Tuesday, December 22, 2009

Stepper Drivers and Ball Mice


A4983 Carrier

Stepper Driver

Here's one of my new stepper controllers. They can't put more than 0.5A into my 1.68 ohm motors from the 12v supply without overheating, probably due to current squared losses in the chips' fets. I will have to try a lower voltage. The white stuff is some heatsink compound- I was trying to heatsink the chip with little success. I will try something with a clip that goes under the board to hold it steady.


Quadrature Wheel

Ball Mouse Belly

Buffer

Being Used as Input

I'm using a gutted ball mouse as an input device. I'm picking the quadrature signals directly off the photodiodes, so they need some buffering. Quadrature translates perfectly to step/dir, so this allows precise bidirectional control of my stepper. With the A4983 set to 16x microstepping, I have 3200 steps per rev and therefore a wicked reduction ratio from the mouses' wheel. I'm re-splining my stepper shaft because the other splines weren't close enough to the motor body, and I already had this rig set up to test the drivers with. The arduino is in this circuit for two reasons: 1) to count pulses and tell me where I'm up to, and 2) to debounce the pulses, and provide them at a steady rate to my stepper controller to avoid missed steps and noise getting into it.


Overview

Re-Splined Shaft

5 comments:

  1. Hello :-)

    I bought just the same board from Pololu, I think you suggested it to me on RepRap forums.

    And give a look at the shaft of my motor, I am shure it's not so perfect as yours ;-): casainho-emcrepstrap.blogspot.com

    I do not understand why you want an encoder on this stepper motor! If it misses steps, them you have a problem with force! It should not miss any step and you can test it before!!!

    And I am using the 200 full steps and you can see it works nicely (see the my latest videos). Also I read that full steps have more torque.

    If you go with full 200 steps, will the A4983 get cooler?

    Another problem as you can see on my videos, are the grip... I think a larger wheel (maybe like Makerbot) would help here... but we loose force going with a wheel (larger diameter than the 5mm Nema 17 shaft).

    ReplyDelete
  2. Any chance that the knot in you stepper motor wires is adding some significant inductance at the high switching frequency the A4983 is using? I know it sounds like it won't do anything, but I would suggest undoing the knot and trying again.

    ReplyDelete
  3. >> I do not understand why you want an encoder on this stepper motor! If it misses steps, them you have a problem with force! It should not miss any step and you can test it before!!!

    Seems like a great idea to me... -- If you start skipping steps, you can now add heuristics to the firmware to stop the build and sound an alarm. You've more than likely just crashed your head into a partially detached part.

    ReplyDelete
  4. Casainho, I'm using the encoder as an input device rather than feedback, since there's no computer in my workshop yet. Not planning on using them for feedback, although the possibility is always there

    ReplyDelete
  5. Ok, I understand now.

    Stepper on this application are good because there is no need for feedback, which makes system simple, cheap and less problematic.

    Stepper should have a good margin before start skipping steps.

    ReplyDelete