Re: DIY probing

Hi Chris,

MaxNC makes a nice little probe. 1/4" shaft, I run it in a collet on
the Sherline mill. I like the probe, but the software provided
generates a Gcode file; fine for copying, but I needed a "grid" of data
points for my application. I haven't determined the probe's accuracy or
repeatability (I'd have to include the CNC Sherline mill in that too).

I ended up writing (and still working on) a digitizer program, to get
the data points I wanted. I'll also give it a mode to "copy" parts.

It would seem to me that the probe must be moved quite slow, in order
not to "overrun" when the part is contacted. Kinda hard to decelerate
after hitting (OK, lightly touching) the part; one needs to save the
machine coordinates as soon as possible.

Don't know if/what EMC can do with a probe. Basically, you need to move
in a "search" pattern, back and forth in X, then over in Y to a new
line. OR, Do Y before X. If the probe touches, then withdraw in Z,
(take data?) and move again until desired grid point is reached (I'm
open for ideas here).

For CM? I've asked before, what do they do? How do they interact with
the user? If I'm trying to find the edge of a part with a center
finder, I move first outside of the part up near the edge I want, then
approach that edge. First in X, then come around and approach the
"around the corner" in Y (then G92 and set the appropriate coordinates).

So I could see one Jogging to near the part, then requesting a "X
sense", then moving around to a point where one could request a "Y
sense" (same thing again for Z). Center on a hole? move down in Z at
the approximate center of the hole, then the software could easily move
the probe around to find three points to describe the circle, and then
calculate the center.

I now recall that in Radiation Therapy they digitized Xray negatives,
that is to say they digitized the ends of radioactive needles, and then
the data was used to calculate isodose curves around the needles for
planning treatment.

So being able to "save a point" by pressing 'S' could be useful. With a
"bomb site" in a holder in the spindle, one could locate, for example,
holes on a part. Or edges, or whatever.

I could TOTALLY see moving the probe into the approximate center of the
horn's bell, and then determining the center and diameter, or "walking"
down the sides. Do you assume it's symmetric? Do you digitize circles
at various depths? Do you "walk down a side, rotate around the circle a
defined amount (angle), and then do another walk down the side? Or
combinations of the above?? Lottsa interesting movements!

Then there's the data format to output. MaxNC (the ONLY probe I've
seen) as I've mentioned, just generates a "copy" file. OK, DXF file
would not be that hard to generate either. How about STL? What does
that file format look like? I'm thinking about doing a simple .DEM
format, as used by USGS. Then STLtrans can convert it into STL, and
I'll let STLwork generate the Gcode files!

What will you do with the data on the horn's bell? Just curious.

Pardon my rambling! ;>)

Any Ideas list?

Alan KM6VV
P.S. You won't believe how many times I've digitized a certain little
Gecko pin! Had to keep it away from my wife, she saw it in the box with
the drivers, and wanted to put it on the refrigerator! Thanks M!


ccs@... wrote:

>
> > MaxNc sells probes, have one ,....now who has standalone software?
>
> Yes, I was tempted by that. Devices sound kind of simple, but devil
> is in the details - not sure if I'd rather make the design compromises
> myself and know what they are, or pay them to do it and have no idea
> what compromises were made until I get it. How do you like yours?
>
> As for software, I was planning to investigate the probing capability
> in EMC, and then either use, build on, or rewrite that OR (assuming a
> stepper machine dedicated to probing) write a program to do it in
> BASIC (!!!)
>
> It does seem to me that I would want lots of reduction or
> microstepping drivers in order for vibration to not false trigger
> the probe.
>
> Just to explain, I'm trying to probe the inside of a french horn bell
> (in to the start of the curve) to determine its shape. Current idea
> is a table with a cutout the size of the rim - horn is clamped under
> that with the bell opening up. Gantry arrangement moves over this and
> lowers probe down into the bell. Just occured to me I can probably
> borrow the sherline mill Z axis to carry the probe and maybe even the
> X and Y motors to run the gantry, since it need not go particularly
> fast. Of course the temptation would then be to make it heavy enough
> for milling ;-)
>
> Chris