What's new
What's new

Ball Screw Troubleshooting

Dynatect bought Lead Screw International in Traverse City Michigan. I was in there about a month ago and they have some interesting machines. The link below is a good power point that explains how ball screws are pre-loaded and how to check for lost motion, bad thrust bearings, lose bearing capture, stretching ball screws, and checking alignment.

I deleted the backlash program screens because they were overkill. I write a simple program and use single block to move an axis 2 inches, set an indicator to zero, move .004" in same direction, move opposite direction .004" and the deviation from zero is what you need to set the backlash comp. Then I move 2 inches back to start point and repeat after setting backlash comp. I have found if it is a metric screw the comp is .001mm and inch screws are .0001".

The reason for stretching a ball screw is to help compensate for thermal growth when a manufacturer doesn't use a separate detect unit like a scale on an axis.

Ballscrew Inspection
 
Another note, if a new ball screw is installed, zero out the backlash comp for that axis or it will over compensate.:)

A linear guide axis will most likely have zero backlash and a box way slide will have a couple tenths even with a new screw and bearings if the gibs are adjusted properly.
 
Another note, if a new ball screw is installed, zero out the backlash comp for that axis or it will over compensate.:)

A linear guide axis will most likely have zero backlash and a box way slide will have a couple tenths even with a new screw and bearings if the gibs are adjusted properly.

One thing that might help is to take a straight bar, clamp it to the bed perpendicular to the axis of travel, then set up two tenths indicators at some distance from each other on the machine Z housing. Could be two mag stands or a bar with post for toolholder mounting and holes/set screws for the indicators.

Then once you've got backlash set where you want (using one indicator), you can set two up, jog a couple thousandths back and forth, and see it the two read the same at both extents. Same reading - good gibs/linear bearings. Different readings - gibs worn or out of adjustment, or bad bearing blocks. You can move the clamped bar to check other parts of the slideway.

If you've got a machine with great kinematics such that even a worn slideway shows the same readings, a backup is just to take the table and try rocking it as a second check. Might be easier to do with two people. Then you have to ensure you're not getting Y movement if reading off X (or vice versa).

Sorry if the pros have better ways to do this, just my thought on it...
 








 
Back
Top