[sdiy] drift followup

[Ian Fritz]

Hi gang --

Last week a question came up concerning temperature drift in a 
current-switching tri VCO using an unsaturated current switch, e.g. a 
CA3080 driven at 60 mV, such as in my old ultrasonic VCO (Electronotes # 112).

I've now worked out the math to calculate the (tuning) drift of an OTA with 
finite voltage drive.  The calculation only takes a couple of lines if you 
work with (yuk) hyperbolic functions.  Here is the result:

Define "a" as half the ratio of the drive voltage to the thermal voltage:

a = 0.5 Vin / Vth,  where Vth = 26 mV,

and define T0 as room temperature (300 K).

Then the drift D is given by

D = -a (sech a) (csch a)  / T0

It is easy to verify that for small "a",  D reduces to  -1/T0, i.e. the 
familiar result of -3550 ppm/K.

For a 60 mV drive I find a = 1.15 and D = -1550 ppm/K.  This is consistent 
with the fairly large amount of tuning drift I found in the ultrasonic VCO, 
which was compensated as described in a previous post.  Be sure to note 
that this is a tuning drift -- the scale factor is not affected.

   Ian