0

MPR121 sensitivy, crosstalk, and grounding
Moderators: adafruit_support_bill, adafruit

Please be positive and constructive with your questions and comments.

MPR121 sensitivy, crosstalk, and grounding

by ellindsey on Wed May 15, 2019 2:40 pm

I am working on a haldheld prop with LEDs in it which is a 3D printed cube about 4 inches wide. I am planning to use the MPR121 capacitive touch breakout board with twelve electrodes in the edges of the cube to detect how the cube is being held, sensing which edges the fingers of the person holding it are pressing on. The whole system will be controlled by a Trinket M0, which will connect to the MPR121, an accelerometer board, and a large number of Neopixel LEDs. It will be battery powered.

My first concern is about sensitivity. The cap touch electrodes will be buried under about 1/8" thickness of printed plastic outer shell. Each electrode will be about 2 inches by maybe 1/4 inch, as a strip of foil down the center of each edge. Will it be possible to get a satisfactory indication of whether someone's finger is holding the cube on that edge through that thickness of plastic? I know that it is possible to adjust the sensitivity of the board, which leads to my second question.

Because of the way the internal components in this prop need to be mounted, the wires from the MPR121 need to run in close proximity to each other for a few inches before they fan out to the individual sensing electrodes. Is this going to be a major problem? If I significantly increase the sensitivity, will I need to worry about a touch on one electrode causing a false sensing of touch on others because wires have to run next to each other? I'd rather not have to shell out for fine coaxial wires to connect the board to the electrodes, but I will if that's going to be required.

Also, how important is grounding? This is a hand-held, battery powered item, so there's no way to get a good external ground. Is that going to make reliable cap touch difficult? Will it help if I line the inside of each face of the cube with foil connected to the ground pad on the MPR121? I may do that anyway just to block stray light as the 3D printed plastic isn't nearly as opaque as I would like and there's a lot of leakage of stray light through it.

ellindsey
 
Posts: 2
Joined: Wed May 15, 2019 1:14 pm

Re: MPR121 sensitivy, crosstalk, and grounding

by adafruit_support_mike on Thu May 16, 2019 5:17 am

ellindsey wrote:My first concern is about sensitivity. The cap touch electrodes will be buried under about 1/8" thickness of printed plastic outer shell. Each electrode will be about 2 inches by maybe 1/4 inch, as a strip of foil down the center of each edge. Will it be possible to get a satisfactory indication of whether someone's finger is holding the cube on that edge through that thickness of plastic?

That should be okay. In general, a pad's sensitivity is related to its surface area and the amount of that surface that's covered when someone touches the pad. The sizes you've mentioned are reasonable.

You can test and refine your touchpad dimensions by putting strips under a piece of plastic and seeing how well they work under those conditions.

ellindsey wrote:Because of the way the internal components in this prop need to be mounted, the wires from the MPR121 need to run in close proximity to each other for a few inches before they fan out to the individual sensing electrodes. Is this going to be a major problem?

Nope.. the MPR121's designers thought of that. Only one pad is active at any time, and all the other pads are connected to GND.

Obviously you don't want to twist all the wires into a long bundle, but short lengths of wire running side by side will have a nearly constant mutual capacitance. The MPR121 will just include that as part of the pad's nominal capacitance during the calibration pass.

ellindsey wrote:Also, how important is grounding?

Nearly irrelevant.

The MPR121 measures the parasitic capacitance of the pads by sending a fixed amount of current into them, and measuring how long it takes for the voltage to rise by a certain amount. Human flesh is about 80x better than air as a dielectric, so putting your finger on or near a pad will have a noticable effect on the charge/discharge timing.

Connecting to Earth ground would create a whole new set of electrical problems without solving any.

ellindsey wrote:Will it help if I line the inside of each face of the cube with foil connected to the ground pad on the MPR121? I may do that anyway just to block stray light as the 3D printed plastic isn't nearly as opaque as I would like and there's a lot of leakage of stray light through it.

That will change the behavior, but it's a toss-up as to whether the changes would be good or bad.

Any grounded foil near a touchpad will create another parasitic capacitor. The electrical field from the touchpad will try to loop around and connect to the ground plane, creating what's known as a 'fringing field'. You generally want to keep the fringing field small because it pulls the electric field away from the volume of space your finger will occupy when it gets close to the touchpad.

You can use a grounded backplane beneath the touchpad if you need to isolate the pads from each other. The grounded plane will have much more effect on the fringing field than any crosstalk from another touchpad could do.

There is a 3-layer structure called a 'capaciflexor' that can be interesting though. The main idea is to put another, slightly larger touchpad between the main touchpad and the ground plane. If you connect the main touchpad to the one in the middle of the stack, the pad in the middle will generate almost all of the fringing field to the ground plane, forcing almost all of the outer pad's field to extend straight out from its surface. The overall effect is to focus and strengthen the part of the electric field that detects nearby fingers.

The capaciflexor is kind of an advanced technique, and you probably shouldn't need it for your design.

adafruit_support_mike
 
Posts: 57131
Joined: Thu Feb 11, 2010 2:51 pm

Re: MPR121 sensitivy, crosstalk, and grounding

by ellindsey on Thu May 16, 2019 9:17 am

Alright, sounds like I should be good with just the pads and no extra shielding copper. Thanks for the advice.

ellindsey
 
Posts: 2
Joined: Wed May 15, 2019 1:14 pm

Please be positive and constructive with your questions and comments.