Traditional thermal paste or liquid metal paste?

Yeah, Liquid Ultra was a bit thicker before... Hence the people preferred it over Phobya Liquid metal.(Before Grizzly Conductonaut visit the market). And Coollaboratory Liquid Pro have never been an option.

You have still options as (if you don't have tested all) Prolimatech PK-3, Coolermaster Mastergel Maker, Gelid Extreme, Phobya NanoGrease extreme, Grizzly Kyonaut and ICD. I wouldn't look other places.

 

I have these pastes laying around > Phobya NanoGrease extreme, Grizzly Kyonaut and ICD

ICD was always my prefered choice but I damaged some CPU coatings with it while removing it. Its very abrasive. So switched to Nanogrease. Kryonaut is nice on larger dies. But pumped out on me in my previous machines on the CPU side.

Conductonaut is too liquid like for my tastes, need a perfect fitting heatsink for that. I have a 3~5c core temp difference, I feel that is too large to use conductonaut.

 

I've been doing a lot of research lately in regard with thermal conductivity, as I'm planning to do some major tweaks to my heatsink assembly. Still waiting for everything I ordered online to arrive, so it will be a while until I can share some progress.
A few aspects to consider when thinking about thermal paste/alloys/liquid metal, etc:
-thermal conductivity is a very tricky property: for example mixing together two materials which have a very good thermal conductivity on their own, could very well result in a material which has a very low thermal conductivity. This is because thermal conductivity depends on crystalline structure/molecular structure of the material to transmit heat. For example graphene has a excellent heat transfer along the surface, but is terrible across the surface. The same applies on graphite (very good along the cleavage lines/parallel to crystal structure and terrible across it).
-oxidation or not enough pressure between thermal transfer surfaces can drop the efficiency a lot. Sometimes just adding a bit of more pressure will create a huge difference. - Just make sure you won't crack your chip or board.

Materials with best thermal conductivity:

• diamond (but is brittle and the diamond paste will act as an abrasive material, especially on dilation/contraction cycles)

• graphite/graphene/carbon fiber - the direction of heat conductivity is along the sheet. So. placing a sheet of graphite as thermal conductor will have same efficiency like any ordinary paste, except it may be more brittle and can cause trouble

• silver - this is the best metal for thermal conductivity, but is a bit expensive

• copper - depending on alloy and structure will have very good to silver-like thermal transfer properties, but much cheaper

• aluminum is said to be very good at losing heat (heat disipation), but is rather average in heat transfer efficiency.

• nickel has a thermal conductivity of ~ 90W/mK, which makes it a good material for plating, but could limit thermal transfer in some situations (don't nickel plate your radiator fins)

• solders are mostly crap, especially tin-based. Don't look at alloy composition, because that won't give you anything. Even if it has Bismuth/Indium, it still can have low thermal conductivity. Best is to check the charts for a specific solder (some have even 20W/mK)

• best metal to use as solder with high thermal conductivity is Indium. Has thermal conductivity somewhere around 80-90W/mK (can't remember exact values right now), but the most important part is that it has a very good value even at high temperatures. Plus it keeps its malleability even at -150 Celsius (it's being used in soldering cryogenic systems). Also, even if it makes a oxidation coating (around 100-200 angstrom), it still can be soldered with or without flux. They recommend to not use flux for tin-based solders (as it may not be efficient in dissolving indium oxide), but I couldn't find any specific recommendations.

• Indium will slowly infiltrate into copper, creating an intermetallic compound which is very tough and brittle. That is why Liquid Metals (they contain indium) leave a gray-ish mark on copper. Nickel is a good infiltration barrier, though (a minimum thickness of 5 microns).

Based on the above, I can tell you, Liquid Copper from Coollaboratory is crap. They tried to add copper particles in a silicon suspension, but the overall efficiency of it is rather lame. Coollaboratory Metal Pads are also crap especially if your heatspreader has spring-loaded screws. I noticed on my laptop that it melted and created a thick plate on a side of the CPU, leaving almost half of the DIE uncovered. This happened in time, after few weeks of use.
Until now, best that I used was ThermalGrizzly Liquid metal. Still need to do some testing, as my laptop lost a bit of efficiency in time. It became stable now, but is definitely less than it was before. Possible due to heatsink moving on top of the cpu die, or maybe I didn't put enough LM, or it got oxidized... can't say for sure yet.

 

Could have been too little.
I found that putting too little can be extremely bad if you don't paint the heatsink surface with LM extremely well and press hard on it, on a heatsink that has not been sanded down and smoothed yet (and even then you should paint the HS with LM anyway).

I LM'd my r9 290x, I put a very thin layer on the HS and the usual very thin later on the GPU, but after 1 week, I was reaching over 90C, hotter than normal paste! I then checked and there was literally NO LM at all on the GPU! it had all been absorbed by the copper and was left was just a hardened film, if anything, on the GPU (literally nothing liquid, GPU die almost exposed) and a lot of hardened "crystallized" stuff on the heatsink.

When I redid it, by painting the heatsink with more force and putting much more LM than before (like two rice grain sizes worth), the temps have now held much more stable over several weeks, much better than before.

Because of that, I found it's better to use a little too much rather than too little at all. Just use proper protection and if necessary, foam dam barriers (cutout in the shape of the chip) for extra runoff insurance just in case.

 

@Falkentyne If you'll check this post http://forum.notebookreview.com/thr...ng-dell-precision-m6800.812606/#post-10660323 , the first spoiler shows how I did the TIM for my CPU and what I used for protection. Anyway, I had two sheets of copper 0.2mm thick, which makes a lot of LM surfaces (6 faces), so there is a lot to go wrong. Now I got a thicker copper sheet (0.65mm after polishing) and I'm going to replace it very soon. I needed this copper sheet as spacer, to make for the original paste which was considerably thick. Also, when replacing the copper I will have to reapply the LM.
Question: is it better to have the copper surfaces polished mirror, or sanded, to make them a bit rough?

 

3-5°C core temp difference ain't awful. I would open up and test pressure-flatness with a very soft cream on die (almost as water) and clamp down the screws. Open up and see how the cream looks like on die and HS. I used this medical cream for pressure testing. You can find simular maybe a bit thicker in consistency. But not too thick as thermal paste. Don't give up before you even have tested.

 

Thanks for the tip! Im very careful with liquid metal paste so I worry for all the imperfections. Also because the latest AW laptops are harder to open. The Kaby lake versions have clips around the palmrest which break very easily. Now have to replace the whole palmrest which basically means I have to fully disassemble it. In that regards the previous versions where easier. So I am searching for a long time solution that I dont have to maintain that often.

 

My experience with Kryonaut is good. It lowered stock temps on my P651 with 1070 with 5C on CPU and 3C on GPU. Combined with undervolt and a laptop plate with 2 fans my temps in Witcher 3 is at max 70C on CPU and 73C on GPU. Prior to UV and repaste CPU max was 87C and GPU 78C.

 

Guys, I wanted to pick up some Phobya NanoGrease Extreme after reading this thread but I can't find it on any US or even English sites. Do I have to stumble through some foreign site and import this stuff? Is there an easier alternative for something that isn't abrasive like ICD but is thick enough to last with worse fit and clamping in laptops? This is for a 2011 MacBook Pro and extending its life without having to repaste often is my main concern.

 

See if you can find Gelid Extreme.

 

Agreed. Re-pasting from LM to LM I remove remains of LM, and only make the heat-sink stain as smooth and shiny as possible, I dont try to remove it. This way the heat-sink does not NEED to suck no more LM lol In the long run, maybe trying to remove the stain I would shave off too much copper.

 

I still prefer CLU to conductonaut because it’s so much easier to spread. I think I will have to give liquid pro a try, they may have changed the formulation since it’s now listed on their site as 80 w/mk vs 73 for conductonaut

 

@Raidriar Please check my other post.
Before you pay $20 for a 1g syringe of LM, at least TRY This.

https://www.ebay.com/itm/Galinstan-...ass-Vial-Thermal-Compound-Paste-/323068372053

I'm using it *RIGHT NOW* on my laptop and temps are identical to conductonaut. IDENTICAL. Maybe 1C difference at most.

And look at how much you get.
Would last you basically a lifetime. You'd probably wind up spilling or dripping more out of it from accidents or from the custom syringe you will need to get (usually free at pharmacies) than from what you use on a laptop.

let's be real here, folks.
What would you rather buy?

30g of the SAME STUFF for $22? or 1g of this stuff for $18? https://www.amazon.com/Coollaborato...659&sr=8-1&keywords=coollaboratory+liquid+pro

 

What type of syringe?

 

Any basic measuring/plunger syringe.

Just has to have a small narrow tube at the end.
And you need some sort of cap to cover the tube.

If you don't want a free pharmacy one you can try these:

https://www.amazon.com/Pack-Refilli...e=UTF8&qid=1524035251&sr=1-4&keywords=syringe

 

How does it spread? Conductonaut is very difficult to work with, CLU is very easy to work with. I've heard Liquid Pro is hard to work with and given its high w/mk rating, I'm guessing it's like conductonaut.

 

Spreads the exact same as Liquid Ultra. Another user said it's the exact same stuff, just CLU repackages it and sells it for a gigantic markup.
Btw how is conductonaut difficult to work with? I've had no difficulty with it. Are you using the spreading black swabs? They don't work well if they're wet or worn down (if alcohol residue is on them, for example, liquid metal just winds up sticking to it). Make sure you're using clean objects to spread it. if it's difficult to work with, you probably have an old or contaminated object you are using to spread. I found the more I used the same swab (especially after trying to wipe it with alcohol to clean it), the more it became difficult to spread with it.

The only difference between CLU (OLD CLU, which is better than the new stuff), CLP and Conductonaut is the addition of that one material (I forgot what it's called) which either raises or lowers the melting point. Normal galinstan should melt at about -12C. There is an article on that issue, actually.

 

Why I’ll never use liquid metal paste again
Captin Camerica
Published on May 8, 2018
It finally happened.....I was readjusting some fans luckily and seen some of the thermal grizzly liquid metal on the board... so now I’m going to get a new board and stick with my gelid extreme paste. That **** was a pain in the ass to apply and it ruined my board, it’s not even worth it.


Thermal Paste Investigation - GELID GC Extreme Vs Arctic MX4
Won't use LM conductive TIM's...
Tech YES City
Published on May 7, 2018
Here is a video on some thermal paste investigation, and after having mediocre overclocks and coming into a very solid overclock with some coolermaster master gel pro, I then decided to see what was a highly recommended thermal paste, and stumbled upon Gelid GC Extreme, and after now having tested it on three different GPUs, I am convinced, this stuff is legit.

Original video that helped me out here (feel sorry for the guy): https://www.youtube.com/watch?v=F4TUf...

 

Almost nobody read the disclaimer about Lm and simply straight away jump into results without doing any research.

 

Coolermaster mastergel maker is pretty good especially for the money, less good than kryonaut with about 2 to 3c but i felt it worked better than Nanogrease.

The thing is I woould recommend Kryonaut if it didnt had the potential to go bad so quickly over time. The heatsink needs to have a perfect fit for it to stay good. If there is a slight uneveness it pumps out. Nanogrease doesnt pump out but it is very hard to apply very thinly for the optimal results.

 

Even after a year with CM maker gel nano thermal performance hasn't degraded at all!

 

Yeah papusan recommended me that paste partly based on your experiences.

 

When? I don't remember seeing it.

 

He did in PM to me, he said you had good experiences with it and Falkentyne as well.

 

Kryonaut working great on my on RB completing a year now, just as day one. Thank goodness, if I didn't have the EVGA conductonaut would be the next test drive.