Random reboots

[Lowe]

I have a relatively old machine, a 2014 build made up of an i5 4690K, Gigabyte Z97MX Gaming 5, 16GB Crucial Ballistix Tactical RAM, Palit GTX1080 Jetstream. There's a couple of SSD's and a spindle drive in there. For years it's ran at 4.1ghz without any alterations to voltages etc in the BIOS.

About a month ago it started to randomly shut down and restart on the odd occasion. I put this down to the high ambient temps (was when we were having 30c and above weather) so didn't really think that much of it. But, over time it's started to happen more frequently, to the point now where it struggles to sit in Windows or even idling at the BIOS page without randomly shutting off within 5 minutes.

Initially I ripped out the RAM, tried each stick individually in each of the slots to see if that would make any difference, and it didn't. I then bought a new Seasonic 650w power supply which again made no difference. I removed all cards and reseated them, checked over the mobo for any obvious signs of bulged capacitors or heat marks to no avail. Temperatures on chipset, CPU and gfx card are all well within normal bounds.

I've also tried resetting to default chip speeds and BIOS settings, taking out the BIOS battery, clearing CMOS, turning off the onboard gfx, changing the PCIe slots manually to Gen3, enabling power loading in the BIOS but nothing's had any success.

So any ideas as to what else I can try?

[blueball]

Stating the obvious but I assume you cleaned out all fans and heatsinks to remove fluff/dust.

[Saracen999]

Stating the obvious but I assume you cleaned out all fans and heatsinks to remove fluff/dust.

That was going to be exactly my first idea. Happened to me about 20 years ago. Drive me nuts trying to find it and it turned out a fan was so dusty it was slowing it down, reducing cooling and ....

The fan? Oh yeah ..... inside the PSU. And anyone not electrically-aware needs to be careful messing about inside any PSU, because even when unplugged, PSU capacitors can carry quite a kick.


Similarly, I had a time where a cheap, old case fan had bearing issues. It was one of those 'orrible sleeve-bearing nasty things. At least there was a very slight screeching noise indicating a dry bearing. Replacing that fan with about £3 of ball-bearing fan solved the issue. I bought in a stock of decent ballrace fans after that.

But there are so many things it could be. A dodgy memory stick, a plug-in card not properly seated .... had that wirh a graphics card where the backplate wS a feaction too short for the case and when you screwed the retaining screw it lifted the card just enough to be right on the cusp of not making contact. It woukd then, in days or weeks, work that extra fraction of a mm more loose and .... reboot. Intermittently.

I replaced that backplate with one that looked identical but was about 1/2 mm longer and no more problem.

But it could be so many things.

An obvious question, Lowe, is had you just done anything just prior to the problem starting a few weeks back? If so, and if you can remember, that is a good place to start.

Had you moved the machine? Check everything is properly seated and not wiggled loose.

Had you added, removed or changed hardware? If so, that might point at an issue.

New software? Or new/updated drivers for hardware? Could be that.

Or any number of different bits of hardware could be intermittently faulty. That's hard to nail down, unless you have sufficient "spare" hardware to replace components one bit at a time, from PSU to RAM, to .... almost anything, really. And most people don't.

All you can really do is try to think it through, and work methodically. And if yoy're anything like me, keep notes or three weeks from now you won't remember what you've already tried.

Actually, scratch that. I'd have trouble remembering three hours from now. Now what were we talking about?

[Lowe]

Whole computer's been cleaned out yes. No issues with fans and the CPU has a Corsair AIO water-cooling system over it. Temps are actually pretty good, even under screaming load the CPU isn't going more than about 50c.

Nothing's changed with the machine over the last few months other than me buying a new keyboard. No moves etc other than taking it to a mates house after it got to the point of me tearing my hair out haha! I've ruled the keyboard out however since it still does it with no keyboard/different keyboards plugged in. No real changes in software or drivers other than the forced Win10 stuff.

I don't think it's a card since I've run it on onboard gfx etc and it still packed in. Given I've got two sticks of RAM in I've tried each individually just in case one was dying. I know it's not impossible but it's pretty unlikely to have both sticks dead.

Like you say Saracen, I don't have a parts cannon to hand to throw stuff at it. The PSU was a bit of a wasted purchase but the new one is modular so I'm kinda looking at that being a small upgrade haha!

I'm slowly coming to the conclusion that something on the mobo is on it's last legs

[Saracen999]

It might indeed be the mobo, but that's really about the last thing to replace as it's such a drastic upgrade and affects so many things, not least probably an OS re-install.

At this risk of climbing on one of my hobby-horses, this is one small part of why I detest the notion that an OS can silently upgrade itself .... that "forced Win10" bit.

Do you happen to have a spare HD somewhere?

Could you remove your current OS drive (Hard or SSD) and replace it with a temporary drive? If you have multiple drives, disconnect all the others so there is no possibility of compromising data, etc.

Then, on that spare drive, re-install an OS (not necessarily Win 10) and your major apps?

If it's the mobo, all this will make no difference and you'll still get reboots, which will eliminate and possibility of it being Win10 or any installed software or drivers. That is, at least you'll know it's hardware.

If, on the other hand, you get no reboots and the system is stable for a period in which you'd normall get several reboots, then it suggests the issue isn't mobo, isn't CPU, isn't cooling, in fact isn't hardware at all.



Or, you can just at it as fate telling you it's time for a new mobo, CPU, etc and a full rebuild is in the immediate future. This is especially attractive if you've been wanting to do that anyway.

[Lowe]

It does it just sat in the BIOS, let alone an OS

[Saracen999]

Mmmm.

I did have an issue many years ago that sounds a bit like that. I diagnosed a dead board and bought a replacement. Turned out there was a trIck to complete BIOS resets to 'factory' .... and it was neither obvious nor documented.

I can't remember the details but it was something to do with 'main' and backup BIOSes.

It was an old board, took a Q6600-era chip, but I can't remember the board. Pretty sure it was a Gigabyte though.

That said, I don't want to get your hopes up and it may be nothing to do with your problem.

Trouble is, I no longer have that board (I don't think) and cannot remember what the damn "trick" was. Not even sure it was official, but it worked.

I'll see if I can dredge my memory but ....

[MaddAussie]

Mmmm.

I did have an issue many years ago that sounds a bit like that. I diagnosed a dead board and bought a replacement. Turned out there was a trIck to complete BIOS resets to 'factory' .... and it was neither obvious nor documented.

I can't remember the details but it was something to do with 'main' and backup BIOSes.

It was an old board, took a Q6600-era chip, but I can't remember the board. Pretty sure it was a Gigabyte though.

That said, I don't want to get your hopes up and it may be nothing to do with your problem.

Trouble is, I no longer have that board (I don't think) and cannot remember what the damn "trick" was. Not even sure it was official, but it worked.

I'll see if I can xredge my memory but ....

Sounds like a Gigabyte board I had a few years ago..

[Kumagoro]

Have you tried stripping it down to the bare minimum and booting either a Linux live CD/usb stick or even just a Windows installer.

So just motherboard one stick of ram, no her or ssd and running off on-board graphics?

I have had SSD's in the past which would randomly crash Windows and they needed a firmware update.

Also try leaving the bios battery out overnight. I once had many years ago a mobo which wouldn't reset after a failed overclock even with the battery out for an hour or two. Also I would try re flashing the bios, even if at the latest version already.

[Lowe]

Have you tried stripping it down to the bare minimum and booting either a Linux live CD/usb stick or even just a Windows installer.

So just motherboard one stick of ram, no her or ssd and running off on-board graphics?

I have had SSD's in the past which would randomly crash Windows and they needed a firmware update.

Also try leaving the bios battery out overnight. I once had many years ago a mobo which wouldn't reset after a failed overclock even with the battery out for an hour or two. Also I would try re flashing the bios, even if at the latest version already.

Haven't tried a complete barebones build, or left the battery out for that long. Will give it a go. Like I say though, haven't bothered with Linux Live etc because it does it just sat in the BIOS screen, let alone some kind of OS.

I did risk it and flash the BIOS to the latest version just in case but that's not helped. Not sure I want to do it again because it's that unstable I doubt it'd stay up long enough to flash it.

[peterb]

Have you changed the CMOS backup battery?

[shaithis]

Sounds like a dodgy mobo, although I have seen the same with a dodgy power cable, always worth swapping that first!

[Lowe]

Have you changed the CMOS backup battery?

Haven't changed it, but I did pull it out and try it.

Sounds like a dodgy mobo, although I have seen the same with a dodgy power cable, always worth swapping that first!

Have tried 3 different IEC cables

[ik9000]

Have you changed the CMOS backup battery?

This^ if it's happening in BIOS and the temps are sound on everything it could well be this. Sure caps do fail, but I'd check the battery first. Only ever had to change one once but it did improve things no end on that machine.

[Lowe]

I’ll pick one up later one just in case ������

[AGTDenton]

Just my 2 pence

Disconnect all peripherals except keyboard.
Remove the GPU, and disconnect all hard drives, front USB/Audio/Firewire headers.
Remove any other daughter boards.
You should have only, Mobo, CPU & RAM.
You will now need to plug monitor into Motherboard instead of GPU.
See if PC still crashes in the BIOS with a barebones setup.

If in a barebones mode it still crashes, I have had a couple of instances where the integrated GPU portion of the CPU has failed on this era of intels. Disable the integrated GPU of the CPU and see if that helps (will need to plug back in the GPU to have a display).

If it no longer crashes in a barebones mode, start connecting the other devices one by one and see when it starts to crash.