Re: Digitizing

If you want to write your own software then it is certainly doable.

As far as a simple way of doing it then it would be possible to use
turbo cnc to do a series of one dimensional digitizes, you would
simply change the start point and length of each one and increment
the other axis between scans. The gcode to do this could probably be
procuded fairly easily with a simple program or maybe even a few
excel formula and a macro or VB script. Turbo CNC has a single gcode
for a digitise which includes the x,y and z limits and step over etc
so it is simply a matter of puting the right numbers in each time.
Every time the command is run the points are just added to the points
file.

I have a similar problem to you, I am digitizing propellers, they are
wide and flat near the tips but narrow and high at the middle. I
think that the thing that really slows things down is all the up down
motion. After all if you were only rising 0.1mm above the object
each time you probed it doesn't take all that long to scan along. My
props take a long time because in the central section they are high
to I set the feedplane high but it is narrow so much of the time is
spent doing deep probes of nothing.

So what I did was to make a simple digitizing program. It is a
little too simple actually as it only has millisecond timing and to
does not drive my steppers all that fast however it is faster over
all. About twice as fast at least. The way it works is the probe
touches down and the position is recorded, then it lifts until the
contact is broken and lifts a further fudge factor (say 0.1mm) it
then goes off on it's merry way to the next probing point, if it
sidestrikes it lifts until the contact is broken and moves off again.

This means that over perfectly flat areas the feedplane is only 0.1mm
above the object. It also ramps up when it comes to the object.

The program is written in C and runs in dos. It needs some work to
make it faster in terms of stepper driving but it does work. Rather
than spend a lot more time on it I want to just do some experiments
to compare the speeds of the Tcnc alorithm and mine and then just
hand it over to be added to Tcnc if he wants it.

Doug Fortune also produced an interesting algorithm that I did
implement, there are some threads on the subject under the
title "digitizer algorithms", again I just need to get around to
doing some more tests. Dougs was quite cool because the probe has
constant contact and is was very fast, I just wanted to check the
results were ok.

I have almost finished my digitizing website (where I also describe
my digitizer) so perhaps that will spur me to finish things off.

Cheers,

Graham

> Hello--Wondering if anyone would know of software or procedure that
> would allow digitizing a grid within a defined profile rather than

an

> XY rectangle? There is a lot of wasted time/motion in scanning into
> the unneeded corner areas of the rectangle.
> Perhaps a polyline of the profile could be exported DXF format and
> then use Excel to merge the profile into the XY rectangle

coordinates

> and throwing out any XY coordinate that is outside the bounding
> profile? This is over my head but thinking that a formula could be
> used to examine each coordinate?
> Or perhaps there is some software availabe to do this?
> I'm digitizing archtop guitar plates which are highly contoured.
> Any and all help is greatly appreciated.
> Thanks
> Nelson