Hello, I'm adding support for a Kossel Pro delta printer with a BBB+CRAMPS controller.
For the extruder PWM temperature control, I used PID values from BeBoPr lineardelta hal files: setp pid.0.Pgain 0.30 setp pid.0.Igain 0.00001 setp pid.0.Dgain 0.9375 I noticed that the temperature overshoots the set point and was wondering why. Here's an example of data I collected after setting the temp to 60 degrees: while [ 1 ]; do halcmd show | egrep 'pid.0.command |pid.0.error |pid.0.feedback |pid.0.output'; echo =========================; sleep 1; done 556 557 float IN 60 pid.0.command 0.000010 0 <== e0.temp.set 556 557 float OUT 33.6 pid.0.error 0.000010 0 556 557 float IN 26.4 pid.0.feedback 0.000010 0 <== e0.temp.meas 556 557 float OUT 10.58001 pid.0.output 0.000010 0 ==> e0.heater ========================= 556 557 float IN 60 pid.0.command 0.000010 0 <== e0.temp.set 556 557 float OUT 27.7 pid.0.error 0.000010 0 556 557 float IN 32.3 pid.0.feedback 0.000010 0 <== e0.temp.meas 556 557 float OUT 8.81001 pid.0.output 0.000010 0 ==> e0.heater ========================= 556 557 float IN 60 pid.0.command 0.000010 0 <== e0.temp.set 556 557 float OUT 17.9 pid.0.error 0.000010 0 556 557 float IN 42.1 pid.0.feedback 0.000010 0 <== e0.temp.meas 556 557 float OUT 5.87001 pid.0.output 0.000010 0 ==> e0.heater ========================= 556 557 float IN 60 pid.0.command 0.000010 0 <== e0.temp.set 556 557 float OUT 6.6 pid.0.error 0.000010 0 556 557 float IN 53.4 pid.0.feedback 0.000010 0 <== e0.temp.meas 556 557 float OUT 2.48001 pid.0.output 0.000010 0 ==> e0.heater ========================= 556 557 float IN 60 pid.0.command 0.000010 0 <== e0.temp.set 556 557 float OUT -4.9 pid.0.error 0.000010 0 556 557 float IN 64.9 pid.0.feedback 0.000010 0 <== e0.temp.meas 556 557 float OUT -0.9700063 pid.0.output 0.000010 0 ==> e0.heater ... I tried modifying PID values after reading this: https://github.com/machinekoder/machinekit/wiki/pid_theory_es , but I still saw overshoot to about 77 degrees. As an experiment, I tried deleting some signals and setting the pwmgen value manually: halcmd delsig pid.0.output halcmd delsig limit1.0.out halcmd delsig e0.heaterl halcmd setp hpg.pwmgen.00.out.01.value 0.2 That gave a slower rise in temperature which made me wonder if the pid.0.output needed to be scaled to a percentage, suitable for input to pwmgen. I noticed that there's a pwmgen->scale value in the machinekit code src/hal/components/pwmgen.c and the description http://linuxcnc.org/docs/html/hal/rtcomps.html#_pwmgen "Scaling factor to convert *value* from arbitrary units to duty cycle." sounded like what I wanted, but hal_pru_generic/pwmgen does not have a scale parameter. I tried adding "setp hpg.pwmgen.00.out.01.scale 100" to my .hal and the temperature rose very slowly and did not overshoot, but does not seem to get all the way to temp. I chose a scale of "100" thinking that 100 degrees below the set point might need full, 100% pwm power. Is this the right way to produce less overshoot or does hal_pru_generic's pwmgen have a scale similar to the non-PRU pwmgen? Thanks! -- website: http://www.machinekit.io blog: http://blog.machinekit.io github: https://github.com/machinekit --- You received this message because you are subscribed to the Google Groups "Machinekit" group. To unsubscribe from this group and stop receiving emails from it, send an email to machinekit+unsubscr...@googlegroups.com. Visit this group at https://groups.google.com/group/machinekit. For more options, visit https://groups.google.com/d/optout.
