Re: CAM Software Advice for Shopmaster CNC

--- In CAD_CAM_EDM_DRO@yahoogroups.com, John Dammeyer <johnd@...> wrote:

>
> Hi Wayne,
>
> See below.
>
> > I have a system that is similar to what John suggests,
> > but differs in the details. Important features such as
> > ESTOP are not currently implemented (or even designed.)
> >
> > The system is broken into to two parts. One piece of software
> > (called Steps) takes a fairly standard set of CNC codes
> > and breaks them into timed step codes. A timed step code
> > is just a timestamp and axis increment/decrement --
> > (0:12:345.670 +X) (0:12:345.890 -Y). These are merged and
> > shipped down to a separate buffer board that that has
> > some memory and an inexpensive microcontroller that can
> > squirt out the pulses at accurately timed intervals.
> >
> > Here are the URLs:
> >
> > Serial => Parallel Buffer Board:
> > <http://gramlich.net/projects/cnc/buffer/rev_a/index.html>
> >
> > CNC => step code converter:
> > <http://gramlich.net/projects/steps/index.html>
> >
> > By the way, this stuff actually works; it has lots of
> > rough edges, but it works. In addition, everything
> > is published -- code, gerber files, parts lists, the
> > works. It needs another round or two of polishing, tho'.
> >
> > Is my approach of any interest to other people?
> >
> > -Wayne
>
> Interesting. Once I finish the Electronic Lead Screw I'd be interested
> in looking further into this. Especially since I have a spare tray of
> 16F877 devices that are sitting around doing nothing but getting older.
>
> I'm curious about the maximum step rate you can get. Your time stamp
> example appears to only be good to the nearest milli-second (1 kHz step
> rate) or was that just typing in the numbers from left to right on the
> keyboard.

John:

The buffer board with the PIC16F877 can generate a step
every 20uSec. However, I'm not sure that there is any
realistic way to ramp a stepper up to that speed. I
think a step every 100uSec is a more realistically
obtainable rate -- call it 10KHz step rate. It won't
win any of the Inch/Min. derbies out there, but quite
respectable for a table top machine.

The software performs all of the computations in floating
point and quantizes at the very end. Thus, as the hardware
step rate improves, the software does not need to be
rewritten.

My next rev. of the buffer board will probably be to
switch over to an Atmel AVR processor (probably an
ATMega168) to pick of a factor of 5 or more in speed.

After that, I intend to tackle the bandwidth bottleneck
and switch over to 10MB Ethernet using a Microchip
ENC28J60 chip. After that, I would speed up the the
microcontroller again, -- either an ARM or an FPGA.
I'm leaning towards the FPGA, since they are kind of
neat devices I've always wanted to figure out.

Does that answer your questions?

-Wayne