NVIDIA on the Cause of RTX 2080 Series Card Failures

FrgMstr

FrgMstr

Just Plain Mean
Staff member
Joined
May 18, 1997
Messages
55,142
NVIDIA on the Cause of RTX 2080 Series Card Failures

If you are a fan of games and computer hardware, especially GPUs, you have undoubtedly seen anecdotal evidence that NVIDIA's new RTX 2080 series cards are failing due to artifacting, simply stopping working, and in some cases actually catching on fire. Now it seems that NVIDIA is owning up to there actually being an issue with its new RTX cards.

If you like our content, please support HardOCP on Patreon.
 
wow...

i was actually expecting them to somehow say these failures are a "feature"

still seems like a rather BS answer to me though...
 
Calling it test escapes could be any component on the board that went bad from the cheapest part to the core. That's not letting the public know what was wrong with the cards and what steps they are taking to improve quality. People want to know so that they can feel confident in purchasing NVIDIA products and know that it will not start a literal fire.
 
The ONE time I jump on the early adopter bandwagon.... Hopefully the 2070's are immune from this.
 
Last edited:
They should know for a fact what range by serial number is impacted. If they want to maintain consumer confidence they should do an optional recall/RMA of sorts for every affected serial number.
 
So the 2080ti is the Chevy volt of gpu?
 

Attachments

  • volt9.jpg
    volt9.jpg
    74.3 KB · Views: 0
I was going to pass on the RTX 2080 Ti until I saw Dell recently released a 49" ultraultrawide monitor - I'd like to move from a 1080 Ti and a 3440x1440 display to an RTX 2080 Ti and a 5120x1440 screen.

But now that it's clear Nvidia and its partners are charging a premium price for a product made with bottom-barrel components, I'll keep waiting...
 
Gigantopithecus said:
I was going to pass on the RTX 2080 Ti until I saw Dell recently released a 49" ultraultrawide monitor - I'd like to move from a 1080 Ti and a 3440x1440 display to an RTX 2080 Ti and a 5120x1440 screen.

But now that it's clear Nvidia and its partners are charging a premium price for a product made with bottom-barrel components, I'll keep waiting...
Click to expand...

All the tear downs show the card is actually overbuilt part wise.

The issue is that some faulty parts were allowed through due to a QC human error screw up.
 
wow I am shocked Nvidia is owning up to it to be honest. Hopefully they will be able to fix the issues they are having for people cause man $1200+ for a video card to have known failures?

I mean shit after BF5 with Ray Tracing showing to be a failure (IMO might be both dice and Nvidia) whats the point of even buying one right now? Too bad Pascal 1080ti cards are becoming harder to find.
 
I'm really nervous hearing about all of these failures. My RTX 2080Ti arrives tomorrow. I'm going to use it with my existing motherboard and CPU for a few days before I install new ones!! If this thing goes up in flames, I don't want my new motherboard, 8700k, and DDR 4 going up with it.
 
Brackle said:
wow I am shocked Nvidia is owning up to it to be honest. Hopefully they will be able to fix the issues they are having for people cause man $1200+ for a video card to have known failures?

I mean shit after BF5 with Ray Tracing showing to be a failure (IMO might be both dice and Nvidia) whats the point of even buying one right now? Too bad Pascal 1080ti cards are becoming harder to find.
Click to expand...

The failiure there is on DICE for not delivering a game with a decent DX12 rendering path that doesn't cripple performance regardless of DXR being used or not. DX12 runs on shit on both AMD and nvidia in all the frostbite engine games.
 
If this becomes a deeper "problem" for nVidia, beyond the little smile on Lisa Su's face, if she (has taken) some risk, and 7nm RX Vega 64 is (happens to be) close to release (liquid as previous gen) - it could be just the gap filler for a RTX recall we could be looking for. AMD may not be able to hit 2080 Ti performance I suppose, however 2080 mark should be attainable for them which in this case may not be so bad (for short term).
 
Last edited:
Every ubergeek teardown done on these cards had the opinion that they certainly "appear" to have been overbuilt as hell.

If I was Nvidia, I'd be pretty demoralized when their reference design is this beefy and yet somehow just plain bad parts somehow end up in a design that strong.
 
Advil said:
Every ubergeek teardown done on these cards had the opinion that they certainly "appear" to have been overbuilt as hell.

If I was Nvidia, I'd be pretty demoralized when their reference design is this beefy and yet somehow just plain bad parts somehow end up in a design that strong.
Click to expand...

Yeah, the teardown showed that the reference build could support fairly insane power levels if you kept the card cooled, and gamersnexus has shown that even the somewhat limited FE blower keeps all the parts (even on the cards that are failing) well within acceptable temp ranges.

Whoever nvidia contracted out to build these things fucked up big time with the QC.
 
NVIDIA havent cut any corners with part selection or build. The fire I cant work out considering all the power inputs are fused.

Probably a bad batch of a part or soldering issue caused by thermal fatigue.
 
"NVIDIA hasn't changed the vram supplier. They had an exclusive contract with Micron for the release cards.
Samsung delivered exclusively vram for the Quadro cards and Micron for the RTX Gaming cards.

There are three manufactures for GDDR6 VRAM - Samsung, Micron and SK Hynix.
We'll probably see cards with SK Hynix GDDR6 as well.

That the new cards are using Samsung VRAM has nothing to do with the failure of the cards with micron memory."
Click to expand...

Quote from the geforce forums linked in the article. Would make sense that the initial cards having problems would have micron memory since they all have micron memory. Of course thats only if what the poster is saying is true...
 
aokman said:
NVIDIA havent cut any corners with part selection or build. The fire I cant work out considering all the power inputs are fused.
Click to expand...

Actually, the fin stack on the 2080 FE (Reference) cards is a bit anemic. Within spec, but only just.
 
Krenum said:
Actually, the fin stack on the 2080 FE (Reference) cards is a bit anemic. Within spec, but only just.
Click to expand...

We speculated early on that it wasn't sufficient - but go watch the gamersnexus video where they look at the temps. The design is keeping everything well within spec.

I don't think Nvidia would have shipped these cards with the ability to push 124% power target if they thought their cooler wasn't good enough.
 
Good on them to stand by the product. If it were Apple it would have been full on excuses on why you held it wrong to install.. oh wait, Apple doesn't even let you do that! ;)
 
Rvenger said:
Good on them to stand by the product. If it were Apple it would have been full on excuses on why you held it wrong to install.. oh wait, Apple doesn't even let you do that! ;)
Click to expand...
Some actuary probably said, this problem is going to burn someone's apartment building down and the lawsuit is going to be horrendous. Credit where it is due.
 
Krenum said:
Actually, the fin stack on the 2080 FE (Reference) cards is a bit anemic. Within spec, but only just.
Click to expand...

Considering people are cooling the VRM passively, I find that hard to believe.

Dont mistake high temps for insufficient cooling, the fans are set to ramp up as needed and are more than capable flat out.
 
aokman said:
Considering people are cooling the VRM passively, I find that hard to believe.

Dont mistake high temps for insufficient cooling, the fans are set to ramp up as needed and are more than capable flat out.
Click to expand...

Who's cooling a reference RTX 2080 series card passively under load?
 
Krenum said:
Who's cooling a reference RTX 2080 series card passively under load?
Click to expand...

I'm cooling a Windforce OC with a Morpheus II and dual 120mm fans on the front and two thin smaller fans blowing on the back. No other heatsinks. So far, I can touch the VRMs or the VRAM chips on the card during a full load bench and it won't burn me. A lot of air is getting blown on the card on both sides...
 
The old saying is: "You can't test in quality". I don't reject NVidia's explanation, but find it strange that they need such strict component testing to weed out bad ones in the lot they receive. Don't their parts vendors take care of all of that? They're supposedly equipped to do a very good job of it, and would want to. Could NVidia be buying cheapo parts? They'd have no motivation to do so on high-end boards, certainly.

I'm surprised at this finding. I thought it would probably be non-conservative board layout or some overloaded power component.
 
And for the fellow whose 2080ti board caught fire and is worried about an RMA: I'd think EVGA would be sending a courier to his house ASAP to get it for analysis, and dropping off a replacement while there. And a refund.
 
Krenum said:
Who's cooling a reference RTX 2080 series card passively under load?
Click to expand...

Everyone was watercooling just the gpu before waterblocks arrived. Everything else was passively cooled and this was overclocked TI cards with shunt mods.
 
Advil said:
I'm cooling a Windforce OC with a Morpheus II and dual 120mm fans on the front and two thin smaller fans blowing on the back. No other heatsinks. So far, I can touch the VRMs or the VRAM chips on the card during a full load bench and it won't burn me. A lot of air is getting blown on the card on both sides...
Click to expand...

I know you know what passive cooling is so I'm not going to insult your intelligence. Your card is being air cooled. I was responding to a card being passively cooled.
 
You must log in or register to reply here.
Back
Top