Thermal paste not as effective after a month or so...

Status
Not open for further replies.

MrMoo252

Commendable
Oct 31, 2016
5
0
1,510
Hi,
I have a gaming laptop with an i7 6700HQ and GTX 1070 that had heat issues, I opened it up and noticed there wasn't enough thermal paste on the CPU so I bought some new paste and applied it to both the CPU and GPU which dropped my load temps by about 30C (no more thermal throttling, yay!).
Then after a couple of months, I noticed the temps had gone up from about 70C to about 80C so I reapplied the thermal paste again using a different method (spreading a thin layer on the CPU rather than putting a blob on there and letting the heatsink spread it) which gave me an even lower load temp of about 65C to begin with.
Now, another month or so down the line and I have noticed my temps slowly creeping up, my CPU can hit about 88C when playing planet coaster on full settings which isn't immediately dangerous for a laptop but it's definitely a concern in the long run, especially if it keeps creeping up.

Should thermal paste lose some of its efficiency this quickly?
I'm using Coolermaster Maker Nano, is it worth trying another brand or should I expect differences between brand new paste and month old paste?

Note: Remember this is a laptop so any slight change in the cooling department has a massive effect on the temps, laptops aren't as forgiving as desktops unfortunately. Also, 80C is a pretty normal max load temp.
 
Solution
MrMoo252,

On behalf of Tom's Moderator Team, welcome aboard!

The phenomenon you're experiencing is known as "pumping" or "pump out". Typical silicon Thermal Interface Material (TIM) will ooze out from between the processor Die and your laptop's cooler over a relatively brief period of time (a few months) due to thermal cycling, which causes Core temperatures to increase, requiring the TIM to be repeatedly reapplied. Your description matches this scenario perfectly.

Liquid metal TIM, such as CoolLaboratory Liquid Ultra, is required instead of silicon based TIM because it's much more resistant to pump out, and will remain in place for several years with stable Core temperatures. There's actually three products that can be used...
MrMoo252,

On behalf of Tom's Moderator Team, welcome aboard!

The phenomenon you're experiencing is known as "pumping" or "pump out". Typical silicon Thermal Interface Material (TIM) will ooze out from between the processor Die and your laptop's cooler over a relatively brief period of time (a few months) due to thermal cycling, which causes Core temperatures to increase, requiring the TIM to be repeatedly reapplied. Your description matches this scenario perfectly.

Liquid metal TIM, such as CoolLaboratory Liquid Ultra, is required instead of silicon based TIM because it's much more resistant to pump out, and will remain in place for several years with stable Core temperatures. There's actually three products that can be used, listed according to thermal conductivity:

Liquid Metal TIM

Thermal Grizzly Conductonaut - 73.0 W/mk (Newest Product)
CoolLaboratory Liquid Ultra - 38.4 W/mk
CoolLaboratory Liquid Pro - 32.6 W/mk

Typical Silicon TIM

Cooler Master Maker Nano: 11.0 W/mk
Arctic Silver 5 - 9.0 W/mk
Gelid GC Extreme: 8.5 W/mk
Arctic Cooling MX4 - 8.5 W/mk

Of the three liquid metal products, Liquid Ultra is prefered because it has a slightly thicker consistency, so it's less runny and tends to stay in place a little better. However, Thermal Grizzly Conductonaut contains the highest percentage of "Indium", which has superior thermal conductivity at 81.8 W/mk.

Intel used Indium solder on desktop CPU's, but switched to TIM on 3rd Generation Ivy Bridge, which has since caused a large percentage of problems with excessively high core temperatures. This has encouraged many users to de-lid their processors.

Keep in mind that liquid metal TIM is electrically conductive, so extreme care must be used when working with it to guard against getting even the smallest stray droplet on nearby components or circuit paths.

Once again, welcome to Tom's!

CT :sol:
 
Solution
Absolute Stability:
Arctic Silver 5 will not separate, run, migrate, or bleed.

During the CPU's initial use, the compound thins out to enhance the filling of the microscopic valleys and ensure the best physical contact between the heatsink and the CPU core. Then the compound thickens slightly over the next 50 to 200 hours of use to its final consistency designed for long-term stability.

Extended Temperature Limits:
Peak: –50°C to >180°C
Long-Term: –50°C to 130°C

http://www.arcticsilver.com/as5.htm


pretty much should be set and forget

I using coolermaster paste on this and its like 5 yrears old / no cooling issues yet [Thermal Fusion 400 ] came with the cooler in the box
 


Thanks for the reply, it's extremely informative. Fortunately, the 6700HQ doesn't have a lid so I can apply the TIM directly to the CPU, also, there are no SMD components on the top of the CPU (like there used to be) so i don't have to worry about leaking liquid metal from the edges of the die. The GPU on the other hand could be a problem as there are components quite close to the chip, i might use silicone based TIM for that, the CPU is causing me more concern than the GPU for some reason, probably because the GPU has 2 cooling fans and a bigger heat sink.

@junkeymonkey I used to use Arctic Silver years ago in the days of the Core 2 Quad, it did wonders with my overclocking capacity along side a massive heat sink, I might have to get some for my GPU
 


Well there's something i never knew existed, all the time I've spent trying to wrap insulation tape around awkward places in hopes that something won't short out and all the time you could get liquid tape!
Anyway... What happens when the tape dries? do you peel it off or leave it on?

EDIT: I have just read your temperature guide and what a guide it is!
To anyone experiencing this issue or any other CPU cooling related issues I strongly recommend putting in some time to carefully study CompuTronix's temperature guide, It's professionally written and extremely thorough.

Thank You CompuTronix, your work is very much appreciated
 


You need to leave the liquid electrical tape in place. Even when liquid metal TIM is applied with extreme care and attention to detail, there will always be some excess material that will flow from between the mating surfaces and remain along the edges.

Unlike desktop computers, since laptops are typically moved, jarred and carried in every conceivable geometric position, it's always possible for some excess TIM, which has the consistency of mercury, to find it's way onto a component or circuit path. This is why you shouldn't remove the liquid electrical tape.

You will need a small, fine tipped brush to apply the liquid electrical tape, as the brush included in the top of the container is much too large for detailed work.

Also, thank you for reading my Guide!

CT :sol:
 


Due to my laptop being quite expensive and still being in warranty (its a custom clevo so the parts are under warranty, not the laptop as a whole so im allowed to open it up :)) I have decided to give arctic silver a try instead of liquid metal, if the same happens with AS-5 then i'll have to take the plunge into the world of liquid metal.

Anyway, i have replaced the paste with AS-5 and as soon as i took the heat sink off i could tell the CPU didn't have enough paste on it, yet i'm sure i put enough on when i applied it, signifying that what you originally said was correct. The strange thing is it has only happened on the CPU, there was a very big difference between the amount of paste on the surface of the GPU die and the amount on the CPU die. It looked almost like it had turned watery on the CPU, as if it had been sprayed with alcohol.

Hopefully Arctic Silver will do a better job at staying where it's needed, my idle temps are hovering around 30C and I've just done a quick 15 minute prime95 stress test which hit 69C.

I'll just have to see how it goes for now, if the same happens, ill probably just stick some liquid metal TIM on the CPU and use AS-5 or coolermaster maker nano on the GPU.
 
I could've saved you the trouble; contrary to junkeymonkey's endorsement of AS5, it will pump out over time. Keep in mind that AS5 is also conductive, so be careful with it.

Pump out is relative to Die size, so smaller Dies are more susceptible. Since the GPU Die is much larger than the CPU Die, pump out is less evident over time.
 
I suspected it was down to the size of the die, I'll get some liquid metal ordered ready for next time, it takes a few days delivery.
Just out of curiosity, does this happen when you de-lid a CPU to put your own paste on? if so why doesn't it happen to Intel's stock TIM?
 
Status
Not open for further replies.