LinuxCNC S-Curve Accelerations

automata wrote: Hi,

It should be noted that, the time to execute the acceleration/deceleration part of the motion segment will not change. Hence the overall trapezoidal trajectory planner need not be changed. Only the playout part of the acceleration and deceleration sections need to incorporate the jerk limitation parts.


-automata

I am not a motion control expert by any means..
I don't understand how what you are describing is possible.

As I understand it:
linuxcnc plans with no jerk limiting so in other words maximum acceleration.
If you add jerk limiting after planning then the acceleration profile must take longer, undoing all the planning.

The only way around that, that I see is to have linuxcnc initially plan with less then maximum acceleration.
This would give you head room to add jerk limiting on the ends and add more acceleration in the middle so the acceleration
profile ends up taking the same time.

So then I would imagine the percentage setting might represent how much over the INI acceleration setting jerk limiting would go to
maintain this acceleration time length.
ie 10% Jerk setting would add as much jerk limiting as you can while using 110% INI acceleration setting.

Is this roughly what you intend to try?
Chris M

Chris M: Yes that is exactly how jerk percentage works. You violate the Max acceleration constraint on the overall path. However, the ratio of the max_a in trapezoidal plan to max_a in jerkpercent plan is related by the formula
max_a_jerk_percent = max_a_trapezoidal*100/(1-jerk_percent*0.005). So you can also limit the max_a_trapezoidal taking into consideration the jerk percentage.

If jerk percentage is kept at 100%, which means the zero jerk time in the figure ( support.motioneng.com/Software-MPI_04_00...n_ty_jerkpercent.gif ) is 0, the max_a_jp = 2*max_a_trapezoidal.

One more point to note is that with jerk control enabled, machines are able to sustain much higher accelerations.

Then again, the jerk percent can also be set to 0, to retain same behaviour as the current planner.

-automata

That sounds very interesting.
IIRC the developers that did the cubic planner was not so interested in jerk limiting because he felt the acceleration curve would be slower (because of ramping time) on most machines that were low powered.
But I also remember the guy working on jerk limiting saying that he could increase acceleration with jerk limiting above what was possible without.
Of course the way you are planning on doing it, one could try either situation.
I'm rooting for you!

Chris M

It would be really awesome to have a Jerk control, i'm available to do any testing you may need on real actual machines, i have machines with steppers and with brush less motors, 3 and 5 axis, other than robot kinematics i can test almost anything else

I was thinking that using s curve in our jogging planner would probably be a good project too.
rapid acceleration is really the name of the game when jogging slowly.
It would surely be easier to to do then the full motion controller.

Chris M

Did anyone have a further look into this?