Re: [CAD_CAM_EDM_DRO] Re: Black box again

Hi All:

Just my two cents, but I think a black box that does a basic move and then
notifies the software of its completion is not a good thing from a software
perspective. It makes it very difficult to write a program for it. It would
mean that the OS for such a chip is almost certainly limited to unixRT or
DOS. The largest untapped market out there is windows compatible software
(sorry unix lovers, but it's really true). The ideal Black Box would have a
couple of Megs's on board and allow windows programs to do buffer fills
while the black box pulses merrily along. The problem with a black box that
can do basic moves, is that with some programs you have millions of very
small moves. The overhead of translation and transmission would make it
impossible to use such a box under Windows type systems, due to the times
involved, and the compexities of velocity computation for each of the small
moves.What you would end up with is a program that suffers from stuttering
between moves unless written on a Real time OS. And if you have a Real-Time
OS, you don't really need a block box.
I think a black box that would be more usefull from a programmers
point of view, would accept a large buffer of pulse data which is put out in
the timing relationship embedded in the pulse stream. I know from
experience, that under windows, the fire and forget it approach would only
work for some types of G-Code programs and would be very hard to use for
others. The Main problem is the wide variety of setups out there. If you
have 30000 steps per inch, you could expect that a small move will take long
enough to complete, but under some setups..ie. 156 steps/inch (I have seen
it), a small move could be processed so quickly that unless the program
responds within uSeconds, the steppers will sputter. So maybe I'm wrong, but
it seems to me that if your gonna design a black box pulseing unit, it
should not know its a CNC device, just an intelligent pulsing device that
can buffer information and handle the buffering well. This would be more
usefull to programmers that anything else. Any computer is more than capable
of doing all the other calcuations involved , embed velocity information in
the pulse information and would have lots of time left over.
Memory is cheap enough these days that 1 or 2 meg's (or even 8) should
be able to be integrated into such a device, as a programmer under windows,
I think the black boxes I see being developed, while unquestionably usefull,
are trying to do too much, and as a result, will be found to have severe
limitations to the windows programmer.

(No offense, just my thoughts as a programmer)
Art
Original Message -----
From: <dkowalcz@...>
To: <CAD_CAM_EDM_DRO@yahoogroups.com>
Sent: Thursday, June 28, 2001 6:39 AM
Subject: [CAD_CAM_EDM_DRO] Re: Black box again

> Hi Alan,
>
> Here's the chip I'm helping with:
>
> http://www.public.iastate.edu/~dkowalcz/rt900_1.pdf
>
> And here it is again on a proto-board with connectors, etc.., as it
> might look to the consumer:
>
> http://www.public.iastate.edu/~dkowalcz/900000_1.pdf
>
> It's a "dumb box" type device, based on a PIC16C74, 3 axes and 16
> bits digital I/O. Tentatively named StepEx.
>
> My role on this project is to adjust TurboCNC to be able to drive
> this device as a user option, so right now I'm a little short on the
> fine details. Have a look at http://rutex.com to see what the
> hardware guy is up to.
>
> Haven't got source yet, I should be getting a couple chips in the
> mail any day now to play with on my workbench.
>
> I'm also working on a "smart box" personally, since a chip that
> accepts just a motion command like "MOVE(params)" via COM/USB port
> and lets the computer know when it gets there seems a bit more
> flexible on the software side. But that hasn't gotten much farther
> than just some rudimentary specs and picking up a handful of PICs and
> a "burner".
>
> BTW, thanks Ballendo for explaining "fire and forget" in this
> context...!
>
> Dave Kowalczyk
> Seattle WA
>
> --- In CAD_CAM_EDM_DRO@y..., Alan Marconett KM6VV <KM6VV@a...> wrote:
> > Hi Dave,
> >
> > "fire and forget"? USB's are getting easier to play with on the
> device
> > end, anyhow. The PC side? Well, I'm not anxious to write a special
> > windoz driver! If a std. driver can be used (must need something
> like a
> > printer dump, right?), then that will simplify the task. Then you'd
> > have to write the application in windoz too! I think ActiveWire
> (was
> > mentioned the other day) has some type of driver for DOS.
> >
> > http://www.activewireinc.com/
> >
> > I don't think I'd put all the smarts of a controller down in the USB
> > (8051), probably just generate the pulses.
> >
> > Open source for PIC code to run 3 axis? Can I see? ;>)
> >
> > Timed updates. Yup, there is a USB xfer mode (isochronous) that I
> think
> > does that within the 1ms frames.
> >
> > Talk about re-inventing the wheel. Saw a model of "Ezekiel's
> Wheel" the
> > other night on one of the "robot wars" type of show (I only watch
> for
> > the technology). Also saw a tube with wheels in it at right angles
> used
> > for steering. Clever, but I didn't get a good enough look to
> figure out
> > how it worked. The tube was rotated to steer somehow.
> >
> > Alan KM6VV
> >
> >
> > dkowalcz@i... wrote:
> > >
> > > To list,
> > >
> > > Well, I'm currently co-developing a dumb box with another semi-
> > > commercial hobbyist. It's an open source, PIC based device that
> > > handles 3 axes via serial port. Drawback is that it requires
> more or
> > > less timed updates still, just fewer and farther between. Still
> in
> > > development - I'll let the list know when it's done, if it flies
> that
> > > is.
> > >
> > > I think a good "fire and forget" USB smart controller on one
> chip
> > > is long overdue in this hobby. Commercial ones are around, sure,
> but
> > > they start at about a G-note and typically aren't scalable. So I
> got
> > > some PIC's to engineer one the other month.... That's still
> cooking
> > > too, but *way* on the back burner.
> > >
> > > Dave Kowalczyk
> > > Seattle WA
> > > not a huge fan of wheel re-invention either!
>
>
> Addresses:
> FAQ: http://www.ktmarketing.com/faq.html
> FILES: http://groups.yahoo.com/group/CAD_CAM_EDM_DRO/files/
>
> Post messages: CAD_CAM_EDM_DRO@yahoogroups.com
> Subscribe: CAD_CAM_EDM_DRO-subscribe@yahoogroups.com
> Unsubscribe: CAD_CAM_EDM_DRO-unsubscribe@yahoogroups.com
> List owner: CAD_CAM_EDM_DRO-owner@yahoogroups.com, wanliker@...
> Moderator: jmelson@... timg@... [Moderator]
> URL to this page: http://groups.yahoo.com/group/CAD_CAM_EDM_DRO
> bill,
> List Manager
>
>
>
> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>
>