Apply Microsteps as a denominator rather than quotient

[sjkelly]

We currently do this:

  // Set the increment across the phase table from the specified microsteps
  wire [7:0] abs_increment = (microsteps == 8'd0 ) ? 8'd64 :
                             (microsteps <= 8'd2 ) ? 8'd32 :
                             (microsteps <= 8'd4 ) ? 8'd16 :
                             (microsteps <= 8'd8 ) ? 8'd8  :
                             (microsteps <= 8'd16) ? 8'd4  :
                             (microsteps <= 8'd32) ? 8'd2  :
                                                     8'd1  ;

In theory non power of two increments should work in LUT traversal. so you could e.g. get 64/3 microsteps. Which might be useful for encoders and also less FPGA logic.

[sjkelly]

This is a proposed solution for #148