System crashing, but only when graphics card installed

M

Murolith

n00b
Joined
Aug 21, 2015
Messages
13
So I'm at the end of my rope outside of a couple of hardware swap tests that remain.

About 3 weeks ago or so my computer hard crashed while playing a game, and it acted a little strangely since, but whatever.

It's a mobo from an old Lenovo with an Intel 6500 CPU, Geforce 1050 TI and 8GB of RAM and a 550W Antec PSU.

The other night I tried firing it up to rip a CD (yeah I still buy music like an old fart) for my wife, and shortly (within a minute or so) after Windows loaded, the computer restarted as though I'd hit the reset switch. When it restarted, I had no mouse or keyboard control, or even power to them (no lights) and shortly after being hung up, it rebooted again and again and again until it launched Recovery mode.

Safe mode boots just fine. I've reinstalled Windows 10, and it loaded perfectly without issue...until the video card driver installed and then the system crashed in the same loop again. Safe mode still works fine.

Remove video card, and use integrated, and it works fine.

Slide in old ATI/AMD card: windows 10 starts, but never finishes loading the desktop, no icons, no input other than Ctrl Alt Del (which gives a black screen that I can hit escape to exit from and that's it) and eventually it reboots (takes longer, though) and when it does and I go into Safe Mode I found that it loaded the default driver for that card on bootup.

Essentially, no matter what card I plug into the PCIe slot, as soon as drivers load beyond Windows basic video driver, the system crashed. Works fine if I totally block all drivers from installing after removing them (which is a LOT harder to do in Windows 10 than it should be).

Tested the 1050TI out in my son's rig, and it works fine even on a burn in test of Furmark for an hour.

TL:DR: WTF?

Motherboard, CPU, and PSU are my only apparent options, right?
 
First step might be to look at the Windows event logs. You can filter by critical events and see what's going on when it crashes. Sometimes you don't see anything useful, sometimes you do.
 
The only thing it's said so far is Kernel Power system crash: it's the not helpful type. Thing is when it crashes it just hard crashes without a dump or BSOD or anything.
 
You have to clean the drivers when swapping out video cards between different manufacturers, otherwise you will run into the issue of Windows never fully booting as you did.

Is this the motherboard?

1611346295737.png


If it is, this motherboard came from a low-power SFF meaning it is possible the x16 PCI-E slot is providing the full 75W of the specification. Custom motherboards from pre-built manufacturers have been known to skimp on power delivery to the PCI-E slot before, delivering as little as 30W.
Murolith said:
The only thing it's said so far is Kernel Power system crash: it's the not helpful type. Thing is when it crashes it just hard crashes without a dump or BSOD or anything.
Click to expand...
That is just saying that Windows detected a power interruption or failure. It means there is an issue somewhere with power delivery. Since both video cards are giving you issues I am going to say it is either the power supply that is bad or the PCI-E slot is bad. I would try putting the video card into a different x16 slot if you have one and see if the issue goes away. If it doesn't then swap in a different power supply if you have one available.
 
The clues are pointing me toward possible PSU issue or mobo. Have you tried using a different PCI-e slot?
 
Hardware related from the get go....but by all means keep reloading Windows. 😬

Sounds like power supply and adding a graphics card then crashing is a symptom. Or bad graphics card/slot but sounds power related.
 
Armenius said:
You have to clean the drivers when swapping out video cards between different manufacturers, otherwise you will run into the issue of Windows never fully booting as you did.

Is this the motherboard?

View attachment 321610

If it is, this motherboard came from a low-power SFF meaning it is possible the x16 PCI-E slot is providing the full 75W of the specification. Custom motherboards from pre-built manufacturers have been known to skimp on power delivery to the PCI-E slot before, delivering as little as 30W.

That is just saying that Windows detected a power interruption or failure. It means there is an issue somewhere with power delivery. Since both video cards are giving you issues I am going to say it is either the power supply that is bad or the PCI-E slot is bad. I would try putting the video card into a different x16 slot if you have one and see if the issue goes away. If it doesn't then swap in a different power supply if you have one available.
Click to expand...
that's not the specific motherboard, but it's in the right ballpark. I'm thinking that's the problem as it was an OEM SFF system I pulled it from. As for the drivers, I used DDU to uninstall them every time I tested, and the same thing keeps happening. I don't know of a more thorough way to clean them out, but I'm open to suggestions.
 
Last edited:
Bowman15 said:
Hardware related from the get go....but by all means keep reloading Windows. 😬

Sounds like power supply and adding a graphics card then crashing is a symptom. Or bad graphics card/slot but sounds power related.
Click to expand...
Only reason I reloaded Windows the first time was from the fact that it was fairly unstable from the get-go, then got better for a month, then went nuts and I had swapped the mobo without doing a fresh install when I installed it so I figured, screw it, i'll go with a fresh install and see what happens. Didn't really take long, and it's looking like I'll need to again if it's the mobo/CPU anyway.

What I wish I could do was swap just the mobo and keep the CPU and RAM handy for now because I'm a cheap bastard who doesn't want to shell out 300-400 dollars for a new core system right now, but I may not have the luxury of being picky.
 
Murolith said:
Only reason I reloaded Windows the first time was from the fact that it was fairly unstable from the get-go, then got better for a month, then went nuts and I had swapped the mobo without doing a fresh install when I installed it so I figured, screw it, i'll go with a fresh install and see what happens. Didn't really take long, and it's looking like I'll need to again if it's the mobo/CPU anyway.

What I wish I could do was swap just the mobo and keep the CPU and RAM handy for now because I'm a cheap bastard who doesn't want to shell out 300-400 dollars for a new core system right now, but I may not have the luxury of being picky.
Click to expand...

Hey, not picking on you. That's how you learn by going through this type of thing. Just keep narrowing it down and you'll find the issue! 😁
 
Murolith said:
I wish I had one, but it's only got the single X16 slot.
Click to expand...
Pretty sure it's the mobo, if integrated works fine then that makes it even more likely. The updated driver for AMD or nvidia is probably enabling the higher power modes and the slot can't provide the needed 75W anymore.
 
Test a yellow molex wire for steady 12.0+ volts with an accurate Digital voltmeter when the vid card driver is installing to see it drops below 11.8V
 
Bowman15 said:
Hey, not picking on you. That's how you learn by going through this type of thing. Just keep narrowing it down and you'll find the issue! 😁
Click to expand...
Yeah, no worries. I wasn't offended. In general I try to exhaust software as the cause when possible before hardware since it's the cheaper fix (and the WAY more difficult to track down to the granular level most of the time).
 
learners permit said:
Test a yellow molex wire for steady 12.0+ volts with an accurate Digital voltmeter when the vid card driver is installing to see it drops below 11.8V
Click to expand...
I have a spare PSU, though, that was working fine last time I tried it. Got the 550W because I was moving to a more power hungry video card, but my system as is is actually less power hungry, so it should be fine to test. If the same thing happens I'll still test the 12V rail and see what's up. I suspect you guys are right in thinking the mobo isn't supplying the power to the video cards correctly, though. Weird that it would be fine for weeks and then crap out, though.
 
Well...so far with my old 450W PSU everything seems fine..

I'm entirely okay with this.

I guess my newer PSU just got itself a weak 12V rail over time and shit the bed.

That's a much cheaper fix.
 
I had a similar problem with my GTX1070Ti, it would have random restarts while playing some games, particularly Fortnite and CoD Warfare. After a while there were random reboots on startup or a few minutes later.

I tested another PSU and the problem persisted. Turned out both PSUs were faulty.
 
You must log in or register to reply here.
Back
Top