Hi everyone there,
I get a headache position drift problem during the stepper motor controlled back and forth motion. I have been wroking on the problem for the last two weeks, and still can not figure it out. So I write down here my problem in hope of getting some clues from those have some experiences in stepper motor motion control.
My goal is to use a stepper motor to drive a beam flage fixed onto the shaft to go back and forth for a certain angle, and this back and forth motion needs to be stable and reproducible within hours. With the SMD1 stepper motor driver, I only need to send in a pulse train to control the number of steps(the pulse frequency defines the stepping rate), and another pulse to controle the direction (i.e., cw or ccw). To do this, in LabView, I used the counter of the DAQ board to produce a finite pulse train to step the motor, and used one digital line to produce a pulse to switch the direction every 500ms, the timing is contolled by the software.(I used the oscilloscope to monitor these two pulses, the pulse train produced by the counter looks very stable, and delay between the pulse edges of the direction pulse and the pulse train is also very stable within 500ms. There seems no software timing problems at such a long delay.) And this labview controlled motor appears to work perfectly as I required, the motor steps a defined number of steps at one direction, and then steps back at the origin position. But after running for a few minutes, the back position shifts gradually(I can not observe a obvious shift at the first few minutes) with repect to the origin position(the position I start running). I looked it up at the manual and also the website, I got a clue that this might because the motor steps at the resonace frequency of the motor and hence misses or takes extral steps sometimes. So I changed the steping rate from 1Hz to 1000Hz to see if the shift disappears or gets smaller at certain frequency range, but the shift always exists.
I did observe the shift got smaller within the same running time when change the step model from full step (1.8degree/step)to half and quarter step model. According to this, the motor did miss steps (or pulses) during running, the quarter step model has smaller shift due to the smaller steps. But I also noticed that the motor always shift at one direction, no matter I started with clockwise or counter-clockwise, this seems really strange to me. Because the motor should shift at each direction with the same probability if we assume the motor missis steps randomly.
There is something must be wrong, either the labveiw program or the motor, but I don't even know a clue of this weird shift problem. And because of this, I haven't made any progress in my lab for the last three weeks. So any suggestions could be very helpful for solving this headache and will be really appreciated.
Thanks a million in advance and Merry Christmas!
Li
