by cvoinescu » Fri Dec 07, 2012 6:30 pm
It's just a shot in the dark, but it could be a form of quantization error or interference (as in waves in general, e.g. Moiré patterns, not necessarily electrical interference). If the DSP's rastering clock was a frequency of the same order of magnitude as the step pulses for the X axis, this could happen. However, I think that that's unlikely, because a DSP is plenty fast and should run at a much higher frequency. Another source could be the high-frequency switching taking place in the laser power supply. That frequency is not that high: it's in the tens of kHz, which is comparable to the frequencies involved in stepping. So either the laser is amplitude-modulated with that high frequency, or the control signal is sampled at discrete intervals of that frequency. The latter seems more likely: it is entirely conceivable that the laser can not be turned on or off except in quanta of one period of the oscillator in its power supply. Staying with the same idea, you may also be seeing the mechanical effect of the stepper drivers being choppers, also working at a finite frequency.
Then again, I may be way off base and it could be something else entirely.