Re: [CAD_CAM_EDM_DRO] Re: Out of Estop.. BUT

Bertho Boman wrote:

> Sorry Jon,
> My comment was not in any way intended to be negative.

no problem.

> I know that going Digital to Analog (+/-10V) is an industry standard.
> All I am saying is that if we can pull together a DSP based motion
> controller,
> we do not need the rather expensive I-O card in the PC and minimum
> analog
> components on the controller board. Most of the stuff would would go
> into
> the DSP and a glue chip. It would be easy to build for the group and
> the servo tuning would be done in software. The less analog stuff and
> the fewer
> components, the less things to go wrong.

Well, really, there's no need for a DSP, I don't think. Really, you
could make the tach
information available from the encoder counter already in the PC, and
then you only
need the motor current, which could be provided with a relatively low
resolution
ADC. With a little very simple digital filtering and deadband, smooth
response and
freedom from hunting should be possible. Although I have written pretty
large
software packages, I was a bit loath to chop into EMC with both hands,
and
other people were reporting success with standard velocity servo amps,
so I thought
that was the way to go. (Also, I started with an Allen-Bradley CNC
control that
was a LOT less flexible, as I had only the binary paper tape at the
beginning.)

> Note, I am not proposing to do a full motion controller. Leave EMC or
>
> whatever as before. All I am suggesting is to use digital handshaking
>
> between the PC and the servo amp card, use a software tach and
> interface
> the encoder through the servo amp card. No I-O Card needed.
>
> Before someone jumps on me for anti-analog, I make my living designing
>
> electronic products and are comfortable with both analog and digital.
> I just think there will be fewer problems when non-electronic hobbyist
>
> assemble the boards.

If you take a look at my Servo Amp board (I know Mr. Boman has), your
point is
pretty well self evident! There are a HELL of a lot of parts on the
analog servo amp,
and without full understanding of the circuit, someone not highly
skilled in electronics
and control theory would have a hard time figuring out where a problem
is coming
from. Having a computer in the loop gives vast possibilities for
self-diagnosis and
auto-tuning.

> If you or Matt are still interested, I probably can dig up the article
> on the software tach.

Yup, I would like to see this. I DO know that a simple analog filter,
whether RC,
LC or active filter, will not do well at all with a frequency range of 1
Hz to 100 KHz,
for instance. But, a digital filter, possibly a very simple one, might
do a lot better.

But, of course, my system DOES work, and has been doing so for nearly 2
years.
Sometimes, a pile of op amps can be debugged totally in a week, while
the DSP
guy is still verifying that he has the right version of the compiler for
the DSP
chip he plans to use.

Jon