The Official Thermal Interface Material thread

[Bones]

I was referring to people who use it normally, like a PS5 system for example. Open a PS5 in 40 years, and it will still be liquid in form. I'd bet money on it anyway, but yes if I knew we were talking about idiots who combine it with metals they shouldn't, my answer would have been different, lol

It may or may not still be there but much, if not most of the heatsinks themselves will at least show alot of damage due to contact with the stuff.
I'm willing to bet at least some PS5's you see up for grabs on fleabay with symptoms due to overheating have this problem, similar to the PS3 "LED's of death" that also being due to overheating.

If I ever get a PS5, LM removal and heatsink inspection/cleaning/repair will be the first thing I do and I don't care if it's new or used.

 

[Waldorf]

@Bones
cant comment on the 5, but the only ps3 overheating i had to deal with (mil/gov customers/myself), were all down to fan controller not ramping up fan rpm under load,
and those (few) units were exchanged for a new/open box one at no charge.

as long as it was applied properly, there is no air trapped (chip/HS), the LM will not degrade.
and one of two reason for using it (1st being heat transfer)...

 

[Bones]

@Bones
cant comment on the 5, but the only ps3 overheating i had to deal with (mil/gov customers/myself), were all down to fan controller not ramping up fan rpm under load,
and those (few) units were exchanged for a new/open box one at no charge.

I agree - It was a fan speed issue when the system was under load but there was also the issue with the solder they used as a contributing factor.
Silver solder is what they used and it's a harder solder type than 60/40 (Tin-Lead) so it's brittle in comparison. It doesn't withstand thermal expansion and retraction cycles as well as 60/40 does so it will develop micro-fractures at the spots the chips are soldered to the point the chip simply cannot function... And that's why alot of these units were having these issues.
Reballing is the only true fix to resolve it when the solder goes bad and that's done with 60/40 solder, once done that's it - It just runs compared to how it was before.

Get that done and the fan working properly, the unit will run for years beyond what you'd think it could.

as long as it was applied properly, there is no air trapped (chip/HS), the LM will not degrade.
and one of two reason for using it (1st being heat transfer)...

It's really not a matter of LM degregation, it's the simple fact LM will literally eat away at the cooler/ cooler surface reducing contact area/surface over time.
Way too many accounts of pitted and damaged coolers from it's use to deny the fact of it.

I mean LM does work but if the cooler itself is slowly being dissolved, what good does THAT do for your cooling over time?
Nothing good I'm afraid.

 

[Shrek]

It reminds me of the old days of copper soldering iron tips; the copper would desolve into the solder, embrittling the solder.

I'd hate for liquid metal to get close to my aluminum heatsink
Gallium Induced Failure of a Heat Sink - Bing video

 

[Waldorf]

@Bones
not sure what materials are used, and correct to say not making sense to use LM in that case,
but i would still see the issue with the (type) of materials chosen/used, not LM (causing it).
e.g. i know the polished nickel stuff never got "damaged", only darkened,
only copper surfaces were affected (where i used it).

@Shrek
lol, an old (previous) chip sold to someone, got fused to the HS, as i wasnt aware of the problem back then,
and used LM for "maintenance" free build when it comes to TP.

luckily the (dead) chip was the reason it wasnt working, so user never knew i couldnt even remove the cooler from the chip.

 

[Shrek]

Liquid Metal caused the heatsink to glue to the chip?

 

[lexluthermiester]

Liquid Metal caused the heatsink to glue to the chip?

BITD, yes it happened. It's not easy to make that mistake these days as most manufacturers are careful about the engineering and metallurgy.

 

[freeagent]

As much as I looove to chase down those C's.. LM has no place in my systems

 

[wheresmycar]

I had 2 options, either read all 28 pages... or just skip-em and ask the following:

I've been using TG Kryo mostly... Is it still one of the best solutions in 2023? What does it better?

 

[lexluthermiester]

As much as I looove to chase down those C's.. LM has no place in my systems

I'm not a fan of LM either UNLESS the system in question is going bare-die, then yes.

 

[freeagent]

I'm not a fan of LM either UNLESS the system in question is going bare-die, then yes.

Its been decades since I ran naked.. it wasn't even a thought.. but you are right in that case I would try it

 

[lexluthermiester]

I've been using TG Kryo mostly... Is it still one of the best solutions in 2023?

That is a very good TIM. If you have a bunch still, use it up.

What does it better?

There are some that are better, but we're talking about margins of less than 4%. So unless you're going to do some extreme OCing, you're good to go.

Its been decades since I ran naked.. it wasn't even a thought.. but you are right in that case I would try it

I did it recently for a client who wanted to go all in on a 12700k. The IHS was actually holding the system back. After we removed it and put on some LM, all was golden. It was pushing 5.8ghz where as before it was only able to get 5.3ghz. In the right use-case-scenario, LM makes sense and can be very effective. Not for daily driving though..

 

[Pictus]

I had 2 options, either read all 28 pages... or just skip-em and ask the following:

I've been using TG Kryo mostly... Is it still one of the best solutions in 2023? What does it better?

Honeywell PTM7950 is a type of polymer with a longer lifespan than normal thermal paste.

 

[Waldorf]

@Shrek
not the LM directly, but the reaction between the LM and the alu part from the fan/heatsink.

only started to use LM when i switch back to LC, and got tired of TM, which (in the past) always seemed to get "repelled"
if applying the more than once, even after proper cleaning with alcohol etc,, especially when cheap/stock stuff was on there before,
and (applying) LM didnt get affected by it.

 

[Shrek]

Ah! one should never use it with aluminum

gallium and aluminum - Search Videos

www.bing.com

 

[Waldorf]

lol, 11y too late for me. dont even remember if the packaging said anything about alu back then.

that (FX) cpu was actually the only one i "lost", as next user didnt game/care for more than stock cooler,
so on all gaming rigs i build it wouldnt have mattered.

just was glad the cpu was dead, and it didnt matter anymore that i couldnt unlock it from socket before removal.

 

[TheLostSwede]

Longtime lurker. Gotta show some love to GELID GC-Extreme. Been using it for years (switched from MX-4) and I find its still relevant (IMO). Great spread, temps, and longevity. Though a bit pricier. It's doing great on my 5950x (w/ mild PBO settings) in conjunction with the Thermalright FC140.

Used to use it as well, but it seems like there are better options out there these days, even though they might not be that much better.

 

[Lewzke]

I have received yesterday the PTM7950 package from ebay (shipped from Korea) and Honeywell has infrastructure in Korea. Maybe the seller sells cutted pieces from large rolls.
The application was difficult, especially peeling off the foil from the pad, even after the fridge. But I succeded finally. On the GPU the first application was good, but on the CPU I had to reapply two times. The results are much better than MX4, around 10 Celsius lower on a 8750H notebook i7 but without undervolt the CPU still hits the limits on the short boost period * but that CPU is hot in every review. With undervolting the CPU is well below the limits. Do you think I got the original Honeywell product or just a well made fake product? I asked the ebay seller and he replied that it is the same product as the description. Considering the performance I think it's original but I'm not sure. I can cut a test piece and try to melt it, that way I can test the material behavior to be sure that it is at least a 2 phase material.

 

[Pictus]

The results are much better than MX4, around 10 Celsius lower on a 8750H notebook i7(...)

It melts with the combination of BOTH pressure and heat above 45°C.
If the surface/geometry of the CPU/GPU/heatsink does not fit properly
or the mounting mechanism does not apply even pressure on all 4 sides
the performance probably will be worst than a normal thermal paste.
If my memory is not bugged, the same applies for Dowsil TC-5888.

People is satisfied with stuff from

PTM7950 SP Super Highly Thermally Conductive PCM Paste

Buy PTM7950 SP Super Highly Thermally Conductive PCM Paste for $19.99 with Free Shipping Worldwide (In Stock)

www.moddiy.com

PTM7950 SP Super Highly Thermally Conductive PCM Pad

Buy PTM7950 SP Super Highly Thermally Conductive PCM Pad for $5.99 with Free Shipping Worldwide (In Stock)

www.moddiy.com

 

[Shrek]

I'm not sure of any virtues with a phase changing material.

 

[Lewzke]

I'm not sure of any virtues with a phase changing material.

It works well in direct die cooling on notebooks. I feel like the heat transfer is noticeably better with this paste. 10 celsius is a lot compared to MX4, and maybe around 4-5 Celsius better than Kryonaut, but the temperature is cooling solution/chip/pressure/surface dependent. There are coolers where Kryonaut and this Honeywell performs the same, but in notebooks it looks like Honeywell is favorable. Probably the non-pumpout behavior and the melting properties just works well on the extremely flat silicon chips where the mounting pressure is high. Consumer thermal pastes just run out of the contact surface at this application after a while. Thermal conductivity is not the whole story, Kryonaut has better conductivity but fails on other aspects. Even if you create a highly conductive gel, if runs out from the surface, you will end up with dry and non filled contact. I have also Deepcool Z5, a very stable and similar paste to MX4, and when I dissasembled the notebook, the whole paste was in the sides of the chip, and at the contact surface only the paste carrier oil remained. It was funny to see the pumpout effect after half year of running. I will try at some point 7958 SP from moddiy, an optimized paste version of 7950, what Lenovo uses in the Legion series. I never tried liquid metal, because of the risk of electrical shorts. It is not a bulletproof industrial solution. This Honeywell is what i looked for.

 

[Pictus]

I'm not sure of any virtues with a phase changing material.

To complement what Lewzke wrote.
Phase Change Materials: Innovations in Thermal Management

Phase Change Materials: Innovations in Thermal Management

Electrolube's Jade Bridges and Ben Han discuss the latest innovations in thermal management materials for electronic devices: phase change materials.

electrolube.com

(...)I will try at some point 7958 SP from moddiy, an optimized paste version of 7950, what Lenovo uses in the Legion series.(...)

I bet it is just a more volatile solvent to achieve faster drying time.

PTM7950-SP | The most thermally conductive stencil printable Phase Change Paste | PTM7950SP

PTM7950SP: Ultra-conductive phase change paste for thermal interface applications. Minimizes resistance, excels in reliability tests, & offers cost-effective application. Stencil printing compatible (down to 0.1mm), but requires drying step (24hrs or 5min @ 100°C). Ideal for power modules...

www.caplinq.com

 

[Lewzke]

Interesting. What if you make the paste version drying process with Cinebench loop? I mean why not to dry assembled with running operating temps? It will run below the suggested 100 Celsius, because every chip is limited to that value. Also the pad version needs settling time, I suppose that the paste version takes even longer to settle. The first hours will be the drying and whet it dries it will start to settle like the pad version. The only concern is that you apply compression before the drying process, it can pumpout like normal pastes so you shot yourself in the foot. It will work or not? Worth a shot if you have this paste. Maybe it can be done with half tightened cooler and after the Cinebench loops you can tighten the screws totally. But yeah, this will be a mess of result probably. Better to make the steps properly as in the description, aka leave the paste on the chip for a day (covered with a paper U turn to shield from dust particles) and then assembly the system and it will work as the pad version. So probably the 7950 SP version not worth it for home usage if the pad version exists and ready to run, but the 7958 SP is an improved version with slightly less thermal impedance so you can further gain a little extra thermal performance with it if you apply it correcty, I can wait 1 day for 2-3C better performance, it can make a difference when the CPU throttles or not.

 

[wNotyarD]

Gotta fix a botched repasting of my 3070. Wish me luck.

 

[lexluthermiester]

Gotta fix a botched repasting of my 3070. Wish me luck.

View attachment 309767

Good luck brave warrior!