I measured my pads. If you're looking for the pads thickness, here it is. 3.5mm 1x 100x10mm 2.5mm 1x 100x6mm 1mm 12x 10x15mm 1x 6x35mm 1x 8x110mm 1.5mm 1x 5x10mm 1x 6x100mm 2x 7x7mm 3.5 and 2.5 for the blacks 1.0 for the darker grays 1.5 for the lighter grays
I applied liquid metal to this and lost around 10-12C, real difference might be around 6-8 as dust and old paste contributed to worse temps. Clocks are more stable and a bit higher as well now. I am going to give a short tutorial on what I did. First cleaned the die, sanded the die down with 600 sandpaper just a little, then 1000 and 2000, so the LM gets saturated more easily into the structure of the die without the shiny finish coating that is at the top. Apllied a bit of LM to both die and the cooler copper plate onto the contact area. Left it there for 48hrs. This is just a saturation phase, the LM isn't for final application. Cleaned LM off, applied 3 coats of clear acrylate nail polish over the pins around the die, first coat directly over the pins without going much to the sides, next coats a bit more wider so they cover over the previous coats and + a bit more area to the sides. I then applied the very squishy mesh fabric over the pins, I just buy ebay ear muffs which are made from such squishy polyurethane material, it acts like a barrier for the LM in case it spills (which it in my case does as I apply a bit more as its better than not applying enough and not seeing results). It goes around the whole die and is higher than the die, around 2-3mm higher which is fine as it can squish down, it's just important that it has space to squish down so it should be pretty much center, otherwise it might get stuck between cooler and the die or the railing thing which is around the die. I secure in place with loctite superglue, and again it goes over the pins, around the whole die in a way that it can squish down easily and not get stuck on anything between the cooler plate. This acts like a barrier obviously for the LM, it protects the gpu pcb board, not and the pins are protected by acrylate nail polish, both can withstand high temps that are regular for gpu. Last thing to do is applying the actual LM coat, I apply a tiny bit on the copper plate, spread it around with ear swap, then apply more onto the die itself so it's nicely watery without any dry spots, its a bit more watery than people usually recommend but I am done applying "just enough" which can sometimes turn into "not enough" where you then see little to no results. At that point is assembly, this has to be done slowly and carefully without forgetting any cables and placing it precisely how it should be, also carefully with any torn pieces of thermal pads as those can stack up on each other creating a gap, if they're torn they should be scraped off and placed on the components so they have around the same height as they should have, I never replace those, rarely if they're too damaged only, as I never saw much improvement replacing all with new ones. I guess that's about it, die and copper saturation, protection of pins, anti-spill protection, application and careful assembly. It takes a bit of time but this way I sometimes had amazing results with both cpus and gpus, and I like to play it safe so I can apply more LM not risking little to no results. After that I would personally never again disassemble the card, its not much worth it unless you notice performance difference what doesn't seem to happen with LM, doesn't seem to be as usual as with standard paste.
Just measured mine. 3.5mm 1x 100x10mm 2.5mm 1x 100x6mm 1mm 12x 10x15mm 1x 6x35mm 1x 8x110mm 1.5mm 1x 5x10mm 1x 6x100mm 2x 7x7mm 3.5 and 2.5 for the blacks 1.0 for the darker grays 1.5 for the lighter grays
i recently got a secondhand 1080 and was experiencing crashes. i dismantled it to see if there were any issues and there was 1 pad and it was broken. this video shows me theres supposed to be more. what was the thickness of the pads you used to replace it?
Thanks for the video. I'm considering flushing the heatsink with water and dish soap. I bought this used pc with a GTX 1080 ti. However, it was not just a "used" pc unit in the shop/store. It was both a used unit and a PC that the employees use to play games until it was sold - to me, and everyone in that shop smoke. So, smoke was pushed into the heatsink almot 12 hours each day constantly. I hope this helps
As long as you don`t put fans or any electronics in the water it will be fine, you might do more washes to get rid of the smell. Glad you liked the video :)
@@samscreativevideos1845 Yes, of course. It's just common sense to not wet any electronics, especially while they're connected to electricity. I did like the video, only I forgot to actually "like" the video until I was reminded by your reply 🫢
I just did regular MSI 1080 with liquid metal similar way I wrote in my other comment (no saturation this time cos dude wanted the card ASAP) the temps at benchmarks at all stock and 100% fan speed went from 73 to 53°C, 75 to 55C, all around it was essentially a 20°C drop, at stock fan control it was around 7-8C drop and 22-24% slower fan speeds. I should say the card was never opened so paste was bad, I applied new gelid ultimate pads (1.5mm ones have to be squished down before you install cooler otherwise they put up too much resistance and pressure on screws). I do multiple tests, firestrike, heaven, furmark, last light, shadow of tomb etc. to test a bit wider variety and all around a huge drop in temps, 20C is pretty much one of bigger drops you can see, quite usual on older cards that werent repasted. This makes me think of that 1080 Ti I mentioned in the previous comment where I explain how I apply LM, and I am pretty sure it was opened before so it probably was repasted at some point, the paste was probably quite alright only 2-3 years old I guess and the 10-12C difference was then mostly the LM performance. It would make sense since the bench temps are more or less the same and these cards are very similar, only around 15-20% more power on the Ti.
i got 2mm gelid ultimate pads , can i use them on the vrms and the capacitors , instead of buying variety of sizes as the 1st comment on this video mentions ?
Se eu quiser acessar apenas a parte que faz a troca da pasta térmica , tenho que remover todos os parafusos do blackplate também? Ou bastam apenas os 4 ao centro, assim como fazem na gtx 1070 ti da MSI ? Eu queria só trocar a pasta termica por outra melhor pra ver se baixava a temperatura.
@@samscreativevideos1845 obrigado amigo. Minha placa da MSI está alcançando 84 graus, o limite configurado no Afterburner em um gabinete novo e fechado mas com boa ventilação e seis fans de 120 mm. Se removo a lateral de vidro temperado a temperatura da gpu cai para 75 graus em média.
You should try phase-change material Honeywell PTM7950. It comes in as a sheet, lasts longer, performs works better than traditional pastes and there is no pump-out effect. For the power section (vrm) and vrams, I advise you to apply thermal putty Upsiren UX Pro. It is also a good idea to do undervolting.
My GTX 1060 6GB that I bought back in 2016 sometimes goes into black screen with the fans blazing at 100% (temp never goes above 70C), I tried repaste it anyway, use another psu, try it on another pc, etc etc to no avail...until I change all the thermal pads on the vram and problem solved!
@@rygaros6669 Where did you get those measurements from? On the MSI page for the GTX 1080Ti Gaming X (which is the one in the video) they indicate that you need the followin pads: MSI GPU thermal pad Information Unit: mm L*W*T 106*7*1 (6*6*1)x2 (14.23*10.3*1)x12 96*6*1.25 36*16*1.5 9*5*1.5
I measured my pads.
If you're looking for the pads thickness, here it is.
3.5mm 1x 100x10mm
2.5mm 1x 100x6mm
1mm 12x 10x15mm
1x 6x35mm
1x 8x110mm
1.5mm 1x 5x10mm
1x 6x100mm
2x 7x7mm
3.5 and 2.5 for the blacks
1.0 for the darker grays
1.5 for the lighter grays
so what ones should i buy
The 1070ti msi titannium had the same thermal pads thickness?
I applied liquid metal to this and lost around 10-12C, real difference might be around 6-8 as dust and old paste contributed to worse temps. Clocks are more stable and a bit higher as well now. I am going to give a short tutorial on what I did.
First cleaned the die, sanded the die down with 600 sandpaper just a little, then 1000 and 2000, so the LM gets saturated more easily into the structure of the die without the shiny finish coating that is at the top.
Apllied a bit of LM to both die and the cooler copper plate onto the contact area. Left it there for 48hrs. This is just a saturation phase, the LM isn't for final application.
Cleaned LM off, applied 3 coats of clear acrylate nail polish over the pins around the die, first coat directly over the pins without going much to the sides, next coats a bit more wider so they cover over the previous coats and + a bit more area to the sides.
I then applied the very squishy mesh fabric over the pins, I just buy ebay ear muffs which are made from such squishy polyurethane material, it acts like a barrier for the LM in case it spills (which it in my case does as I apply a bit more as its better than not applying enough and not seeing results). It goes around the whole die and is higher than the die, around 2-3mm higher which is fine as it can squish down, it's just important that it has space to squish down so it should be pretty much center, otherwise it might get stuck between cooler and the die or the railing thing which is around the die. I secure in place with loctite superglue, and again it goes over the pins, around the whole die in a way that it can squish down easily and not get stuck on anything between the cooler plate. This acts like a barrier obviously for the LM, it protects the gpu pcb board, not and the pins are protected by acrylate nail polish, both can withstand high temps that are regular for gpu.
Last thing to do is applying the actual LM coat, I apply a tiny bit on the copper plate, spread it around with ear swap, then apply more onto the die itself so it's nicely watery without any dry spots, its a bit more watery than people usually recommend but I am done applying "just enough" which can sometimes turn into "not enough" where you then see little to no results.
At that point is assembly, this has to be done slowly and carefully without forgetting any cables and placing it precisely how it should be, also carefully with any torn pieces of thermal pads as those can stack up on each other creating a gap, if they're torn they should be scraped off and placed on the components so they have around the same height as they should have, I never replace those, rarely if they're too damaged only, as I never saw much improvement replacing all with new ones.
I guess that's about it, die and copper saturation, protection of pins, anti-spill protection, application and careful assembly. It takes a bit of time but this way I sometimes had amazing results with both cpus and gpus, and I like to play it safe so I can apply more LM not risking little to no results. After that I would personally never again disassemble the card, its not much worth it unless you notice performance difference what doesn't seem to happen with LM, doesn't seem to be as usual as with standard paste.
Hi, thanks for the video. I have a question can you tell me the correct thickness of thermal pads?
Just measured mine.
3.5mm 1x 100x10mm
2.5mm 1x 100x6mm
1mm 12x 10x15mm
1x 6x35mm
1x 8x110mm
1.5mm 1x 5x10mm
1x 6x100mm
2x 7x7mm
3.5 and 2.5 for the blacks
1.0 for the darker grays
1.5 for the lighter grays
i recently got a secondhand 1080 and was experiencing crashes. i dismantled it to see if there were any issues and there was 1 pad and it was broken. this video shows me theres supposed to be more. what was the thickness of the pads you used to replace it?
Thanks for the video.
I'm considering flushing the heatsink with water and dish soap. I bought this used pc with a GTX 1080 ti. However, it was not just a "used" pc unit in the shop/store. It was both a used unit and a PC that the employees use to play games until it was sold - to me, and everyone in that shop smoke. So, smoke was pushed into the heatsink almot 12 hours each day constantly. I hope this helps
As long as you don`t put fans or any electronics in the water it will be fine, you might do more washes to get rid of the smell. Glad you liked the video :)
@@samscreativevideos1845
Yes, of course. It's just common sense to not wet any electronics, especially while they're connected to electricity. I did like the video, only I forgot to actually "like" the video until I was reminded by your reply 🫢
can i use arctic mx4 or mx6 as thermal paste , is it a good choice for this hefty card ?
Quais thermalpd você usou ?
I just did regular MSI 1080 with liquid metal similar way I wrote in my other comment (no saturation this time cos dude wanted the card ASAP) the temps at benchmarks at all stock and 100% fan speed went from 73 to 53°C, 75 to 55C, all around it was essentially a 20°C drop, at stock fan control it was around 7-8C drop and 22-24% slower fan speeds. I should say the card was never opened so paste was bad, I applied new gelid ultimate pads (1.5mm ones have to be squished down before you install cooler otherwise they put up too much resistance and pressure on screws). I do multiple tests, firestrike, heaven, furmark, last light, shadow of tomb etc. to test a bit wider variety and all around a huge drop in temps, 20C is pretty much one of bigger drops you can see, quite usual on older cards that werent repasted. This makes me think of that 1080 Ti I mentioned in the previous comment where I explain how I apply LM, and I am pretty sure it was opened before so it probably was repasted at some point, the paste was probably quite alright only 2-3 years old I guess and the 10-12C difference was then mostly the LM performance. It would make sense since the bench temps are more or less the same and these cards are very similar, only around 15-20% more power on the Ti.
i got 2mm gelid ultimate pads , can i use them on the vrms and the capacitors , instead of buying variety of sizes as the 1st comment on this video mentions ?
Se eu quiser acessar apenas a parte que faz a troca da pasta térmica , tenho que remover todos os parafusos do blackplate também? Ou bastam apenas os 4 ao centro, assim como fazem na gtx 1070 ti da MSI ? Eu queria só trocar a pasta termica por outra melhor pra ver se baixava a temperatura.
Yes you'll have to remove the backplate screws as well since some of them do hold the heatsink
@@samscreativevideos1845 obrigado amigo. Minha placa da MSI está alcançando 84 graus, o limite configurado no Afterburner em um gabinete novo e fechado mas com boa ventilação e seis fans de 120 mm. Se removo a lateral de vidro temperado a temperatura da gpu cai para 75 graus em média.
why didnt u change the vrm pads ?
Are 1080 and 1080 ti basically the same for this?
Best graphic card Nvidia ever did....
Thx
Also what pads did you use?
3.5mm 1x 100x10mm
2.5mm 1x 100x6mm
1mm 12x 10x15mm
1x 6x35mm
1x 8x110mm
1.5mm 1x 5x10mm
1x 6x100mm
2x 7x7mm
3.5 and 2.5 for the blacks
1.0 for the darker grays
1.5 for the lighter grays
You should try phase-change material Honeywell PTM7950. It comes in as a sheet, lasts longer, performs works better than traditional pastes and there is no pump-out effect. For the power section (vrm) and vrams, I advise you to apply thermal putty Upsiren UX Pro.
It is also a good idea to do undervolting.
is it always necessary to replace the thermal pads?
My GTX 1060 6GB that I bought back in 2016 sometimes goes into black screen with the fans blazing at 100% (temp never goes above 70C), I tried repaste it anyway, use another psu, try it on another pc, etc etc to no avail...until I change all the thermal pads on the vram and problem solved!
@@deny_kurniawan I will take that as a yes. Thanks!
These kind of video's are quite useless if you don't show before and after temps! 🙄🙄
And tell me what is the thickness of the thermal pads
3.5mm 1x 100x10mm
2.5mm 1x 100x6mm
1mm 12x 10x15mm
1x 6x35mm
1x 8x110mm
1.5mm 1x 5x10mm
1x 6x100mm
2x 7x7mm
3.5 and 2.5 for the blacks
1.0 for the darker grays
1.5 for the lighter grays
@@rygaros6669 Where did you get those measurements from? On the MSI page for the GTX 1080Ti Gaming X (which is the one in the video) they indicate that you need the followin pads:
MSI GPU thermal pad Information
Unit: mm L*W*T
106*7*1
(6*6*1)x2
(14.23*10.3*1)x12
96*6*1.25
36*16*1.5
9*5*1.5