Hello Everybody, what Sebastian so eloquently described is what I originally meant, however what the other people that replied added is also relevant I think. The linear velocity phase results in a constant acceleration that will have a Dirac delta function in its first order derivative, and that is not desirable, due to the frequencies introduced.
I must admit that my understanding of DC motor internals is somewhat buried, also the last time I implemented a state control is 15y ago...So I ll not follow up on these suggestions, although they would probably tackle the problem at the root. Many years ago I implemented a sin^2(x) acceleration profile that was jerk free, I wondered if anybody else had implemented such a thing in the linuxcnc's TP. But using the feed forward terms is a very nice suggestion from Eric, ty for that, I ll do that today. BR Max. Am 22.03.2017 um 20:03 schrieb Sebastian Kuzminsky: > On 03/22/2017 12:50 PM, Nicklas Karlsson wrote: >>> Hello, >>> >>> I want to ask if anybody has worked on different acceleration profiles >>> for linuxcnc ? I am looking for something that would remove the >>> discontinuity in the first derivative of the acceleration ? >>> >>> BR >>> Max. >> >> I guess to start with the DC motor model for example here >> https://www.electrical4u.com/torque-equation-of-dc-motor/ would be a good >> idea. Driving this with a limited voltage and current would probably be a >> good model for the limitations. > > I think Max is asking about changes to the LinuxCNC trajectory planner > that implement variable (non-constant) acceleration, aka "limited jerk". > > Currently when you start from a stand-still, command a move, and come to > rest at the end of the move, LinuxCNC implements a trapezoidal velocity > profile. This starts with a constant-acceleration phase until it > reaches the maximum speed for the move, then enters a constant-velocity, > zero-acceleration "cruise phase", and finishes with a > constant-deceleration phase back to zero velocity. Each phase > transition has an instant of infinite jerk, when acceleration goes from > one constant value to another in a single motion cycle. > > I think Yishin Li (https://github.com/yishinli) has done some work in > the area of jerk-limited trajectory planning, but AFAIK it has not shown > up as a pull request to our repo yet. > > ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot _______________________________________________ Emc-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-users
