Re: Calm down guys

At 09:07 PM 10/12/99 -0500, you wrote:

>Hey guys, calm down a bit. Gar- I think that I'm the guilty one that
>invited Hoyt to join the list a long time ago.

Thus perishes another of his assumptions. Hope he gets the point.

>Hoyt- I know that you are doing great things with your programming, but you
>need to communicate things better. Why don't you revise your web site to
>explain things a bit more and add some pictures of your setup. Are you
>running a mill or lathe or both?

Good idea. I have some shots I was thinking of tossing up there, but
they're related to other biz too and am not sure which site is the best
place for them. The manual is always available to the truly interested too.
That BTW seems to have led to me getting knocked off for emergency tool
change with auto-return to machine position and running file pointer, what
an honor even if they did it in their own code. But you understand my
primary emphasis isn't programming or selling software, I'm a practical
machinist and make my money by making chips. I am also getting biz doing
machine conversions to CNC, presently doing a build for a guy who wants to
make a million parts.

As for what I'm using it on myself, I have presently a Maximat Compact
which is older than Moses but still a fine chunk of iron. It's the
three-axis with a decent sized table for smaller milling work. I do turning
with it too of course. Having ultimate control over the program means
altering the source for special work as needed. For example, I spliced two
routines together and made it emit a file which generated a female
semi-cylindrical surface that couldn't be machined from end with a boring
bar but had to be done from the side with ball-nose cutter. That meant
writing a linear loop that moved tool on X, picked it up, moved it back and
set it down. This file was opened and run to the output file as a step in
the other routine, which handled generating small arcs adding up to the
whole semi-circle. So each time it made an arc it called the other file and
made another ruled line in the work. That proved so handy I formalized it
in principle as a 'synthesis' menu entry, allowing any file to be run at
machine and accumulated into whatever work is being programmed, a poor
man's subroutine call if you like.

>but wouldn't mind reading more about how you went about programming and
>building the setup.

I started writing the code with examples given by AS with the package. I
had minimal other C++ experience and that so long ago I forgot it. But if
folks like open architecture in here they should love C++ as it requires
you to define everything that isn't in libraries already ... I once looked
for hours for a function to increment an array and finally wrote one
myself, which took only about 8 lines.

>Recently, you added a digitizing routine to it. That
>was really cool.

Thank you I thought so too and it was so simple! You can see that module's
code on alt.machines.cnc if you're quick enough. But what's really cool is
that if you have a number of modules and control over them in source, you
can do more and more with less and less extra effort. I already had a
routine to step through user defined XY array and at each stop call a
peck-drilling routine. The pecking is a separate module which is also
called by routines that handle polar (bolt circle) or spiral patterns, and
in fact both of those use the same math as the helical interpolation; they
just accept extra parameters and fall through to it, with some if()
statements along the way to switch in or out things peculiar to one or the
other. For example pecking is disabled with a config menu setting, so a
user with two axes get prompted to drill manually while the machine steps
the array, pausing at each pos and writing the pause to file too, but
three-axis users get auto pecking at each place and re-run files have no
pauses. And helical loops through a series of downsteps and small arcs,
spiral uses a radial increment added each time it loops, and polar and
spiral use moves from point to point instead of arcs, but both w/o
downstepping ... So why not have the XY array routine call the digitizing
routine? It took about 40 minutes to patch that up and now it digitizes the
XY array from Z axis as well as doing it on spindle and Y axis, picking
which type of scanning to do by which axis is selected for the probe.
Shortly I'm going to patch it to the tool table and let it find and record
my offsets for me.

At the same time, don't rule Linux/EMC out. It will get

>easier to install. You might like it.

Yeah I had thought about it long before I picked the LPT method. I'm still
wondering about the source code. Open source is so convenient. You know
what language it's in?

>We all probably need to set up a page with a comparison of
>OS/setup/cost/features for the different systems.

I got mine going with I-LPT, three #34 NEMA steppers in 450 and 600 oz/in,
10A PS, three stepper amplifiers, printer cards, cables and all ball
screws, nuts and drive hardware for about $24-2500. The hardest thing about
wiring it all up was doing pin-charting on the cables; everything else was
clearly diagrammed. I chose goods from a number of vendors: Microkinetics
(good people, they fixed PS once when I thought it was OOW), Eastern Air,
Applied Motion, Techno-ISEL and Berg. Have a good one.

Best wishes,

Hoyt McKagen


Belfab CNC - http://www.freeyellow.com/members/belfab/belfab.html
Best MC Repair - http://www.freeyellow.com/members/batwings/best.html
Camping/Caving - http://www.freeyellow.com/members/batwings/caving.html
May you live in interesting times