Hey lads, having some graphics card trouble

I’m at the point of suspecting I’ve a dead graphics card (980ti) on my hands, but I thought I’d pop the full set of problems on here in case anyone has a bit of brainstorming on potential alternatives.

So, earlier today I was playing a little Witcher 3 when the computer crashed without warning or error. Automatic restart, but no display signal and the vague smell of smoke. Uh oh.

First things first, tried to work out where the smell was coming from, but every other part of the computer seemed to be booting correctly including all the fans (GPU fans as well). Bit tricky to work out where a smell is coming from when you have 7 fans blowing air all over the place. Oh well.

First things first, tried a different video cable and connection type. No output with hdmi, same with DVI. Okay.

Tried a different card from another computer. Computer boots up fine, and now there is display. Uh oh. Sounds like a new graphics card is on the…cards (oh ho!). Annoyingly, can’t test the dodgy card in the other computer, because the card is massive and won’t fit.

The card was adequately powered (650w PSU, gold-standard), I’ve ensured all connections are good. I’ve done a scan for fried capacitors or anything along those lines, there’s nothing apparent.

While GPUs are expensive and that’s annoying, my major concern is considering why there might have been a failure, just in case another element of the PC is playing up and might fry a replacement. The computer is showing no signs of unusual behaviour otherwise, so I’m not inclined to suspect the PSU of being dodgy. Having said that, we’ve had problems of poor grounding in the wiring of this apartment, so that’s always a concern.

It’s a bit late, so I apologise for this message being rather haphazardly structured :smiley:

Anyone have any thoughts?

Sounds like a good enough reason to sell a kidney, remortgage your house and turn over a local post office in exchange for being able to afford a new RTX card…

But yes. Swapping over the card and having the system boot does kinda point to the problem. Give the card a good sniff too when it is out the system… you can tell when it’s melted. Chances are it’s just a cap or on-board fuse or something simple. Does the fan feel seized at all? Did you experience any performance downgrade or throttling just before failure?

What make of card is it. Some cards have silly long warranty periods on them… and as gfx cards aren’t cheap, it is worth RMAing if you can.

I’ve been in the same boat when I wasn’t sure the cause of taking my card out and fearing the psu. Are you running a surge protector? Have you had random reboots or spikes? Do you have a decent multimeter to be able to test the outputs? Does the PSU smell at all?

1 Like

Try running the pc with the other card in it for a ew hours to see if it crashes?

1 Like


Let’s put the kidney+remortgage plan down as ‘backup’, though I like your gumption!

It’s clearly not a big melt if so, as there’s a light smokey hint but nothing profound. The caps seem good, but it’s a little tricky to visually spot problems with the fuses. The central fan ‘might’ be running a little sluggishly. I think it is, but it’s not like I stared at the thing when it was running normally and took RPM measurements.

No performance downgrade or throttling before failure, it was a sudden death.

Inno3D x3. 3 year warranty which ended almost to the day last month. Pardon me while I shout into a pillow…thank you for your patience.

No surge protector. Silly I know, but I found them strangely hard to find in Belgium when I was setting up the computer and, as is the way of things, it drifted out of mind. Paying the price now, perhaps.

We did have the on the living room circuit trip a week or so ago and it seemed like it might have been connected to me removing something from a USB socket. Given those things push less than an amp, I’d figured it was probably just a coincidence of timing, but given this perhaps there was more to it.

I have a decent enough multimeter - Extech 330. Nothing spectacular, but should do the job. Bit of an annoying card to get into, unfortunately, they’ve plated the thing up like a tank with the tiniest screws known to man.


I would, but given that the spare card is Morrow’s and if I fry that as well she’ll actually eat my liver, I’m a little hesitant.

I’d say that the easiest way to know for sure is to strip the graphics card down and check the board for fried components.

You might be lucky and find its just a cap that exploded which you could replace or get replaced.

Did the other graphics card you test with use the PCIe additional power cables? Its possible that you may have lost a 12V rail on your PSU if it is the multiple rail type. Most decent power supplies are single 12V rail though and something popping in the PSU would likely result in it just shutting down which doesn’t seem to be happening for you.

1 Like

Yeah, that’s next on the list. I’m doing a hail mary warranty request in case anyone at Inno3D support is particularly nice or unobservant about dates, but if that comes up no then I’ll pull it apart and see what happens.

Yeah, same power cables were used and it’s a single 12V rail PSU. Good thought though, cheers.

1 Like

Try checking the 12V rail, just measure if its within specs. Otherwise sounds suspeciously like you got a fried GPU, still the PSU would be a candidate! With no way to check the gpu though any discussion is a bit pointless.

Checked the 12v rail, voltage difference of 12.18V. Sounds okay to me. Agreed?

Sounds good, bit weak of a test to be fair. Would need to run it under load for some time and see if flactuates a lot even if not it might still be broken. Still with the other gpu working it points pretty harshly to a broken gpu.

Yeah, was thinking the same. There’s always the chance there could be some instability in the PSU, but don’t really fancy standing there for a couple hours looking for voltage flickers :smiley:

Mm, probably true, sadly.

Could still be the PSU, was the other card you tested with similar?
I’m just thinking that the PSU might not work properly when under high load such as gaming. Could have then blown the card. So risk is that you might get a new card and that is going to blow.
I’m fresh out of ideas on how you would test this without an actual psu tester.