casting resin question

[dmh]

I was wondering about the properties of casting resin when used to encase electronic componets. Has anyone tried this or looked into its thermal qualities? I know that some plastics like teflon have good heat transfer and can withstand high temps. but what about ployester casting resin? would it be ok to encase the capacitors? Basicly what i was wanting to do is encase a gainclone amplifier board in a solid block of resiin.

[SY]

Usually, casting resins are polymethylmethacrylate (PMMA) or some version of polyester. They can be used for encasing a circuit, but they're not optimal- they're not terribly pure, they've got poor electrical properties, and they can't transfer heat very well.

You really should use electronic-grade resins- unfortunately more expensive- and make sure that you don't encase any components that need to dissipate heat.

Electronic-grade resins are generally epoxies, silicones, and (occasionally) polyurethanes.

[tiroth]

I think Bondo might actually be decent in terms of heat transfer, at least comparable to epoxies.

http://www.diyaudio.com/forums/showt...ht=#post402302

I don't know what it's dielectric properties are though, since I was only looking at it for potting transformers.

[jeff mai]

Originally posted by dmh
I was wondering about the properties of casting resin when used to encase electronic componets.

RS Components here in Australia sell potting compounds - both polyurethane and epoxy types. They are the "mix in the bag" types and some come in different sizes so there is no waste. There are several different types with differing properties.

I wouldn't use just any old plastic. You could have problems with exotherm while it's curing or it may absorb moisture. You need something rated for electrical use.

Digikey doesn't seem to sell this stuff, but I'm sure someone in the US does. Maybe Newark?

[jneutron]

Originally posted by dmh
I was wondering about the properties of casting resin when used to encase electronic componets. Has anyone tried this or looked into its thermal qualities? I know that some plastics like teflon have good heat transfer and can withstand high temps. but what about ployester casting resin? would it be ok to encase the capacitors? Basicly what i was wanting to do is encase a gainclone amplifier board in a solid block of resiin.

There are several things to consider.

1. Thermal transfer. Many of the components are cooled by the leads. They transfer heat along the lead, to the pc pads, which then transfer heat via lateral conduction, or air convection.. Diodes and resistors typically do this. If you cast around the pc board, you can cause the assembly to run quite a bit hotter.

The epoxy I am the most familiar with is Emerson and Cuming 2850 and 2851. Both are alumina filled, and are among the highest conductivity epoxies around. 2850 is a 2 parter, with a choice of hardeners...24LV, 11, and 9. 24LV is room temp cure, but has a T_g of about 45 C, so is not very good for high temp operation...it is excellent, however, for cryogenic operation. 11 and 9 are higher T_g systems, but require oven curing. 2851 is a one part, thermal cure system, with a T_g of 117 ^C. This is the standard choice for encapsulated bridges, single and 3 phase..

2. Thermal expansion coefficient. Many unfilled epoxy systems expand and contract far more than pc boards, components, and metals...as the circuit heats up, the epoxies can rip the components to shreds..This can happen also, when the unit is cooled down from initial cure. the 2850 and 2851 systems expand at 29 to 35 ppm/ ^C, whereas lots of epoxies have TCE's of 80 to 120 ppm. All copper alloys are about 16, aluminum is 25. Fiberglass is 16, steels are also..the filled epoxies, like the 2850's I mention, also will start to expand at about 90 ppm once the T_g has been exceeded.

To counter the forces of expansion/contraction, the components can be coated with an RTV (be aware of acetic acid outgassing), or silastics made for the job. This keeps the epoxy away from the delicate parts..it unfortunately, also provides a thermal barrier.

Encapsulation will also trap flux and contaminants...meaning, the board had better be very very clean.

Teflon is not a good thermal transporter, it is more a thermal barrier.

Bottom line?. If you have to cast for shock or acceleration specifications (home audio??), then encapsulation is certainly attractive..but there is a bit of homework to do for the thermal aspects.

Cheers, John

[dhenryp]

McMaster Carr sells several electrical potting compounds (along with everything else in the world):

http://www.mcmaster.com/

[dmh]

Hmmm.. a lot to consider. Thanks for the info. I was just thinking the esthetics of a solid block of clear resin with the components (maybe point to point construction) suspended in the middle would be very nice. Could I address the shrinkage/expansion issues with a slower cure mixture? What about including a heat sink that extends outside the block?

[jneutron]

Originally posted by dmh
Hmmm.. a lot to consider. Thanks for the info. I was just thinking the esthetics of a solid block of clear resin with the components (maybe point to point construction) suspended in the middle would be very nice.

It would look very cool indeed.

Originally posted by dmh
Could I address the shrinkage/expansion issues with a slower cure mixture?

No, that would only address the exotherm one..the TCE is not necessarily tied to cure rate. All you can do is try casting some dummy assemblies, and see if they survive..clear, it'd be easy to tell..

Originally posted by dmh
What about including a heat sink that extends outside the block?

That will help, but the heat also has to get to the sink. This means making sure all components are somehow tied thermally to it..
Given a milling machine, I'd hog a cavity out of a sink, and pot into the cavity..

Cheers, John

[tiroth]

Why not just place the IC on one side of the block? Your floating block of gainclone just attaches directly to the sink then, and there is no thermal penalty.