Posted By Greg Elmassian on 18 Apr 2011 04:08 PM
One of the problems that has lead to abandoning stepper motors in many applications (and using servo motors and some sort of position feedback) is that if the motor is commanded to move and does not, your "program/controller" has no way to know this has happened.
conversely, if you move the motor armature it will not "know" either.
When I worked with some of the first digital plotters at HP, there had to be periodic "position resets" or recalibrations to know that the plotter arm was where we thought it was.
Gearing the motor down to increase the torque (and holding torque) means pretty darn slow speed unless you have a sophisticated controller that can vary the slew rate.
Regards, Greg
Yep, all very true.
Another reason for having a sloppy gear train... when power is turned on, to get the armature position sync'd to the controller (since the controller won't know where it is) the controller will have to step the motor through at least one full revolution, of which the armature may make a full revolution -- or something less, even maybe only one step... But if the gear train is tight the motor might be able to turn, if for example, it is driving a turntable and the turntable is locked in position.
As for speed, in the application of moving a RR turntable, you don't really wanna turn it like a weedwacker so an ultimate speed of "slow" at the turntable end of the gear train is quite acceptable.
Still for precision positioning, an external position sensor makes better sense, even if using a stepper motor for motion.