Liquid metal repaste 2 months ago - temps really high

Hi all,

I repasted my laptop a couple of months ago - and now the temps are really high again.

I'm guessing I need to do it again, but any ideas what I did wrong that it didn't last?

Thanks!!

 

I don’t have special powers to see what you did last time and how much material you applied


Sent from my iPhone using Tapatalk

 

Lol of course, but in your experience do you think it could be because I didn't put enough on?

Thanks

 

Which cores are hot? Have you had a read through the repaste guide by iunlock? Lots of good stuff in there.

Sent from my SM-G925I using Tapatalk

 

It could be. Could also be the heatsink isn’t level so not getting full contact with all cores.

Really need to see some data.


Sent from my iPhone using Tapatalk

 

core 0,1,2. I'll have to have another go I think. Triple check the heatsink is level..

 

Check your pads to see if they all have about the same level of indentation on them. If not that’s good evidence your pads are causing the issue


Sent from my iPhone using Tapatalk

 

Hmmm fair enough. I didn't change the thermal pads in fairness...

 

well I repasted today, it basically looked like it completely dissolved...

Any ideas based on that what I did wrong last time?

Thanks all!

 

@LaptopGamingGeek

It looks to me that (if 2 months ago it was 1st time LM applied) part of LM has reacted with copper and lost its volume a bit.

During 2nd time LM application I leave the stain on copper, only smooth it, not remove. Now your LM should last much longer. as the copper stain has the proper mix inside.

 

Ah I see, that's good to know. Yes that was my 1st time.

I did actually clean the "stain" a little bit, but found that I couldn't completely get rid of it and as such reapplied on top of it. Temps are lovely and cool again and I turned my OC down to 4.1Ghz (from 4.3Ghz) but it's good to know what happened rather than think it was in incompetence.

That said, I struggled to remove the Tobii connector this time and around and now eyetracking doesn't work, I now need to replace that! Doh! No idea where to find a replacement cable though (that's in the UK...)

 

Besides the second repaste, the reason it completely dissolved is because of lack of pressure allowing air to get inside the heatsink and cause oxidation. All it takes is a few molecules to make it happen. Then when you combine high heat, it just increases it.

If you apply LM to a copper heatsink, have high pressure with a completely airtight seal, you will see that it will remain liquid even after a year, because it couldn't oxidize.
However if you take some spare or exposed copper surface and apply a thick coating of LM and leave it exposed directly to the air at room temp, it will be completely hardened in 2 weeks because most of the gallium will be absorbed, leaving the remaining to lose its eutectic properties and harden completely.

Sanding off the old LM (you need 3000 grit or polishing sandpaper + alcohol; just using napkins will take -forever- and the surface has to be completely smooth) and leaving the silver stain (don't sand down to the copper) will not fix the problem if you don't fix the pressure issue and heatsink balance first. But having enough psi and even balance is already hard enough on these laptops.
Make sure you have a layer of super 33+ tape around the chip, this will help a lilttle with blocking air as well as providing insulation around the die (this won't act as a dam though). Polyurethane foam dams (10-20 PPI, maybe 15 PPI is best) cut out and trimmed helps even more, as this can help block air and act as a complete failsafe from random LM ever escaping the chip. But balanced, firm and even pressure and an airtight seal is what makes LM last without having to repaste.

http://forum.notebookreview.com/threads/after-1-year-of-thermal-grizzly-conductonaut.799343/

@Papusan had LM last just fine on a copper heatsink for 3 years, because of proper balance, insulation and prep work.
If you are not sure if you have a decent fit, you can buy this stuff and test your heatsink balance to avoid unnecessary work and repasting over and over.

https://www.amazon.com/d/Computer-Heatsinks/Innovation-Cooling-Contact-Analysis-Heatsink/B07CKHRKHQ/

This stuff is fujifilm prescale (Ultra Low). It's possible to get free samples from "Sensorprod" if you ask them very nicely. Keep in mind they want your business money, so be discreet about it.

 

Bang on I've often thought this but not posted it as well. It's why I'm a bit more generous than most with first time LM repastes

Don't bother it's ****

 

Going with what c69k said, did you apply the LM to the die and a very light layer to the heatsink itself (like in the @iunlock guide)?

 

Absolutely mate. Followed it to a T.

All is fine now though.

 

My first LM application didn't last that long either.

On the second attempt, beside finding a suitable "spongy" material to seal around the cpu and gpu chips, i did use K5 Pro instead of thermal pads to be sure that there was no resistance whatsoever from the pads and that the heatsink would seat as tight as possible over CPU and GPU and it seems to have worked fine so far. Two months have passed since the LM application and there has been no difference from day one temps.