Troubleshooting PSU and PC Shutdown

[beavereater]

I leave my PC on 24/7. Have had zero issues with maybe the PC restarting itself in the middle of the night which could be that it's installing something, IDK.
I turn it off because I left Japan for 2 weeks (Christmas vacation)
Came back, and it wouldn't turn on. I tried unplugging it, cleaning it out, heating the room up (I don't know).
I bought a new battery for the mobo and a new PSU (Going from 650 to 750)
New mobo battery didn't do the trick. New PSU had me up and running. I started to render a video, very low quality.
About 30 minutes into watching some Youtube it just randomly shut off, completely off. No BSOD or reset, just OFF. My video didn't render (Obviously)
I was able to turn it on right away but about 2 or 3 minutes in it turned off again.

At this point I thought it was a heat issue, but CPUID and HWMonitor show normal temp ranges. I haven't tried this while rendering. However the second shut off happened with nothing running except Youtube.

I'm writing this from my PC (Please don't turn off) in the hopes that someone here might give me some insight into what is happening and what I should do going forward.

[JVRaines]

Bad solder joint(s) on a circuit board? Maybe they cracked when you powered off for the first time.

[vhelp]

I can tell you from my experience (desktop pc) in this same issue--the computer would shut down for no reason, or will not turn on again, or will turn on and then off intermittently.

I had this trouble for at least two years before I finally did something about it. I had an idea what it could be but was too lazy to do anything about since I am mostly on my laptop. But finally, I needed to desktop pc for some files on the hdd and the unit would not turn on to a full boot. It would bail after a few moments or seconds or less on power-on.

What the problem was is that the hdd ribbon cable was frail, worn or dried out. Hey, it was built in 2008, I think. Anyway. I learned it was the ribbon cable and replaced it, and sure enough the pc booted up. Mind you, I only did this one. I haven't had the need to use the desktop pc since. And that was about a month ago. I should give it a test run, just to be sure.

My problem, (I believe) is that I keep my apt cold during winter. My apt is all electric and expensive, so I do whatever I can to cut the costs down, including lowering the heat, like turning it off during the day (just before work) and turning it partly on when I arrive from work around 7pm or so, and I mainly just keep the computer on all weekend. Did this sort of practice since i moved into this apt. Anyway.

I believe that this and what jvraines stated is happening to you. If the motherboard or psu does not solve it, then look to anything cable-wise or ribbon, and replace it. You might be surprised when it boots right up w/out problems.

[larioso]

I had that graphics card became faulty at one time, and blow fuses in apartment even.
It was consistent in that every power up, it just blew fuses in apartment again.

This is if psu is not fast enough to discover overpower on MB or similar.
So thinking some component or card is at fault, or about to - and you psu is faster in discovering that.
This results in termination, no questions asked.
Check event log in administrator tools if any sign of message.

Or which also happend to me - a pci card is bent forward a bit and loose contact on some part of connector.
Pulling a cable from a card, like soundcard made it leave it's slot a bit.

One thing that might assist, turn off fast boot in bios so more memory tests are done - might give info.
But memory failure normally result in BSOD, but doesn't hurt if any strange message.

[Nelson37]

First, you need a good quality surge protector / battery backup unit, I highly recommend APC. Could be a building wiring fault. Improper building ground a major cause of such issues.

Next, disconnect everything not needed for testing. Printer, speakers, USB sticks, all external devices, include KB and mouse, because you do not need them to turn it on. Internally, disconnect power cables to ALL drives, ALL internal peripherals except the mobo. Disconnect ALL data cables from mobo. Remove ALL cards not needed to boot. If multiple RAM sticks, remove all but one, or two, depending on board. If you have a video card and also mobo video, remove card and connect to mobo.

Turn it on, leave it on, don't touch it, and see what happens. Recommend 2 to 4 hours. If nothing happens, shut down, and start re-connecting items ONE AT A TIME, while repeating test for each individual device.

Knowing the PC's age, also make and model, is often useful for diagnostic purposes.

Edit - When testing in minimal configuration, after test with new PSU, re-connect and re-test old PSU. EITHER old PSU failed on it's own, AND/OR some other mobo component also failed for no real reason, OR a power surge took out the PSU and additional components were made flaky, OR the new PSU is bad, OR the old PSU is fine but some other component failed in such a way that the new PSU can deal with it and the old one could not, OR the other component just elected to be dead when you first tried the old PSU. Knowing make and model of new PSU might be useful for diagnostic purposes. Can you borrow another PSU?

[beavereater]

Late getting back...

I've had zero issues with my PC and the new PSU...until:

I started rendering videos. It would near the end of the render and the the PC would turn off. I monitored CPU temps this time around and it went up to 50C and that's it. From there, it hovered and again right at the end it shut down on me.

I used Win Performance Re Assessment to put stress on the PC but AGAIN it just decided to crap out near the end. CPU temps never went above 50C.

What should I do?

ETA: New PSU is Thermaltake TR2 Gold 700W

[Nelson37]

go thru the bios and check for a thermal shutdown setting, go to event viewer and look for any pre-shutdown messages, stop rendering videos for a while and if it still shuts down, then do what I told you a month ago.

How's your heatsink paste, and do you have access to a digital thermometer to double check cpu and chipset temps?

[Krispy Kritter]

Shutdowns are typically a hardware fault. Most commonly the power supply or the motherboard.


Haven't seen that power supply model rated yet. Based on previous models, it could be good or not. I recommend Seasonic.

Reference here: http://www.tomshardware.com/forum/id-2547993/psu-tier-list.html

[beavereater]

OK. I'll go through the process when I get home tonight. I've encountered other random errors such as my usb devices not connecting etc. Wondering more and more if it's a mobo issue.

I haven't put new paste on the CPU since I got it. The heatsink is a massive one (Can't remember the brand/model off the top of my head)

By digital thermometer you mean software right? I do and the temps are going only as high as 50 degrees celcius. If you mean something else, then I don't.

Will post again in the evening.

ETA: 50 degrees when rendering. Below 30 when idle.

[redwudz]

That doesn't really sound like a thermal problem, at least with the CPU.

And putting in a new power supply should have removed the PS as the problem.

If you really think it's a CPU thermal problem, you could run a CPU stress test like ones here:
http://files.extremeoverclocking.com/file.php?f=205 Monitor the temps, though.

I would normally unplug all unneeded cards, USB devices and internal drives to rule these out.
You can run a stress test with minimal setup as it will also stress some other parts of the MB.

I have had drives, cards and RAM overheat and cause problems with shut downs, but those are rare.
Sometimes just unplugging all connectors and re-plugging them can help. RAM also.

If you rule out all those above, then the motherboard may be the problem.

EDIT: Sorry, Beavereater, I didn't edit your post, just hit the wrong key.

[redwudz]

That doesn't really sound like a thermal problem, at least with the CPU.

And putting in a new power supply should have removed the PS as the problem.

If you really think it's a CPU thermal problem, you could run a CPU stress test like ones here:
http://files.extremeoverclocking.com/file.php?f=205 Monitor the temps, though.

I would normally unplug all unneeded cards, USB devices and internal drives to rule these out.
You can run a stress test with minimal setup as it will also stress some other parts of the MB.

I have had drives, cards and RAM overheat and cause problems with shut downs, but those are rare.
Sometimes just unplugging all connectors and re-plugging them can help. RAM also.

If you rule out all those above, then the motherboard may be the problem.

[pandy]

Try this one http://www.ocbase.com/

[Nelson37]

I asked for the PC's age and make and model. There is a reason for this. Post the information needed to assist in solving your problem.

Describing the other random errors with something other than "etc" should also be useful. Any one of them might point specifically and clearly to the problem.

Random guessing will eventually hit on the correct solution, YOU providing additional information that only YOU possess might solve the problem that YOU have in a faster, more efficient manner.

By digital thermometer I meant a separate piece of hardware.

Currently needed is information related to shutdowns and other errors when NOT encoding or CPU testing in order to separate thermal issues from just general flakiness, which thermal stress will often exaggerate. The other issues you mention need to be described in detail.

Or, YOU can fail to provide information necessary to solve the problem that YOU have, or take another three weeks to respond, in which case I and possibly others will conclude that YOU simply do not have any strong desire to solve the problem that YOU have and lose interest.

[beavereater]

@Nelson37

Sorry, tge tardiness to my replies here are work related.

My PC is home made. Mobo, CPU etc naje/model information I'll post when I get home tonight. I built the PC in 2011.

[beavereater]

Specs:

AMD FX8350
Gigabyte 990-FXA-UD5 (8 GB RAM)
GForce GTX 660
3 SATA HD (2TB, 1TB, 500 GB)
1 SSD Drive (OS Win 7 64bit)

Not sure what else to list.
I'm thinking maybe it's time to start building a new PC anyways. Still, would be nice to have a working spare to do some tests with down the line.

I also noticed, when my case was open and PC turned on, that a cord was dangling into the fan of the PSU. So, there wasn't any fan going on the new PSU for I don't know how long.

I have to put some of the pieces back together tonight. I have my case and parts all over.

[Nelson37]

Numerous questions were asked for the purpose of solving your problem. You have failed to answer nearly every one of them. I dislike wasting my time to no purpose.

Good luck to you.

[beavereater]

Thought I would update this:

I pulled everything apart. Saw that my cable management was getting out of wack, cords dangling into places they shouldn't be etc. Replace the SATA cables for the HDDs, plugged everything in (Nice and clean) and now I've got no issues. I've been rendering 4K all day without a problem. No shutdowns, no missing drives, nothing.

While I can't say with absolute certainty that the cables were it, I believe my troubles are over. However, I have been looking at new PC parts and now I'm thinking about building a new machine and keeping this one on the side for testing purposes.

[jollyjohn]

I had similar problem withe two PCs. The power switch got stuck from dirt build up inside. After dipping them in acetone and pressing the button many times they were OK, but the first PC that had the problem got the MB ruined from trying and trying the wrong fixes.