Zotac Nvidia 1080 Ti AMP Extreme Edition concerns

[StormLightningSL]

Hi I was using an AMD Vega 64 up to a few days ago. Just sold both (I had two, bought for Crossfire, I know I was an idiot) for $750. Swapped an old 1080 Ti I had in my son's PC into mine, so I am now using the Zotac Nvidia 180 Ti Amp Extreme Edition. The money goes towards saving for a new 3080 or 3070 Ti, whichever I can buy first.

Now, this 1080 Ti card may have some issues. My son has complained that after long gaming sessions of a couple of hours, the GPU crashes the system. I believe it is due to overheating at the high clock speeds of the card. I tried down-clocking the memory by 200MHz, and my gaming sessions seem to go on without any hitch. Of course, I don't play long sessions of FPS shooters like him. But, RDR2 / CP2077 seem to be playable for an hour or more for me, though I don't have the settings too high.

Though the GPU is only 4 years old, and Zotac has a 5 year extended warranty, this card is out of warranty because I forgot to send in the registration for extended warranty within the stipulated 28 days of purchase. I had asked my vendor to check the thermal paste on the GPU chip a while back, and he had said it seemed perfect. Is it necessary / suggested that I re-do the thermal paste, and replace the thermal pads on the card? Any ideas or suggestions? Should I be concerned about the life of the card?

 

[Space Lynx]

Hi I was using an AMD Vega 64 up to a few days ago. Just sold both (I had two, bought for Crossfire, I know I was an idiot) for $750. Swapped an old 1080 Ti I had in my son's PC into mine, so I am now using the Zotac Nvidia 180 Ti Amp Extreme Edition. The money goes towards saving for a new 3080 or 3070 Ti, whichever I can buy first.

Now, this 1080 Ti card may have some issues. My son has complained that after long gaming sessions of a couple of hours, the GPU crashes the system. I believe it is due to overheating at the high clock speeds of the card. I tried down-clocking the memory by 200MHz, and my gaming sessions seem to go on without any hitch. Of course, I don't play long sessions of FPS shooters like him. But, RDR2 / CP2077 seem to be playable for an hour or more for me, though I don't have the settings too high.

Though the GPU is only 4 years old, and Zotac has a 5 year extended warranty, this card is out of warranty because I forgot to send in the registration for extended warranty within the stipulated 28 days of purchase. I had asked my vendor to check the thermal paste on the GPU chip a while back, and he had said it seemed perfect. Is it necessary / suggested that I re-do the thermal paste, and replace the thermal pads on the card? Any ideas or suggestions? Should I be concerned about the life of the card?

I would replace the thermal paste dust it a little bit too. get a high quality paste. arctic mx-5 is about to come out, or the noctua paste forget name of it.

also set a custom fan curve in MSI afterburner, stock fan curves are usually garbage. the default custom fan curve in msi afterburner should be fine. i usually make mine slightly more aggressive.

also why was it is stupid to sell your other two cards for 750 bucks? thats a great price for such old cards, good job on cashing out on the crypto craze. now take that 750, and buy your son a PS5 for 500. its the most economical way to game moving forward with the crypto craze not going away for at least several years this time... sadly...

 

[StormLightningSL]

"also why was it is stupid to sell your other two cards for 750 bucks? thats a great price for such old cards, good job on cashing out on the crypto craze. now take that 750, and buy your son a PS5 for 500. its the most economical way to game moving forward with the crypto craze not going away for at least several years this time... sadly..."

Hehe! I was an idiot to buy two of those cards for Crossfire instead of a 2080 Ti, which was just a little bit more at the time, but I've always been something of an AMD fanboy and user of Crossfire since the R9 290 days. Still have those Sapphire R9 290s, in fact, installed in home office PCs.

Anyway, those days are over for everyone now. Single cards FTW I guess.

I've been thinking on those lines, but PS games are very expensive where I am. Steam is much cheaper. Ofc, I would be spending a lot more on the PC itself, but then it's useful for a long time and for so many other things. We did have a PS3, but skipped the PS4. Not ruling out "going back" to a PS5.

 

[qubit]

Crashing due to overheating is very bad for any card. Definitely redo the thermal paste and then use a hardware monitoring app like AIDA64 or similar, to monitor the temps while gaming.

Note that contrary to popular belief, it's ok to stress test a graphics card (AMD or NVIDIA) with Furmark, which has its own temperature monitoring as long as it's done carefully. Just makes sure to only do it with a card that's got a beefy cooler like this one and then not more than 3-5 minutes and not too often. My cards simply throttle down to a safe speed and power level, with the temperature topping out at something reasonable like 70-75C. Remove the power limiter and you'd be in trouble though.

Using Furmark on a lower end card with a weak cooler and that doesn't throttle can easily result in burnout as it overheats in a few minutes.

 

[droopyRO]

You should use MSI Afterburner + RTSS, it allows you to see the temp, power draw, voltage, frequency, GPU load etc. while you game with an overlay.

 

[StormLightningSL]

@qubit - I tried MSI Afterburner's Kombustor stress test for about 3 minutes and reached around 65-66* C (145-155* F). Didn't experiment too much though. GPU seems to be doing fine, afaict. I will change the thermal paste and thermal pads as soon as I can though, just to be safe.

@droopyRO - Thanks for the tip. I was not able to get the OSD for MSI Afterburner until I saw that clip. Now I'll monitor the stats for a bit and see if they are getting out of hand.

Edit: I did get a crash after about 30 mins of CP2077 at all high settings at 3200X1800 resolution (my quirk). But, I was watching the temps on MSI Afterburner + RTSS, and the GPU was at about 60* C (140* F), which should not be a cause for concern. I had also set the memory at the full OC speed of 5600 MHz, which may have contributed to the crash as I now believe it is the memory that is giving issues of over-heating. I probably need to replace the thermal pads or contacts used for the stock GPU cooler.

Am I on the right track? Or, should I do some more research? Is there any app that allows me to check the memory IC temperatures specifically? Does this card even support that?

 

[qubit]

What I also did was to set the memory at the full OC speed of 5600 MHz. I now believe it is the memory that is giving issues of over-heating.

An overclock will certainly cause it and this is the root cause. Sounds like the card is actually working normally. Doesn't hurt to ensure that the thermal pads are doing their job properly though, even at stock.

 

[StormLightningSL]

I think I was not able to explain properly.

The card has a stock OC speed of 5600 MHz as per its specifications. I was running it with the memory a little slower at 5400 MHz just to be sure not to run into errors or crashes. I got that idea after running into display errors while playing Kingdom Come Deliverance, and searching the web for solutions to the errors. It was suggested that lowering the memory speeds would help, and that seemed to get rid of the issues as such.

However, when I set the card to run at full stock OC speeds and was stressing the card with high settings on CP2077 at a higher resolution than I normally play at, I had a crash. I will be trying to repeat the error, and to solve it by lowering the stock OC on the card memory. In any case, I am already looking for decent thermal pads and thermal paste online.

I will do whatever is necessary to salvage the card because I also feel that the card is not broken, but just over-heating due to cooling problems. I live in a place where the ambient winter temperature is 25* C (75-80*F) during the daytime, and summer temperatures go up to 33-35* C (90-100* F) when inside. Plus my cabinet is a mid-tower and though I've added 2X 200mm and 5X 120mm fans, I still think the PC is not cooled too well. In any case, replacing the thermal pads should help, though I don't know if the thermal paste needs replacing because my vendor did say it was perfect.

 

[FreedomEclipse]

Its well documented that some manufacturers release their cards with unstable clocks and it doesnt matter if its bone stock or factory OC, clocks are still unstable. Some manufacturers like Zotac do like to deny that its an issue or if issues exist. I remember seeing a few threads on here regarding issues with Zotac cards and Zotac not being very helpful in finding a solution.

Also heres one thread -- i did some searching around. same card. unstable factory OC clocks

 

[StormLightningSL]

Wow, almost exactly the same problem, though his was from the start, and mine has happened recently, plus my card is out of warranty because of my carelessness Oh, well!

I'll try the thermal pads thing, and reduce the overclock on the memory and hope for the best. Is there any way to change the BIOS settings to lower the memory clock like that guy was asking for? Since mine is out of warranty, I have nothing to lose by trying everything I can to fix it and make it work without hassles. Even if it works at stock Nvidia 1080 Ti speeds, I would be more than happy, hehe Especially with current prices for new cards being where they are. Ouch!

 

[FreedomEclipse]

The only thing i can suggest is flashing a bios from a card with lower clocks and seeing if that works out. It is fairly high risk in the sense that it goes wrong, it could brick your card but that is reversible if you have a spare graphics card or iGPu

There should be 3 bioses on the TPU database that has lower clocks.

While this is a suggestion -- you do this at your own risk. It might be worth googling and reading up on how to flash a bricked card just as a precaution if you go ahead with it.

Technically it shouldnt brick your card if the flash is successful. It just wont work properly...


And most importantly: Dont forget to backup your original bios

 

[StormLightningSL]

The only thing i can suggest is flashing a bios from a card with lower clocks and seeing if that works out. It is fairly high risk in the sense that it goes wrong, it could brick your card but that is reversible if you have a spare graphics card or iGPu

There should be 3 bioses on the TPU database that has lower clocks.

While this is a suggestion -- you do this at your own risk. It might be worth googling and reading up on how to flash a bricked card just as a precaution if you go ahead with it.

Technically it shouldnt brick your card if the flash is successful. It just wont work properly...


And most importantly: Dont forget to backup your original bios

Thank you for the effort taken Appreciate it!

I currently have the second BIOS in the list, with the highest clock speeds of 1646/1400/1759. I think the GPU chip speed is not the problem as such, and the memory of my particular card is either OC too high at stock, or the thermal pads have lost some conductivity. I am still trying to get my hands on reliable thermal pads from my local vendor who usually supplies my stuff. If that fails, I will use Amazon for my purchase.

The alternate BIOSs sound good to me, but I am using a -50 MHz memory downclock at the moment, and the available ones are -24 MHz. Will that be enough? If I still have to downclock using MSI Afterburner, then I would rather not re-flash the BIOS and stick to the app being loaded with Windows on start-up.

I'll wait for a few days until the thermal paste/thermal pads issue is sorted out. If nothing else works, I think the Zotac AMP edition BIOS would make most sense, just to be on the safer side.

 

[FreedomEclipse]

The alternate BIOSs sound good to me, but I am using a -50 MHz memory downclock at the moment, and the available ones are -24 MHz. Will that be enough? If I still have to downclock using MSI Afterburner, then I would rather not re-flash the BIOS and stick to the app being loaded with Windows on start-up.

I am unable to answer this question. Maybe the other bioses have tweaks to the power handling or changes to memory timing that would stop the card flipping out.

Nothing is known till you try it and find out for yourself.