The Official Thermal Interface Material thread

[ThrashZone]

Hi,
AMD thermal ink

 

[Vario]

Might be a bit thin

View attachment 288587

What was the brand? Had similar thin results with NTH1 before. Presently using NTH1 though (because it came with the HSF). AS5 stays thick (no pump out), but you have to use spread method on it. I use spread method on all pastes (finger in plastic bag).

 

[80-watt Hamster]

Might be a bit thin

View attachment 288587

Still squeezing out; better dial it back a bit.

I'm only sort of kidding.

 

[freeagent]

That is Arctic Silver Ceramique

Temps were a tad warmer than what I was used to.. almost out now lol.. gotta wait till friyay

 

[Mazer]

I use SYY 157 for everything it's pretty solid even for bare die laptop CPUs.

 

[Sithhy™]

I have a ~9 year old Dell laptop, which probably has been using the same thermal interface material since whenever it left the factory. Few days ago I've bought Genesis Silicon 851 as it was quite cheap & supposedly offered almost same thermal conductivity as the best non-liquid metal TG thermal paste, which even though seemed somewhat fishy, I still went with. After applying the whole tube (0,5g) to the two dies in my laptop & then checking thermals, I am honestly not sure if it helped... kind of feel like the temperature at the hot end of the spectrum is worse than it was before as I think it heats up easier & a bit more than it did before but also manages to lower the temperature quicker, however the idle thermals have been improved by at least 5*C at the minimum. The stock thermal paste my laptop has been using might've not even been a paste, but rather something akin to thermal glue or similiar compounds, as it was not crumbly or cracked when I took off the cooler & was stuck to it as well as the CPU die quite well despite its potential age

 

[droid-I]

huh~ 9yr aged thermal paste, totally dried up, would think. But good you updated it if the laptop otherwise works OK.
Thermal glue, will not correct/help and I would not put that stuff in my laptops (reminds me of some older desktop stuff, heat sink superglued to the CPU, not easy to "update".

 

[nomdeplume]

I have a ~9 year old Dell laptop, which probably has been using the same thermal interface material since whenever it left the factory. Few days ago I've bought Genesis Silicon 851 as it was quite cheap & supposedly offered almost same thermal conductivity as the best non-liquid metal TG thermal paste, which even though seemed somewhat fishy, I still went with. After applying the whole tube (0,5g) to the two dies in my laptop & then checking thermals, I am honestly not sure if it helped... kind of feel like the temperature at the hot end of the spectrum is worse than it was before as I think it heats up easier & a bit more than it did before but also manages to lower the temperature quicker, however the idle thermals have been improved by at least 5*C at the minimum. The stock thermal paste my laptop has been using might've not even been a paste, but rather something akin to thermal glue or similiar compounds, as it was not crumbly or cracked when I took off the cooler & was stuck to it as well as the CPU die quite well despite its potential age

Is GS 831 a paste that requires a few hundred hours and good amount of thermal cycling before stabilizing? That would make sense here if so.

 

[droid-I]

holy moly, what kind of minerals and heat conductive stuff do they put in those pastes. Thermal cycling for stabilizing? I thought this was a home computer.
In some laptops and motherboards, it is just a "silly rubber" pad... btwn heatsink, CPU, VRM, bridge etc.

 

[Bones]

Is GS 831 a paste that requires a few hundred hours and good amount of thermal cycling before stabilizing? That would make sense here if so.

Even if that's true I doubt there would be enough stabilization to drastically change temps from what they would be when first used.
Thicker pastes are harder to break in anyway because they are thicker in the first place, thinner TIM's like MX-4 are thin enough to settle in quickly and do their job, plus MX-4 lasts a long time too.
I'd not worry about it so much and go with what's known to work in the first place.

 

[freeagent]

I like SYY-157, I have used about 50-60g. I will probably grab some more because it is uber cheap.

 

[droid-I]

Interestingly, some paste just dry up within 3 yrs, depending on duty use of the PC, otoh it if still works, dry up might not mean the paste does not do it its job,
but yes, I have been disassembling /assembling stuff in my PCs, GPUs for years. Good to check&update the paste once in a while.

 

[lexluthermiester]

Hi,
Yeah mx-5 was greenish and wet or at least mine was mx-6 is gray and on the thick side.

MX-5 is Blue not green. You really did have one of the bad batch tubes didn't you, eh?

Might be a bit thin

View attachment 288587

It's supposed to be thin. You're ideally only suppose to have enough to fill all the gaps, any more can be counter productive. That looks like a very excellent application to me.

is this good or bad? I have never seen this before!!! wow

This is excellent!

That is Arctic Silver Ceramique

Temps were a tad warmer than what I was used to.. almost out now lol.. gotta wait till friyay

I'm curious, what were your temps?

 

[freeagent]

MX-5 is Blue not green. You really did have one of the bad batch tubes didn't you, eh?


It's supposed to be thin. You're ideally only suppose to have enough to fill all the gaps, any more can be counter productive. That looks like a very excellent application to me.


This is excellent!


I'm curious, what were your temps?

My temps were fine, in the upper 70s for OCCT.. I just didnt like the spread between the cores, die, and ccd. The ccd was running a bit warmer than usual. I used the last squeeze in my tube.. its still thin, just not quite as thin..

I thought it was pretty sweet and had to take a pic

 

[lexluthermiester]

My temps were fine, in the upper 70s for OCCT.. I just didnt like the spread between the cores, die, and ccd. The ccd was running a bit warmer than usual.

To be fair, for OCCT, upper 70's is damn decent for a 5900X.

 

[droid-I]

This Akasa 455 works just fine, no need to judder around with thermal heat paste "capacities".

 

[freeagent]

To be fair, for OCCT, upper 70's is damn decent for a 5900X.

No that was my X3D

 

[Veseleil]

@freeagent not as good as yours, but still my personal best. MX-4 for more then a year. And it was a peanut squash technique.



Though I'm not sure how did that circle appeared, as I didn't twist the cooler when unmounting.

 

[Pictus]

@freeagent(...)
Though I'm not sure how did that circle appeared, as I didn't twist the cooler when unmounting.

 

[ThrashZone]

@freeagent not as good as yours, but still my personal best. MX-4 for more then a year. And it was a peanut squash technique.

View attachment 288781View attachment 288780

Though I'm not sure how did that circle appeared, as I didn't twist the cooler when unmounting.

Hi,
Cooler cold plate is likely flat and chip is not flat so cooler does not touch the middle of the chip.

 

[squall831]

Well, seems like I may try that GD900, it's an obscene quantity for what I use but is dirt cheap. A shame that it isn't that good for direct die applications, but 0.2 $/g can't do miracles

 

[lexluthermiester]

No that was my X3D

Ah, ok, yeah might have been a bit warm. Still, not bad.

 

[Bones]

@freeagent not as good as yours, but still my personal best. MX-4 for more then a year. And it was a peanut squash technique.

View attachment 288781View attachment 288780

Though I'm not sure how did that circle appeared, as I didn't twist the cooler when unmounting.

Most likely it did when the cooler was mounted but it's really nothing to worry about. Since MX-4 will dry out over time, it had to be when the cooler was first placed on the chip while it was still in a semi-liquid/paste form, hence the circular pattern that's "Set" into the TIM after it had dried.
I've had my coolers to twist a little when mounting them but it never caused any issues with temps and so on.

The contact pattern I'm seeing looks good to me, nice and even overall.

 

[Veseleil]

Most likely it did when the cooler was mounted but it's really nothing to worry about. Since MX-4 will dry out over time, it had to be when the cooler was first placed on the chip while it was still in a semi-liquid/paste form, hence the circular pattern that's "Set" into the TIM after it had dried.
I've had my coolers to twist a little when mounting them but it never caused any issues with temps and so on.

The contact pattern I'm seeing looks good to me, nice and even overall.

The circle probably comes from a coldplate design itself, but once I did twist the cooler and left a scratch marks on it, and the CPU IHS. The MX-4 isn't fully without hard particles I've learned that day.

 

[freeagent]

I think AMD IHS is soft compared to Intel.