Re: Characterizing NEMA17 stepper motor/driver micro-stepping accuracy
On Wed, Jan 8, 2020 at 09:09 AM, tomofreno2000 wrote:
The steppersonline precision gearboxes spec is < 25 arcmin. The built in encoders minimize the small increase in positioning error with load due to changing motor reluctance shown in Howard's data, and prevent loosing steps, but now you will add a much larger 25 arcmin of slop downstream of that.I got my G11/Gemini-1 used and PE was pretty bad (why the guy sold it), After properly adjusting the stock worm it started to shine. then the crummy $2 (yes the original servo motor cost $2 w/o encoder) quit working, with my extensive brush motor racing history, i replaced the brush system and working well for 8 years. Then most were replacing the gearbox with McLennan gearboxes and the PEC was no longer even needed IMO (subjective). The McLennan gearbox is spec'ed at "max backlash +/-1 degree". With 250 encoder, 25:1 gearbox and 360 tooth main gear, this is 9 million clicks per rotation, maybe the gemini-1 has some anti backlash compensation? But RA never really stops or backs up, so should not be a issue on that axis. My point is, having the encoder on the motor increases its resolution and seems most use this method.
I autoguide now and get under 1/2 arc sec rms and 1/4 rms often on PHD guiding graphs.
Of course my second stock motor failed and I replaced it with a ebay 5W Maxon motor (half the size of the 3W stocker and $13), added the McLennan gearbox, fabricated a ALU custom mount to tuck the motor back in 2010, still working well with much added clearance. Only took Losmandy 8 years to catch up :)