3080 FE Thermal Pad change: Mem Temps still 100c!! Help!

[Rhaos]

So, got up eartly this morning, brewed a coffee and took the card apart...again.

Reapplied thermal paste, closed her back up and ran Heaven Benchmark.



A lot better temps now then before

I think this is as good as I can get it. Going to now work on improving airflow in the case.

Thanks to everyone who has helped or replied to posts in this thread. I hope others out there who are attemptign to do the thermal pad change, learn something from my mistakes here.

 

[andybird123]

I have 3080fe 2mm gelid Xtreme die side and 3mm on the back ( I think you could even use 2mm extreme gelid for the back ) and I don't have such problems how do you explain that ? Different heatsinks ?

My temps at highest at stock been running for 2 weeks 75c GPU with hotspot Delta 11-12 . My temps are lower with undervolt of 1905@ 893mv which mostly boosts to 1920 and max just barely hits 70c

Vram has had massive difference before would go high into the 90s now stays in the 80s , mining before instantly hit 110c now 82-86c

I used thermal grizzly I will keep an eye out if temps worsen if they do I will try TFX but for pads I will use the same as I've had no issues on what you reported

I used 1.5mm from the get go and mine also worked fine first time. I know the natural human response is to get defensive, but there are any number of reasons why you managed to get 2mm to work, maybe you cut them small and squished them down, we don't know as we aren't you, but I think looking at the body of evidence, its far more likely that people are having issues with 2mm than with 1.5mm, so the prudent course of action is to advise other people to use 1.5mm in the first instance and if they tried 2mm and have massive issues, to suggest they try 1.5mm.

 

[uscool]

I used 1.5mm from the get go and mine also worked fine first time. I know the natural human response is to get defensive, but there are any number of reasons why you managed to get 2mm to work, maybe you cut them small and squished them down, we don't know as we aren't you, but I think looking at the body of evidence, its far more likely that people are having issues with 2mm than with 1.5mm, so the prudent course of action is to advise other people to use 1.5mm in the first instance and if they tried 2mm and have massive issues, to suggest they try 1.5mm.

It's more on wanting to know the reason why some have success or not with 3080fe and 2mm gelid Xtreme, I didn't squash them much at all just enough to get them to stick and from looking side on I can see how much higher they were to the metal plate around the gpu so I know they would be compressed

My assumption was gelid Xtreme squeeze a lot so are more forgiving I don't know about other pads how they compress do also think 1.5mm gelid Xtreme would work as well just don't compress them a lot when applying on

I think you have to be more precise with mm if going with less forgiving pads that don't squeeze as well

 

[uscool]

So, got up eartly this morning, brewed a coffee and took the card apart...again.

Reapplied thermal paste, closed her back up and ran Heaven Benchmark.



A lot better temps now then before

I think this is as good as I can get it. Going to now work on improving airflow in the case.

Thanks to everyone who has helped or replied to posts in this thread. I hope others out there who are attemptign to do the thermal pad change, learn something from my mistakes here.

Fair play you must be expert taking it apart now, did you re apply the pads when applying paste again or left the old pads ?

 

[Trox]

prudent course of action is to advise other people to use 1.5mm in the first instance

I would be concerned that using 1.5mm in the first instance means some of the areas covered fail to contact the heat sink plate.
Does the temperature monitor all the chips or only a specific area? If it's just that latter then you would never know.

2mm extremes at least give some reassurance that you get the coverage and the heat sink will compress it.

Using 2mm first then changing out to 1.5mm on a specific area if issues occur to me sounds the best way, obviously I'm no expert and these are just my thoughts on it.

 

[Rhaos]

Fair play you must be expert taking it apart now, did you re apply the pads when applying paste again or left the old pads ?

Not sure what you mean by old pads. But basically yesterday my new repalcement Gelid Ultimate's turned up (1.5mm and 2.0mm) and I fitted those as per instructions from Falkentyne https://i.imgur.com/8jR7hfP.jpg.

I followed what you did second time round using your diagram for the Backplate.

However it seems I did a **** poor job on the GPU thermal paste, so redid it all this morning.

 

[uscool]

Not sure what you mean by old pads. But basically yesterday my new repalcement Gelid Ultimate's turned up (1.5mm and 2.0mm) and I fitted those as per instructions from Falkentyne https://i.imgur.com/8jR7hfP.jpg.

I followed what you did second time round using your diagram for the Backplate.

However it seems I did a **** poor job on the GPU thermal paste, so redid it all this morning.

I meant the gelid pads from the previous attempt took it apart only redid paste ?

Myself might look into changing paste with tfx just to see if get any difference but concerned will need to change pads again because will be compressed and they don't bounce back and maybe pads will have less contact pressure

 

[Rhaos]

I meant the gelid pads from the previous attempt took it apart only redid paste ?

Myself might look into changing paste with tfx just to see if get any difference but concerned will need to change pads again because will be compressed and they don't bounce back and maybe pads will have less contact pressure

So yesterday I added the new Gelid Ultimate Pads after removing the Gelid Extreme pads I had on there.

You are right, cleaning the old Gelid Extreme pads is a nightmare, however I found a great soultion, use a toothbrush and a bit of IPA. Works wonders.

I must have put too little thermal paste on the GPU. Hence why the GPU temps and hotspot were really high.

This morning I took the card apart, and just reapplied the thermal paste o the gPU (MX4), and adjusted the Gelid Ulitmate pads that may had slightly moved or stuck to the backplate. The Gelid Ultimate pads on the Front (Gpu side) were fine and did not need adjsuting. I could see indents on the pads where they had made contact with the heat sink.

Too add when I removed the old Gelid Extreme pads yesterday on the Front, the ones which I squished down, there were no indentations on the pads. So I am 100% sure that they were not making contact with the Heatsink and why my temps were super high.

Hope the above explains it, so basically no I did not replace the Gelid Ultimate's that I had applied yesterday with new ones, no point as they were fine.

 

[Skynet5]

So, got up eartly this morning, brewed a coffee and took the card apart...again.

Reapplied thermal paste, closed her back up and ran Heaven Benchmark.



A lot better temps now then before

I think this is as good as I can get it. Going to now work on improving airflow in the case.

Thanks to everyone who has helped or replied to posts in this thread. I hope others out there who are attemptign to do the thermal pad change, learn something from my mistakes here.

really please you got it sorted and importantly better mem temps.

not sure what your gpu temps were before, but they are fine.

 

[uscool]

So yesterday I added the new Gelid Ultimate Pads after removing the Gelid Extreme pads I had on there.

You are right, cleaning the old Gelid Extreme pads is a nightmare, however I found a great soultion, use a toothbrush and a bit of IPA. Works wonders.

I must have put too little thermal paste on the GPU. Hence why the GPU temps and hotspot were really high.

This morning I took the card apart, and just reapplied the thermal paste o the gPU (MX4), and adjusted the Gelid Ulitmate pads that may had slightly moved or stuck to the backplate. The Gelid Ultimate pads on the Front (Gpu side) were fine and did not need adjsuting. I could see indents on the pads where they had made contact with the heat sink.

Too add when I removed the old Gelid Extreme pads yesterday on the Front, the ones which I squished down, there were no indentations on the pads. So I am 100% sure that they were not making contact with the Heatsink and why my temps were super high.

Hope the above explains it, so basically no I did not replace the Gelid Ultimate's that I had applied yesterday with new ones, no point as they were fine.

the die side not difficult to remove and clean as they are on smooth surface vram/vrm, just the back gets stuck on the PCB with the little chips I would be worried using toothbrush incase I chipped one away looking back I would probably use tooth pick and slowly chip away carefully with IPA , with my OCD I would change the pads even just re applying the paste because they have been compressed , mine seems fine but at some point much later on I might redo mine and use TFX paste just to see if any difference

 

[Rhaos]

the die side not difficult to remove and clean as they are on smooth surface vram/vrm, just the back gets stuck on the PCB with the little chips I would be worried using toothbrush incase I chipped one away looking back I would probably use tooth pick and slowly chip away carefully with IPA , with my OCD I would change the pads even just re applying the paste because they have been compressed , mine seems fine but at some point much later on I might redo mine and use TFX paste just to see if any difference

Thats why i didnt go with Extreme pads and used Ultimate instead. less squidgy.

You have nothign to worry about using a toothbrush, especially as they are designed for our teeth gums, and i wasn't applying a lot of pressure or scrubbing super hard. Just being gentle with it.

Its just like cleaning you teeth. It definitley was easier to get rid of the stuck down bits of Extreme Gelid pads.

 

[RavenXXX2]

Just done a 3080 FE, used extreme pads and mx4, just the ram side. Nicehash 96Mh/s, max junction temp 80c at 50% fan, happy with that.

 

[Skynet5]

Just done a 3080 FE, used extreme pads and mx4, just the ram side. Nicehash 96Mh/s, max junction temp 80c at 50% fan, happy with that.

what did you use 2mm?

 

[RavenXXX2]

Yeah just on the ram and VRM. Cross method and blobs for the paste application, a load of it though. Temps are great on the core and ram. IMO doing the back of the PCB is not necessary at all.

 

[Skynet5]

Yeah just on the ram and VRM. Cross method and blobs for the paste application, a load of it though. Temps are great on the core and ram. IMO doing the back of the PCB is not necessary at all.

I'm gonna get 2mm extreme then and mx-4. I'm in no rush, will wait to see for a decent price for both...I think the prices have risen for these based on demand.

 

[TaKeN]

Yeah just on the ram and VRM. Cross method and blobs for the paste application, a load of it though. Temps are great on the core and ram. IMO doing the back of the PCB is not necessary at all.

I also ignored the backplate on mine, what paste did you use? Can you share your Core & Hot Spot temps please?

 

[RavenXXX2]

I also ignored the backplate on mine, what paste did you use? Can you share your Core & Hot Spot temps please?

MX4 and in gaming I'm getting lower core temps to the tune of 5c, same for hot spot. With an undervolt for 260 watts at 1840Mhz temps are like 55c on core and 70 mem in control, well impressed. This with fan set to 45%.

Tested port royal, 5 loops stabilized at temps shown in screen.

 

[FlyingFish]

How did you get those junction and hot spot temps in Afterburner?
When I last looked they weren't options

 

[Skynet5]

MX4 and in gaming I'm getting lower core temps to the tune of 5c, same for hot spot. With an undervolt for 260 watts at 1840Mhz temps are like 55c on core and 70 mem in control, well impressed. This with fan set to 45%.

Tested port royal, 5 loops stabilized at temps shown in screen.

How you get those measurements to display? MSI afterburner doesn't give them all...you using something else?

 

[TaKeN]

MX4 and in gaming I'm getting lower core temps to the tune of 5c, same for hot spot. With an undervolt for 260 watts at 1840Mhz temps are like 55c on core and 70 mem in control, well impressed. This with fan set to 45%.

Tested port royal, 5 loops stabilized at temps shown in screen.

Nice! Thanks, my delta between core and hotspot is 10 at idle and gets between 14 and 15 under load, I'm wondering if I didn't clamp the screws down tight enough on the X bracket, but I didn't want to strip the Hex head so I went easy on them.