Complete information on working SM style device.

[orbs]

Circuitry :
Make a decoded digital count to 3 and reset counter w/CMOS or equl.
Phase the circuit dimensionaly too.

Looking at the picture, the duty cycle of coil A is about 31% (8/26). That means you could for example use a 74HC4017 with Q9 connected to MR so that it divides by 9. Then Q1, Q4 and Q7 provide pulses for coils X, Y and Z, and for coil A you use an OR gate like the 74HC4075 (or 3 diodes and a resistor) connected also to Q1, Q4 and Q7 to get about a 33% duty cycle. Then you could use a programmable clock generator module based on CY27EE16 or ICS307 or similar that is connected to a PC to generate exact frequencies, or use some other higher frequency oscillator .

At room temperature the HC4017 might be fast enough for a divide by 9, but these are relatively high frequencies, and the OR gate also introduces a slight delay/phase shift. Dividing by 3 would allow to generate a higher output frequency but it would require to form short pulses from long pulses.

Most flexible would probably be to have a microcontroller or DSP that has several programmable counters that are directly connected to output pins. That would allow to change the duty cycle and also to compensate for hardware tolerances and especially if its connected to a PC all parameters could be recorded and reproduced exactly and also of course control them dynamically.

Maxim even has an IC (DS1094L) that can generate 3 phases but it might not be flexible enough.

[zerotensor]

spherics' description is somewhat vague regarding the angles.  We're working in 3-d here, so actually each coil is oriented 109.5 degrees from each other, the magic tetrahedral angle.  Of course you can find 30-60-90 triangles everywhere, especially when doing projections onto 2-d...

Anyhow, I had a different take on the placement of the coils:  Why not make them concentric?

Check it out:

[tao]

spherics' description is somewhat vague regarding the angles.  We're working in 3-d here, so actually each coil is oriented 109.5 degrees from each other, the magic tetrahedral angle.  Of course you can find 30-60-90 triangles everywhere, especially when doing projections onto 2-d...

Anyhow, I had a different take on the placement of the coils:  Why not make them concentric?

Check it out:

The coils are 120 degrees from each other, based on what he said, and it all fits with close-packing and what I modeled in 3d...

[zerotensor]

The coils are 120 degrees from each other, based on what he said, and it all fits with close-packing and what I modeled in 3d...

Three of them are 120 degrees from eachother as viewed from the pole along the axis of the fourth.  This is a projection into 2d.

Any two coils make an angle of 109.5 degrees with the absolute center in space, and it all fits with close-packing and what I modeled in 3d.

[zerotensor]

These angles:  (I left two of them out for clarity)