Avid Pro Rack and Pinion question - losing steps while jogging.

I have an Avid Pro4848 and have it working with a new controller I am making. One problem I had was with the rack and pinion mechanism.

If I run a GCode program, it works perfectly and returns to the same 0,0 location it started from. Even fairly complex 30 minute programs seems to not lose any steps. Even running at a 20K mm/min (~400 IPM) feed rate. I am very happy with that.

However, jogging even with low speeds (1000 mm/min) loses steps like crazy. I traced it down to de-energizing the R&P steppers. If I keep the power on to the steppers, it jogs fine and does not lose steps. If I de-energize the steppers between jog commands, it loses steps. There is a "thunk" when it energizes them and I assume that is where steps are being lost.

I have tried increasing the tension of the pinion against the rack but to no apparent effect. This happens with both Y (dual motor gantry) and X (single motor) axes. It is possible the Y is still slightly racked as I haven't had a chance to tune it 100% square. But because it also happens with X, I don't think that is a cause here. I don't see the problem with Z - it uses a ball screw (though good idea to keep Z motor energized anyway). It could be backlash but I see it if I do multiple jogs in the same direction (ie not reversing direction) and running long programs do not cause any lost steps. The lost distance lost is proportional to the number of jog commands issued - around .3 mm per jog.

Is this just an aspect of rack and pinion? With the holding idle current reduction setting on my drivers, I don't think it is a problem to keep the motors energized. I never saw this with screws, though. Would helical R&P act differently?

Yes, you are correct. If you remove power from a stepper drive, it will not hold position. What ever made you think that it would?

Well, my screw based machine doesn't. I get Z but for X & Y, there are no opposing forces so why would they lose a couple of tenths of a MM each jog? I don't think it is so far fetched. And oddly enough, Z with a ball screw doesn't.

It's simple. The friction/stiction is greater than the force on the moveable parts.

Once powered down and then back up, there is no guarantee that a stepper will "jump in" to the same torque detent that is was in during shut down. They will most often jump to the nearest 1/4, 1/2 or full step detent. Even with your ballscrew machine there was no real reason for the stepper to startup in the same position as it was in when shut down. This may have happened and you just didn't notice it. Ballscrews will have a ~2.5 to 5 turns per inch and rack will be in the .9 to 1.5 range, so the error woul;d be there just maybe one fifth of the tenths you saw.

To repeat, if you wish your stepper machine to hold its position, leave the drives powered up, its that simple. There is no reason, other than an impending electrical storm, to shut it down.