[CarbideBob]

Could you expand on your experience? I'd like to add a scale and dual closed loop. I've read up on the two methods explained in the Galil literature.
Karl

Unfortunately I can't disclose some of the details in our dual loop system as we hold them as "trade secrets" and our competitors are still trying to figure out what we have done.

Most importantly do everything you can to remove any backlash in the system, otherwise you will have to "soften" the loop tuning up so badly that you'll wonder why you tried this method.

Another thing that helps is to locate the reader head as close as possible to the ballnut. This helps remove instabilities resulting from abbe error (twisting of the table).

Be warned that this will appear to be tuned ok with an unloaded axis but have problems when cutting. Be sure to apply a load by pushing on the table when setting up your tuning.
Also make sure you try a wide range of speeds when tuning. You may also encounter problems with high decell rates if you have very free slides as the screw transfers from tension to compression. We use a gain break which changes the tuning parameters at higher speeds and standstill.

Very close inertia matching of the motor and reflected load helps quite a bit. It's a lot of math but it's best to calculate the values for the PID loop before starting the tuning. If you don't have a copy I'd recommend Dr. Jacob Tal's "Step-by-Step Design of Motion Control Systems" (he's the founder of Galil).


On the original posters question on the sine wave scales it it very easy to reshape and change the sine wave to square waves with an op amp but you may not like the resolution that you get.
Hiedenhain and Mitutoyoboth both incorporate interpolation units which chop up the sine wave into little pieces to give the final resolution. This is why their boxes cost so much.
Bob