Issues with Graphics Card

Is this a SC series card ? ... SC series are almost always reference PCBs and they tend to have issues with VRMs. See if you can duplicate it running Furmark ... take note of power, hat and voltages as it runs ... and if exceeds 82C end the test

http://www.ozone3d.net/benchmarks/fur/

Uh, everything I see about your PSU says it's trouble. it could also very well be the source of your GPU issues as 3D requires a higher power draw than 2D. That PSU could damage parts as it's known to use a low-end OEM to make the PSU. It might appear to work fine, but it could be stressing parts, especially as certain parts demand more power.

Try switching out your CPU with a known good quality PSU (borrow if need be) and see if the problem clears up. Also try a different GPU as well, just to see if the problem lies there. (I'd try both in the order listed, keeping the test PSU for the replacement GPU test.)

I've had a similar problem with GPU long time ago. It would work fine in 2D, but any attempt to run 3D game would end up quickly in artifacts and freeze of whole machine. Although it was possible to run some older 3D games (I remember 'Enclave' worked fine). This led me to assumption that it is some weird kind of hardware problem, and indeed swapping GPU solved problem. So I think if you already tried all software side solutions and nothing worked, it is also hardware problem with GPU. Although, if your motherboard has more then one PCIe x16 slot, I would move the card to another slot for a test to rule out bad motherboard slot before going for new card.

I guess you would need an expert that knows well how Nvidia cards work to answer the question why it happens. Perhaps you could try at Nvidia forums? (https://forums.geforce.com/).

If your GPU is still under warranty, contact the manufacturer and see about an RMA to get it fixed. (I'd keep an eye on the PSU too though.)

Thing is, some of the garbage PSUs sometimes will still seem to run quite well, that is until you dig deeper and see tons of deadly noise on the lines. JonnyGuru showed that with the worst Bestec PSU out there. It ran for them, delivered right voltages for the most part, but there was so much noise from bad parts that the power supplied would have been deadly to anything connected to it. Maybe not immediately, but it would be sooner than later.

Ok, assuming all fans are the same size and rpm, lets think about this.

1. Intake fans will have their flow impacted by air inlet filters; if ya not an anal nerd / geek (guilty as charged), these filters get clogged with dust. Fog machine testing has shown it can get up to 30% reduction. so lets call it a third to make matheasy. So what do we get

... 2 fans blowing in x 2/3 for air filter restriction = 1.33 air flow in
... 3 fans blowing out = 3.00 air flow out
... that 980 Ti is also exhausting some air thru the slot but lets ignore that

So we have 3.00 / 1.33 or more than twice as much air going out than in. Since the case hasn't implo2ed from the negative air pressure, air must be getting in somewhere to make up for that loss, right ?

2. At this point you are probably thinking "I heard of that, negative case pressure brings dust into t he case". And while that is true, it's the least of your concerns. Loom at pictures of ya case and see where the largest openings ... unfiltered openings are . This is almost always going to be the rear case grille and the vented slot covers. So we have identified the most likely places for air to get in ..

3. What do we know about the air back there ... up against the wall or perhaps under your desk ? That's where heat from the 250 watt GFX card is exhausting ... that's where your 850 watt PSU is exhausting ! And yes, much of that hot exhaust air is being sucked right back into your case thru those big openings.

So the 1st step in solving the problem is getting rid of that negative air pressure. With 5 fans either 3 in and 2 out or 4 in and 1 out is the only way to go. If ya try it otherwise, invest $33 in a garage band fog machine from Amazon, blow the fog behind the case and then watch the case fill up w/ fog.

I know, we all learned in 8th great earth science that hot air rises. And it does ... But if ya can understand the expression "don't pee into the wind" you will immediately understand why hot air rises is meaningless. Hot air does not rise up against a fan blowing the other way and the effect of that force on the fan is too small to measure.

The 980 Ti when installed w/ proper case ventilation should top out at 77C



At load that's hotter than any air cooled 980 Ti they tested other than the Strix, Colorful and reference cards.

Gigabyte GTX 980 Ti XtremeGaming 65°C
ASUS GTX 980 Ti Matrix 68°C
ZOTAC GTX 980 Ti AMP Extreme 67°C
MSI GTX 980 Ti Lightning 65°C
Colorful iGame GTX 980 Ti 82°C
Palit GTX 980 Ti JetStream 71°C
ASUS GTX 980 Ti STRIX 76°C
ZOTAC GTX 980 Ti AMP! 70°C
MSI GTX 980 Ti Gaming 74°C
EVGA GTX 980 Ti SC+ 76°C
Gigabyte GTX 980 Ti G1 Gaming 70°C
NVIDIA GTX 980 Ti 84°C

Click to expand...

2. After addressing the fan situation, You can take side panel off and use a desk fan to blow tons of air in there as a temporary measure) another thing we can check is this ... Run HWiNFO (Sensors only) and check the max voltages on the three power rails (3.3, 5.0 and 12.0 volts) when Furmark is shut down for reaching 84C

The ATX standard requires that voltage not waver beyond 5% ... in reality, for stability we want much more than that. Enthusiasts like to see , < 1%. If voltage on the 12v rail drops 5%... that means that to supply the same power, amperage must increase 5% ... more amperage means more heat Unfortunately, if it not more than the ATX standard of 5%, then you have no warranty recourse.

3. The card could be faulty. It is an SC series and, historically, that's a problem in and of itself. Call EVGA and tell them ya story, and what you have don to narrow the cause.... be prepared for "it's the [name 5-6 other PC components] fault responses tho.

Yeah, trying to get an estimate on power put it at around 567W with the 1080Ti... awfully close to max for that PSU... I'd actually say bump up a replacement/test unit to 650-750W. The R9-270X doesn't pull as much power (497W vs 567W) as a 1080Ti, so that might explain why it didn't crash the way it did previously. I personally would just get a replacement PSU anyway...

https://pcpartpicker.com/products/compare/R7V48d,HHx9TW,9q4NnQ,Rp8H99/

If you're in the States, one of those four in the link should make good replacements with a little more breathing space with a 1080Ti added.

OK I realize that ya spend a lotta time cleaning out dust but it has no impact. If you pulled out the dust filters completely, you still have the same problem. 3 is still bigger than 2 so more air out than in, the filters provide resistance all on their own even when new. Its might not be 3 > 1.33 with dirty dust filters ... its still 3 > 1.66 w/ brand new never used / freshly cleaned dust filters.

The purpose of Furmark was not to address the crashing issue, it was to address the temps issue. You really shud not be seeing anything over 77 and i expect recycling all that hot air thru ya case is the cause. The question is, if ya reverse the top fans, squished in as it is between furniture, will it suck down air climbing yo the wall.

That's where i was going here:

The ATX standard requires that voltage not waver beyond 5% ... in reality, for stability we want much more than that. Enthusiasts like to see , < 1%. If voltage on the 12v rail drops 5%... that means that to supply the same power, amperage must increase 5% ... more amperage means more heat Unfortunately, if it not more than the ATX standard of 5%, then you have no warranty recourse.

Click to expand...

Did you ever use HWiNFO to take the voltage readings ? If voltage drops, amps go up and certain circuits mat shut down or misbehave.