Nvidia RTX 2080 Ti graphics cards are dying on a lot of users

Why are we even talking about lighting though? I Think all the lights in my house combined are less than 150W ( 6-10W LEDs)
 
Why are we even talking about lighting though? I Think all the lights in my house combined are less than 150W ( 6-10W LEDs)

Lots of people still run regular bulbs at 60 watts, but yeah the conversation is silly to some extent. But if you do come close to overloading a circuit the voltage can drop and it will be much harder on the power supply which can cause them to fail. Reality is Nvidia has a problem with their 2000 series that is causing the cards to fail, figure how few there are and how many reports of failures and yeah not everyone had a bad power supply as most people that bought such cards are running sufficient power supplies.
 
Everything ran fine until today, now I’m having trouble booting into windows and the Device Manager under display adapter shows a “Windows has stopped the device because it has reported problems. (Code 43).”

Anyone else with this issue? Using a Zotac AMP 2080Ti.

Here's what GPU-Z looks like, notice the memory size and clocks..
View attachment 115959
Code 43 is usually associated with a dead piece of hardware.
I imagine some of the issues maybe related to PSU, especially Corsair PSU and how it delivers the power to the GPU. Also, shouldn't dual 8-pins cable able to handle 300 watts?
The FE has one 6-pin and one 8-pin. Combined with the 6.25A from the slot it can handle 300W. But the PCI-E cables can be overprovisioned if need be. I ran a Maxwell Titan X for years with a custom 450W BIOS without issue, and that was also run with one 6-pin and one 8-pin. So long as the +12V rail from the PSU isn't being overloaded it's fine. My power supply can go up to 70A on the +12V rail, and I was using around 55A.
makes me wonder if the crud Nv gave AMD in regards to RX 4xx and in flip sight some 5xxx were "breaking" specs has not turned the tables on them (once again, not first time they failed SIG spec themselves)

75w for 6 pin 150 for 8 pin plus the 75 from the "slot" itself, but yep, very much depends on quality of the PSU being used (how it handles the power delivery etc) as well as if the GPU is delivering the power properly as well, fancy voltage regulator chips can only handle so much, hard to regulate something that is too irregular from the get go ^.^

(even though your using Ngreedia) I personally feel sorry for anyone having issues with their system (especially with mucho expensive GPU like 20 series) the downtime and extra money as well as frustration sucks ass :(
The issue with the RX 480 was that it was overprovisioning the PCI-E slot, pulling 8-9A. Testing of the 2080 Ti shows it pulling no more than 5.5A from the slot. The slot is specified to handle up to 6.25A, or 75W. The cables have overprovisioning room built in. A proper 8-pin mated to a strong PSU has no problem handling 15-18A. AMD's own R9 295X2 used an average of 450W while gaming on two 8-pin cables, or 150W + 150W + 75W = 375W max according to PCI-E specification. The difference is AMD had well-defined power requirements for the card while I don't see anything similar from NVIDIA, but in stock condition the FE is still well within spec.
 
Are watercooled cards failing as well?


I don't think this is a cooling issue--it seems to be vendor specific to some degree but it's hard to nail down. I've seen more reports of the Nvidia FE cards failing than others. I'm watercooling 2 EVGA Ti XC Ultras and haven't had issues.
 
ZOTAC refuses to pay shipping for RMA. Just a heads up to anyone considering purchasing from them.

Serious question.. does anyone pay for RMA shipping? I thought standard practice if you didn't pre-pay for something extended was you pay to get the card to them, and they cover shipping to you.

I mean last two times I've tried to RMA, I've been given an obstacle course (Sandisk SSD), or talked out of it (MSI) by telling me how long and painful it'd be.
 
EVGA RTX 2080ti OC Edition headed back this morning. Running super hot, 83-90c on 100% FANS. Lot's of stuttering almost like a throttling on/off on/off kind of feeling. My case is very well ventilated. My prior 1080FE ran for years at max of 76c. I'm starting to see weird pixelated shadows in gaming as well...Assassins Creed Odyssey, We happy few, and more. I'm not running any OC other than the small stock OC from EVGA. I have a new card arriving today direct from NVIDIA this time but after visiting their forums I'm pretty damn concerned that I'll be part of the RMA game soon. I'll update this afternoon after some heavy gaming.


[/QUOTE

The new card direct from Nvidia arrived today. So far here's what I notice compared to the EVGA 2080ti OC

1. Runs 5c cooler
2. Fans are quieter
3. 7hrs of heavy gaming and so far so good
4. The stutters are gone that the EVGA card had
 
Serious question.. does anyone pay for RMA shipping?
Nvidia did an advanced replacement for me, sent me the new card in about a week and also sent a pre-paid label for me to print out.

I still have both cards, I'm doing more stability testing to be sure the new card is working but things look good so far.
 
Serious question.. does anyone pay for RMA shipping? I thought standard practice if you didn't pre-pay for something extended was you pay to get the card to them, and they cover shipping to you.

I mean last two times I've tried to RMA, I've been given an obstacle course (Sandisk SSD), or talked out of it (MSI) by telling me how long and painful it'd be.
Most places I deal with over RMAs for customer stuff, it's you pay to ship item to them, they pay to send the RMA back. Unless there is some advanced RMA going on. I don't RMA stuff often (maybe once every 6 months if that. Last item I sent back was a EVGA power supply, only sent literally the PS itself back, got back a fully shrink wrapped entire boxed PS. Had to pay shipping to get it to them. Sent it USPS cheapest option, heavy power supplies can get a bit $$$ to send back.
 
Most of these cards are new enough that you should be within the return window for most places.

If a manufacturer wanted me to pay shipping for RMA i'd laugh and just return the card to Amazon/Newegg/whoever. Fuck that noise.
 
Most of these cards are new enough that you should be within the return window for most places.

If a manufacturer wanted me to pay shipping for RMA i'd laugh and just return the card to Amazon/Newegg/whoever. Fuck that noise.

I’m sure most people want a card back sometime soon, rather than go back to fishing for a deal. My time is worth more than that.
 
You can't send it back to the retailer after 30 days.

Also, all the RTX cards sold on Newegg were "return for replacement" only. You can't get your money back, just another card.

Also power supplies. I go through a LOT of Corsair power supplies. I rarely have a bad one but I'm sending one back right now. It's just under 2 months old. I have to pay shipping to them. That's how it is.
 
My only experience with a Corsair power supply was from around 8 years ago at this point, and it gave me headaches for years across two different builds. The first build was a nForce chipset and I thought it was freezing because it was.. nForce. Then when I got an intel chipset and was getting the same freezing issues I realized it was the shitty corsair power supply. I only buy Seasonic now.
 
It's just one and a half memory modules effected. Releasing a driver/firmware to utilize only 9.5GB of memory soon. :p
 
Lots of people still run regular bulbs at 60 watts, but yeah the conversation is silly to some extent. But if you do come close to overloading a circuit the voltage can drop and it will be much harder on the power supply which can cause them to fail. Reality is Nvidia has a problem with their 2000 series that is causing the cards to fail, figure how few there are and how many reports of failures and yeah not everyone had a bad power supply as most people that bought such cards are running sufficient power supplies.

I run my 2950x on Corsair HX1000i on 240v. No chance in hell of voltage dipping unless its utilities fault
 
Checked my PSU, 99.9A on the 12V rail available. Not near maxing it. Card working fine, I think other than some specific driver bugs related to RTX, I've been good.

For RMA, I've been a fan of EVGA for years with their system.
 
2080 ti ref 11.PNG
2080 ti msi 11.PNG


Saw this on another forum. The left side is the stock 2080ti FE, the right side is an MSI 2080ti. The voltage regs are snuggled up against the memory chips on the stock FE. MSI moved them away. The micron chips will self destruct past 90c apparently. This *could* be what we're witnessing. If so, even if my new 2080ti FE is working ok, my thought is that long term survival is pretty low....
 
View attachment 116657 View attachment 116658

Saw this on another forum. The left side is the stock 2080ti FE, the right side is an MSI 2080ti. The voltage regs are snuggled up against the memory chips on the stock FE. MSI moved them away. The micron chips will self destruct past 90c apparently. This *could* be what we're witnessing. If so, even if my new 2080ti FE is working ok, my thought is that long term survival is pretty low....
The power phases should not be getting that hot unless you're doing some serious overclocking. I did see a Reddit post that mentioned slightly underclocking the memory fixed their issues, though. Maybe GDDR6 wasn't ready for prime time, yet.
 
View attachment 116657 View attachment 116658

Saw this on another forum. The left side is the stock 2080ti FE, the right side is an MSI 2080ti. The voltage regs are snuggled up against the memory chips on the stock FE. MSI moved them away. The micron chips will self destruct past 90c apparently. This *could* be what we're witnessing. If so, even if my new 2080ti FE is working ok, my thought is that long term survival is pretty low....

interesting. for those of us with Zotac AMP 2080 Ti's, our cards use the reference board design (same as the FE): https://www.techpowerup.com/reviews/Zotac/GeForce_RTX_2080_Ti_AMP/5.html
i'm sure a ton of other AIB cards fall in the same boat.

relevant to this (link was posted by someone else in another thread): https://www.hardwareluxx.de/index.p...leidet-offenbar-an-erhoehter-ausfallrate.html

when overclocked the FE can see ~92 C near the memory, which is rated for 95 C. I don't know if this is making a mountain out of a molehill but it's interesting nonetheless. for the time being i'm removing the fairly low overclock i've got on my card. it will be interesting to see how all this shakes out.
 
Last edited:
The power phases should not be getting that hot unless you're doing some serious overclocking. I did see a Reddit post that mentioned slightly underclocking the memory fixed their issues, though. Maybe GDDR6 wasn't ready for prime time, yet.

Or Micron is just a shitty RAM producer?
 
ZOTAC refuses to pay shipping for RMA. Just a heads up to anyone considering purchasing from them.

Zotac is a shitty brand with horrible quality issue in the past.

Not sure why people want to purchase from them. Even in China (where they came from) they are consider as the budget brand like Colorful with high failure rate.
 
Could these problems result in a class action suit?
 
Ugh, I really want to upgrade, but not to this. Guess I'm waiting until this is resolved.
 
Could these problems result in a class action suit?
What would be the point. They are being replaced with in warranty. With Nvidia they are being quick and saying shipping to replace their cards. Seems Nvidia fucked up and is going to cost them a lot to keep replacing cards already.
 
So any indication that there would be an issue here with the 2080 FE cards? or is it just the ti variant?
 
Evga customers are complaining about high mem temps on there 2080ti ICX cards as they can see the temps on there cards.
 
Back
Top