[REVCAM_Bob]

O.K. for the electronic gurus.... Here are a couple links to some simple ir reciever units i.e. one you can buy for $15....

Could these be used to receive a signal from an electronic edge finder i.e.
the kind with the red light emmiting diodes on them? And if so, are there
any cheap probes out there that this could be used with also?


http://www.factorydirect.co.uk/facto...rreceiver.html

http://lnx.manoweb.com/lirc/

[Meganick]

Need to consider a few things.

-TV remotes modulate the LED source at frequencies from 38-60kHz to form pulse streams. See http://www.everlight.com/pdf/Irm-8601s.pdf . So it is a little more complicated than a simple on/off touch/no-touch indicator

-An example of the actual pulse stream that the a remote sends out is shown half way down this page http://lnx.manoweb.com/lirc/?partTyp...&partName=lirc . Note that the time scale is 5msec/unit and that some of the commands take almost 30 units or between 100-200msec. This is not a fast protocol. You probably would not want to use this for a CNC edgefinder unless you were going VERY VERY slow.

-Also need to come up with a way to tell a CNC program that you Are Touching, Not Touching, or Not Sure incase the IR connection is interrupted.

[REVCAM_Bob]

Can't it sense just on and off?

[REVCAM_Bob]

I think I can tell cnc to stop o.k., and already have the ability to read the
machine postiion..... Just need the signal...

[Meganick]

Some exampes of transmitters can be found here in the Home Brew section: http://www.lirc.org/

What you'd be looking at could be a small PIC or other microcontroller that would be built into the edge finder that would modulate the LED to generate two different button push pulse sequences depending on whether or not it is touching.

So again the real question is can you tolerate a delay of say 1/4 second from the time contact is made to when the controller is actually signaled that you have contact? During that time, the motor will continue to advance the table. Say you had were advancing at 4mil/sec to an edge, you would overrun by 1mil in the 1/4 second.

What are typical edge find approach rate numbers?

Maybe a better way would be to have the edge finder put out a button sequence that is interpreted by CNC controller as a jog/advance signal. So form the time you say go, each pulse output is used to advance the controller one jog step until you touch. If the IR gets interrupted, it stops.

This is sounding interesting. Tie the IR receiver into a ModBus module/Mach3

[Meganick]

And if you designed your own custom decoder, you could should much shorter code sequences.

[REVCAM_Bob]

I guess maybe I am overcomplicating this.

My bad, not really thinking infrared, just red led, or are they one in the same?

Isn't there a simple way just to detect if and led on a simple electronic
edge finder is on or off?

[REVCAM_Bob]

Meganick, I really do not even know what approach speeds would be...

Maybe some of you guys that have used probes could clue us in?

I would imagine a sequence could be used to do a rough approach at a higher
speed, then a slow approach to minimize overshoot....

[Meganick]

You can hack into it and just run out an extra wire...

Even better would be a pair of wires coming off the edge finder run thru an optoisolator.

[REVCAM_Bob]

Trying to do wireless, so edge finder can be stored in turret for unattended
operation...

[chuckels]

If I understand you I think what your after is a photo transistor which is basically a light operated switch to turn on and off when the LED turns on and off right?
the electronics gurus should be able to draw a quick circuit to interface with your inputs.
chuck

[REVCAM_Bob]

Chuckles, yes that is it! Something simple, but could the ir receivers i.e. these things already built that plug into a serial port do it also?

Of coarse if a photo transistor could do it, then I could tie directly into
the input of a galil card or ethernet i/o unit which could be easier.....

Could a photo transistor pick up a red led on an electronic edge finder from
the corner of a machining center??

[REVCAM_Bob]

I am not opposed to the infra red method, I just thought that if the edge finder already exists and is "off the shelf", that is might be easier than
reinventing the edge finder itself.....

[REVCAM_Bob]

Not to change the subject, but
2 more questions, what is a Mach3 and do tool setters really cost 8-10K
but "real" i.e. renishaw heidenhein etc. touch probe systems only $750?
Something seems out of line here..

[REVCAM_Bob]

O.K. this place in th U.K. sells infra red transmitter fobs and the reciever
is cheap, but anyone know where to get this stuff in the u.s.?

http://www.factorydirect.co.uk/index.html

[CJL5585]

Chuckles, yes that is it! Something simple, but could the ir receivers i.e. these things already built that plug into a serial port do it also?

Of coarse if a photo transistor could do it, then I could tie directly into
the input of a galil card or ethernet i/o unit which could be easier.....

Could a photo transistor pick up a red led on an electronic edge finder from
the corner of a machining center??

Although you desire to build something that seems so simple, it is not simple by any means if your desire is to have a system that is accurate and is also repeatable.

If it was simple and inexpensive, it would be available for purchase, or a freebe added by some manufacturer.

Example:
Most RED LED's operate at a wavelength at or near 830 nanometers. Also, the infrared (presence of heat) spectrum begins near 835 nanometers.

To use a phototransistor detector mounted on the machine would require the detection of ONLY the LED illumination from the edge-finder. This gets complicated due to the fact that the photodetector will detect any light or heat type sources within the phototransistor's detection spectrum which could be from infrared and upward in the spectrum. The built in lens system on a photo transistor covers an angular area of 15 to 40 degrees, depending on the model and part number.

A lens system with focus capability would be required to focus only on the LED of the edge finder. A 830 nanometer tunable bandpass filter would be required in order to keep out any signals picked up from flourescent, incandescent, and other light sources which also emit light in all frequencies.

A better approach would be to have a very small normally un-modulated FM radio transmitter mounted in the edge finder which emitted a modulated signal only when the touch probe made contact. The detector circuit would look for the un-modulated FM carrier signal to verify the transmitter was operational. The modulated signal (tone applied to un-modulated signal) would indicate that the probe contact had been made. This could be monitored from any location within yards of the machine.

[pastera]

Simply modulate the LED at a few hundred kilohertz and tune the receiver to only receive that frequency. A more robust solution would modulate on two frequencies, one for not touching and one for touching. That way the system could detect a loss of signal and not crash the probe

Use a visible diode - you can see if it is working

Aaron

[REVCAM_Bob]

What about robbing an ifra red transmitter from a fob? as in the below website?

Fob:
http://www.factorydirect.co.uk/facto...s/irkeyfob.htm

Receiver
http://www.factorydirect.co.uk/facto...rreceiver.html


Would these items take care of all the interferance issues?

[Eurisko]

Random thoughts on probing...

I use a Renishaw probe daily at work. The probe is used to zero the work coordinates using a finish bore from a previous operation.

The probe typically positions at 75 ipm, encounters the part, then reduces the feedrate to 25 ipm and probes the same spot again (fine tuning). The probe takes 8 hits on the bore diameter (4 rough & 4 finish).

The probe sends an IR signal to a base unit mounted on the machine, less than 6 feet away. The probe is generally accurate to .0005 true position. It would be better, but that is about the positioning accuracy of the machine.

The probe uses a standard 9V battery. The battery life is approximately 1 week of continuous use.

Radio Shack used to sell an IR detector which would convert a modulated IR signal (40 kHz?) to TTL logic levels.

It takes about 240 microseconds for a machine , travelling at 25 ipm, to travel .0001 inches. Keep that in mind when programming the interrupt code .

*** Updated ***
The second probe hit is at 1.2 ipm , NOT 25 ipm.
That gives us exactly 5 milliseconds to travel .0001 inches.

[Eurisko]

I think that the easiest way to do it would be to use pulse width modulation.
The probe would output a train of modulated pulses of a predetermined width when not touching, and a train of shorter pulses when contact is made.



No pulse=no probe. Stop the machine.