Re: [CAD_CAM_EDM_DRO] Re: CNC Controller Software

><snip>

I have not read the entire thread (yet), so this may have been
discussed. I have found that a clean pulse stream is extremely important
when trying to get all the performance out of your steppers, especially
with marginally sized ones. If you think about it, a pulse that comes a
little sooner than expected is asking the stepper to instantly accelerate,
and one that comes a little later than expected is asking the motor to
instantly decelerate. At lower RPMs, where you have more torque available,
that is not as important. But as you increase the speed of the stepper,
your torque falls off, and you don't have as much to waste by pulse
jitter. It seems to me that the gecko pulse multiplier helps smooth out
pulse jitter. In my experiments it seems like I can get better performance
out of steppers with software indexers and the pulse multiplier, but when I
asked Mariss, he said it should not make any difference. Maybe he didn't
understand my question, but I didn't want to take up too much of his time.

>Many people are now using step and direction controlled DC servo
>systems. These are capable of much higher resolutions and rotational
>speeds than stepper systems. When you use an encoder at these higher
>rates, it is imperative that you have a clean, fast pulse stream, and
>that you can output steps to match the enhanced resolutions available.

No doubt hardware can give you a faster pulse stream than software. I was
all too painfully aware on my x-y table, using surplus Compumotor M drives,
that have 25,000 steps/rev. That's 125,000 steps/inch! It made it really
slow. I don't want to start an argument, but it's been my experience that
with servos it's not nearly important to have a 'clean' pulse stream. The
PID (or PD, or PI, or whatever) loop inside the drive smooths out the pulse
raggedness. Just like the pulse multiplier option for geckos.

- Steven Ciciora

>Fred Smith - IMService
>Discounts and specials are at:
>http://www.cadcamcadcam.com/hobby
>