physical contact sensing alternatives

Fred Hawkins

What different alternatives are there to sense a physical contact with an hollow non-conductive cylinder? Think soda bottle, pill bottle, or a bit of hosing.

I would like forgo putting a micro switch that would interrupt the surface.
I want to sense a person or any part of person, with or without clothing or gloves.
I want a cheap solution, without much fuss or circuitry. If I can whip the prop into doing the work of another chip (adc, capacitive sensor chip), I would be happiest.
I'd like it to be simple enough to replicate 3 or 4 times and still leave·lots of·propeller/cogs/pins.

I·haven't found a capacitive solution, though I'm sure it exists.

I have looked at using a·LED in reverse to·notice a·touch, and may do that if it isn't·particularly sensitive to·ambient light. But it does disturb the surface and·may·not be·able to withstand misuse.

Non conductive material is a poly plastic. Tape solutions applied to the surface·are ok, if I can find one.

Maximum frequency looks under 50 hertz. Duration may be as little as 1/10 of second, but on the order of a couple pounds force.

Can in place; kick it. (Thanks)

Ken Peterson

Fred: I put a capacitive sensor object on the OBEX. You could try to play with that. Just cover the inside (or outside) of your cylinder with foil or foil tape and experiment. Needs only one pin and a resistor. Works by sensing the change in RC time delay. Automatically calibrates itself and senses a sharp increase in C.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔


Post Edited (Ken Peterson) : 5/31/2008 3:02:23 AM GMT

Mike Huselton

Think Ultrasonics...

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
JMH

Phil Pilgrim (PhiPi)

Parallax carries the QT113-D Touch Sensor. It might do what you want.

-Phil

Beau Schwabe

Fred Hawkins,

Along similar line as Ultrasonic, you might be able to do something with the resonate frequency of your container. More description of your container might be in order, but the basic idea is this.... Use two piezo transducers, one as a microphone, and the other as a speaker and sweep the frequency to the piezo speaker. Monitor the "sound signature" across the swept frequencies on the piezo microphone. Any difference or change within the container either by adding something, taking away something, or even nearby changes can affect the characteristic "sound signature"... If there is a significant change between the sound signatures from one sample to the next, you can use that information to determine your actions.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beau Schwabe

IC Layout Engineer
Parallax, Inc.

Fred Hawkins

Ken Peterson said...
Fred: I put a capacitive sensor object on the OBEX. You could try to play with that. Just cover the inside (or outside) of your cylinder with foil or foil tape and experiment. Needs only one pin and a resistor. Works by sensing the change in RC time delay. Automatically calibrates itself and senses a sharp increase in C.

Why I asked. Thanks Ken. I'll give it a try.

Phil, I like that chip but want to save my bucks for a qfn propeller -- I want to make a prop system board that's narrower than 3/4 inch and·as cloase to that long as well. Plus I'm thinking about a mems 3axis·chip already and and and and...

What recession?

Beau: I'd say, but people would just make fun of me. Or worse, not. I don't think ultrasonic would work, as once it's picked up, it's picked up and stays there. I just want to know when its picked up in the right places.


Post Edited (Fred Hawkins) : 5/31/2008 4:17:12 AM GMT

Fred Hawkins

Smoke tested Kens routine without the prop_tv.

Got it working, sort of. But after I tried it with a 6 foot wire, it became noisy. The indicator LED would fire often as long the prop plug was connected. Taking the plug off, it would light up about every 10 seconds. And I had to set the threshold much higher to $77 or so.

I think I know now why the internet has lots of moaning about how temperamental this is. High fiddle factor, methinks.

Fred Hawkins

Food for thought, two pdf's from TI that have good basic information about capacitive touch sensing.

PCB-Based Capacitive Touch Sensing With MSP430

http://focus.ti.com/lit/an/slaa363a/slaa363a.pdf

MSP430 Capacitive Single-Touch Sensor Design Guide

http://focus.ti.com/lit/an/slaa379/slaa379.pdf

Fred Hawkins

Ken, TI's material has the resistor connected to ground:

Sensor Design Guide said...
The capacitive sensor forms the C part of the RC circuit. A single I/O line is used to charge, discharge, and generate an interrupt when the voltage of the capacitor crosses a threshold. The sequence of events is as follows:
1. One side of the resistor is connected to the sensor, the other side to ground. The sensor is connected to the I/O line.
2. The I/O port is set to output high. This charges the capacitive senor to near VCC very quickly....

That's opposite from what your spin program uses. What advantage, if any, comes from tying the resistor to vcc?

Fred

Ken Peterson

@Fred: I don't know if it makes a difference. You can tie the resistor to ground, charge the cap with a high and then wait for it to go low. Mine was doing the opposite - tie the resistor to Vcc, discharge the cap by pulling low, then wait for it to go high. In either case you charge the cap with the pin and then wait for the resistor to discharge / charge the cap to a different voltage across the logic threshold. Since the phenominon is primarily AC, I don't think it matters which DC bias you use.

If you want to run it 6' away from the Prop, have you tried connecting with coax? The shielding might help.

Ken

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔

Fred Hawkins

Ken, Yes, that's where my thinking ended up too. That, and that 'high fiddle factor' goes hand and hand with cheap solutions.

My next pass will be an optimized pad -- a 1/2 inch copper disk surrounded by a ground field. If I read Quantum's Touch Secrets (ap note AN-KDO2) rightly, the field is generated at the margins heading for the ground. TI suggests using a 5M resistor which may be pass 3.

Erik Friesen

I am currently using the quantum touch 60248. It a a matrix style interface for 24 keys. It is around $6 from digikey. If you are disposed I recommend these chips as they come with good documentation etc. They are a slam dunk in my opinion. They will work with no calibration right out of the box, if you have a decent xy electrode shape. They work even better with a little adjustment which is not difficult. Some how they have taken the fiddle factor out of their chips in my opinion. I have code if you want it but it is not hard to interface to their spi protocol. I have wondered however if the prop could emulate their style on a few keys. They send a pulse out on the y and then read the effect on the x.(or vice versa). Typical key touches on my setup from a baseline of 400 would yield a 15-20 difference.

Fred Hawkins

re: capacitive touch circuits and chips. The June 2008 issue of Circuit Cellar has two articles that consider capacitive touch sensors. One surveys the differences between touch sensors from Quantum Research, Silicon Labs and Cypress Semicondutors.

This pdf may be useful: http://www.silabs.com/public/documents/tpub_doc/anote/Microcontrollers/Small_Form_Factor/en/an338.pdf

Post Edited (Fred Hawkins) : 6/7/2008 7:23:04 PM GMT