I used to do compound radius boards and built a router contraption to do that. There is also analysis in the book (Appendix II.3) for "conical" boards, although any swept line will give you "flat" string lines which means the geometry can be pretty arbitrary.
However, once above ~16" fretboard radius, which is mostly what I build these days, the difference between a compound board and a cylindrical board is so small that it is easily buried in manufacturing tolerance, relief, drop-off, cold creep and seasonal movement. I've found that there's not much point to compound radius boards if the nut radius is greater than 16" and only do them (on request) if the nut radius is to be 10" or less.
There was a good article in American Lutherie maybe about 3 years ago called something like, "...cylinders almost do it..." which discussed the pros and cons of various surface geometries and came to much the same conclusions.
|