On 5/16/25 11:32, Leonardo Marsaglia wrote:
Hi guys, hope you're doing well!
I'm pretty much successfully using external offsets for some of the parts
we manufacture on the lathe.
But today, while the spindle was turning doing an eccentric I got a
following error at the moment I went from MDI to Manual control. I'm using
servo motors but with mesa stepgen in velocity mode with it's internal PID.
I guess this is a configuration problem.
Could also be hdwe. I, sevreral years ago, converted my Sheldon to be
driven the then new 3phase closed loop stepper/servo's. Bare in mind
also that my controller is not a wintel box, but is an rpi4b. The
individual controller in these has a decoder that is fed from an encoder
on the rear of the motor, putting the PID's in linuxcnc out of a job
because the PID is in the controller, giving truly instant control, so
my stepgens are driven directly from the trajectory planner and there is
no feedback to linuxcnc. That change plus around 45 volts to the
controller makes the motor control at least 2x faster than pure
steppers. Because the error signal also instantly controls the motor
current, feeding the motor no more than it takes to null the error, the
motors run 70C cooler and the workpiece shows nothing but a mirror finish.
I am now using this same motor tech several other places, and am slowly
replacing the steppers with these stepper/servo's, including in my 3d
printers. The printers run 10x faster than OOTB, with much improved
accuracy. Loss of a step is totally historical, it just doesn't
happen. Ever. If the error is that big, the motor current rises to move
he motor to reduce the error. Hanpose drivers can work with up to 90
volts for nema 17 motors, pulling the motor into position in
microseconds but that overdrive is so short. there is not any heating to
dangerous levels. I made some rotary axis things I carve screws with, a
3NM motor driving a 5/1 worm, 500 revs output speed is 2500 revs on a
nema 23 motor, 2 hours of runtime per hard maple screw, the motor is
warm but not burn you hand hot.
But a warning, only motors with optical encoders. the smaller "42C"
motors have a hall effect encoder, which can be just as accurate when
stopped but take far longer to do the AD conversions, so they only move
AFTER that AD conversion is complete, rendering what should be slow
smooth motions as sawtooth shaped motions and parts printed will have
shingles superimposed on the finish. Trust me you do NOT want that.
I'm using LCNC PRE 2.8 version on this lathe because it works. But I
suspect this is a config problem more than a bug.
Thanks a lot!
Check your feedback paths, velocity and position modes are conflicting
critters.
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users
.
Cheers, Gene Heskett, CET.
--
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author, 1940)
If we desire respect for the law, we must first make the law respectable.
- Louis D. Brandeis
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users
