Measuring rate of change of capacitance

Thread Starter

goutham1995

Joined Feb 18, 2018
90
I am working on a project which has a capacitive touch sensor. I want to measure the force with which I hit the sensor by finding the rate at which the capacitance changes. Is there any way to sample the capacitance value it reads every few microseconds and log it in an array? Is there any other way to find the rate of change of capacitance so that a harder hit indicates a steep slope compared to a softer hit?
 

nsaspook

Joined Aug 27, 2009
6,595
Yes. The pic CTMU module works by measuring the voltage across a capacitance as a measure of time while connected to a constant current source. With a fixed discharge, charge, hold time period, the voltage at the end of the timing period is a function of the total capacitance. For relative touch measurements you need to adjust the current/time equation to generate a linear voltage slope (in the ADC measurement range) with most values of the expected capacitance changes.


Some old demos of CTMU sensing.
https://www.flickr.com//photos/nsaspook/sets/72157628477730187/show/

For enhanced stability and range sensing you will need an active/mutual drive design to direct the E-field to the desired direction as shown inside the black box in the demo video.
http://ww1.microchip.com/downloads/en/appnotes/00001334b.pdf
 
Last edited:

BobaMosfet

Joined Jul 1, 2009
891
Most capacitive touch sensors are about distance, not intensity. It cannot tell a sharp hit from a subtle press. It just knows that you are affecting the field charge based on proximity. Not to say that you couldn't use the duration and location of a press to perhaps determine a strike .v. a press.

You could measure the voltage drop across the cap with an OpAmp. Now, go! Read, learn more!
 

nsaspook

Joined Aug 27, 2009
6,595
Most capacitive touch sensors are about distance, not intensity. It cannot tell a sharp hit from a subtle press. It just knows that you are affecting the field charge based on proximity. Not to say that you couldn't use the duration and location of a press to perhaps determine a strike .v. a press.

You could measure the voltage drop across the cap with an OpAmp. Now, go! Read, learn more!
Using the CTMU module I could easily set thresholds to increase the sample rate (16 capacitive touch channels takes approximately 12 μs on a PIC24F) to give reliable velocity rates for an average finger press. While you don't have a pressure reading of actual touch/press you can calculate the momentum of a uniform velocity and mass and then the force as the rate of change of momentum on the surface. That said, a inductive touch system is much better for actual touch force measurements.
 

MrAl

Joined Jun 17, 2014
6,806
Hi,

Rate of change measurements require at lease two samples over a known time interval.
If y1 and y2 are the samples at times t1 and t2 respectively, then the rate is:
v=dy/dt=(y2-y1)/(t2-t1)

That is still an estimate though.

The acceleration is found by doing two such measurements so if v1 and v2 are two such rates at times t3 and t4 then the acceleration is:
a=(v2-v1)/(t4-t3)

The jerk is found by doing two of those measurements so if a1 and a2 are two such measurements at times t5 and t6 then the jerk is:
j=(a2-a1)/(t6-t5)

This is probably the measure you want as the higher 'j' is the harder it must have been hit especially if there is significant mass involved.

There's one more with another unusual name, but you probably dont need that unless you have to impose a limit to the jerk.

Another requirement though is these measurements must be done all between the time from where the button first starts to move to just before it stops and the rate of movement must be smooth. Otherwise you'll get a jerky jerk :)
 
Last edited:
Top