Homebrew PCBs

--- In Homebrew_PCBs@yahoogroups.com, mpdickens <md30022@y...> wrote:
> 
> --- Ron Amundson <ron_amundson@h...> wrote:
> > 
> > >It's been my experience that epoxy is your friend
> > >against harsh environments and climates...
> > 
> > I've used epoxy too, and have filled potting boxes
> > with it. However it can
> > raise havoc with heat producing parts, as well as
> > creating additional
> > thermal and conductive paths. However if the
> > economic conditions allow it,
> > its quite workable except when it comes to service.
> 
> Heat was never an issue when the box is in the water.
> The entire box acted as a heat sink. Hoever, we also
> did it like this when devices had heat issues and the
> unit was not submersed in water:
> 
>      --------------------------------
>      |epoxy potting                 |
>      |  ------------           ------- 
>      |  |           ----------| heat |
>      |  | PCB       |componet | sink |
>      |  |           ----------|      |
>      |  ------------           -------
>      |                              |
>      --------------------------------
> 
> The heat sink protrudes from the expoxy and dissapates
> the heat.

The epoxy itself can dissipate the heat as long as you design enough
surface area between the heat sink and the epoxy, and from the epoxy
to the water. Water is a great conductor of heat and seawater is
pretty cold.

Long ago I fixed those overheating C64 power supplies by sawing off
the plastic top and epoxying an aluminum heat sink right on top of the
epoxy encapsulating everything. I knew it worked well because the base
of the heat sink got warm within 20 seconds of turning it on but never
got any warmer. In that case things like conductivity made no difference.

> > On key thing with some epoxies is that they can
> > cause stress related
> > failures, eg op amp Vos drift is exagerated with
> > temp, and ferrite beads
> > really hate the stuff unless you balance their
> > coefficient of expansion with
> > the epoxy.
> 
> Extruding the heat sink as diagramed above will fix
> the op amp problem. Regarding ferrite beads, This is
> strange to me. I've hever had an application where
> they conducted heat and I cannot think of one. 

He meant the different rates of expansion as the entire device changes
temp. This will not be much of a problem with something in seawater,
should be at a pretty constant temp as long as the power producing
components are designed to give up their heat pretty efficiently to
the seawater.

> > The underwater stuff sounds really cool. What did
> > you have to do for cases,
> > as I imagine the pressure loads were really high.
> 
> The expoxy potting *was* the case. In fact, it was
> perfect: Cheap, reliable and never leaked and good to
> 1000 Feet. Woods Hole Oceanographic Institute still
> does it this way. Pressure was never an issue because
> epoxy contracts very little under high pressure.  

My favorite kind of project box.

For my side branch of the thread on low leakage, I may just pot the
high impedance part of the circuit if I can find an epoxy encapsulant
that has very little conductance.

Steve