Dynomotion

Group: DynoMotion Message: 9248 From: phil loder Date: 2/22/2014
Subject: Kmotion step response not showing actual current or position.
I am running a kflop and kstep with 3 axis no feedback and am wanting to tune the motors.
When I test run the axis it shows command in blue as a bell curve and the red actual is flatlined at 0. also all 4 currents are flat lines at different levels.
The motors are running and sounding ok but I have no idea how well they are following the command. 
They all seem to run g code correctly in kmotion cnc. 
the motors are not mounted at this time, I want to get this method worked out before mounting.
Group: DynoMotion Message: 9249 From: Tom Kerekes Date: 2/22/2014
Subject: Re: Kmotion step response not showing actual current or position.
Hi Phil,

KSTEP is an open-loop micro-step pulse and direction type of drive.  There are no real-time current measurements sent back to KFLOP (as with a SnapAmp).

Basic open-loop steppers have no feedback so there is no means of plotting or knowing if or how well they are moving.

If you add encoders they can be wired back to KFLOP encoder inputs to plot actual position, detect following errors, stalls, loss of steps, correct errors,  etc... 

Regards
TK


Group: DynoMotion Message: 9250 From: phil loder Date: 2/22/2014
Subject: Re: Kmotion step response not showing actual current or position.
Tom, thanks for the info.
If you were tuning a open loop stepper running from the kstep, how would you do it?


On Saturday, February 22, 2014 8:55 PM, Tom Kerekes <tk@...> wrote:
 
Hi Phil,

KSTEP is an open-loop micro-step pulse and direction type of drive.  There are no real-time current measurements sent back to KFLOP (as with a SnapAmp).

Basic open-loop steppers have no feedback so there is no means of plotting or knowing if or how well they are moving.

If you add encoders they can be wired back to KFLOP encoder inputs to plot actual position, detect following errors, stalls, loss of steps, correct errors,  etc... 

Regards
TK


Group: DynoMotion Message: 9252 From: Tom Kerekes Date: 2/22/2014
Subject: Re: Kmotion step response not showing actual current or position.
Hi Phil,

I would basically push things until miss steps or stalls are observed then back off by a significant margin (30~50%).  Of course under load things will change.  You might want to mark a position on the motor shaft at a known position so you can tell if steps were lost by moving back to that reference position.  Use low Jerk and Acceleration on very long moves in order to determine the best case top speed.  Top useful speeds on size 23 steppers are usually under 1000RPM which would be 1000/60*200*16 ~ 50000 microsteps per second.  Then back off Velocity by some margin.

Then test acceleration.  Set Jerk to a large value (1000X Acceleration) so that acceleration is instant.  Increase acceleration to see where problems begin.  Then back off acceleration by some margin.

Then reduce Jerk to reduce the disturbance/shock caused by instant acceleration.   Plotting Velocity is a good way to see the effect of limiting Jerk.  For most systems the Jerk Limit should be a value about 20X the Acceleration Limit to that the Acceleration Torque is ramped up over about 1/20th of a second.  Faster than that is unlikely to reduce the disturbance significantly.  Slower than that will increase move times significantly.

Please note that some motions are performed limiting Jerk (3rd order motion) and some with unlimited Jerk (2nd order motion) so it is a good idea to characterize your system both ways.  You can temporarily simulate unlimited Jerk motion by setting the Jerk value to a huge value (1000X the acceleration).  Motions performed in KFLOP (MPG, Jogs, Homing, etc..) and Rapid G0 moves are 3rd order moves using KFLOP's parameters.  KMotionCNC coordinated motion moves (G1,G2,G3) are 2nd order moves using the Trajectory Planner parameters.

HTH
Regards
TK