Brushless Motor

I redesigned the motor we made earlier.  I figured the brushes were creating a lot of friction and keeping its top speed down.  I wound four new coils onto spent wire spools, 1,000 wraps each of 32 AWG enamel wire, and mounted the new coils around the bottle with the magnets taped to it. 

The magnets are mounted 90 degrees apart (or as close as I could get them to 90 degrees using the pattern printed on the bottle).  If the electromagnets were small enough I could put them 22.5 degrees apart within a 90 segment and pulse them in a row.

 
However, they wont fit this close together, but I can place them around the full 360 degrees in multiples of 22.5 degrees +/- 180 or 90 degrees--in other words 67.5 degrees--and if I get the right placing it will work the same way (with, in this case, lots of tweaking and taping).

(Originally I tried to do this with three coils in multiples of 30 degrees, which in theory should work, but it was too sensitive to the placement.)  In this second case the magnets are pulsed in the opposite direction from the way the rotor turns.  I hooked up a controller and wrote a short program to drive the motor.  The LEDs on the breadbord indicate which coil is powered and to what degree.  It worked better to move adjacent coils up and down in voltage in a triangle wave pattern to trade off the rotor.

Here is a code snippet to give an idea.

All I had at the time were some PNP transistors to hook it up so everything is "drawkcab".  A "HIGH" signal means "LOW" voltage, etc.  I need to clean up the code by putting the pins in an array and having a single loop.  I am also thinking of using a smoother sine wave function to trade the rotor between coils and increase the voltage to the coils (currently it is running on 5 volts off the controller board, but the control of the rotor is very loose).  One surprising side effect, as the power is adjusted along the triangular wave the board can not do a true analog transition, but approximates this by varying the pulse of an on/off signal very rapidly, this causes the magnets on the bottle to vibrate and produce sound.  The motor "sings" as it turns.  Also, contrary to my original goal, I can not get to high speeds with this setup, but I can turn the motor as slow as I like, pause for as long as necessary, and easily reverse direction, kind of like a step motor, which gives me some ideas for applicaiton if I can build a more rigid structure, run it with more power, and gear down for a mechanical advantage without too much friction.