On Fri, May 11, 2018, at 11:16 PM, Chris Albertson wrote: > I looked at the unit. It seems like a good idea. It uses inputs from > both the motor's shaft encoders and another encoder on the object that is > being moved, like a linear encoder on the table. > > Question: Let's say I wanted to do this myself. Is there a method that > "everyone" in the machine tool industry uses for combining the reading of > multiple encoders? If not it seems like the perfect application for a > Kalman filter. > > But maybe you don't combine them but use the linear DRO for position loop > and the motor shaft encoders for velocity. > > This is a common problem I think with robot arms. The joint has an angle > sensor but the motor has a shaft encoder. So the control loops might be > nested. > > It reminds my the old saying the "A man with a watch knows what time it is, > a man with two watches is never sure of the time." >
We did something like this several years ago at Stuart's shop in Wichita, on a big Giddings and Lewis boring mill. We used two PID loops, with their outputs summed. The position command went to both loops. The feedback for one loop came from the motor encoder, and the feedback for the other loop came from the linear scale. The motor loop was tuned as normal, except that the I-gain was kept at zero. The linear scale loop was tuned using ONLY I-gain. So the linear scale loop corrected the fairly small steady-state errors due to things like the lead screw heating up (10 foot long screw, it adds up). It also compensates for backlash in the screw, and if there is much of that it leads to disturbances on direction reversal. Isn't going to fix a clapped out machine, but can improve the accuracy of a tight machine. -- John Kasunich [email protected] ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot _______________________________________________ Emc-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-users
