Adding my 2 cents worth in as well.  I've personally had good luck with a
product from Hysol (P/N: ES 4412).  Have used it for YEARS and have not noticed
any problems (yet).  Sets up hard enough that you don't need a enclosure.  We
pour into a form and pop it out afterwards.  Used in the automotive industry.
There is a certain amount of infant death associated with it (probabally stress
due to the curing process) so we always burn in the units but few die after the
first week.

Dean Biddle wrote:

> Hi,
>
> I need to encase an assembled PIC device in a polymer block.  For example,
> place the tested PCB+components in a mould and add a polymer so it sets
> around the board and components to make a waterproof case.  I also need to
> encase a sensor which needs only be used up to 100 degrees Celsius.
>
> I am concerned that the polymer, whether it be epoxy, neutral cure silicone
> or some other compound may change in resistance with time (the device uses
> 15 bits for each measurement, so small changes count).  This change could be
> due to water absorption or structural reorganisation.
>
> Please share your successes and failures.  One problem with epoxy is the
> exothermic polymerisation, however, the setting temperature should not
> exceed 100 degrees Celsius.