Schmidt trigger as capacitive sensor

[boylesg]

After advice that this circuit will be too sensitive to environmental inputs I have been playing around with it a little.

I found that twining together, even loosely, to two input wires of two of these circuits, I could weakly activate one Schmidt trigger by touching the input wire of the other one - clearly not desirable.

If I untwine the input wires then this effect disappears or at least become undetectable with an LED indicator.

However I found that if I bunch the input wires together, wrap them with Al foil tape and then connect the tape to ground, then the effect also disappears. If I remove the ground connection then the effect of activating one schmidt trigger with the input wire of the other becomes stronger.

So I may have discovered a way to eliminate significant environmental effects on this circuit when using it as a capacitive sensor.
**broken link removed**

 

[FvM]

So I may have discovered a way to eliminate significant environmental effects on this circuit when using it as a capacitive sensor.

I still think that this kind of stray voltage detector doesn't deserve the designation "capacitive sensor". By the floating input, it's behaviour strongly depends on the used OP/comparator type.

"Enviromental effects" are also battery versus AC power supply of the circuit and presence of 50 Hz power cabling in the vicinity. Did you e.g. try to operate the circuit battery powered in your garden? (Provided you are not standing under an overhead power cable)

 

[boylesg]

I still think that this kind of stray voltage detector doesn't deserve the designation "capacitive sensor"

Well perhaps not given that these generally detect a change in frequency rather than a change in voltage.

But it never the less seems to serve my purpose and is a great deal simpler to solder up.

By the floating input, it's behaviour strongly depends on the used OP/comparator type.

I suppose I could tie the input lines to GND via a 1M resistor or something. Then they wouldn't be floating. Is that likely to result in any significant improvement?

Enviromental effects" are also battery versus AC power supply of the circuit and presence of 50 Hz power cabling in the vicinity.

The circuit will be powered via a 12V gel cell.

Did you e.g. try to operate the circuit battery powered in your garden? (Provided you are not standing under an overhead power cable)

I have been testing the circuit on my bench with soldering iron power cables etc all over the place and so far these stray voltages have not been enough to activate the circuit to a level detectable with a LED indicator.

Stray voltages may activate the circuit enough to be detectable on the analogue pins of an arduino but, given the difference between this and finger activation of the circuit that lights a LED, I can't see any reason why I could not filter out any such minimal activation in the arduino program.

- - - Updated - - -

What is the purpose of the 470k resistor?

If I leave it out it doesn't seem to make any difference to the circuit?

 

[FvM]

The 470K resistor increases the hysteresis (difference between on- and off-threshold). But I can imagine that the difference isn't essential in your application.

I suppose I could tie the input lines to GND via a 1M resistor or something. Then they wouldn't be floating. Is that likely to result in any significant improvement?

Good question. I suppose, you have a LM339 comparator, in this case the input floats up to a a few volts below supply voltage. Connecting a bias resistor would decrease the required AC voltage to trigger the input. On the other hand, it's reducing the input impedance.

Personally, I don't like floating inputs and similar undefined circuits that only operate by chance.

 

[boylesg]

The 470K resistor increases the hysteresis (difference between on- and off-threshold). But I can imagine that the difference isn't essential in your application.

I don't get it Brian. I thought the hysteresis came from the output pin through the 1M resistor.

The voltage from the output pin does not pass through the 470K resistor, only the voltage from the voltage divider does?

Good question. I suppose, you have a LM339 comparator, in this case the input floats up to a a few volts below supply voltage.

Well not according to my multimeter it doesn't. More like 0.2V, or something like that.

Connecting a bias resistor would decrease the required AC voltage to trigger the input. On the other hand, it's reducing the input impedance.

Personally, I don't like floating inputs and similar undefined circuits that only operate by chance.

Perhaps another resistor in series with the input pin as well to compensate for the reduction in input impedance?

Can you measure the input impedance of the LM339 with a multimeter? Where would you place the second multimeter lead?

 

[boylesg]

I tried tieing the input pin down to GND with a 470k resistor but my finger no longer generates enough voltage difference across the laminated paper to fire the schmidt trigger.

But I noticed that if I touched the Al foil on to the metal sheet that I am soldering on top of then the circuit was triggered.

I happened to check the metal tray with the multimeter and found that it was at Vcc 5V - the lead from my power source is touching it some where along the line.

It got me to thinking that if I was to put a strip of Al foil on the board, tie this to Vcc and then lay my laminated paper with its individual Al foil sensors stuck to the back on top then pressure from a finger would result in a sensor making contact with the Vcc strip which would trigger my schmidt trigger.

Then my sensor circuit becomes a straight voltage detector as you say and will be less sensitive to stray 'envrionmental' voltages.

Might have to reduce the width of the Al foil sensors so that it is less likely that adjacent keys are not triggered simultaneously.