0

Im working with an EAP (electroactive polymer) thats capacitance changes when deformed. It varies between the ranges of around 0.5 - 3 nF. Im trying to connect it to an arduino to take readings that I can then do something useful with.

I managed to alter the CapacitanceMeter tutorial too to get things working a bit better, but it doesnt seem to be accurate enough as Im getting values that swing by around 500pF when un-stretched.

Im using a 10M Ohm resistor, and realise I could go to a G ohm but surely that will increase noise?

Patrick
  • 1
  • 1
    What is your code? –  May 30 '13 at 11:57
  • Are you timing the charge cycle and deducing capacitance knowing R = 1Mohm? – Andy aka May 30 '13 at 12:18
  • Try using your setup with some known (fixed) capacitors over the range you require and see if it is measuring their values correctly. – JIm Dearden May 30 '13 at 12:27
  • 1
    The R/C charge measurement technique will not give useful results as the resistor size gets so large that the current through it starts to get within an order of magnitude of the leakage and bias current of the MCU input pin. In these cases it may be more suitable to measure capacitance by using the oscillator technique and measure the resulting frequency. For oscillators that end up being higher frequency that an MCU can handle effectively you can use an external counter chip to divide down the oscillator by a factor of 16 or 256. Then measure the slower frequency with the MCU. – Michael Karas May 30 '13 at 15:25
  • See http://electronics.stackexchange.com/questions/19607/reading-very-low-capacitances-with-arduino?rq=1 – pjc50 May 30 '13 at 16:17

1 Answers1

2

In the capacitance meter tutorial it calculates the time taken to charge to 63.2% - it counts in milliseconds from what I can see.

500pF x 1Mohm is a CR time of 0.5 milli seconds - too low for the resolution

You need a substantially bigger resistor to get decent resolution

Andy aka
  • 434,556
  • 28
  • 351
  • 777
  • 2
    And if the resistor gets too big then the leakage current and bias currents of the MCU's input pin will take over and measurement will not be feasible using the R/C charge measurement technique. – Michael Karas May 30 '13 at 15:16
  • @MichaelKaras - quite possibly – Andy aka May 30 '13 at 16:49