CBATML: Parallax's 360 Feedback Servo as a Stepper Motor:)
rjo__
Posts: 2,114
CBATML(craven behavior at the machine level)
This was a successful experiment... but no fancy graphs here.
You have to take my word for it.
As you probably know the 360 has feedback that is accurate to about .08 degrees. The problem
is that when you tell it to stop... it will, eventually, plus or minus a couple of degrees.
It all has to do with inertia ... boring. The question was... if you know not the inertia, can
you tune your commands to get the thing to drift to the right place?
Of course you can!!!
I hooked the 360 up to one of ServoCity's gearboxes at 7:1. And used logic to
pass a single PWM signal as sort of a ramp and drift technique... wait a bit and then do it again.
Works out to be +/- .5degrees before the gearbox and therefore about +/- .07degrees after the gearbox,
Did it for my Propeller Eye project, but I'm guessing that if you hooked it up to the right lead screw, and
then tuned the whole getup again, it would give you some precision sweetness.
Didn't intend to post this... so not so many notes:)
This was a successful experiment... but no fancy graphs here.
You have to take my word for it.
As you probably know the 360 has feedback that is accurate to about .08 degrees. The problem
is that when you tell it to stop... it will, eventually, plus or minus a couple of degrees.
It all has to do with inertia ... boring. The question was... if you know not the inertia, can
you tune your commands to get the thing to drift to the right place?
Of course you can!!!
I hooked the 360 up to one of ServoCity's gearboxes at 7:1. And used logic to
pass a single PWM signal as sort of a ramp and drift technique... wait a bit and then do it again.
Works out to be +/- .5degrees before the gearbox and therefore about +/- .07degrees after the gearbox,
Did it for my Propeller Eye project, but I'm guessing that if you hooked it up to the right lead screw, and
then tuned the whole getup again, it would give you some precision sweetness.
Didn't intend to post this... so not so many notes:)
Comments
I might add logic for self-tuning, haven't decided yet. I think it would make a good control project for an undergraduate.
By the way, actual target is 42000 clocks on the high signal... kind of hard to figure that out from the program