Mesa stepgen motion profile and joint following error.

I'm using mesa 5i25 to control a yaskawa motor.
Joint following error often happen when motor accelerate and decelerate.
I made a quick test and see this happen because mesa motion profile (see below figure).


I want to leave this task for yaskawa drive so is there a way that i can change mesaprofile (to simple one) or lower jerk time?

1. The displayed following error (error proportional to velocity) is most likely caused by
incorrect FF1 tuning (or feedback scaling).

2. The acceleration mismatch is expected. This has nothing to do with step generation or hardware but
rather the fact the LinuxCNCs trajectory planner is a second order planner and has no jerk control.
That said, it looks like your drives have _WAY_ too much lag in acceleration
(they seem to have a almost 50 ms time constant) If the drives have any built in jerk limiting,
or command input low pass filtering, you may want to adjust these settings.

the feedback signal is taken from hm2_5i25.0.stepgen.00.position-fb not from the motor encoder. It's an open loop system but following error keep anoying me

There is no jerk limiting in the stepgen hardware or driver so I suspect you have either an
instrumentation issue or something odd in your hal file.

Can you post you hal/ini files?

Here is an example of a Mesa stepgen running a profile with 20 IPS velocity and
somewhat greater than 1/2G acceleration:



Notice that the residual stepgen errors are less than 10 uinch

Your plots show a large following error proportion to velocity which suggests that either

1. FF1 is not set properly,
2. You have a thread order issue
3. You do not have pid.N.error-previous-target set true
4. You are using a new LinuxCNC installation (with fixed PID) and you have the PIDs command-deriv connected

i thought there is an motion interpolation in mesa card that slow down accel, maybe large following error happens because my acceleration was so high and physic limit does not allow it.

If you are open loop then you should not get a following error unless the stepgen
acceleration and velocity limits do not have enough headroom vs your machine limits.

If the feedback is from encoders or the drive then it looks like your drive has some filtering
or jerk limiting applied.

The main ferror in your plot however is the velocity dependent error and this is likely caused
by a setup error in your hal file.