Man!!! I am not having any luck tuning my servos thru Galil.

I have the Galil Tools program with the scope. I can get good movement if I set the servo drives up in Velocity Mode, but in Torque Mode (recommended by Galil for positioning systems)  I get bad oscillations no matter how I setup the PID.

Anyone have any pointers? What are the best signals to watch in the scope? I have been using Reference Position vs Encoder Position and tuning till I get a nice crisp rise with a little overshoot and quick settling out. Have the movement set to 2000 increments in 300ms. I can get a good tune in Galil, then when I go to Mach and run the axis it may be ok moving fast but oscillates badly moving slowly - or vice versa.


Should I just leave it in Velocity Mode? It does seem to overshoot about .003 before settling in Velocity mode. Maybe I could tune that out, I haven't tried.

What are you using for PID numbers?


Mike

The best I could get it dialed into was around KD367 KP30 and if I add any KI at all it gets unstable. However without KI it comes to rest a little short of it's target position. I don't understand why I can't add any KI and still maintain any stability. It seems as if even a KI of 1 is too high.

With the above settings and KI set to 0 I can get it so it will rapid ok, but running a part program causes it to go into oscillation at some point in the run. I thought maybe I was trying to run the part program too fast so I dropped the feed down to a mere 10ipm but it still went into oscillation.  I did try using IL with a setting around 2.5 but that didn't help any. Less then 2.5 rendered KI ineffective.

I also tried 328, 19, 1 and several other variants.

what do you think?

I am not sure where to go with this but these videos from Galil may give you some insight.

Mike


http://www.galilmc.com/learning/tutorials.php#tuning

what type of servos are you currently using if there are digital drives it will be best to run them in stepper mode but if you have to run the drives in velocity mode the PIDs from the drive will fight the PIDs from the galil card and that means tuning Will have to be done by setting the digital drive PIDs low and fine tuning with the Galil card sounds simple but its a real "PITA" also auto
doesn't work well that way so setting the K P I values one at a time

The drives are Danaher model 20360. I don't know if they are "digital". I never noticed any place in the setup (software) where you would do PID tuning. I know they have a way of telling what motor is connected to them and that pre-configures the drive to run best with that motor. I think that only sets up the motor specs, like how many poles, motor voltage, things like that. There is no "autotuning" feature like some of the drives have. I suppose I could run them in step/direction mode, but isn't that kinda the same end result as running in velocity mode? In velocity mode the drive is doing most of the control and the Galil just makes sure it gets where it's supposed to be? In step mode the drive is doing all the control? I'm not sure I understand the diference between velocity and torque mode.

I know the Galil PID settings are nowhere near as touchy when the drives are in velocity mode. i can find a large range of settings that seem to work ok in velocity mode. In torque mode a small PID change makes a big difference. In velocity mode I still can't add any KI though.

In the Galil Tools program I can watch the scope and change the PID settings and the motor will make a move each time which in turn shows on the scope. So, what should the correct wave form look like? i know it should shoot up to position as quickly as possible and settle. But how can i prevent the motor from going in oscillation (which happens when I go back to Mach and try to run a G Code file). What should I be looking for? How should the motor sound? Should there be any overshoot before it settles? should I be looking at anything other then Reference Position vs Encoder Position?

is it possible the vibration from the drive belt teeth riding on a flat pulley combined with poor tuning could be causing the system to go into oscillation? That's one thing I was considering.

Thanks