Merry Moulder Moved A Mass Of Metal

Hi guys!

So I've got these vertical grinders, four of them, with pretty heavy mechanical backlash in the drive screws on the Y axis (this being the axis that moves them back and forth towards the material being worked).

The recurring problem on these is that the load on the y-axis weighs somewhere on the order of a small buick. The servo driving the screw loads up and breaks the load loose, moves it, then overshoots before coming to a halt. Sensing this, it tries to drag the load back to target, resulting in the exact same occurence in the opposite direction. We've lubricated the bejeezus out of the slides for this axis, and the screw is packed with grease, all to reduce the amount of force required to break the load loose.

But now another thought occurs - inertia. Given the mass involved, we may be building up so much force in the break-away that no amount of effort will stop it on a dime, particularly with the lousy drive mechanism in place.

I wonder if, given the circumstances, is it possible we've got TOO MUCH lubrication? Maybe we need a little drag on this thing to reduce the overshoot?

Any thoughts?

TM

Are you sure all the slides and are inline? Does the problem only happen when there is material on the slide or does it do it when empty?

I would probably start with getting it to work with no material.
To me it sounds like the slides might be binding a little. The backlash in the screw shouldn't effect it much unless the slack allows the screw to buiild enough momentum the when the slack is gone it causes the load to "jump". That "jump" is beyond desired movement so it goes the other way with the same conditions and same results.


Just my thoughts.

Drewcrew6

Tim (ya little Irish ****!)...

I would definitely say there is an inertia issue.

Are you using a "canned routine" for positioning? What is it?

What are you using to control/detect "positioning"?
... limits? encoder? LVDT?

Is it the case that you are only either accelerating or decelerating?

Is there any time that the speed is constant?

How fast and how far are you trying to move it?

If your process is constantly aware of the position then you can use a "look ahead" to anticipate your slow-down and stop points. Of course, that would require that you start with a new screw with minimum lash.

I could describe a half-dozen ways to do this... but the appropriate method depends on your answers to the previous questions.

What is the bearing material in the slide? Do you have a better weight estimate than "a small buick." You seem to have a lot of "stiction", that really shouldn't be there.

Another thing could be is that the servos are not properly tuned. If your gains aren't high enough, your slide won't move until the integral gain winds up and breaks the slide free.

A lot of servo systems can compensate for backlash as long at the system is stiff enough. If you accelerate gently enough, but with enough force, you can keep the load at one side of the backlash, and you only have to "let go" when you decel and the load transfers to the other side of the lash. Sometimes positioning involves stopping short of your target and creeping up on the final position to always approach from the same direction.

What kind of grinders are these? Are they machine tools or something else?

What kind of screw? Ball or Acme? Can you put another nut on the screw and set up a pre-load to remove the backlash? Have the grinders always acted like this or is it an age thing?

Hi again!

I'll dig into these questions and get back with you in a few days. Too many sticks in the fire

Thanks!

TM