Self driving piezo used as CNC touch probe?

[harvie]

Hi! I am thinking about building piezo probe for my CNC. Classic piezo probe design is basicaly "piezo knock detector" circuit using either opamps or uC with ADC sensitive enough to pick the piezo signal. That has one disadvantage: It detects when probe crash into workpiece, but can't tell if it has left or not. knock detectors also can't tell if you left your hand on the door after last knock or not.

I want to build piezo detector that would tell me if something is touching it or not even when it's not moving.

I've just found that there are "self-driving piezo transducer". It has two crystals. One big and one small for feedback. Think about it as of that little feedback winding that some transformers have. It look like this:



My idea is to sample the feedback at arduino's ADC and close the feedback loop by outputing it as hi/low pulses back to bigger crystal on that transducer. That should make the buzzer resonate on it's frequency (eg. 3kHz or whatever it's built for).

Now do you think it will get out of tune if something touches the piezo? I am thinking about glueing some shaft (which will act as probe) to that transducer. Adding such weight to transducer will probably change it's tuning so i will need to recalibrate for new resonant frequency. Now when something touches it, the resonant frequency should change and i should be able to detect it in the arduino.

Do you think this will work? Do you have some tips on implementing this for me?

- - - Updated - - -

I've just found iteresting document on this topic! http://www.intricad.com/touchprobe/intricad_gettingthefeelforit.pdf

 

[betwixt]

The first thing that came to my mind was piezo discs used as alarm sounders but they are already mentioned in the article. The image is of one I salvaged from a smoke alarm a few days ago. The article mentions it wasn't very effective but the author says they attached the probe to the edge of the disc which is not the way to do it. The operating mode of these things is they 'dish' one way or the other when supported firmly around the edges, normally this is done in a tuned acoustic chamber to amplify the sound but they oscillate just as readily without the chamber.

I can't do any experiments at the moment but I would suggest if you make an oscillator using such a disc (the small silvered area is the feedback pickup) and support it around the edge but attach the probe to the center rather than the edge it might work. The frequency will change slightly if the probe touches but these things have quite a high 'Q' so it might be better to monitor for changes in the current drawn by the oscillator to check whether the probe is loading it or not.

Brian.