One of the things that I find annoying about most of the probe designs is that the rods must be insulated from each other because they form part of a series circuit connecting all of the balls in a chain.

It occurred to me that the rods could all be grounded if logic connected to each of the balls could tell when they were grounded through the rods. So, a probe hit would be recognized if any of the balls was not grounded.

I think that the total cost of this would be a handful of diodes and resistors.

I'd like to find some PC board material that is thicker than the usual 1/16 inch. Does anyone know where I could find some material that is 3/16, or so? Single sided would be most convenient.

As an alternative, I could glue some standard laminate to some aluminum (in some ways, that would be even better). Any suggestions on an adhesive to use? The main requirement is that it must be stable (It would be convenient if it were an insulator, also.)

HOOD -- if you read this, please feel free to put your stuff on this thread.

Ken

We use a lot of G-10 where I work, and there are always decent size scraps to be had. We have it in fractional thicknesses from 1/16 to 5/8 inch. I don't think we have any copper clad except for the 1/16 inch. How big a piece of 3/16 do you need?

- Joel -

We use a lot of G-10 where I work, and there are always decent size scraps to be had. We have it in fractional thicknesses from 1/16 to 5/8 inch. I don't think we have any copper clad except for the 1/16 inch. How big a piece of 3/16 do you need?

- Joel -

I would need a piece the diameter of the probe -- which is yet to be determined. But I do need copper clad. That's the whole point of using that material. Do you have any suggestions on the proper adhesive to use with it.

Thebest approach would probably be to fasten some copper clad to some aluminum. I suppose I could just screw it down, but adhesive seems cleaner.

Ken

Ken
My probe has an OD of 50mm, The balls are 6mm Dia, Rods 5mm Dia x 16mm Long, centre hole for stylus holder going through is 25mm Dia and stylus holder itself is 22mm Dia. The Balls sit in solid copper 3mm thick (which sits in pockets 2.9mm deep machined into the base) and they are 8.1mm apart (ball centre to centre) The balls are on an 35mm Dia in the probe body. Need any more dimensions just shout.

Hood

So help me out.

Is the function of the probe to simply provide a binary hit on the work piece? Does it spin? If not, why not?

Seems like if it is spun and some rate that would ensure a hit in between a motor step then you'd be good to go. No? Then the runout is not an issue as long as you can measure it and assuming the software allows you to figure it into the ball diameter. I guess you'd just add it to the ball diameter.

A spinning probe is a totally different design issue.

Actually, if the feed rate was low the RPMs become less of an issue. It's the relation between RPMs and fee rate. So 3000 rpm and 6 ipm = 500 hits per inch assuming the eccentricity of the rig allows only one hit per rev. Slow it down to 3 ipm and you get 1000 hits per inch. Pretty slow I admit. Or spin it up to 5000 revs.... as my Bridgy begins to scream for relief.

But if this is just for finding home, who cares? 500 rpm would do nicely.

So...then you need a commutator rig to maintain electrical contact while it is spinning. Hmmmm.

-jd

Runout really shouldnt be an issue if using as a probe for digitising as all that you are doing is making a point cloud and it doesnt really matter if it starts 2 thou or 10 thou out as the next point will be exactly the same and so on. The reason I am bothered about runout is I want to use mine as an edgefinder so it needs to be spot on true with the spindle.
Running the spindle with the probe might cause other problems such as the contact being lost between rods and balls as spring pressure is not much and it would obviously be out of balance if its not dead true.
Just my thoughts but probably not even worth the 2c you USA guys usually give ;)
Hood

Sir I have no idea how it really works and if you have a time could you gime me a short explanation /
REGARSD cLAVE




One of the things that I find annoying about most of the probe designs is that the rods must be insulated from each other because they form part of a series circuit connecting all of the balls in a chain.

It occurred to me that the rods could all be grounded if logic connected to each of the balls could tell when they were grounded through the rods. So, a probe hit would be recognized if any of the balls was not grounded.

I think that the total cost of this would be a handful of diodes and resistors.

I'd like to find some PC board material that is thicker than the usual 1/16 inch. Does anyone know where I could find some material that is 3/16, or so? Single sided would be most convenient.

As an alternative, I could glue some standard laminate to some aluminum (in some ways, that would be even better). Any suggestions on an adhesive to use? The main requirement is that it must be stable (It would be convenient if it were an insulator, also.)

HOOD -- if you read this, please feel free to put your stuff on this thread.

Ken

Hood do you have a link showing the probe you built?

Mike

Hood do you have a link showing the probe you built?

Mike
I have been waiting until I finished my Bridgeport BOSS retro fit before I made any more incarnations of my probe as it is a lot tighter than the manual mill I have retro fitted so it will be easier to make the probe more accurate,
I have decided on a new design which will be capable of being used as a toolsetter/edge finder and probe. It can be used manually (LED indication) or hooked up to my CNC controller (Mach3)
Below is a link to the first probe I made and also the first test of the new one.
http://mysite.wanadoo-members.co.uk/hoodsmachiningpages/Probe3.jpg

http://mysite.wanadoo-members.co.uk/hoodsmachiningpages/Probe.jpg

http://mysite.wanadoo-members.co.uk/hoodsmachiningpages/Toolesetter.jpg

Hood

it looks nice and for sure going to work out great,but I''d like to know how it really work.

best
Clave I have been waiting until I finished my Bridgeport BOSS retro fit before I made any more incarnations of my probe as it is a lot tighter than the manual mill I have retro fitted so it will be easier to make the probe more accurate,
I have decided on a new design which will be capable of being used as a toolsetter/edge finder and probe. It can be used manually (LED indication) or hooked up to my CNC controller (Mach3)
Below is a link to the first probe I made and also the first test of the new one.
http://mysite.wanadoo-members.co.uk/hoodsmachiningpages/Probe3.jpg

http://mysite.wanadoo-members.co.uk/hoodsmachiningpages/Probe.jpg

http://mysite.wanadoo-members.co.uk/hoodsmachiningpages/Toolesetter.jpg

Hood

The way it works is simple, the balls make contact with the copper and the space between the copper is bridged by the pins (which are insulated from each other). The pins are held against the balls by a spring and any side movement or end movement against the pin holder causes it to break contact with the balls and thus the circuit is broken.
I adjusted my first one by means of the four setscrews you can see sticking out the side of the metal holder and got it true to within 0.005mm and its repeatability at picking up an edge or coming down against the table or workpiece was again within 0.005mm
Hood

Thought I better just say that my exploded drawing in the above link is not too clear as really the pins/in holder should be at the other side of the copper/balls.
Hood

How about a probe rod attached directly (superglued) to a pizeoelectric crystal? Any contact would generate an easy to detect voltage. Take apart a clicker type butane lighter and attach a voltmeter to the wires; it takes the faintest pressure on the crystal to generate a voltage. Nothing moves so the contact/no-contact point hysterisis is negligable.

Mariss

How about a probe rod attached directly (superglued) to a pizeoelectric crystal? Any contact would generate an easy to detect voltage. Take apart a clicker type butane lighter and attach a voltmeter to the wires; it takes the faintest pressure on the crystal to generate a voltage. Nothing moves so the contact/no-contact point hysterisis is negligable.

Mariss
Very interesting, will have to try this out, would save a lot of hassle and make things easy to get true and make for a very compact probe.
Hood

Mariss, what size voltage are we talking about with the crystals?

Hi Mariss, funny I should find myself in the same forum as you after months of not visiting.

I have been looking at this myself, it is posible to buy small "bimorph" piezoelectric devices for very little, the ones I have are about 2mmX15mm and 1mm thick. They give a voltage when bent and can also be caused to vibrate with an applied voltage of about 65v. I am using them for laser doppler experiments but saw the potential for digitizing.

I think it is possible to either do what you suggest or use a pair of them. Use one to vibrate the tip and couple that to the other one, when the tip touches an object it will dampen the vibration and reduce the voltage measured. Certainly I can create a nice sine wave out of one if I touch it with another also driven by a sinewave.

Graham

The bimorph I have gives 4v with 10um of deflection

got a link to this bimorph so I can get an idea of what you are talking about.
Hood

go to www.rswww.com and stick in the word bimorph into the search box, no real data as such

Thanks, already there lol, should learn to look before I ask. I have also found some info on them here http://www.mide.com/prod_oem_applications.html and http://www.fuji-piezo.com/Bimorph.htm
will do some reading.

Hood

thank you very much.
clave

The way it works is simple, the balls make contact with the copper and the space between the copper is bridged by the pins (which are insulated from each other). The pins are held against the balls by a spring and any side movement or end movement against the pin holder causes it to break contact with the balls and thus the circuit is broken.
I adjusted my first one by means of the four setscrews you can see sticking out the side of the metal holder and got it true to within 0.005mm and its repeatability at picking up an edge or coming down against the table or workpiece was again within 0.005mm
Hood

Well, if you strike it with the built-in hammer mechanism it's thousands of volts (enough for a spark). Very little energy involved so clamping would be easy. I would think of it as an insensitive contact microphone.

Mariss

One thought is that unlike a normal probe once you have stopped moving the voltage will disappear so you won't know you are still in contact, in most cases this doesn't matter I guess.

The thing I like about the vibration method is that there is a chance it will work for very gentle contact on soft materials, if you could detect the clipping of the sinewave you would know you had come into contact with something. I saw a picture of a probe digitizing some cloth and they said it was peizoelectric so I guessed this is how they do it but I don't know.

Graham

it doesn't take much of a tap either my electronic drums are using piezo for pick up all i have to do is tap the head with my finger and its enough to get the data to the puter the setup is just a piece a foam the piezo sits in the head stretches over the top of it

http://www.ems.k.u-tokyo.ac.jp/morita/Papers/10_SA_Kanda2000.pdf

Exactly! You know when you are moving, the pezio device sends a signal, you stop moving and log the coordinates when you get the contact signal. It tails off afterwards alright but you still get a sharp rising edge at the instant of contact. That's what you are after.

Mariss

Mariss would this be any safer on the probe itself than one of the units that have been discussed here? I suppose I'm asking if the axis would continue to move slightly after receiving the signal or would this still be software related?

Mike

I always like to simply. I saw the break-away drawing with the 4 balls and everything for a contact probe. I thought what an elegant but complex solutiuon. You jostle the balls, they break contact with a circuit. It does what it's supposed to do.

Meanwhile I'm getting hung up with the "complex" part. Shouldn't it be easier where something gives an electronic "clang" when contact is made? Something with no moving parts at all? Like a toothpick in your ear touching your eardrum? Your'e OK until the faintest pressure touches the end of the toothpick. Lateral or on-axis, it doen't matter: you'll scream.

Same thing here but very simple mechanically. Just a probe rod afixed to a pezio-electric crystal. Touch it slightly and it screams "contact!".

Sorry for the gruesome analogy but it works.

Mariss

just did some playing around with a buzzer i had from ratshack i noticed i get lower readings if i tap like the center was under 1 volt then if i have it push and bend the crystal bending was sending anywhere 12 to 15 volt

Oh Mariss.

That is so romantic! Perfect... You are, of course, describing the subtle sliding back-n-forth of a piezo-electric elements intermeshed current producing chrystalene structure..

I havent spent much time with digital edge finders and 3d probes, but i have played drums for years. Electronic drums, especially those from before 1998 are based on simple piezo transducers and op-amp based amplification circuitry.

If the 10uM deflection for a 4.5v pulse is a real rating, a simple probe based on a pizeo element seems very interesting, yeah.. No moving parts is a bit of a oversight, considering some sort of actuator must still impact the pizeo disc...

Why dont you do some experiments.. It could be fun...


Everyone likes new products. Especially if they are appropriately engineered, and priced for the man on the street.


Robert

Exactly! You know when you are moving, the pezio device sends a signal, you stop moving and log the coordinates when you get the contact signal. It tails off afterwards alright but you still get a sharp rising edge at the instant of contact. That's what you are after.

Mariss

Fine for bed of nails but not so good for adaptive feedplane when you need to know when the probe has left the surface.

You need some give in the system for axis deceleration I don't suppose that would be too hard to do but it needs to give in all directions.

Jostle the balls? There clamped!

Graham

Mariss and all,

Just use one of the piezo disks that have 3 terminals. You conect it to a single npn transistor and a couple of resistors and you have a ceramic resonator-oscillator.

Glue the probe to it, tweak in the gain to get it to run at some frequency.

Monitor the input current or the oscillator (ac) output voltage. When the probe is touched (any axis), the gain is reduced, the oscillator stops and you detect the current draw change or the AC output.

I'd use a light weight probe rod and a rigid glue to get good coupling to the ball. The piezo disk mount could be the part that had the flexture that prevents destruction in the event of a crash or over-deflection.

Where do I get the ruby ball?

Barry

Barry,
Why does it have to be ruby?

-Brady

Dunno - maybe the tempco. Seems like they are always ruby on CMMs. Maybe it has something to do with Kansas.

Ruby is hard (almost as hard as diamond) and cheap.

Get Replacement Styli and Accessories for Renishaw CMM Probes from www.mcmaster.com -- $35 for a 5mm ball end probe.

Quoting from the web page.
(E-J) Styli— Have Grade 10 ruby balls with a sphericity tolerance of ±.000025" and a diameter tolerance of ±.0001

I consider that pretty reasonable.

Ken