Re: Homebrew STG card - KISS

["Bert Eding" <b.eding-at-chello.nl>]

With this, I answer the question how many people are interested, I am.

By the way I am following this thread for some weeks now, it is very
interresting.
I work at a company in the Netherlands who makes motion controllers,
they are connected by firewire to the PC. My company thinks that this is the
right way.
Next year all big PC mother board companies will have this interface
standard on the board.
Asus alreay has a motherboard with firewire, ofcourse Sony has it on almost
all the laptops.
We use it because we require high speed data transfers from one motion
controller to another.
It is also deterministic.
Our motion controller box is not just a simple IO device, it has a Texas DSP
to run the motion
controller. At the host side we supply a C-library as interface, currently
only for windows NT.
Our motion controller now supports 4 axes per box, but we can connect up to
48 axes currently.
It is not suited for us home hobbists because it has another price category.

Although I here different stories, I think that a box outside the PC is the
best solution,
-No problem with connectors, you can put as many as you want in the external
BOX.
-EMC safe

To develop a cheap solution, I think that using the bi-directional parallel
port could also be a solution.
Just make to transfers possible to the box, one to read all data in fixed
sequence, one to write all
data in fixed sequence, and there should be a latch command ofcource.

Regards,
Bert Eding,
The Netherlands.


----- Original Message -----
From: "Paul Corner" <Paul.Corner-at-tesco.net>
To: "Multiple recipients of list" <emc-at-nist.gov>
Sent: Saturday, March 04, 2000 6:07 PM
Subject: Re: Homebrew STG card - KISS


>
>
> > The problem is to plan for some sort of an overall architecture.  This
creates
> > so much trouble - that I guess the best thing would be to just keep
things
> > stupid simple.  Any homebrew card that would have what STG boards have,
will
> > not be cheap and simple,  no matter what you do.  (IMO, cheaper Yes,
simple No)
>
> When I initiated this thread, I made the assumption that EMC users would
> dedicate a PC to machine control. Therefor, using firewire, USB, or what
ever,
> would not be part of the design brief. I agree with Arne (hi Arne,nice to
see
> you back). Keep thins simple and the cost will be much reduced. The basic
> design currently has a minimum chip count of six. Five of these are low
cost
> TTL beffers and address decoders and one LS7266R1.
> This gives a basic two axis DRO and/or feed back for steppers - cost,
around
> $20, plus board. By adding some more encoder inputs, a couple of 8255's,
DAC
> and buffers, the toal cost shouldn't be much more than $160 to $200. To
add PCI
> would just escalate the cost and complexity.
> I should have some preliminary files ready by the end of the week if
anyone is
> interested ( HPGL plots only, at present. GERBER to follow later) -
Subject to
> spare time at work.
>
> Regards, Paul.
>
> BTW How many people are interested in this project - enough to justify a
small
> run of PCB's ? - And is there any takers to write the software module ?