Looking at it again you're on the right track if you really want to use the R-0-F thing. It's pretty hard to help you with this because of the way your files are written. Your expectations of what pid is supposed to do is probably off. What you saw happen with the pid output is exactly what should have happened given the circumstances.
(pid Igain = 0, Bias = 60, Pgain = 1, scale = 120, min =-115, max = 115) > > ask M4 S3300 = 5% pwm so value are 4.95 > ask M4 S1 = ~50% pwm so value are 60 > ask M3 S1 = ~50% pwm so value are 60 > ask M3 S3300 = 95% so value are 114.95 > *Pgain *is the *Propotional* change to the output vs the input. *Igain* is the *Integral* response of time vs the Error. Here you ask for a speed of 3300rpms and since there is no encoder it always sees 0rpm. The output of the PID (supposed to be controlling a motor) has to increase to try to get the input (supposed to be an encoder attached to the motor) to increase. This is the whole point of "closed loop'" control. Since you have a proportional gain that is much higher than the integral gain, it just shoots to the max and stays there since the pid output cannot satisfy the input (no motor, no encoder). What you have here is large changes made with no time consideration. the problem come when i change Pgain = 0 and Igain = 0.1 or 1 etc > > ask M4 S3300 = 0% pwm so value decrease quickly to -120 > ask M4 S1 = ~50% pwm so value are 60 but decrease slowly to -120 > ask M3 S1 = ~50% pwm so value are 60 increase slowly to 120 > ask M3 S3300 = 100% so value are 120 > This is not a problem, this is what you are asking for. You have no, or very little proportional gain set so the output will not act aggressively to the input, it will make very small changes incrementally. Couple that with the fact that you have now included Integral (time) gain to the pid. What you wound up with here is that it is making very small changes slowly. This is backwards from how you tune a PID but it doesn't matter because you have nothing to feedback and close the loop. The only thing you can get from this is what the minimum and maximum outputs were, but that's not really relevant because you're limiting the pid. Because of the way you have the hal/ini files written it took me a while to realize what you were trying to do. I decided to write it myself because I enjoy a challenge, which this certainly is because of the hardware you chose to use. It's funny because by the time I was done I realized you were really on the right track. Do yourself a favor and don't worry about the low pass filter and the 5%-95% thing until after you have a working spindle. Your hal file is super hard to read and confusing....and I look at hal files all day. Your main problem I think is using bias of pid for the offset. This is written like a legacy (linuxCNC) hal file but it works fine with MK. You can change the loadrt's and addf's to newinst if you want. Double check for errors because I can't load these files since I don't use the same hardware. ##### HAL FIle ########################### # Spindle ########################### loadrt pid names=pid.spindle loadrt offset names=offset.pwmgen loadrt near names=near.at-speed loadrt scale names=scale.spindle-vel-rpm addf pid.spindle.do-pid-calcs servo-thread addf offset.pwmgen.update-output.funct servo-thread addf near.at-speed servo-thread addf scale.spindle-speed-rpm servo-thread # scale RPS to RPM setp scale.spindle-vel-rpm.gain [SPINDLE]VELOCITY_RPM_GAIN # setup Spindle PID setp pid.spindle.Pgain [SPINDLE]P setp pid.spindle.Igain [SPINDLE]I setp pid.spindle.Dgain [SPINDLE]D setp pid.spindle.FF0 [SPINDLE]FF0 setp pid.spindle.FF1 [SPINDLE]FF1 setp pid.spindle.FF2 [SPINDLE]FF2 setp pid.spindle.bias [SPINDLE]BIAS setp pid.spindle.deadband [SPINDLE]DEADBAND setp pid.spindle.maxerror [SPINDLE]MAX_ERROR setp pid.spindle.maxoutput [SPINDLE]MAX_OUTPUT setp hpg.encoder.00.chan.00.scale [SPINDLE]ENCODER_SCALE setp hpg.encoder.00.chan.00.counter-mode 0 setp hpg.pwmgen.00.out.00.pwm_period [SPINDLE]PWM_PERIOD setp near.at-speed.scale [SPINDLE]AT_SPEED_SCALE # offset the pid output the value of max output to shift the output above 0 setp offset.pwmgen.offset [SPINDLE]MAX_OUTPUT net DAC-ctrl-in <= pid.spindle.out net DAC-ctrl-in => offset.pwmgen.in # scale max expected input * 2 to equal a PWM duty cycle of 1.0 setp hpg.pwmgen.00.out.00.scale [SPINDLE]PWM_OUTPUT_SCALE # set the pwmgen output pin setp hpg.pwmgen.00.out.00.pin 921 # encoder connections net spindle-pos <= hpg.encoder.00.chan.00.position net spindle-pos => motion.spindle-revs net spindle-vel-fb-rps <= hpg.encoder.00.chan.00.velocity net spindle-vel-fb-rps => motion.spindle-speed-in net spindle-vel-fb-rps => scale.spindle-speed-rpm.in net spindle-vel-scaled-rpm <= scale.spindle-speed-rpm # Speed command in RPS net spindle-vel-cmd-rps <= motion.spindle-speed-out-rps net spindle-vel-cmd-rps => pid.spindle.command # offset output to feed to pwmgen output net DAC-ctrl-out <= offset.pwmgen.out net DAC-ctrl-out => hpg.pwmgen.00.out.00.value # Spindle control signals net spindle-enable <= motion.spindle-on net spindle-enable => pid.spindle.enable net spindle-enable => fur.do-6 net spindle-vel-cmd-rps <= motion.spindle-speed-out-rps net spindle-vel-cmd-rps => near.at-speed.in1 net spindle-vel-fb-rps => near.at-speed.in2 net spindle-at-speed <= near.at-speed.out net spindle-at-speed => motion.spindle-at-speed ##### INI File ############# # SPINDLE ############# [SPINDLE] # PID, input is RPS OUTPUT_PERIOD = 10000000 P = 0.1 I = 0.0 D = 0.0 FF0 = 0.0 FF1 = 0.0 FF2 = 0.0 BIAS = 0.0 DEADBAND = 0.0 MAX_ERROR = 0.0 #Max desired spindle speed in RPMs (3300) / 60 = Spindle Speed in RPS. MAX_OUTPUT = 55.00 # quadrature mode (mode-0) is encoder PPR * 4 ENCODER_SCALE = 1600 # scale pid output to pwm control value. Max duty cycle of 1.0 / (Max speed * 2) ---range of both directions PWM_OUTPUT_SCALE = 110 # PWMgen output frequency in Hz PWM_PERIOD = 10000000 AT_SPEED_SCALE = 1.01 VELOCITY_RPM_GAIN = 60 -- 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. To view this discussion on the web visit https://groups.google.com/d/msgid/machinekit/bb111ed2-ed85-40f7-afea-fd3a6b57eee6%40googlegroups.com.
