Re: Interface. Gauging interest.

Why the persistent mindset that step pulses be sent over the serial
interface be it USB or ethernet? That's like trying to use a hammer
as a screwdriver.

The other end of the cable has to have a box. That box has to have
electronics in it to:

1) Try to unsnarl step pulse timing created by serialization while
the PC is still doing donkey-work creating step pulse timing/quantity.

2) Accept much lower bandwidth commands and generate step pulses
locally.

2a) Send ASCII coordinate data, have the box run the motion control
algorithms and generate step pulses. This uses the least bandwidth.

2b) Send velocity commands; have the box turn them into step pulses.
If the velocity command is a 16-bit value and staircasing is 1,000
velocity updates per second, the data rate is 2kB/sec/axis.

Once (2a) was called "smart mode", (2b) was called "dumb mode".

Mariss

--- In CAD_CAM_EDM_DRO@yahoogroups.com, "yracg" <gcary@...> wrote:
>
> --- In CAD_CAM_EDM_DRO@yahoogroups.com, Mike Pogue <mpogue@> wrote:
> >
> > yracg <gcary@> wrote:
> > > My gut feel says that bit-banging a USB interface isn't going
to give
> > > the performance that most will expect (I could be wrong).
> >
> > Dennis Schmitz wrote:
> > > And you're also right about the USB interface.
> >
> > What kind of performance are you guys expecting out of a USB
interface?
> >
> > Mike
> >
>
> If commands are buffered, I would expect very high-frequency pulse
> trains without glitches. Otherwise I would expect something
> comparable to the parallel port with occasional glitches.
>
> Greg
>