Getting a large change from small change

[edmondng]

Need some help or suggestion here.
I'm trying to collect some data resulting from very small changes. What i want is for example a change from 40pF to 40.5pF to result in a bigger voltage change say 0.5V to 0.8V, enough for me to see some difference. In other words, a very big SNR.

The way i have my setup now is using a cap reference and comparing it with the variable, and going through a differential amplifier.

I was thinking in the sense of circuit manipulation where for example if you have 2 caps in series, where 1 is very very large like 1000pF and the variable cap is only 40pF, Then the resulting capacitance from 40pF to 40.5pF will result in only a very small change.

Same is true for resistance in parallel. Put a 52 Ohm in parallel with 1100 Ohm, but now change the 52 to 53 ohm and the resulting signal is still small. So is there some way for me to manipulate such that i can get a bigger signal change with my small input change.

Ideas, suggestion would be great. Thanks

 

[Emicro]

What is the reasoning for changing the hardware?

 

[berkeman]

Maybe you can use some sort of a bridge setup to get a bigger response to changes:

http://en.wikipedia.org/wiki/Wheatstone_bridge

I think you can do something similar to the Wheatstone resistive bridge, but use capacitors and AC excitation instead, but to be honest, I never really understood how bridges worked anyway, so I don't think I'll be much more help beyond pointing you at the subject matter. Good luck!

 

[mgb_phys]

Generally you are looking at some sort of balanced system
the classical example for resistance is a wheatstone bridge http://en.wikipedia.org/wiki/Wheatstone_bridge


edit - snap !

 

[berkeman]

Hey, great minds think alike!

 

[Bystander]

Oliver & Cage, Electronic Measurements and Instrumentation, McGraw-Hill, 1975, the AC bridge "bible;" also, you'll find tricks under application notes from various manufacturers of lock-in amplifiers.

 

[dlgoff]

Hey, great minds think alike!

Some just a little faster. Maybe we need to display posting time seconds.

 

[edmondng]

hmm i don't know how much of a bridge would help cause i already have something like that.
Cap ref in series with resistor, parallel with cap series resistor. Measure the voltage differential at resistor. If all values are same, difference is zero.

What I'm more concerned is the change. In other words changing the values from 40pF to 40.5 pF can give me enough resolution to see that change. I could change the ref cap to give me different slopes of change but it all still comes out about the same in terms of max change. Is there maybe some way to convert the linear relationship to some exponential relationship or other equation that can be related to it?