Help with Faulty MSI RTX 2070 Gaming Z!

[Headcrash77]

Greetings!

 

This is my first post here, so apologies if this ends up in the wrong section.

 

On the 23/09/2018 I bought an MSI RTSX 2070 Gaming Z on ebay from a company called CCL computers. I had previously been using an MSI GTX 1060 Twin Frozr. This was in a new PC that I built in late August with an 8700k (specs and picture at bottom of post).

 

The GTX 1060 never went above 69°C even under full load. To my dismay, even while having a much larger cooler on it, the RTX 2070 was reaching 80°C in my case. I looked around online and read that it might be the thermal paste causing the issue.

 

So... like an idiot, I decided to change the thermal paste on the card to some Noctua NT-H1 to see if it would improve temps. It did not. I decided to give it one more try with some MX-4 just to put the idea to rest that it was the thermal paste. No change either (which I kind of expected). I was going to simply let it run hot, but this made my PC bluescreen several times, which the 1060 never did.

 

Of course, in changing the thermal paste, I damaged the "MSI Warranty Void" sticker when removing the screws for the cooler. And, when cleaning the thermal paste off of the cooler heatsync smudged some of the numbering on a sticker attached to the cooler. (not enough to make the numbers unreadable, but enough to know that something has been done) So when I asked to send it back to the company who sold it to me they accepted my request, tested the card, said it was faultless and that it never rose above 71°C, but would not accept it for a refund because of the damaged sticker. This was a bit irritating since I DID mention the sticker to them when requesting a solution.

 

At this point, I tried to salvage the situation by looking into Undervolting with MSI Afterburner, since changing to a custom fan curve alone wasn't helping much. By default, the card was clocked around 1965Mhz while pulling around 1013mV. So I tested lower voltages to see if they were stable in a few games (Vermintide 2, Overwatch, Heaven Benchmark) and had success at 875mV for a while, but ultimately had to increase the power to 900mV to stop the display driver from crashing. While it lasted, this worked a treat, lowering the temps from 80°C when gaming to 67°-71° and no higher.

 

However, after playing overwatch for several hours without a problem, the display driver crashed on me. So I tried it again, only for it to crash minuted after getting into game. I tried adjusting the curve to increase the voltage a bit, but to no avail. And then tried running the card at stock to see if it would still crash. And it did.

 

Now, no matter what I do, any 3d application crashes the display driver within 3 seconds. I've tried shutting down the pc, re-seating the card, testing the low voltage and the stock settings again. I'm at the end of my rope. I spent £585 on this card, and now it seems like the most expensive PC component i've ever bought, is now the biggest mistake i've ever made with PC hardware.

I've always had mid-range PC's and the one time I save up and decide to go for more premium components, I get hit with problems, and then start making mistakes. Honestly I just wanted to put the card in my PC and for it to work, not exceed temps in reviews by 11°C and crash the system. I get that reviews are usually done on open air test benches, but for the temp difference to be more than 10°C? I wasn't happy about that, especially since this card is advertised as being "Cool and Quiet".

 

So, after that long idiotic story. I have 3 questions:

#1: Is there ANY way to fix this problem I'm having with the card? Since its Crashing (and sometimes artifacting) no matter what I do.

#2: Is it worth trying to RMA the card with MSI?

#3: Is it just worth selling the card on ebay for a loss as a "Faulty, buy at your own risk" item?

 

Any help would be greatly appreciated. I can't really afford to waste £600. And I assume because of my stupidity. I have.

 

-Headcrash

 

PC SPECS:

CPU: Intel i7-8700K @ 4.6Ghz

RAM: 16GB DDR 4 G.Skill Trident Z 3000Mhz (no XMP profiles enabled so its running lower than 3000Mhz)

MB: Gigabyte Z370P D3-CF

GFX: MSI RTX 2070 Gaming Z

PSU:EVGA 650W GQ 80+ GOLD

OS: Windows 10 Pro 64-bit

[emosun]

tldr but just as a comment we don't need a picture of the machine to diagnose an issue , it just makes a long post even longer

[Jurrunio]

What a bad time, because Nvidia came under fire for dying Turing GPUs

https://www.gamersnexus.net/industry/3387-hw-news-dying-2080ti-investigation-1080ti-stock-almost-gone

 

7 minutes ago, Headcrash77 said:

#1: Is there ANY way to fix this problem I'm having with the card? Since its Crashing (and sometimes artifacting) no matter what I do.

GN suspects overheating memory to be the culprit. Try improve their cooling.

 

7 minutes ago, Headcrash77 said:

#2: Is it worth trying to RMA the card with MSI? Or is the smudged sticker under the heatsync and the warrenty screw from another card a dealbreaker there? (I assume it is)

it's not a bad idea to send it back to MSI, but be prepped to pay for the repairs. Or alternatively (no advertisement intended), give GN a look, as they are collecting cards for investigation atm.

 

9 minutes ago, Headcrash77 said:

#3: Is it just worth selling the card on ebay for a loss as a "Faulty, buy at your own risk" item.

I won't. After the whole 20 series GPU dying thing, 20 series cards sell at a lower price then they did.

[Headcrash77]

2 hours ago, emosun said:

tldr but just as a comment we don't need a picture of the machine to diagnose an issue , it just makes a long post even longer

Apologies, I only put it there to address potential questions about airflow.

2 hours ago, Jurrunio said:

it's not a bad idea to send it back to MSI, but be prepped to pay for the repairs. Or alternatively (no advertisement intended), give GN a look, as they are collecting cards for investigation atm.

 

Thanks for your quick response!

 

I was considering trying to RMA the card with MSI, but I'm worried that because i've removed the cooler to change thermal paste, and in doing so smudged a sticker under the cooler and then replacing screw with the warrenty void sticker with another, that this would cause problems. Since after reading MSI's conditions, they seemed really strict, to the point where if you've done anything but USE the card then it wouldn't be accepted.

 

Strangely enough though, after you mentioned it possibly being memory overheating, i tried using MSI afterburner to underclock the memory as much as it would go (-502mhz) and it was able to get through a Heaven benchmark without problems. Unfortunately it still crashed after 20 minutes in Overwatch.

[Jurrunio]

5 minutes ago, Headcrash77 said:

I was considering trying to RMA the card with MSI, but I'm worried that because i've removed the cooler to change thermal paste, and in doing so smudged a sticker under the cooler and then replacing screw with the warrenty void sticker with another, that this would cause problems. Since after reading MSI's conditions, they seemed really strict, to the point where if you've done anything but USE the card then it wouldn't be accepted.

I'm not sure if they will reject repairs if you clearly state you're willing to pay.

[Headcrash77]

Well after contacting MSI (uksupport@msi.com) I got this message back within minutes:

 

"You have contacted the Notebook & AIO repair center. There is no direct warranty support from MSI for end users who have purchased our components. The supplier is obligated to provide support for any MSI components that they sold. Thank you."

 

I'm completely at a loss of what to do now. Since the company I bought the card from has already rejected further dealing with this situation because of the warranty sticker.

 

I'm going to try posting a ticket on the msi site, hopefuly that will yeild better results.

[asand1]

Disable gsync if you're using it. There are currently driver issues causing bsods with Gsync.

[Guest]

Poster above is right, try to disable Gsync, and also maybe try to use a single monitor if you're plugged to more than one. But your problem probably has little to do with the card reaching high temps, GPUs can run at 80c without any issue. Instability is not necessarily tied to temperatures.

[Headcrash77]

I do not use Gsync, and have tested the card with one monitor, even with another system, and the display driver crashes almost instantly when entering any 3D application.