[sdiy] Temperature Compensated Exponential Converter Using SSM2164

[David G. Dixon]

Oops!  Please disregard the following statement from my last message:

> It bears noting that the temperature compensation is not perfect, and
> deteriorates rapidly as control voltages go above 0V (i.e., the sum of
> square errors gets geometrically larger with increasing control voltages
> above ground).

I made the classic mistake of failing to non-dimensionalize my errors.  When
I divide the square error by the square of the desired value, I find that
the tempco gets progressively worse the further out the control voltage is
from 0V, on either side, as expected.  High-frequency compensation may still
be necessary in this circuit because of this fact.

Also, it bears noting that an algebraic solution of the equations gives the
following values:

Tempco voltage: 0.301V

Feedback resistance: 56.8k

These values give perfect conformance at 25 degrees C, but the conformance
deteriorates significantly at lower voltages, and much less so at higher
voltages.  The best fit values of 0.289V and 54.4k give better conformance
overall, and the deterioration is symmetrical about 0V, but there is a local
maximum conformance error at 25 degrees C!

In other words, this design is still a compromise.  However, it is probably
superior to using a tempco resistor, but I need to do more analysis to
determine this for sure.