[sdiy] Tempcos

[Ian Fritz]

At 05:09 AM 12/17/2004, Paul Perry wrote:

>I expect high is better, because it only takes two extra resistors to pad
>the
>coefficient down (one in series, one in parallel).

Usually you just need the one in series, as the V/Oct adjustment can take 
care of the increase in total resistance.

>  But there is no way
>to INCREASE the coefficient.

Oh yes there is!  :-)  <http://home.earthlink.net/~ijfritz/sy_circ.htm>

>Which, curiously enough, is very close to that of pure copper, IIRC.

Not so curious, really.  All pure metals have a linear temperature 
dependence at high temperature.  The only trick for a tempco is to get a 
linear-through-zero response.  This can easily be produced with a simple 
active circuit.  <http://home.earthlink.net/~ijfritz/sy_cir7.htm>

   Ian