Unstable touchRead

Status
Not open for further replies.
S

Somnitec

Member
Hey all,

I'm working on a musical instrument in which I'm using the Teensy 3.2's Touch pins. I'm having some problems though keeping it calibrated in electronically busy environments.

The Teensy pins are hooked up to 11 4x4cm squares of aluminium tape on the inside of a wooden bowl with about a 5mm wall thickness. The user then interacts with the device by touching or hovering over the bowl.
The Teensy is powered with a 12v switching mode going to a simple 4A DC step down module.
I'm using the DAC pin with a 10uF electrolytic capacitor to ground going to a 6.35mm jack output.

After calibration in an environment with few electronics, the instrument works fine, but when I move it to an electronically busy (next to speakers, amplifiers and other sound equipment) the calibration goes way off and isn't even stable after calibrating it.

I've noticed that when I touch the outer ring (ground) of the jack, that things seem to get stable, so I guess a grounding problem could be the culprit. Yet it is a bit impractical to run a separate wire to ground.

Is there a way to stabilize the touch pins on the Teensy, or should I be looking at other solutions?

Thanks!
 
Just to clarify, I made a drawing:

pnDSHrz.jpg
 
It may not just be the environment. Cap sense needs a connecting to ground, so your better readings may have a great deal to do with a good ground path from the user to the device, especially if the device is grounded through a PC case.

The best solution is to stake your user with a ground wire as you noted, but in the absence of a harpoon looking at the floor material and making sure the instrument is solidly grounded to the same potential as the floor (to control interference) may help.

Will still always be the case that if there is enough EMI to energise the user as an antenna cap sense will start acting as an RF receiver though.
 
Capacitive sensing can be disrupted by presence of fast switching signals. The output of a class D audio amplifier (like the one on the Prop Shield) or unshielded switching power supplies can really mess up the readings, if its wires run close to the sensing electrodes. They can also be a huge problem if people are able to touch them (capacitively couple to the fast signals) while also touching the sensing electrodes.
 
I ran a wire to ground, which seemed to ease things at least a little, though still not 100% stable behavior.

Looking at this text, I learn that a mutual capacitance setup might work better for my application. I don't seem to find anyone having implemented that in Arduino; something for on my todo list ;)

Thanks for the help!
 
Last edited:
Status
Not open for further replies.
Back
Top