CNC Lathe - X AXIS "losing" POSITION

Well gents, it has gotten to the point of desperation where I have come here, for the first time, looking for some direction.

We run a 35 year old Warner & Swasey Titan CNC lathe and have been having an ongoing problem for the last three months with the x-axis losing it's position. Both my operators and millwrights are just about ready to waive the white flag... yes, I know this is an old machine and it has produced hundreds of thousands of parts and has more than paid for itself at this point... but with few capital expense dollars coming down the pipeline, we need to find a solution.

The problem.... the lower turret on the x-axis seems to have a mind of its own, and the occurrence is unpredictable. We machine external male threads onto steel alloy. So, at the start of the shift, if the x-axis offsets are at 0.0, after a few hours of running the problem starts... the operator will check his dimensions and he will be 0.015" oversize, and adjust his offset -0.015" accordingly. As the day goes on, the frequency of this size deviation seems to occur more often, however it seems to (almost exclusively) go in the direction of creating an oversize part, i.e. the turret isn't travelling far enough. The other day, throughout the course of the day, my operator had to move his offsets nearly -0.100"! Thankfully, these movements (right now) aren't large enough, and are in the direction of an 'under travel', so concern over a machine crash isn't high right now.

My millwrights have checked everything mechanical: ball screw, linear scale, backlash, etc, etc.... I am beginning to think this is an electrical problem.... because... when it happened today, I had the operator shut off the machine for 30 minutes... when he turned it back on, it returned to making an accurate part.

Now I am no engineer or technician... so this may be a stupid question... is it possible that one of the electrical connections somewhere in the encoder/servo motor/linear scale loop is starting to 'fray' inside the cable and as production time goes on, the heat from the electric current is causing the wires to intermittently lose connection, and thus we are losing electric pulses somewhere?

If the turret is not traveling far enough then that would mean the the computer thinks that it got enough pulses to be in position, when it actually has not reached position and thus is seeing phantom pulses. Since your maintenance staff have pretty much eliminated mechanical issues as a possibility, the most likely cause is electrical.

This can happen with a bad encoder, a loose connection, or electrical noise due to poor shielding. In this case since it's been working for 35 years, worn cables or a bad encoder is a good possibility. An oscilloscope is the easiest way to track this down. Look at the encoder inputs at the computer end. Ideal is a nice crisp 5V square wave, if it is ''fuzzy'' then it's time to dig a bit deeper.

The fact that is seems to go back to normal after a shutdown would indicate that something is heating up over time and goes back to normal after a short shutdown. Maybe a capacitor in the encoder?

Hey WStitan,

I'm actually looking for one of these machines. Any interest in selling?