advice about touchsensor setup

[AlainD]

Hi

I'm using the touch sensor capacity of the teensy LC and also plan to use an adapted version of fasttouchRead for extra pins.
I have some questions about pulling the touchpads to ground and a ground pad to ground. All sensors have a plastic foil over them, no direct bare touch. Teensy is embedded in a plastic enclosure.
The system is connected by USB keyboard emulation and USB cable to a computer, which can be a battery powered laptop.

A) FastTouchRead uses :

pinMode(pin, OUTPUT_OPENDRAIN);
digitalWrite(pin, LOW);

--> I suppose that this gives a short peak current which is way above 5mA. Given the low capacity of the pads this doesn't seems to given problems. Works without resistor and I also tested it with a 100 Ohm resistor in series directly on the Teensy.

B) When using touchRead() I use :

pinMode(pin, INPUT_PULLDOWN);
delayMicroseconds(5); // Test gave about 5T= 3µsec for not touched sensor and about double for a touched one (2T is about 13% V)

Slower, but still very fast for a HMI device. (I want the not sensing pad to be pulled to ground instead of floating.)

C) I have also a ground "pad/surface" connected directly to ground.

My questions :

1) Is it safe -for the Teensy- to keep using the output open drain method without a resistor in series to limit the peak current? If not any suggestion about the size, 660Ohm seems a bit over secure?

2) after the initial pull down, which is safest : A) or B) and would adding a small resistor in series give significant extra protection?

3) Should I add a resistor in C) aka between the ground pad/surface and the Teensy ground pin?

[AlainD]

Has someone expirience?
I know not enough of the hardware side of things.

I founds some references to adding a 50 Ohm resistor as protection for touchpads...

[AlainD]

Unfortunately no answers. Is this so exotic?

[drjohn]

Anecdata: I have done a few touch keyboard interfaces (PCB with solder mask, no resistors) and I have never had a pin go wonky from static discharge or having to source "high currents" to bring the pin up to the reading state. I'm sure one could design a case study for any micro involving kids, scuffed feet, and wool carpet in a 0% humidity environment (or Tesla coil adjacent) which would require more ruggedization. One could also design any PCB with a spot for a resistor, then place 1 ohm resistor or a wire if it turned out not to be needed.

I used this library; it is hardware generic, doesn't specify external resistors, and responds fast enough that I don't have to faff with any board-specific libraries if I change the underlying micro: https://github.com/adrianfreed/FastTouch