rocket science, just basic electronics. If the caps are bad, they must be replaced or the board must be replaced there is no getting around it. That is why I supplied the link to pictures of bad caps.

Nasty

Nasty,

I understand where you're getting at, but you can't expect any OP to take on the task of replacing the caps. Unless OP had a buddy who knew how and was willing to pay for a few beers.

Well either way, if OP can check the board and it clearly has bad caps, then we can call it a day. If not, then let's not point at that direction until all other possibilities are looked at.

...beer is fattening, a good cognac will do...

joel96, did you check the ram with memtest86+, one stick at a time. That way you will know which stick to RMA if you have no bad caps that is...

Nasty

nsvader and Nastyman: I cleaned it out, the problem persisted. I ran Memtest86+ with all four sticks in place, and it resulted in 13,000 write errors. Since I tried running with sticks of RAM from another rig in different slots and still had the random reboots, I doubt that it's specifically the sticks, but I could run it one slot with the same stick to get further suggestion that the mobo is faulty.

kleinkinstein: OCZ 600W Z600SXS. http://www.ocztechnology.com/ocz-600w-s ... y-eol.html. You and LatiosXT may be correct about the PSU being the culprit. Since it crashes just as easily with sitting on the desktop as it does with running games, I'm more prone to think it is the motherboard. If I were to replace the power supply, I might throw money at it with apparently the best PSU ever according to testing (the Seasonic X-1250). The power draw of tasks doesn't seem to be linked to the crashes. As wazkatango mentioned, the GPU probably isn't the cause, since it crashed even with a 4850 from another rig. I updated my 4870's video card drivers and CCC to 13.1. I've seen fewer crashes after doing an upgrade install of Win7, and especially after reseating the main 24-pin power connector. The loose connector was directly responsible for power turning off and for it not turning on, confirmed by reseating everything but that connector, so that was at least part of the problem. It could be that the power supply is to blame just because of the connector might have breaks keeping it from staying in place, but that wouldn't explain reboots in the middle of nothing happening.

Nastyman and LatiosXT: In the course of doing this troubleshooting, I've spent several hours staring at the motherboard and didn't notice any unusual caps. I'll mention it if I see anything unusual. Two of the SATA ports were partially dislodged from the board a while back, but I don't think that would cause random shorts.

I hope you aren't using the ports that were "dislodged" from the motherboard!

You didn't test the ram one stick at a time like I told you to do. ANY errors RMA the stick of RAM, it is no good! This is causing your system to go crazy!

Nasty

Nastyman: I did Memtest86+ over. I started with one stick in slot 1, and it rebooted partway through the test. For slots 2-3, it would freeze or reboot during boot, and would not get past the BIOS. The codes from the motherboard LEDs were different each boot. Then I tried a different stick, and the issues continued in the same way. I think the RAM sticks can be eliminated as the cause since I tried a completely different brand of sticks from another system with the same results. I'm fine with replacing the motherboard or PSU, but if it's due to loose connections or something that will not be fixed by buying new parts, I'd like to make sure sure I know what the cause is. The randomness of it is what makes me uneasy about buying any parts as a solution.

As a side note on an unrelated problem (eliminated as a factor due to trying out a 4850 from another rig), my 4870 is an EAH 1GB made by ASUS, and has a problem common to that model from that manufacturer: the D601 LED on the reverse side of the card is almost always on, which indicates a "critical core power fault." This is worth mentioning since it seems just as random that the fault will be present as it is that the computer will reboot. I might be able to cram the card into my other system to test it (space restrictions were part of the reason I built the newer faulty rig), but it is more likely an isolated fault compounded by the instability of the faulty rig. The solution is to flash the card's BIOS, but it sounds riskier than your usual motherboard BIOS flash, especially given the unofficial nature of the fixes (this is assuming that all of the files are still available for download).

If I replace the motherboard with another LGA775 board, which one should I get? It's been four years since the system was built, so I expect the prices of the socket to have dropped enough to where I can afford whatever the best board created for the platform was. I have an M45 Platinum, which was a popular and recommended board at the time (though apparently it has not held up to long-term stressed environment testing).

I might start up the troubleshooting routine over at overclock.net. One thread mentioned driver conflicts from Ethernet port adapters. There's only so much time I can put into it before it starts becoming more time and cost efficient to buy a new board and sell the old one as-is.

After isolating the freezes down to the motherboard, I decided to put the EAH4870 HD into the other rig and run the cards in Crossfire, which worked for about a week. I had the 4870 intake fan close to the sound card, which I think led to a heat failure on the 4870. There I was, one minute everything fine and browsing YouTube, next minute the computer instantly shut down. After trying a bunch of stuff, I tracked the error down to a smokey-smelling 4870. I opened it up, and one of the black rectangle thingies was bordered by smoke stains and black goo. Image: Asus said it would cost $120 to repair the card I bought for about $125 (and can only sell for about $40 if it worked). They said I couldn't get it fixed at a service center, and GeekSquad said in their typically useless fashion that they didn't offer that service. Asus said that I couldn't buy the part from them or anyone else. Forum users, do you know what capacitor it uses, and if I can buy a replacement from anywhere else? Otherwise, I'm going to sell the thing as-is for about $4, which would figuratively stink as much as the card literally does right now.

The black rectangle thingie is called an Integrated Circuit and can't be bought on the open market because it was used/designed for a specific purpose on that video card....time for a new card that can run DX11.

Nasty