• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

Repasted laptop GPU 1660 TI ends in higher temps than before

Joined
May 24, 2020
Messages
47 (0.03/day)
Hello. As explained, my laptop's GPU 1660 TI has higher temps than it did before i repasted it. I know that shouldn't happen, but i did the repaste twice and after i checked how it was after my first application it seemed ok for the GPU atleast, but the paste hadn't covered the CPU fully, so just for good measure i did the line method on them both. The CPU went from about 93-96 in Cyberpunk to about 83-85 or so. Not the best, but i've already done it twice. But the real problem is with the GPU which seems to have gone from it's max 83 with stock paste to 86 max with the repaste, with fans on max. it doesnt even seem to be moving 1 degree at all. but when i close the game, it instantly drops to about 50 and stays there, even on auto fan mode. And with the line method, there is almost a guarantee that the GPU is covered, since i also tightened the screws with the criss cross method and did it slowly, to make sure it spread out properly. But it still ends in higher temps than before. How come? can anybody help? i really dont want to open it up again and repaste the third time in just 2 days. Any help is appreciated.
 

Attachments

  • GPU data after GPU-Z 1.6.21.PNG
    GPU data after GPU-Z 1.6.21.PNG
    28.2 KB · Views: 267
usually some good noctua paste, or AS5 works great. it might be a case of something else getting disturbed, maybe a thermal pad was pushed or knocked out of place, or a fan is obstructed, i knows its not rocket science, but it wouldnt be the first time. i dont see a way around opening again, unless you make peace with those temps,, or if the curing time for the TIM is a factor & you havent waited long enough

id suggest monitoring the temps while not in a game, then running a game that is more 'established' driver wise to determine if thew GPU is running hotter than it shoudl be.
 
Well I just got done repasting my rig 6! times... ended up having 1 bad tube of Kryonaut, and the other tube dropped temps by 10C.

So PSA - paste does, infact go bad, and can PIT/scratch your heatsink base when it does so. Maybe try a different tube of paste?
 
good noctua paste, or AS5 works great
well... i used MX-4 since i had heart it was only a couple degrees behind the NT-H1, but the problem is that i could get more MX-4, for the same price as the noctua paste. and with only about a couple degrees difference it seemed to make sense to use that one... but alright, i'll keep monitoring it a bit and if it really is done for i'll open it again.
if the curing time for the TIM is a factor & you havent waited long enough
you mean the burn in time? in that case, from what i've gathered, there should be almost no burn in time on MX-4. But im not sure. anyways, thanks for your help though. i'll try and see what happens later on. if nothing good then there's nothing to do... i can dread it, i can run from it, but destiny will still arrive all the same. have a good one mate.



Maybe try a different tube of paste?
well... not gonna lie, im using the MX-4, so the chance of it being bad as the first one (if it really is bad) is a bit high since its a bit old... it doesn't even say it is the 2019 version. and from what i can see online, the 2019 version doesnt come in a small box but kind of a bag. so it looks like i got the old version which should atleast work a bit. i dont want to go out of my way and get another tube... but if it gets to it, i'll try to sort something out. thanks for your help and experience, mate.
 
Shouldn't have re-pasted your GPU if temp isn't above 90°C as modern GPU has a much higher thermal tolerance than that.

How much paste did you apply on the GPU?
Did you also cleaned/scraped off any previous paste before applying new paste?
 
first, have you cleaned the old paste perfectly? there's a chance old paste ruin the new paste
2nd, how much you use the new paste, some prefer small grain, since more paste is not great for the temp
3rd, how good the contact area, and how you tight the heatsink
 
paste does, infact go bad
The white stuff, seems the most likely to do that! It's likely that the Antec Formula 7 paste, OTOH, would last a good while.
 
I'm guessing you've applied either too much paste or you're not screwing it tight enough
 
MX4 is not suited for laptop, the reason is it pumps out at high temp (the TIM liquidfy and spread out so there is almost nothing left on the die after some gaming session)

For Laptop application I suggest these paste
Batman ZF-EX (pretty cheap, rebranded Thermalright TF-X)
Kingpin KPx (expensive, but the best performing)
 

Puget Systems make great computers too. I upgraded the server at work using these guys and their customer service is mind blowing.
 
@Empecial Just dab a small amount on the Die and use a plastic card to smooth it out.. also the key here is to properly clean both surfaces.. Some use Isopropanol alcohol, I get the good stuff, 99%er ;)

But to do a proper job I use ArctiClean 2 part cleaner.. It works extremely well.. W1zzard here on TPU uses that stuff religiously the last time I asked him about it.

Here's an Amazon link, Christ they want double what id generally pay at my local store..

ArctiClean 60ml Kit (includes 30ml ArctiClean 1 and 30ml ArctiClean 2): Amazon.ca: Computers & Tablets

 
Last edited:
@Empecial Just dab a small amount on the Die and use a plastic card to smooth it out.. also the key here is to properly clean both surfaces.. Some use Isopropanol alcohol, I get the good stuff, 99%er ;)

But to do a proper job I use ArctiClean 2 part cleaner.. It works extremely well.. W1zzard here on TPU uses that stuff religiously the last time I asked him about it.

Here's an Amazon link, Christ they want double what id generally pay at my local store..

ArctiClean 60ml Kit (includes 30ml ArctiClean 1 and 30ml ArctiClean 2): Amazon.ca: Computers & Tablets

Definitely, my type of application - very thin and smooth layer. Screws' tightening is also very important for properly distributed pressure - follow the numbers beside them.
 
Last edited:
Some laptops use thicker paste than normal, you may be better off with a thermal pad
 
How much paste did you apply on the GPU?
well, i did a moderate amount for the line. probably a bit more than what people would like.
Did you also cleaned/scraped off any previous paste before applying new paste?
yep. i got some alchohol that was about 85% or so i believe (the best in the house at the time) and did it mostly with cotton swabs to even get the small pieces out properly. and then did the dies a couple more times to ensure they were as clean as possible.

first, have you cleaned the old paste perfectly? there's a chance old paste ruin the new paste
did the best i could and made it clean.
2nd, how much you use the new paste, some prefer small grain, since more paste is not great for the temp
i did the line method to ensure it was spread out and all. i experienced my first application didnt cover the CPU fully so tried this to guarantee it

3rd, how good the contact area, and how you tight the heatsink
im not sure how good the contact area is, theres a copper pipe and a copper plate to make contact. its an acer nitro 7, maybe you can see what i mean then. and i tightened it with the criss cross method slowly to ensure it spread out

I'm guessing you've applied either too much paste or you're not screwing it tight enough
well,i did kind of read that i shouldnt screw it overly tight. but i did do it tight though.

MX4 is not suited for laptop, the reason is it pumps out at high temp (the TIM liquidfy and spread out so there is almost nothing left on the die after some gaming session)
well i saw some reviews and it seems to work for them, but i dont know at this point. i'll see about it then

Here's an Amazon link, Christ they want double what id generally pay at my local store..
well i'll try and check it out if it's worth it. thanks

very thin and smooth layer
well i was kind of hoping the heatsink would help in doing that
Screws' tightening is also very important for properly distributed pressure
i did the criss cross method slowly to ensure it spread out but i dont know how well i did it.

follow the numbers beside them.
from what i could see i didnt see any numbers on there for the way to do it, but next time i'll try and see if there are any

Some laptops use thicker paste than normal, you may be better off with a thermal pad
thermal pad? heard those have worse temps than paste, but at this point im not really too sure. i'll try researching it. thanks
 
Well I just got done repasting my rig 6! times... ended up having 1 bad tube of Kryonaut, and the other tube dropped temps by 10C.

So PSA - paste does, infact go bad, and can PIT/scratch your heatsink base when it does so. Maybe try a different tube of paste?
there was a huge bad batch of kryonaut. and some vendors on ebay AND amazon are still selling the bad batch vials. I had one and that was it for me. I use conductonaut still in my delids but i will NEVER use kryonaut again. I wont risk my hardware on their shady product. I run Noctua NT-H1 and my temps are the same and sometimes even better than with kryonaut and its a heck of a lot cheaper. https://www.amazon.com/Noctua-NT-H1...d&qid=1610032566&sprefix=nt-h1,aps,162&sr=8-3


 
is it seated properly?
 
MX4 is not suited for laptop, the reason is it pumps out at high temp (the TIM liquidfy and spread out so there is almost nothing left on the die after some gaming session)

For Laptop application I suggest these paste
Batman ZF-EX (pretty cheap, rebranded Thermalright TF-X)
Kingpin KPx (expensive, but the best performing)
my 2080 Ti AMP Extreme (350W Powerlimit) was repasted with MX4 over a year ago and the temps are exactly the same as on day one.

just need some proper full coverage manual application and not a pea on the die.
 
well, i did a moderate amount for the line. probably a bit more than what people would like.
i did the line method to ensure it was spread out and all. i experienced my first application didnt cover the CPU fully so tried this to guarantee it
i did the criss cross method slowly to ensure it spread out but i dont know how well i did it.
Well, that one part of the problem you made worse. You applied too much paste. It should be no more than a large drop-size or a pea-size amount. Thermal pastes are heat conductors that pass on heat to the cooler but too much paste like you did, won't pass heat optimally. Not to mention that the paste could leak and spread around the die which could possibly short circuit your electronics. Small amounts of paste is enough to spread on the die as there needs to be the shortest most direct contact possible with the heatsink.
yep. i got some alchohol that was about 85% or so i believe (the best in the house at the time) and did it mostly with cotton swabs to even get the small pieces out properly. and then did the dies a couple more times to ensure they were as clean as possible.
Don't use alcohol or any other type of fluid. Use a plastic scraper such as cards for the hard/dried stuff and use dry toothbrush for all the small bits and pieces.
thermal pad? heard those have worse temps than paste, but at this point im not really too sure. i'll try researching it. thanks
That depends on the product themself. Sometimes it's better, sometime their worse. Check for reviews. Or you can check this graph for comparison:
 

Attachments

  • yqSq6qVDxvcsSQc3kFUvfG.png
    yqSq6qVDxvcsSQc3kFUvfG.png
    411.5 KB · Views: 165
You applied too much paste
well i mean... i did do it a bit to ensure it, but then again, we have results and proven tests that too much paste shouldnt make THAT much more of a difference, except if its an unreasonable amount. and since i didnt do it to an unreasonable amount, i cant see how the paste would end up making it hotter than before
 
The paste (or TIM) is supposed to fill only the tiny pits and gaps between the two surfaces (CPU/GPU - Heatsink). Too much paste and it would become an insulator instead of conductor.
 
The paste (or TIM) is supposed to fill only the tiny pits and gaps between the two surfaces (CPU/GPU - Heatsink). Too much paste and it would become an insulator instead of conductor.
Ninja'd!!! LOL! :laugh: :slap: :nutkick:

@Empecial There is your answer above.
 
Too much paste and it would become an insulator instead of conductor.
you're not wrong, but as said, it was more of a line instead of just the pea. but then again, i am new at this a bit so i'll check it again. But im pretty sure you're aware that, that scenario doesn't have to be the case, even if you use a bit much. but i will check it. thanks man
 
Last edited:
you're not wrong, but as said, it was more of a line instead of just the pea. but then again, i am new at this a bit so i'll check it again. But im pretty sure you're aware that, that scenario doesn't have to be the case, even if you use a bit much.
There's no problem to be more IF the pressure manages to spread it even (up to zero) in which I doubt. A fellow forum member (@nguyen) suggests heating up the paste with a hairdryer or hot gun before screwing the heatsink and it definitely makes sense. It would make the paste less viscous, allowing it to spread easier.
This is the exact reason I don't like the 3-5-whatever number of dots or lines method. For me what works best is a very tiny and evenly spread layer of paste and lots of attention on screwing and tightening. What we're aiming at is the least amount of TIM, with as little as possible air trapped within (it expands when heating, making gaps, acting as a jack and besides is a very good insulator). Using the dots method is prone to leaving air where the paste from different dots gathers in my opinion.
 
Last edited:
  • Like
Reactions: Rei
well, i did a moderate amount for the line. probably a bit more than what people would like.

yep. i got some alchohol that was about 85% or so i believe (the best in the house at the time) and did it mostly with cotton swabs to even get the small pieces out properly. and then did the dies a couple more times to ensure they were as clean as possible.


did the best i could and made it clean.

i did the line method to ensure it was spread out and all. i experienced my first application didnt cover the CPU fully so tried this to guarantee it


im not sure how good the contact area is, theres a copper pipe and a copper plate to make contact. its an acer nitro 7, maybe you can see what i mean then. and i tightened it with the criss cross method slowly to ensure it spread out


well,i did kind of read that i shouldnt screw it overly tight. but i did do it tight though.


well i saw some reviews and it seems to work for them, but i dont know at this point. i'll see about it then


well i'll try and check it out if it's worth it. thanks


well i was kind of hoping the heatsink would help in doing that

i did the criss cross method slowly to ensure it spread out but i dont know how well i did it.


from what i could see i didnt see any numbers on there for the way to do it, but next time i'll try and see if there are any


thermal pad? heard those have worse temps than paste, but at this point im not really too sure. i'll try researching it. thanks


try using multi quote, i got 16 emails for this thread alone

thermal pads can be worse, but if your laptop requires a thicker TIM that wont squeeze out it'll out perform regular TIM every single time. mounting pressure and gaps are something we cannot see or measure, only you can.
 
  • Haha
Reactions: Rei
Back
Top