c2 smooth velocity profile
Trying different trajectory models is a whole new can of worms and IMHO best POCed and developed outside of linuxcnc. I personally like scipy for this kind of stuff.
I can't really help with git or running development branches, but these things are covered elsewhere.
I didn't pay attention and upgraded all my machines to ubuntu 20.04, now I'm a bit in a pickle regarding linuxcnc development... probably will need to setup a VM with debian buster.
If somebody wants to look at it, I think the first thing to do would be to rebase it onto recent master and look for acceleration excursions in chips.ngc. On the "theoretical" side, jogging (should be easy) and feed override (probably some tedious math involved) need to be figured out.
rmu wrote: I didn't pay attention and upgraded all my machines to ubuntu 20.04, now I'm a bit in a pickle regarding linuxcnc development... probably will need to setup a VM with debian buster.
This should work for Ubuntu 20.04 also, might want to give it a try.
rmu wrote: Just a warning, it does exceed acceleration limits while it should not, and the linuxcnc version it is based on is outdated. Do not run on something that moves..
I had thought it might be possible to do something similar to the way external offsets is managed. That is allocate a portion of max acceleration (say 20%) to the s curve which reduces the machine's acceleration to a lower limit (80%). Maybe that override can be calculated?