Arctic Silver Epoxy for F5 mosfets?

[khundude]

I had inended to use Arctic Silver to bond the fets directly to the heatsink. Apparently the silver stuff is capacitive. They have a ceramic one as well, but they don't give any data on its performance.

1. Will I have any issue with the silver Arctic Silver if the fets are two inches apart on the same sink?
2. How does the ceramic perform against the silver?

[Nelson Pass]

It might be conductive.

[marchel]

The Artic silver is designed for CPU use , Which dont need insulating pad, As Nelson Pass said, It might be conductive. One way of making it more efficient.

[khundude]

Originally Posted by Nelson Pass

It might be conductive.

They say it is electrically isolated but capacitive? This confuses me. They don't say how the ceramic performs or any of it's thermal properties. The ceramic isolates completely.

[sangram]

It's capacitive. In amps where the drain connected directly to the supply rail it would not be a problem, but I suspect it would do funky things in the F5 as effectively it would shunt the drain resistors to ground (or wherever the HS was connected). The magnitude would be very small (a few pF AFAIK), but a simulation would help.

AS Epoxy is not removable, and is not advised for any serviceable installation (it's mostly used on graphics cards, which are use and throw). The regular AS is a better bet if you ever think of removing/replacing the MOSFETs. In practical tests, Ceramique performed just as well as Silver, except with very high - >100W - heat loads.

[Nevin House]

The larger problem than the electrical characteristics of the epoxy or grease is that both are in a fluid state as the fets are mounted. Our greases are designed for minimum bond lines which in the best case will have the thermal material only filling the voids between the thermal surfaces while still allowing physical contact between the heat sink and the heat source.

The epoxies are made with somewhat larger conductive fillers than our greases as a bit more than an absolute minimum bond line is needed for best adhesion. For example, while the D50 of Ceramique thermal compound is approximately 0.4um, the D50 of our epoxies are in the 5-10um range. Even with the larger filler particles however, there is still a chance that actual physical contact will exist between the two thermal surfaces.

Nevin House
Arctic Silver, Inc.

[Nelson Pass]

Thanks Nevin. Let me ask you, since I was speculating -
any chance that the silver in it is electrically conductive?

[Diomedian]

I've had a silver-content silver paste become conductive. It took about a month, but when it did I smoked a few big MOSFETs. Well, obviously it was conductive all along, but what took a month was to create a conductive bridge from the metal back of the MOSFET to the bolt that was holding it in place. The bolt conducted into the heatsink which was being used as chassis ground.

-j

[Nevin House]

The silver particles themselves would be electrically conductive, however they are encapsulated by the suspension fluid or epoxy to where a bridge of any significant length is avoided. (Hence the capacitance, electrically conductive particles separated from each other by a non-electrically conductive oil or epoxy.)

In a grease, it is more likely that vibration and flow of the viscous material eventually allows parts that need to be separated to come into actual physical contact with each other.

Once the epoxies are set, you don't have to worry about things moving. If the parts are separated, they will remain separated.

[frags]

Use MX-2 or MX-3. They have better performance and are not conductive/capacitive.