Palit 3080 ti gamerock issue

[BLUE-ANGEL]

Definitely go with Thermal Grizzly like @KURT_DERMS mentioned.

Just bear in mind, if your gpu is still under warranty, confirm with the warranty holder if you are allowed to replace the thermal paste and pads while still under warranty or if they will rather RMA the gpu for you.

Palit themselves are completely against this, BUT, as per their website, depending on the local distributors of the country for RMA, the distributor might be willing to allow this within the warranty period. Just to be cautious in case a re-paste and pads replacement still does not solve the issue.

 

[DiceAir]

Definitely go with Thermal Grizzly like @KURT_DERMS mentioned.

Just bear in mind, if your gpu is still under warranty, confirm with the warranty holder if you are allowed to replace the thermal paste and pads while still under warranty or if they will rather RMA the gpu for you.

Palit themselves are completely against this, BUT, as per their website, depending on the local distributors of the country for RMA, the distributor might be willing to allow this within the warranty period. Just to be cautious in case a re-paste and pads replacement still does not solve the issue.

I bought the card on 11th Jun 2021 so still warranty left. Like I said so far looks like I sorted my throttling issue but will have to see when ambient temps is a bit higher. I was thinking of just accepting it until the issue is getting worse and for example not making my card boost properly. Just makes it easier to claim warranty. I do play warzone a lot on lowest graphics with DLSS set to quality it doesn't run near 100C more like 80C and that is fine for me now.

Do you guys think GPU sag and the fact that my power cables is coming from the bottom side of the gpu kind of pulling towards the bottom a tiny bit can be cause higher temps over time? Tonight I will push the gpu upwards a bit to see if my temps will improve.

 

[DiceAir]

oops double post

 

[StangV2_0]

I know it's an old thread but starting to have another issue. my hotspot temps on the card is around 105C. I started noticing it when my gpu fans went to 3500rpm.

Yup, that's normal for 3080/3080ti. Under load they often pass 100degC. Its one of the 3080's major flaws, in fact its only flaw IMO.
I had this issue on a Gigabyte Eagle 3080 (which admittedly is one of the 3080's with poor cooling) and I replaced the thermal pads myself. I used the Grizzly pads as mentioned by @KURT_DERMS. Exceptional service!

Only issue is it didn't give the results I wanted. Yes they dropped a few degrees but these cards still run hot no matter what you try. The thermal pads aren't necessarily cheap to replace and then you may affect your warranty or worse if you dont know what you are doing. And you will have to list that you have opened the card when you want to sell, since its the right thing to do. Fortunately the person who bough the card from me soon after I did the mod trusted my work and he wasnt concerned about any warranty issues. He still has it I think.

If it was me and I wanted to keep a 3080/Ti no matter what I would try trade it for an Asus TUF or EVGA. They are known to run cooler than the other boards and will be less of a headache for you in the long run. Just my 2c.

 

[DiceAir]

Yup, that's normal for 3080/3080ti. Under load they often pass 100degC. Its one of the 3080's major flaws, in fact its only flaw IMO.
I had this issue on a Gigabyte Eagle 3080 (which admittedly is one of the 3080's with poor cooling) and I replaced the thermal pads myself. I used the Grizzly pads as mentioned by @KURT_DERMS. Exceptional service!

Only issue is it didn't give the results I wanted. Yes they dropped a few degrees but these cards still run hot no matter what you try. The thermal pads aren't necessarily cheap to replace and then you may affect your warranty or worse if you dont know what you are doing. And you will have to list that you have opened the card when you want to sell, since its the right thing to do. Fortunately the person who bough the card from me soon after I did the mod trusted my work and he wasnt concerned about any warranty issues. He still has it I think.

If it was me and I wanted to keep a 3080/Ti no matter what I would try trade it for an Asus TUF or EVGA. They are known to run cooler than the other boards and will be less of a headache for you in the long run. Just my 2c.

ok Some more info after some testings last night. Don't ask me how the fk it happened but seems like my heatsink on the cooler shifted.

What i was doing last night was playing cyberpunk using gpu-z on a second screen to see hot spot temps and it went to 105c with clock speed drops. Then I decided fk this I have to look into this. I opened the case and pushed gpu up to compensate for the sag. nothing same temps. I looked again and though hey I'm just moving the whole card up so contact between die and heatsink contact will not change much so When i pressed on the side of the heatsink I felt it shift a bit and temps went down to 97-100C. I decided to push the heatsink upwards by the heatpipes towards pcb and can get it closer to 97C. Weird that something like this can happen.

 

[BLUE-ANGEL]

ok Some more info after some testings last night. Don't ask me how the fk it happened but seems like my heatsink on the cooler shifted.

What i was doing last night was playing cyberpunk using gpu-z on a second screen to see hot spot temps and it went to 105c with clock speed drops. Then I decided fk this I have to look into this. I opened the case and pushed gpu up to compensate for the sag. nothing same temps. I looked again and though hey I'm just moving the whole card up so contact between die and heatsink contact will not change much so When i pressed on the side of the heatsink I felt it shift a bit and temps went down to 97-100C. I decided to push the heatsink upwards by the heatpipes towards pcb and can get it closer to 97C. Weird that something like this can happen.

Well THIS changes the look of things.
That sounds like a manufacturing defect to me. I can only suggest 2 things even if it's a "easy fix".

1.
Rather ask for a gpu replacement from the supplier explaining what you have found after experiencing the different temps and after the heat sink moved the temps dropped. Because the gpu is still under warranty.

2.
If you don't want to RMA and there are no "Tamper proof" stickers on any of the screws willing to risk the warranty, you could open the gpu yourself and ensure the heat sink is properly fixed tightly at every screw. BUT, I would then also recommend to do a repast while at it since there was movement of the heat sink and you don't know what the current state of the paste is.

Being able to move the entire heat sink means possibility for bad conductivity between the die and the heat sink thus you are experiencing the bad temps IMO.

This is just my opinion and what I would've done if I was you since the gpu is still under warranty.

 

[StangV2_0]

Send the GPU back as soon as possible to get RMA.

 

[DiceAir]

Well THIS changes the look of things.
That sounds like a manufacturing defect to me. I can only suggest 2 things even if it's a "easy fix".

1.
Rather ask for a gpu replacement from the supplier explaining what you have found after experiencing the different temps and after the heat sink moved the temps dropped. Because the gpu is still under warranty.

2.
If you don't want to RMA and there are no "Tamper proof" stickers on any of the screws willing to risk the warranty, you could open the gpu yourself and ensure the heat sink is properly fixed tightly at every screw. BUT, I would then also recommend to do a repast while at it since there was movement of the heat sink and you don't know what the current state of the paste is.

Being able to move the entire heat sink means possibility for bad conductivity between the die and the heat sink thus you are experiencing the bad temps IMO.

This is just my opinion and what I would've done if I was you since the gpu is still under warranty.

no way of replacing thermal paste/pads due to warranty stickers on the card . I will not tell them cooler shifted as they might say it's been tampered with. So what do you think will be the best way to convinced them the card is overheating? Don't want to get there and they say don't look at hot spot temps etc and then thye only look at core temp in thundermaster (palit version of msi afterburner) and then see it doesn't drop clock speed to below 1665mhz so therefore nothing wrong.

 

[BLUE-ANGEL]

no way of replacing thermal paste/pads due to warranty stickers on the card . I will not tell them cooler shifted as they might say it's been tampered with. So what do you think will be the best way to convinced them the card is overheating? Don't want to get there and they say don't look at hot spot temps etc and then thye only look at core temp in thundermaster (palit version of msi afterburner) and then see it doesn't drop clock speed to below 1665mhz so therefore nothing wrong.

Okay seems like you are hesitating on what to do, so this is what I would do since the gpu is still under warranty:

Picking up the phone and call the supplier,

Good day,
I purchased a 3080Ti from you guys in June 2021. I recently cleaned out my PC case and when I put the gpu back after cleaning everything I felt the heat sink moved. Obviously this is not supposed to happen. After this I suddenly started seeing higher temps under load compared to what I normally would get. This is obviously not supposed to happen and must be a manufacturing defect for not tightening the screws or to ensure that the heat sink is properly fitted to the gpu. What can you guys advise, replace my gpu or bring it in for you to look at?
This is quite urgent as I am using my PC on a daily basis and need a resolution asap.

Voila, simple as that, no fuss no hassle. It's either they tell you what to do or you decide stuff it and re-paste it yourself since you know the heat sink moved and you even saw a difference with lower temps after moving it slightly.

 

[goldfritter]

ok Some more info after some testings last night. Don't ask me how the fk it happened but seems like my heatsink on the cooler shifted.

What i was doing last night was playing cyberpunk using gpu-z on a second screen to see hot spot temps and it went to 105c with clock speed drops. Then I decided fk this I have to look into this. I opened the case and pushed gpu up to compensate for the sag. nothing same temps. I looked again and though hey I'm just moving the whole card up so contact between die and heatsink contact will not change much so When i pressed on the side of the heatsink I felt it shift a bit and temps went down to 97-100C. I decided to push the heatsink upwards by the heatpipes towards pcb and can get it closer to 97C. Weird that something like this can happen.

This could also have happened due to sag, I guess, as unfortunate as that is. The PCB could have bent away from the heatsink as it sagged down. Maybe an unintended consequence of having the PCB shorter than the heatsink and backplate.

But it should still be RMA-able, I think, so that is a good avenue to try. Send screenshots/screen captures of the high temps and downclocking. If they return it and say it's tampered with, then at least you can open it up and replace pads/paste and tighten at your leisure.

You can also do what a friend of mine did and add actual heatsinks on top of the backplate where the VRAM is. This shouldn't void the warranty, but assumes decent contact between VRAM and backplate. He is happy with temp results:

https://imgur.com/a/hrS6x4r

 

[DiceAir]

I asked palit now for the latest bios as well. This is what they told me

This could also have happened due to sag, I guess, as unfortunate as that is. The PCB could have bent away from the heatsink as it sagged down. Maybe an unintended consequence of having the PCB shorter than the heatsink and backplate.

But it should still be RMA-able, I think, so that is a good avenue to try. Send screenshots/screen captures of the high temps and downclocking. If they return it and say it's tampered with, then at least you can open it up and replace pads/paste and tighten at your leisure.

You can also do what a friend of mine did and add actual heatsinks on top of the backplate where the VRAM is. This shouldn't void the warranty, but assumes decent contact between VRAM and backplate. He is happy with temp results:

https://imgur.com/a/hrS6x4r

looking at this Palit GeForce RTX 3080 Ti GameRock OC review the backplate only helps little bit with vram cooling and not the actual core. Might still help but don't hink heatsink mod will help. I have some old heatsinks from old motherboard like this one just to test and see if i see any improvement even if it's just to see if I can slow down the rise of temps

 

[StangV2_0]

This could also have happened due to sag, I guess, as unfortunate as that is. The PCB could have bent away from the heatsink as it sagged down. Maybe an unintended consequence of having the PCB shorter than the heatsink and backplate.

But it should still be RMA-able, I think, so that is a good avenue to try. Send screenshots/screen captures of the high temps and downclocking. If they return it and say it's tampered with, then at least you can open it up and replace pads/paste and tighten at your leisure.

You can also do what a friend of mine did and add actual heatsinks on top of the backplate where the VRAM is. This shouldn't void the warranty, but assumes decent contact between VRAM and backplate. He is happy with temp results:

https://imgur.com/a/hrS6x4r

I am not even surprised it is a Gigabyte, exact same 3080 I had I think. This idea also takes into account that there is already a great contact between the Memory modules and the backplate. Because if there isn't then this is all null and void. And I would not trust the factory pads on the Gigabyte to be doing the job. Even after changing the thermal pads, and then trying this I still wasn't happy with the temps so decided to let it go.

I stand by my original post. However now that the card appears to have the loose cooler issue I would RMA it ASAP.

Edit: If you do try the heatsink resting on the back plate, please put thermal paste or thermal pads between the heatsinks and the backplate. And not prestick

 

[JeffTheApricot]

This could also have happened due to sag, I guess, as unfortunate as that is. The PCB could have bent away from the heatsink as it sagged down. Maybe an unintended consequence of having the PCB shorter than the heatsink and backplate.

But it should still be RMA-able, I think, so that is a good avenue to try. Send screenshots/screen captures of the high temps and downclocking. If they return it and say it's tampered with, then at least you can open it up and replace pads/paste and tighten at your leisure.

You can also do what a friend of mine did and add actual heatsinks on top of the backplate where the VRAM is. This shouldn't void the warranty, but assumes decent contact between VRAM and backplate. He is happy with temp results:

https://imgur.com/a/hrS6x4r

Available ?

 

[goldfritter]

I am not even surprised it is a Gigabyte, exact same 3080 I had I think. This idea also takes into account that there is already a great contact between the Memory modules and the backplate. Because if there isn't then this is all null and void. And I would not trust the factory pads on the Gigabyte to be doing the job. Even after changing the thermal pads, and then trying this I still wasn't happy with the temps so decided to let it go.

I stand by my original post. However now that the card appears to have the loose cooler issue I would RMA it ASAP.

Edit: If you do try the heatsink resting on the back plate, please put thermal paste or thermal pads between the heatsinks and the backplate. And not prestick

Ja I confirmed with him and apparently in addition to those new sinks, he also not only added more pads between the back of the VRAM and the backplate, but also replaced the thermal pads between VRAM and heatsink. So not as straightforward as I thought, and probably a warranty-voiding procedure.

Available ?

wut

 

[DiceAir]

for me I don't think the backplate mod will make a huge difference.

 

[Aokiji]

I know it's an old thread but starting to have another issue. my hotspot temps on the card is around 105C. I started noticing it when my gpu fans went to 3500rpm.

chances your thermal paste on gpu die is dry cuasing the spike in temps , have a look @ that first

 

[DiceAir]

I know it's not the best solution but to get better contact now is by using zip ties between the heatsink and backplate to get better contact and now when playing cyberpunk on stock performance bios with stock limits I see much better clock speed and temps seeing 95C. Will have to test for a few days so far it's a bit better without voiding warranty I think.

I'm sure 95C is within spec for hot spot.

chances your thermal paste on gpu die is dry cuasing the spike in temps , have a look @ that first

I know that is what I must do but it voids warranty my sollution might just work for now untill things get worse then I can take the card in for replacent.

 

[DiceAir]

Ok just an update. my problem seems to be fully solved. Seems like there wasn't enough contact and with the card heating up a bit made thee thermalpads or thermal paste settle in or something as my temps have dropped even more to about 85-90C when playing cyberpunk. And doing the gpu-z render test where it used to go 98-99C easily it even manages to go only to 90C max

This is all with power limit pushed to 114% and voltage set to +100 on stock bios no undervolt.

 

[Bullistic]

Hey all,

Quick question... I have the 3080TI Gamerock Palit GPU, I was wondering if anyone would know where I should be able to find an identical set of screws for it as mine has started to break in half (Like split)

Any sort of links or an exact place with prices would really help out as I need a new set of screws to screw down the GPU to the cooler.

Thanks all

 

[DiceAir]

Sorry for posting in an old thread.

Just an update. So I decided to open up the card seeing my hotspot easily go up to 105C thermaal throttle. So I repasted the gpu core only not replacing the thermal pads.

Temps max when playing alan wake, cyberpunk and pretty much any game at full graphics pulling 440W the hotspot now only goes to 73-74C and normal core temp is reading 66C and that is really good. I didn't expect it to make such a huge difference and my performance even went up a bit.