🚀 go-pugleaf

Hello all,
	I am involved with a project that has at it's centerpiece a
very small device with a changing capacitance that we would like to
measure. The chief problems have been signal noise and other unknown
capacitances. I was wondering if anyone had any suggestions or further
ideas for improvement.

	Currently, our variable capacitance is about 1 pF, and it will
change about +/- 10% (what we wish to measure). It will have some
unknown input impedance added from input switching
(multiplexing/demultiplexing) network which should be around 10 pF
(and hopefully will be constant for pass transistor style muxing). We
are currently trying to measure it using something like a 100 KHz sine
wave input into the Capacitance and known resistance in series to
ground, and measuring the voltage across the resistor. But we would
like further sensitivity and less 60 Hz noise.

	We have two ideas:

1) Use a current-to-voltage converter (see below). Where Vout=R*I.
Here we are concerned about the output offset incurred and the limited
gain thereof.
	        +-/\/\/\-+
	        |     R  |
	        |  |\    |
	I->   --+--|-\   |
	           |  >--+-- Vout
	       /---|+/
	       |   |/
	      ===
	 GND   =

Advantages: No problems with small input currents for good amp. Input
held at virtual ground.

Disadvantages: Big amplification of 60 Hz noise. Clipping. Large
offset voltages.

2) We were also considering a bridge configuration, but we have
unknowns in our switching system which greatly outweigh our variable
unknown. Chiefly, a bridge configuration really sounds good only for
its offset nulling, but this might be worthless since the variable
might be swamped by unknowns.


Is there a better way to do all this (measuring multiple
capacitances)? Also, one last idea we've got is to try to make the
input a tuned LCR series circuit. It would be similar to what we're
using now (just CR seiers circuit), but we might also be able to
eliminate 60Hz noise. Any experience/wisdom would be much appreciated.
Thanks!

shishin "squish" yamada
squish@endor.harvard.edu
division of applied sciences
harvard university