[bfreedland]

Has anyone tried to make a probe utilizing a piezo element instead of mechanical switches?

[andy55]

the piezos I have seen react only to very quick actuation, i.e. hit them with a hammer and they will show an output voltage swing.

for probing I assume you would want a signal even when the probe very gently and slowly begins touching the probing surface.

If you can find a piezo that reacts to these slow small changes then why not...

[alnicov]

The piezo element in a musical instrument pickup is quite sensitive, especially the 'ribbon' type which I believe is actually a polymer film of some sort.

-al

[macona]

I have a piezo element off of a ultrasonic die bonder. Already probe shaped. Maybe hooking it up to a high gain amp and see...

You know they make accelerometers that are piezo based and are rather sensitive. They can pick up pretty low g's.

[10bulls]

I have a piezo element off of a ultrasonic die bonder. Already probe shaped. Maybe hooking it up to a high gain amp and see...

Darn it! there goes another idea someones already thought of!

Oh well. As I just mentioned in the touch probe thread, how about using gramaphone needles (with a softer tip)

I think you'd need some pretty good filtering or noise reduction though.
Especially with the steppers whirring away.

Another idea i hazily recall is a method for making bump sensors for
robots using a bent fiber optic (one of those cheap Tina Tuner lamp sort of ones) bend around to a sensor (something to do with interference patterns?). They were supposed to be quite sensitive.

And last (but not least) mount a probe to the centre of a thin Y shaped metal thing with small strain guages on each arm of the Y. They are very sensitve and read absolute (piezo just pulses).

[svenakela]

I've already tested the piezo version, it was investigated long time ago. The output signal is just a quick peak and it will need some sort of amplifier, possibly also an echo device that makes the signal longer.

Regards,
Sven

[10bulls]

I've already tested the piezo version, it was investigated long time ago. The output signal is just a quick peak and it will need some sort of amplifier, possibly also an echo device that makes the signal longer.

Regards,
Sven

Alright then, how about this...
http://catalog.sensing.honeywell.com...ce&PN=FSG15N1A
They are quite pricy (~£40) but I think you'd only need one (if probe tip was epoxied to the sensor).
0.2mV per gram force (up to 1500g), 30 microns full scale deflection.
Maybe as a cheaper alternative, pull apart one of those micro-scales you can get fairly cheaply these days and use the sensor from that.

[10bulls]

Just another thought on the piezo idea. I know the piezo will give a voltage pulse if stressed (piezoelectricity), but does not also it's resistance change with applied force (piezo resistance)? If you put one in a wheatstone bridge I wonder if that would fly (using it as a strain gauge in effect).

[lerman]

Don't forget that there are requirements other than the low force requirement. My requirements list looks something like:

1 -- It must have a mount that "self centers" in three dimensions.

2 -- It must detect displacement in three dimensions.

3 -- It must permit a large motion (so that it won't break when the machine overshoots) in three dimensions. It should be possible to overshoot and break a probe without destroying the probe body. It must have replaceable probes so that when it crashes, the probe can be replaced.

4 -- It must be repeatable.

5 -- It must apply a minimal force to the object being probed (not required for all applications).

6 -- It must give repeatable measurements, independent of temperature, barometric pressure, humidity, etc. It shouldn't wear significantly in use (ruby tips are inexpensive and generally used).

7 -- The measuring probe must be concentric with the spindle axis; either machined that way or adjustable. (Adjustable is probably better so that cheap probes can be interchanged.)

Whether mechanical contacts (rods and balls), strain gauges, piezo for sensors, optical, or other means of sensing is used, the above requirements still exist. I don't think it likely that the simple rod and ball kinematic mount where the rods and balls complete the circuit can be beat for my application.

Ken

[ViperTX]

Maybe you should consider a ring gyro setup.....

[braidmeister]

This has already been successfully done by Roland DGA with their Picza series of scanners. The one I have has what appears to be a very long sewing needle attached to the piezo by means of solder. Instead of reinventing the wheel, you might want to just buy a replacement piezo/probe sensor from Roland directly. They've already done all of the research to engineer a working probe...you'll have to come up with your own electronics.

-Brady

[macona]

Look at their site. Ohh.. Active piezo. Hadnt thought of that, Much more sensitive. But much more electronics.

Argh.. One project at a time... One project at a time...

[CNC_Tec]

Anybody thought of using Quantum Tunnelling Composite as a means of measuring the change of force in a probe?

[rokag3]

Piezo sensor work!
I have made a program with arduino.
Use a diode zenner 5V and a resistor of 700k to 1Mohms
BUT
because this device is very sensitive you must set a trigger level which is dependant of the vibrations generated by your machine to avoid false trigger, this is simply done through a little potentiometer that your arduino will read during the setup .

Due to this level of triggering, there is a minimum speed to sense the tools arriving on the probe, if the speed of crash is too slow then you will not ear it 1m/min its perfect you can go as low as 500mm/min slower is uncertain.

This mean that you will have an overshot, this overshot is dependant of your machine.

Remember that a piezo make a peak at the time of the crash, but the peak disappear and do not reappear at the lost of contact so , when the value of trigger is reached, you must add a time constant that will give the time for your machine to go reverse at the exact moment of the crash. This implied that your machine is very accurate in timing.

Also because of high speed your probe must be able to make huge deflection and come back to a perfect position.

this mean that you cannot avoid kinetic joint 3 cylinders 6 balls which if realized properly using an analogic measure of the lost of contact with the analog entry of arduino might be very accurate).

Actually I am testing pressure sensor .

for the tool setter we are using a lazer and quadri phototransistor.

The use of arduino is specially interesting because we can use wifi transmission so no wire with the touch probe, the arduino used as a recepter is the same that is used for tool setter which is fixed by definition and the cable is not a problem

[79TigerPilot]

Funny, I have been thinking about doing some sort of probe off and on for several years. I think Ken has a good list of requirements. I would add a couple more, 8- low false triggering rate 9- possibly wireless with auto turn on and turn off based on some sort of tilt switch -- no cords to get ripped out.

Piezo was one technology, like others have indicated it is an impulse type of detection and I suspect somewhat prone to false detection. Piezo plastic films would have less risk of damage from a crash than quarts or ceramic. One could also look at resonating a crystal in an oscillator and use the mechanical strain to alter the frequency as a means of detection.
Magnetic restriction is another not contact method, that would be similar to to the piezo.

A third option is to use thin film strain gauges. Another I like would be to do some sort of capacity change but it may be hard to get the sensitivity. I also considered doing the sensing optically with a high definition mouse sensor IC. Also considered driving the probe tip ultrasonically and sensing the damping when the ruby ball hit a surface.

This is not an easy problem, if it were there would be lots of import probe systems. Once you decide on a sensing technology then you have to worry about how to do a good kinematic joint and some re-centering system that is repeatable after a displacement in any of the 3 axises. It must have some sort of off axis breakaway system or plenty of no-damage movement area. Rare earth magnets possibly? The probe could possibly issue an E-stop if its travel was about to exceed the capability, this works for normal probing but not for rapids into your stock.

Just gots to be lots of other ideas???

Always thinking and tinkering.
Craig