Re: Scanner/Digitizer skip function/probe details?

Mariano, Mat, Ballendo and the List,

Thanks for the example. I haven't gotten that far yet! Which "system"
of Gcode is this from? It IS a good subroutine model. I think I'd
prefer a 'C' type function call: f(x).

Yeah, the single letter variable (with a variable number) would simplify
the parsing. Mine currently wants single letters! OOPS! I guess I
"simplified" the GetToken() routine too much! CNCPro uses Vnn. Would
the assignment for the variable system you mention be: "V59=2.34"? If
this is one of the standards, then I might as well follow it.

NIST defines G57,58,59 for something else, probably nothing I need.
I'll probably stay close to CNCpro, MaxNC, and NIST/RS274. ;>)

Got G61 implemented last night in my controller code, I open a log file,
run gcode with G61's, and each pin 13 switch closure stops the axis and
writes a data point to the log file. Not hard at all.

Ballendo and I had been talking about a "learn mode". Opening a log
file on the MDI could allow that easily. Next thing to implement.

I'm still going over the recently implemented interpolation routines, I
need to test for any "special cases". And I also want to add more
"error checking" to detect invalid gcode cases. NIST says .0002" is
used for arc radii, I wonder if there is one for linear moves?

The hardest part is the delay required to hit the feed rates. It needs
to run on everything from a 33 Mhz '386, on up to pentiums. I am
currently using a "calibrated" delay table, but I'd like to get a system
timer driven delay working. Good learning experience!

Best regards,

Alan KM6VV


diazden wrote:

>
> Alan, Matt and the list;
> back in 1983 when I started c.n.c programming the controller that the machine
> used to came with, had what I believe the best way of defining subroutines and
> variables, for example you could define your subroutine at the beginning of the
> program , as per example
>
> g57r1 ( definition of subroutine)
> G00 X0.Y50.
> ZV1(use variable 1 for Z)
> G1 G42 HV2 Y10.
> G1 X30.
> G00Z10.
> G40X0Y0
> G58( subroutine ends)
> T1 m6
> f.1s1000m3
> G00Z25 T1 01
> G55V1A-10. V2A2.(Z = -10. remember ZV1 above, HV2 offset for G42=2MM )
> G59R1( exec the subroutine)
> I firmly believe that the above method is much more easier to learn vs the
> other systems like Fanuc as you need on a separate program for each
> subroutine with the disadvantage of having to move between programs every time
> that you have to alter something on the subroutine.The advantage of the Fanuc
> system is that you can alter most of the machine parameters within the program
> that is executing via variables but I think that it is a bit dangerous? and I have
> got no need of doing that on 17 years
> hope that this helps
> mariano
>
> Matt Shaver wrote:
>
> > Alan Marconett KM6VV wrote:
> >
> > > I have been thinking of adding variables. NIST
> > > uses #nnn for a variable name. Not very intuitive (no disrespect
> > > intended as I admire the efforts of NIST; but variable NAMES are a lot
> > > easier to work with). CNCpro has variables, might be same/similar.
> > > What's the standard (ha)? I'm inclined to "do my own", you know how us
> > > programmer types are. ;>)
> >
> > Most commercial controls have three or four digit numbered variables
> > (usually called parameters in the documentation). They're used for all
> > kinds of purposes including configuring the behavior of the control and
> > displaying the state of internal logic terms. I figure that's where NIST
> > got the idea to use numbered, rather than named, variables.
> >
> > Matt