Electrically preloaded Dual motor drive for R&P

raadu wrote:

I was wondering if LinuxCNC has the capability to drive 2 stepper motors per side of X axis ( 2 pinions per rack for reduced backlash, electrically preloaded ).

It isn't something built-in, and I am not sure that it would work all that well.

If I was building a big machine I think I would try copying the Bell-Everman "Servobelt" arrangement. Or I might try double helical rack-and-pinion displaced sideways to take up backlash.

It is possibly worth mentioning that I think I have heard of systems that use opposed motors to achieve backlash compensation.

I think it would be done by asymmetric PID controllers, with the asymmetry switching depending on direction of motion.
The problem is likely to be that the direction that the cutting force acts is not necessarily the direction in which the axis is moving. This is the problem with software backlash comensation, and the dual-motor arrangement is likely to suffer the same problem.

With servo control you might be able to have a small bias current to take up the slack, then a slow I-term in a PID controller to adapt to the encoder-counts offset between the pinions. Any change in counts-offset could be seen as a "backlash rattle" and counteracted.

I can't see a workable way to do it open-loop with stepper motors, though.

raadu wrote:

I'm trying to stay away from belts for now.

Why? Belts are excellent.

If the motors take turns to drive and the teeth of the R&P are pre-engaged for the direction change, wouldn't this be the same as one motor driving both ways with a high precision R&P?

The issue here is that the cutting force direction and the axis movement direction are not the same thing, and each can reverse without the other reversing. The controller only knows the motion direction.

This is why our suggestions all tend to come back to closed-loop systems where the motor encoders can be used to deduce cutting force direction.