Best electrically conductive thermal paste ?

[Doctorbass]

The problem i see with soldering the fet directly on the bar is :

to align properly all fets while soldering them and keeping them in teh right direction to not have to bend the legs to insert them on the pcb holes.... I know that the legs of to-220 are pretty fragile and can crack easy if not bended progressively along the lengh

It seem a bit difficult to me to first solder them on the pcb to get them properly aligned in and in place in their holes and than hold it at 90 degree and try to press all 6 or 18 fets on the horizontal copper bar :?


I've read on the datasheet of the IRFB4110 that they recommand 300 celsius MAX for 10 sec... at 1.6mm forn the case

Now i wonder hot does it translate when we solder them directly on the copper bar.

Doc

 

[tostino]

That is a good question... I know I got some leaded 63/37 solder paste to try it with, so it should have a melting point of only 187c... Much below the 300c for 10 sec stated. What is the temp they will start to see damage though?

 

[texaspyro]

If you go with copper bars, then you can probably use one of the "liquid metal" thermal compounds. Basically they are low melting point alloys (many people think that they are gallium, but most are bismuth alloys). They will dissolve aluminum in seconds.

I just bought some very fine silicon carbide powder for making some thermal epoxy. It has 4 times better conductivity than aluminum oxide. Also a couple hundred carats of diamond powder (about $1/gram)

 

[oldpiper]

If you go with copper bars, then you can probably use one of the "liquid metal" thermal compounds. Basically they are low melting point alloys (many people think that they are gallium, but most are bismuth alloys). They will dissolve aluminum in seconds.

I hope they're not gallium - gallium melts at just a couple of deg above room temp. (30 C, 85 F). Won't hold anything together on a hot Arkansas (or Texas) afternoon, anyway. "Melts in your hand, or in your mouth."

Cameron

 

[Doctorbass]

If you go with copper bars, then you can probably use one of the "liquid metal" thermal compounds. Basically they are low melting point alloys (many people think that they are gallium, but most are bismuth alloys). They will dissolve aluminum in seconds.

I just bought some very fine silicon carbide powder for making some thermal epoxy. It has 4 times better conductivity than aluminum oxide. Also a couple hundred carats of diamond powder (about $1/gram)

Yes.. that's a good idea, some mechanicla engineer at my work already talked about making their own silicon or epoxy with their own filler like diamond or carbon or alumina powder.

Teh metal liquid you talk about seem like what Luke used already ( the solder-it aluminum paste)

I ordered 2 tube of 7g last hour for testing or maybe using..

Doc

 

[texaspyro]

I hope they're not gallium - gallium melts at just a couple of deg above room temp. (30 C, 85 F). Won't hold anything together on a hot Arkansas (or Texas) afternoon, anyway. "Melts in your hand, or in your mouth."

Yep, gallium melts in your mouth and in your hand... neat stuff

I think the idea for these liquid metal compounds is you want them to stay liquid at your operating temperature. That way you get rather intimate contact. This is one of the more popular brews: http://www.performance-pcs.com/catalog/index.php?main_page=product_info&products_id=4373 The most likely gallium based thermal alloy is http://en.wikipedia.org/wiki/Galinstan

 

[Doctorbass]

I hope they're not gallium - gallium melts at just a couple of deg above room temp. (30 C, 85 F). Won't hold anything together on a hot Arkansas (or Texas) afternoon, anyway. "Melts in your hand, or in your mouth."

Yep, gallium melts in your mouth and in your hand... neat stuff

I think the idea for these liquid metal compounds is you want them to stay liquid at your operating temperature. That way you get rather intimate contact. This is one of the more popular brews: http://www.performance-pcs.com/catalog/index.php?main_page=product_info&products_id=4373 The most likely gallium based thermal alloy is http://en.wikipedia.org/wiki/Galinstan

0.15ml for 14$

that's 93$ per ml !!!

 

[texaspyro]

0.15ml for 14$

that's 93$ per ml !!!

You can buy the raw materials for a LOT less and make your own. Or buy it for around $1 a gram at http://www.rotometals.com/product-p/lowmeltingpoint-2.htm

 

[toolman2]

maby its to late now but, this stuff http://www.jaycar.com.au/productView.asp?ID=NM2013&CATID=45&form=CAT&SUBCATID=752 is just a higher grade of heatsink grease that has silver and is therefore electrically conductive, prolly like arctic silver.

and this stuff http://www.grainger.com/Grainger/CHEMTRONICS-Silver-Conductive-Grease-1UYE2 seems to be aimed at contacts and connectors for maximum electrical conductivity, and secondly as a heatsink grease. -id use it if you dont want to solder.

 

[spinningmagnets]

Here's a review of 17 thermal pastes from the computer industry, a few are NON-electrically conductive, but perhaps useful when someone searches "Thermal paste" on ES:

site:endless-sphere.com best thermal paste

edit: Oooops! Here's the real link:

Thermal paste test, 17 brands
http://www.maximumpc.com/article/features/geek_tested_17_thermal_pastes_face

 

[amberwolf]

site:endless-sphere.com best thermal paste

Here's an actual URL you can stick in there:
http://www.google.com/search?q=site:endless-sphere.com+best+thermal+paste

 

[agniusm]

What about Coollaboratory Liquid Pro? Thermal conductivity:
Arctic Silver 5 - 8.7 W/(m·K)
Coollaboratory Liquid Pro - 82 W/(m·K)

 

[h0tr0d]

What about Coollaboratory Liquid Pro? Thermal conductivity:
Arctic Silver 5 - 8.7 W/(m·K)
Coollaboratory Liquid Pro - 82 W/(m·K)

Incompatible with aluminium, Beware!
Other than that, looks perfect, amazing, etc.