Re: minimilltask V/S bridgeporttask

[Jon Elson <jmelson-at-artsci.wustl.edu>]

Joel Jacobs wrote:

> ----- Original Message -----
> From: "Jon Elson" <jmelson-at-artsci.wustl.edu>
> To: "Multiple recipients of list" <emc-at-nist.gov>
> Sent: Tuesday, May 16, 2000 3:07 PM
> Subject: Re: minimilltask V/S bridgeporttask
>
> >
> > Ahh, wonderful.  I have a variable speed drive that takes 0-10 V
> (unipolar)
> > signals for speed control.  So, when you are working on this, please think
> > about a .ini file parameter so that the speed can be made unipolar, and
> > the direction set with the forward and reverse bits.  Looking forward to
> this!
> >
>
> Something I've been seeing on most of the microprocessor controlled consumer
> electronics when they need a cheap-n-dirty analog voltage is to output a
> variable duty cycle frequency on a port pin and use an RC filter.  Of course
> the response time would not be that great but I think it would work fine for
> stuff like spindle speed control.  You could generate a 0-5v unipolar output
> with one parallel port bit, one resistor and one cap - and a few CPU cycles

Well, that really is no solution.  I already have 8 analog channels (14-bit
resolution), and am only using 3 of them so far.  This technique, with
appropriate hardware, makes one of the best and most accurate DAC's,
without the speed penalty.  R2R ladder DACs always suffer from variations
in the resistors, while a DAC built out of a counter and comparator will have
extremely good linearity (both differential and integral) with no laser
trimming.
This was used to make a 14-bit plus sign DAC on the old Allen-Bradley
7320 CNC control, and in other gear as well.  they only needed 100 Hz
bandwidth on the AB 7320, so even slow 74xx chips could do it fine.
To generate high resolution analog output with a parallel port bit would
burn up all the available CPU time, which is not an option.

This whole exchange is way off the point, anyway.  The original problem
is to get the speed info to the existing DAC, and then hook the variable
speed drive to it, which is pretty easy.  EMC contains the code to do
most of this, but some critical piece, probably about 5-10 lines maximum,
was left out during some reconfiguration several years ago.  This probably
happened when EMC was hacked up to run a Bridgeport Boss
machine with an air motor that shifted the variable speed drive mechanically.
That's still an option that should be available, but analog control of a
variable speed drive or DC spindle servo should also be available.

Whew!  All that just to explain what we are doing here.

Jon