Reducing fringe effect, measuring values

[edmondng]

Hi,

I'm trying to measure some low capacitance using parallel plates i built. We're probably looking into the pF, say maybe 20pF (with the pcb board etc). The change in capacitance that i am really interested is very small. For the moment, i have use water as my dielectric. The capacitance that i am measuring is measured in terms of voltage by using a control capacitor, then goes through my 16bit A/D. The differential voltage from 0ml to 250ml is approx 20mV. This is a really small value.

One thing i did notice is that there seems to be some sort of fringing effect field surrounding the plates. Put you hand within 1 feet of the plates and the voltage swings away. Stay 3 feet away and its more stable. The stability increases with distance, so the nearer you are the more the voltage swing. I take it because the capacitance and e-field is being disturbed, small capacitance change in the pF results in my voltage change.

I was thinking of using some kind of guard material to isolate my test area. What sort of material should i use? Plastic doesn't seem to have any effect. Also since i am looking at small values, is it possible to see a bigger change or amplify my signal? Its 16bit A/D, amplifying the voltage would not help. What i need is a bigger change like from 0ml to 250ml with maybe 1V change rather than 20mV. Any help, suggestion, books to look at would be appreciated.

Thanks

 

[berkeman]

If the value changes when your hand is 1' away, that sounds more like AC mains interference into your test setup. And yes, grounded metal shielding would be good to have around your setup where you are measuring small values. Make a metal cylinder around the whole setup and ground it, keeping it spaced away from your capacitor by several multiples of the maximum plate spacing. The fringe field will not extend out more than a couple times the plate spacing, I wouldn't think.

 

[dlgoff]

"The differential voltage from 0ml to 250ml is approx 20mV."
"What i need is a bigger change like from 0ml to 250ml with maybe 1V change rather than 20mV."

Try a different dielectric material. Maybe some sort of oil. You could check different things?

 

[edmondng]

i tried with factory capacitors and based on the values i use, i plotted a graph of voltage vs capacitance. seems that at a certain region the rate of voltage change is more and then it saturates. Since my plate resulted in only 20mV change the most and the voltage fell somewhere between the region where i could get a high rate of change (using factory caps), i think it has something to do with my whole setup measurement. I must be losing my field elsewhere. Not to mention calculated values were nowhere close.

I'm going to redo my whole setup, enclose everything, put a guard material to gnd around it as well. Any other things i should look at?

Thanks

 

[berkeman]

Make sure your dielectric is non-conductive. What kind of "water" did you use?

 

[edmondng]

i tried deionized and just regular tap water. would adding salt make it more conductive be better? i thought adding conductive would increase the capacitance so if i don't see a lot of 'action' could try it

 

[berkeman]

No, do not make it more conductive. I'm also not real comfortable with water as a dielectric anyway. Can you research oil-filled capacitors, and try it with an appropriate oil? At least I've heard of oil-filled caps.