Does this sound like MOBO or PSU?

[beavereater]

PC restarted as I wrote...f@#$ me...

The short: I have 4 drive, 1 SDD drive with OS and programs on, 1x 2TB drive with important files, 1x 1 TB drive with photography files, 1x 250GB with my Lightroom Catalog (Photography related).

I didn't realize until today that my 1TB drive was missing. I've had issues since my new build. Slow to start, stuck on Startup screen, etc. even with a SSD drive. But, I decided to do some editing and it took me a couple of hours to realize 'my 1TB drive is missing'.

Is shows up in my BIOS, all 4 drives. A few times, after some cable swapping and port swapping, it didn't...but it seems I've found the perfect combination (Or so I thought). Now, all drives are up but anywhere from 20 minutes to 2 or 3 hours I get a sudden restart without a BSOD. It just restarts OR shuts off/down completely (Kinda like my wife...I can only say that here)

I wrote this once, thinking I could get by, but since I couldn't without a restart I decided to unplug my 1TB drive (I've got it all copied to an external drive).

I don't think it's the drive, I can't figure out if it's the MOBO...but I wonder if it's because of the PSU since I've got all this hardware (Although I would think I had the power...only after unplugging the 1 drive will I really know...kinda...)

Is there an app to test PSU quality? How about HDD quality (The Western Digital software says my drive is fine...even though it caused major issues before) How can I truly know??

Worst case, I see myself sending in my HDD for a new one (Free)....then getting a new PSU (Money)...then a new MOBO (Money and more time than I care to waste...but will anyways)...

Any advice?

[jagabo]

You can check the CPU voltage levels in the BIOS. Or with any number of programs like Hardware Monitor.

[usually_quiet]

Random crashes may be caused by driver problems or bad RAM. Have you run WhoCrashed to eliminate the possibility a driver problem, or tested the RAM with memtest86 or memtest86+?

[ranchhand]

I tend to agree with Usually_Quiet as to the most common reason. I would add overheating as a real possibility. For testing, I have the Apevia Power Supply Tester; it also gives voltage readouts on all rails/connector cables. It has saved me headaches at least two times in the past, I always test all components on an open bench build before mounting in the case.
http://www.newegg.com/Product/ProductList.aspx?Submit=ENE&DEPA=0&Order=BESTMATCH&N=-1&...+supply+tester

[nic2k4]

If you feel your PSU is undersized for your rig, you can easily verify by adding up the Amp requierements of all the devices. All you need to know is how many Amp a device needs for each Voltage it uses. For HDD and DVD drive, the info is on the sticker (the datasheet has more details), add-on cards should have the info in the manual (if not default to 0.5A @ 5V). The device with the biggest draw might be the videocard, if that's the case check the manual, it should state a minimum PSU size. The CPU also is a big load, but on the 3.3V rail and if it has a special PSU requirement, it should say so in the manual.

Just add-up all the values and compare the result to what's on the PSU sticker. To be safe no rail should be running above 75% capacity, unless you have a PSU that is rated for full load operation. HDD's need more power to startup, don't use that value other than to check if the total exceeds 100%. If it did you would have problems at startup. The delayed start option in some BIOS's is meant to deal with that.

Try CrystalDiscInfo to check the actual S.M.A.R.T. data for your drives. That will give you an idea whether a drive is accumulating errors. Re-allocation errors are a sign of bad data cable, but a bad controller will do the same. In that case the errors will keep increasing after replacing the cable. When troubleshooting don't cheap out on the cable and remember price does not equal quality.

[jagabo]

Keep in mind:

Bargain basement PSUs typically can't even output half their rating.

Testing by Google found that SMART only predicts about 1/3 of hard drive failures.

The symptoms you described could be caused by pretty much anything in your computer -- hardware and software.

[wulf109]

Dis-connect the 3 mechanical drives and boot with the SSD only and see what happens.

[beavereater]

I did memtest (The one that comes with Win7 that does a test on reboot) and Win7 says memory is fine.

WhoCrashed had 2 events from April both said:

Error: DRIVER_POWER_STATE_FAILURE
Bug check description: This bug check indicates that the driver is in an inconsistent or invalid power state.
This appears to be a typical software driver bug and is not likely to be caused by a hardware problem.
The crash took place in the Windows kernel. Possibly this problem is caused by another driver that cannot be identified at this time.

Using Event Viewer I was able to see what time my system restarted. The event ID was always something about 'Win couldn't shut down properly.' 'Win rebooted unexpectedly.' But no reason that I can figure...

My current PSU is a quality brand, however I might be pushing it beyond that 75 percent load. CPU heat is low, 12c idle (It's cold as fook up here) and it's got a nice coolmaster heatsink on it.

The SSD drive is using a new, quality SATA cable since I had to buy it separately from the drive. The other 3 cables are just the cables the drives came with.

I'm not sure if it's the controller port, the cables or the PSU. I'm guessing that the first step would be to do as wulf109 says, see if I can keep things normal on just the SSD drive. If that fails, what do I do (New PSU?). If the SSD on it's on works, then move all the old SATA cables to newer cables. If that fails??

[jagabo]

My current PSU is a quality brand, however I might be pushing it beyond that 75 percent load.

From your description of when the crashes occur that doesn't seem too likely:

I wrote this once, thinking I could get by, but since I couldn't without a restart I decided to unplug my 1TB drive

If I understand correctly, the computer crashed while you were editing a post. Unless you were also running a video conversion or game in the background your CPU and graphics card usage (the two biggest power consumers) were minimal.

CPU related problems are pretty easily tested using programs like CPUBurn. Monitor temperatures and CPU usages while the CPU is running at 100 percent.

Speaking of graphics cards: Microsoft (from crash reports) indicates ~70 percent of Windows crashes are caused by graphics card driver bugs.

[beavereater]

Alright...I don't know what to think at this point. I went into my case and made sure all the cables were on tight. Might have been a few that were loose, I don't know. With my 1TB still unplugged I've been running things fine. I haven't done anything major but there have been no restarts/reboots.

My cable management isn't that great on. I wonder when shutting the case if I'm knocking something loose. Gonna try CPUBurn. I haven't done anything with my graphics card/driver yet. Hopefully I can get everything running with my 1TB plugged in. I'll pick up some new cables tomorrow. I'm going to leave my PC tonight and see if anything forces a restart/reboot.

All reboots happened while idle really. Nothing running in the background, no photoshop, premier pro or any CS6 app. No games...nothing.

I hate not knowing what is causing it. I'd rather it be the PSU (Easy to buy, swap out). Putting in a new mobo is like building up a new PC. Anyways, I'll post back when I know more from CPUBurn and the over night wait.

ETA: CPUBurn for about 10 minutes. No stability issues. Core Temp monitor never went higher than 35c. Does this at all rule out PSU problem?

[jagabo]

With my 1TB still unplugged I've been running things fine.

That makes the drive suspicious. Is that a USB drive? Maybe an IRQ conflict?

ETA: CPUBurn for about 10 minutes. No stability issues. Core Temp monitor never went higher than 35c. Does this at all rule out PSU problem?

No, but it makes a PSU problem less likely. Did you see 100 percent CPU usage while running CPUBurn? I mention that because I have an old version which is single threaded. So I need to run multiple instances to get 100 percent CPU usage.

[ranchhand]

I suggest using (free) Seagate Seatools for DOS drive tester (I assume your 1TB drive is mechanical not SSD). Download the file, burn to a CD and reboot. It runs from outside Windows and will give you a pass or fail. You have to start testing these components realtime, otherwise you are guessing.

[jman98]

I had a problem a couple of years ago with 2 internal 2 TB SATA drives vanishing from my newly built PC. I was booting from SSD so it wasn't killing me but it made my PC very unstable. I checked my SATA cables and they were fairly cheap ones that because the two 2 TB drives were further away from the mobo connection than the SSD was, the physical connection on the 2 TB drives wasn't very secure. I replaced my cables with ones that had connectors bent into something like an L shape and the connections stayed solid and my disappearing drive issue never happened again. This may not be your problem, but it could be.

[nic2k4]

Another couple things to look at, is your BIOS up to date, what about the SATA driver and check if there are any firmware updates for the HDD.

[deadrats]

I had a problem a couple of years ago with 2 internal 2 TB SATA drives vanishing from my newly built PC. I was booting from SSD so it wasn't killing me but it made my PC very unstable. I checked my SATA cables and they were fairly cheap ones that because the two 2 TB drives were further away from the mobo connection than the SSD was, the physical connection on the 2 TB drives wasn't very secure. I replaced my cables with ones that had connectors bent into something like an L shape and the connections stayed solid and my disappearing drive issue never happened again. This may not be your problem, but it could be.

i second this; there was another guy on here years ago that had a similar problem and i suggested he get himself those SATA cables with the clip that keeps them on tight and his problem vanished as well. i also recall back when SATA hard drives first came out i was having a problem similar to the OP's and it took a while to realize that with the cheapo SATA cables even though it seems that they are on nice and tight they tend to work themselves loose as the drive vibrates from the spinning platters, switching to locking cables solved the problem and believe me when i say i damn near lost my mind diagnosing everything else from mobo to psu to ram to running drive check before i figured it out.

[p_l]

Yup, definitely recommend those locking SATA connector cables, whether straight or right angled.

[redwudz]

Sorry, joining this discussion a bit late, but I agree with the previous posters.

Most times, running down a problem consists of eliminating everything that isn't the problem. Then what's left over usually is.

First unplug all drives and all cards not absolutely necessary to boot. If you still have the problem, go into 'Safe Mode' and see if you still have the problem.
That can sometimes rule out software drivers.

I tend to substitute a PS right off if I suspect it may be problematic. But not all of us have a spare PS setting around. I have very rarely seen problem caused by a PS that was overloaded unless there were some warning signs. An especially hot exhaust from a PS or the smell of hot plastic are a couple of clues. I run a couple of my servers with 400W PSs and 9 HDDS and a DVD RW drive. True, I don't have a heavy duty video card, but they do run cool.

I hate to say bad things about Seagate if that's your problematic HDD, but I do have 4 dead Seagate drives on my bench. I still have ancient Maxtor drives that run OK. Not too encouraging. Test out that drive.

Cables should always be checked, especially SATA data cables. I also have a stack of those I have thrown out.
The SATA data connecter is a terrible design. I find with continued usage they quickly become loose.
Using the ones with the metal attachment clips helps a lot.

Sorry for rambling. It's late here.

Other than that, keep us appraised. If it does turn out to the be MB, I hope it's still under warranty.