Best Thermal Paste According to Independent Studies?

Weird the link will not attach so here it is plain text

Not for aluminum! Which rules out most flashlight applications.

From the Toms Hardware review - image from them.Lower is better.

here is some data that includes cheese

http://www.coollaboratory.com/en/products/liquid-copper/

No mention of issues with aluminum. It might be worth trying.

Has anybody tested… paste… you know, the tasty favorite snack of pre-schoolers everywhere? I’d try it but I’m gonna go and huff some freshly mimeographed tests…

There gave been quite a few test involving toothpastes and surprisingly enough it wasn’t worse than some thermal pastes.

I wouldn’t be surprised if the liquid metal is an alloy of gallium which by itself isn’t that expensive.

Gallium and/or indium are used. Neither is a friend of aluminum or aluminium.

I am not even going to post in this thread.

no problem, can i ask why?

Probably because he doesn't think the hot rodding efforts are worth the marginal gains, so nobody else should bother doing something they might enjoy doing anyway.

No, it's not that at all. Those strawmen sure are easier to knock down though, aren't they?

Show me a test that shows any change at all in light output between a copper MCPCB with really bad thermal paste and a copper MCPCB soldered to the pill. Show me. I would think you would want to know if these things are real too, why am I the silly one by asking for the proof that these things actually work?

I assume the laws of thermodynamics apply so increased thermal transfer rate means the heat gets from one surface to another more easily when the interfaces are more thermally conductive. How much does that affect flashlights is a legitimate question, your saying not much, if someone wants to test this and provide some data then no one will object.

As the links i posted above demonstrate don’t use cheese in place of thermal compound

Share a Coke with Bort!

In a synthetic test where you have a tiny 1cm x 1cm patch and pumping 200 watts through it, those thermal paste charts are probably perfectly true. But change that to 30 watts through a 16mm diameter MCPCB and you can't measure any difference between the worst and the best. The difference is less than the margin of error of the measuring equipment.

Now, how much time, effort, & money (this is still 'budget light forum', right?) are you willing to invest for zero gain?

This does not mean the bottom of the MCPCB & top of the pill do not need to be flat. This does not mean a hollow pill is just as good as a solid pill. This does not mean a non-direct aluminum MCPCB with a dielectric layer in the middle is as good as a direct-thermal copper MCPCB. Do the proper prep work, that's far from 'none of this stuff matters'. The stuff that makes a difference matters. The stuff that does not make more light come out does not matter.

Bort is not one to change his avatar but this photo was impossible to resist :slight_smile:

your probably right, as an unknown person said “In God we trust; all others must bring data”
and:

I've been using GC Extreme, but comfy is probably right. I'm interested in sustained reduction of temperature at the LED, and I have higher confidence in GC Extreme right now for longevity (both long runtimes and life cycle), based on various sources/reviews from real users online - of course it's all from usage on CPU heatsinks, but they measured apparent real differences in temps. Does that apply to the lower power levels of LED's? Maybe, probably not the same though.

Problem is I'm not aware of anyone that has done comparative tests on our modern LED's and MCPCB's. Even so, again, I'd be interested in how they perform over time, and also what the temps are at the critical spots. Keeping the LED cooler at high amps will probably extend the lifetime of the LED (not only brightness is at stake), but again, anything over 3A to the LED is beyond manufacturer's limit.

Best thermal paste for what purpouse? CPU cooling, Led cooling?

You see, my PC is watercooled, back in 6 months i changed the CPU to 4770K…and i didnt got the cooling results i wanted, then i have decided to change the thermal interface between the CPU core and the protective cap -voila, temperatures dropped with 10 degrees!
My great surprice came after 2 weeks, the CPU temp not only rised, but rised alot! Up to 95 deg C under prime 95( while watercooled!), and the CPU became unstable…i used Prolimetech PK 2 by that time…then i swapped it with Pk3… with Gelid….same results after 1-2 weeks of heavy usage
Then i found a topic about that matter in Extreme systems forum, it seems that under heavy load and constant high temps most CPU thermal compound are degrading pretty fast especialy when put on a naked CPu core, exept the pure sillicon based thermal paste and of course liquid metal, now my CPU`s thermal interface is Coolaboratory Pro liquid metal, and after already 5 months i can say its rock solid
The GPu core on the other hand do not degrade thermal pastes that fast! Its bigger, and usualy works cooler, under heavly load and OCed temperatures never pass beyond 55 deg C
Since LEDS are always under high temperatures, maybe results there are simmilar- liquid metal cant be used on Al headsinks , so i dont have that huge experience exept with 2x30w Bridgelux emitters mounted on huge CM copper headsinks

P.S
When i removed the CPU cap the default Intel thermal interface was a pure sillicone glue of some sort
Have pics, if someone is interested

Btw, i have never tested AS5….its 99% micronised Silver, maybe it wont degrade , but it can oxidise

I used to be on extreme and evga forums :slight_smile: Yeah been using that for years Mitko… *Just note it doesn’t work on aluminum which is like all the pills

Yeah i mentioned that, but on copper it works! Plus, AS5 issue that i mentioned earlier…have to test it though

My point here is that PK1-2-3 or even Gelid wont do any good here, vise-vesra, they are worst that a cheap pure sillicon paste, the emitter in our case is always working at its max( most of the times at least), temps are above 85 deg, those CPU pastes arent designed for that

Yet…AS5….unfortunately my CPU is permanently metal- glued, and the LED die dont have build in thermal sensors

Have to ask someone from Xtreme if he tested with AS5, they do experiment alot