Thermal Conducting "epoxy" for medium power pcb

[Jon Slaughter]

Is there some type of non-electrical epoxy that can encase a pcb to remove
heat? (something like a large heatsink(although not necessarily as good)
that you can just form around a pcb?

The reason I ask is that I've seen some circuits encased in this epoxy like
block(all you see is some connectors) and I'm trying to figure out if its
for heat reasons or they just don't want people to get in it. (This thing is
running around 20 amps max so the heat has to go somewhere)

Thanks,
Jon

 

[Jim Yanik]

Is there some type of non-electrical epoxy that can encase a pcb to
remove heat? (something like a large heatsink(although not necessarily
as good) that you can just form around a pcb?

The reason I ask is that I've seen some circuits encased in this epoxy
like block(all you see is some connectors) and I'm trying to figure
out if its for heat reasons or they just don't want people to get in
it. (This thing is running around 20 amps max so the heat has to go
somewhere)

Thanks,
Jon

encapsulation is so you can't get at it.

 

[Jon Slaughter]

encapsulation is so you can't get at it.

So your saying that the compound has in no way anything to do with heat
dissipation?

 

[Jon Slaughter]

Thanks, I'll try to find some of these products and see what they do.

 

[Jon Slaughter]

You could brute-force it and dunk the whole thing in thermal epoxy, but
what
a mess! 3M makes several types. I've used their TC-2810, but it's
expensive.

Copper is over 400x more thermally conductive than this stuff ;/ But maybe
if its a large enough block than it wouldn't be too bad (maybe use a copper
or aluminum heat sink coated in a block of this stuff? (since it is still
much better than air and will help dissipate the heat from the heat sink
into a larger volumn quicker)

 

[gearhead]

Copper is over 400x more thermally conductive than this stuff ;/ But maybe
if its a large enough block than it wouldn't be too bad (maybe use a copper
or aluminum heat sink coated in a block of this stuff? (since it is still
much better than air and will help dissipate the heat from the heat sink
into a larger volumn quicker)

Encapsulation conducts heat "better than air" if you don't take
convection into account.

I worry about encapsulant holding the heat in, myself. A datasheet
may give me an idea how much power a particular package can dissipate
in air, but what happens with that package encased in solid glop?
And make sure you don't get the device trapped inside an air bubble,
then you have dead air acting as an insulator and that would kill you.

 

[Jim Yanik]

So your saying that the compound has in no way anything to do with heat
dissipation?

it probably hinders it more than it helps.

 

[Phil Hobbs]

That makes no sense... if it has more thermal conductivity than air it must
be better than just air alone.

No, because the air moves. Transporting hot material is the most
effective way of moving heat. It's air or water usually, or sometimes
liquid sodium as in nuclear reactors and the exhaust valves of racing
engines.

The thermal conductivity of your average plastic is around 0.1 W/m/K,
which is only 4 times higher than still air. If you have half an inch
of epoxy all around the module, that's more or less equivalent to an
eigth of an inch of Styrofoam. That's probably not going to improve
matters.


Cheers,

Phil Hobbs

 

[Jon Slaughter]

it probably hinders it more than it helps.

That makes no sense... if it has more thermal conductivity than air it must
be better than just air alone.

 

[Jon Slaughter]

Copper is over 400x more thermally conductive than this stuff ;/ But maybe
if its a large enough block than it wouldn't be too bad (maybe use a
copper
or aluminum heat sink coated in a block of this stuff? (since it is still
much better than air and will help dissipate the heat from the heat sink
into a larger volumn quicker)

Encapsulation conducts heat "better than air" if you don't take
convection into account.
--

But even if it does it shouldn't matter too much because the block is still
exposed to that convection. So the air will take heat away from the
material. Now sure it matters to some degree and I might be wrong to just
how much. Maybe a fan would be better.

I guess one needs to know how much convection increases the thermal
conductivity.


---
I worry about encapsulant holding the heat in, myself. A datasheet
may give me an idea how much power a particular package can dissipate
in air, but what happens with that package encased in solid glop?
And make sure you don't get the device trapped inside an air bubble,
then you have dead air acting as an insulator and that would kill you.
---

heh yeah. I was thinking about that to some degree too. The "glop" might not
form a tight seal with the components if it shrinks on solidifying(although
maybe that isn't likely).

Would be nice if there was some scientific proof what works and what
doesn't. Maybe a fan is just as good(and probably a hell of a lot cheaper.
The nice thing about the epoxy is that it protects from other things too.
Cement isn't that bad of a thermal conductor either (about 10x that of air).

My main point with the convection of air is that its there no matter what
you do... except of course if you have to use a fan, then the question is
which is better. Convection + "glop" is better than either alone.

 

[gearhead]

Jon,

The power will come out regardless of the type of heatsinking (10 watts in
means 10 watts out). The issue is what the component temperatures will be.

You could brute-force it and dunk the whole thing in thermal epoxy, but what
a mess! 3M makes several types. I've used their TC-2810, but it's expensive.

The simplest thing to do is to identify those components that will be
getting too hot and treat them individually. If you don't know how to do
this then you won't be sure if any technique will solve your temperature
problems (if you even have any).

Bob

He could use the circuit board itself as a (relatively lousy)
heatsink, like using d2pak devices with the tab soldered down. FR4
conducts sideways better than straight through, but with vias he could
transfer heat effectively to a copper pour on the other side of the
board or internal layers.

The olde seek and ye shall find turned up a couple of links about heat
management.

http://www.electronics-cooling.com/articles/2004/2004_nov_a3.php

http://www.mininglife.com/Miner/general/thermalconductivity.htm

 

[Nemo]

Jon Slaughter writes

Is there some type of non-electrical epoxy that can encase a pcb to remove
heat? (something like a large heatsink(although not necessarily as good)
that you can just form around a pcb?

The reason I ask is that I've seen some circuits encased in this epoxy like
block(all you see is some connectors) and I'm trying to figure out if its
for heat reasons or they just don't want people to get in it. (This thing is
running around 20 amps max so the heat has to go somewhere)

It's generally for complying with safety regulations, usually because
they're hot, or high voltage. For example you may use encapsulation in
products intended for use in potentially flammable atmospheres (in
petrol stations, down mines etc). However although you can choose the
encapsulant to be a relatively good conductor of heat (last time I
checked, silicone-based rubbery ones were best), you're still putting a
heat barrier between the components and the dissipative medium (air) so
although the heat eventually gets out, it does so more slowly and the
components under it get a bit hotter. When forced to resort to designing
encapsulated circuits, you generally end up building a sacrificial PCB,
encapsulating a thermocouple probe next to the hot components, and
running the thing on max load for a couple of hours to check you ain't
going to overheat the devices under there.

Oh and you want the encapsulant to be UL recognised / flame retardant.

I've also seen encapsulation used in satellite electronics, to protect
against the vibration and g-forces when launched. They probably use it
in military stuff for the same reasons, but I don't know about that
field.