On Sunday 10 January 2021 11:38:42 Jon Elson wrote:

> On 01/10/2021 01:36 AM, Gene Heskett wrote:
> > On Saturday 09 January 2021 20:57:24 Jon Elson wrote:
> >> On 01/09/2021 06:39 PM, Gene Heskett wrote:
> >>> Yes so idb drives it closer by overdriving it an "amount" so that
> >>> the friction locked point is closer to the desired point without
> >>> going over. The trick is to null the error without changing the
> >>> sign of the error. If the sign changes, then it oscillates, hard. 
> >>> So you can narrow the friction caused deadband by a small amount
> >>> of overdrive, but you also cannot change the sign of the error.
> >>
> >> Yes, that's the tricky part.  Any time you introduce a
> >> discontinuity into the transfer function,
> >> it makes things unstable at that point.
> >>
> >> Jon
> >
> > Thinking along a different line, I am running the pwmgen at 4khz,
> > mainly because my ears have a "carhart notch" at least 120 db deep
> > so theres not a chance in hell I'd ever hear it. What would happen
> > if I dropped the pwmgen frequency to 400 hz. that same percentage of
> > a pulse coming into the controller would have 10x the length of time
> > to actually move the motor, which may drive it closer to a null than
> > tickling it with a 1% pulse 4k times a second. The trouble with that
> > is that both pwmgens run at the same frequency, and Jon's servo,
> > driving the spindle, may trip its current limits with the longer
> > pulses.
> >
> > Jon?  Can you opine on that? Is there a minimum frequency for those
> > toroids?
>
> Well, at 50 KHz, which is what it was designed for, the
> motor current is always continuous, but could look like a
> sine wave with low inductance motors.  At 4 KHz, it is
> likely to already be discontinuous except maybe at high duty
> cycles.  So, it really won't make a difference to the
> toroids.  I suspect the
> motors will "sing", maybe pretty loudly, with 400 Hz PWM,
> but maybe you don't care.
>
> Yes, a long pulse with 400 Hz rep rate could allow current
> to build a lot.  But, in a PROPERLY tuned PID system, there
> should not be sudden excursions in pulse width.  Now, you
> have to watch out for
> excessive acceleration parameters in the .ini file.  My PWM
> servo amp has TWO current limits.
> The first (adjustable by resistor) limit just cuts off the
> PWM pulse early, and resets on the next PWM pulse.  The
> second limit (not adjustable) trips around 20 A and latches
> the servo amp into E-stop.

Requiring the pushbutton reset. With a nominally 17 amp limit set 
resistively, the other. triggering the shutdown can be counted on one, 
maybe 2 hands in all these years of using it for a spindle controller.

> Oh, now one issue is with the PWM frequency being LOWER than
> the servo period, that might make for some delay in updating
> the PWM generator.  Is this with a Pico controller or a
> Mesa?  I know my controller will immediately update the end
> of the PWM pulse if the width is changed in the middle of a
> cycle.  I can't say for Mesa.

mesa, and that possibility had not occurred to me, my bad. So I'd 
consider 2 khz as being the lowest workable minimum, and when the 17 amp 
kicks in I do hear a chirp from the iron in the motor so it obviously 
slower than the 4 khz pwm. And its possible its going to 100%. But IIRC 
theres a .98 max output set. But it is not set there. PWM_SCALE is 46 in 
this case. How does one figure that value to be 98% mathematically? 

> Jon
>
>
> _______________________________________________
> Emc-users mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/emc-users


Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
If we desire respect for the law, we must first make the law respectable.
 - Louis D. Brandeis
Genes Web page <http://geneslinuxbox.net:6309/gene>


_______________________________________________
Emc-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-users

Reply via email to