tested thermal pastes results inside

[purplegrower02]

I tested a few thermal compounds.
I applied all the exact same amount and spread the same way and held equal pressure for same time.

out of 12 compounds I found ic diamond 24 carat to be the best.
this had the lowest Temps by 3-4 degrees Celsius.

Imo that is a huge drop in temp for doing nothing but using a different paste.
it is a little bit more difficult to spread and slightly more expensive but we'll worth it when you figure the price of cobs, leds, heatsinks, drivers.

ic diamond also make a 7 carat this was just about the same as the rest where the big difference came from the 24 carat.

 

[purplegrower02]

I forgot to add that ic diamond 7 carat and 24 carat are supposed to be the exact same just the amount you receive is different but from my tests the 24 carat seemed slightly thicker and had lower temps. I went back and tested the two and 6 different cobs and every time the 24 carat had a lower temp so that can't be just coincidence. ..

 

[Positivity]

Did you try prolimatec?

 

[Red1966]

Did you use different tubes for each test? It could be just that the quality varies somewhat from tube to tube. Like the 24 came from a better batch than the 7.

 

[Red1966]

Did you try prolimatec?

View attachment 3201626

Wow! Not a single one was twice as good as no compound at all. And only a little better than toothpaste. I would have expected better performance

 

[purplegrower02]

I tried prolimatec a few kinds and a few kinds of artic silver.

yes that is what i was thinking about the 7k vs 24k or it could of been the 7k was leftover from a comp build about 14 months ago... maybe it slightly degraded.

forgot to add all finishes were lapped with diamond compound until it looked like a mirror and was flat within .00015 of an inch

 

[epicfail]

Locally I had access to Arctic Silver 5 and Noctua NT-H1. I chose the Noctua because it was rated better in this chart which differs a bit.

 

[purplegrower02]

Imo you can only use the charts as a base for your decision. Howyou apply the product the amount of pressure tthe length of pressure and type if alum. Or copper all this will effect the end result. So using the exact same product but one only .001 thinner will mostly give a different readibg.

 

[Positivity]

The charts show perfectly how well the different products transfer heat.

Thinner application on all will improve performance in all. Thicker application will worsen performance of all.

Using best performing product with thinnest application will provide best performance of all.

 

[GroErr]

Good post, I just saved myself $20, toothpaste it is - lmao

But seriously, any tips from folks who have done this on best methods to apply these thermal grease products before I get into it? I've seen "thin" referenced many times, is that like barely visible thin as in microns?, 1/32"?, 1/16"?

 

[Positivity]

I think sds uses a stencil with his for a more accurate amount.

I just use the spatula the pk3 comes with. Can get a pretty thin layer if you take your time.

Here are some pics if you haven't seen them already. A cxa 3590 I just did. The more I use pk3 the thinner it's been getting. My first few had a lot more TIM than the latest.

 

[GroErr]

Thanks, the visual is great, a picture's definitely worth a thousand words That looks like the easiest method too, I'll be using the same holders, so mount on COB holder, spread on COB, screw it in... seems straight forward.

 

[purplegrower02]

You want as little as possible. I sand down and lap with diamond compounds all my stars, heatsinks, and even the ceramic backing to the CXA3070. Most people think they are flat but even the ceramic backed cobs are extremely uneven by as much as .005 total flatness. Aluminum stars I've seen one off by as much as .027 of an inch. It's usually low or hollow in the middle.

 

[Hosebomber]

...forgot to add all finishes were lapped with diamond compound until it looked like a mirror and was flat within .00015 of an inch

I would love to know what tool and where you got it to measure to this accuracy.

 

[purplegrower02]

We have optical flats, monochromatic light sources, and precision flatness gauges.

I used just a typical pink granite stone with a precision flatness gauge to check. But with the MLS I can check flatness to a couple millionths of an inch. A half tenth is not difficult to measure at all. You can measure a half tenth flatness with a 0-5-0 indicator and a pink granite stone.

You can pick these tools up in most tooling catalogs or shops, just depends how much you want to spend.

 

[Hosebomber]

PFG's and Optical flats do not measure to that accuracy, only monochromatic sources can ... Are you a precision cnc machinist by trait? That's a multi-thousand dollar tool. I work in aviation and we never get close to those tolerances.