Re: stepper pulse rate idea

[jmkasunich-at-ra.rockwell.com]

> Jon Elson wrote:
>
>> jmkasunich-at-ra.rockwell.com wrote:
>>
>> Ray wrote:
>>
>>>Those of you who have applied the EMC to steppers that run
>>>from the parport know that you have to design into the machine
>>>a compromise between speed and resolution.
>>>
>>>This compromise is fixed in part by Jon's PPMC stepper rate
>>>generator card.
>>>
>>>
>In part?  What part is NOT fixed by it?

The money part.

>You get 300,000 steps/second, extremely smooth step pulse trains,
>including the accel./decel. part. The only thing is, hardware
>doesn't come free.

And there you have it - $200.  Your board avoids the speed/resolution
compromise by adding cost.  If highest performance is your priority,
then spend the $200 and have at it.

Ray's comments were targeted at people who want to run open loop with
a parallel port and the smallest, cheapest external hardware possible.
They don't need 300KHz, and may be happy with 20KHz, or 40KHz, or
whatever.

There is space in the world for lots of different price/performance
tradeoffs, and Ray simply proposed one that fits in the gap between
pure software step and direction and high performance external
hardware.  I'm all for more options, even if they aren't what I
would personally choose.

in another message, Jon wrote:

>jmkasunich wrote:
>>
>> 1)  Instead of setting the step bit in one pass of the
>> code and clearing it in the next, set and clear it in
>> the same pass.  This would result in a very short step
>> pulse (1uS or less).  Then use an external one-shot to
>> stretch the pulse enough to make the drive happy.
>
>
> Or, put an edge detector (XOR gate and an RC delay for
> one input) and a one shot on the step signal.  Any change,
> of either polarity, causes a step pulse of whatever length
> you set the one shot for.
>
> This would require the most trivial change to the software,
> and would require only one I/O per step.

Yes!  That's the kind of input I was looking for.

Paul wrote:

> With freqmod it is already possible to output a quadrature
> signal by setting STEPPING_TYPE to 2.

Cool!

> A number of other binning algorithms are in place for various
> types of output. These are :
>
> Type 0 - The standard step/dir freqmod output.
> Type 1 - A two phase grey code output for interfacing
>   directly to Bridgeport stepper drives.
> Type 2 - A quadrature output for The Motion Group drives
> Type 3 - A four phase output for Taig and Microproto packages.
> Type 4 - An implimentation of the original steppermod algorithim.

How practical would it be to add:

Type 5 - Output one transition on STEP for each step desired,
for use with the edge detector scheme that Jon came up with?

Also, what's the difference between Type 1 (two phase grey code)
and Type 4 (quadrature)?  Seems to me two bit grey code and
quadrature are two names for the exact same signals.


John Kasunich