Fretboard Leveling

[Cody]

When leveling a fretboard the way wood is removed or frets are leveled depends on if the fretboard is cylindrical or compound radius to my understanding. I realize that it takes considerable experience and understanding of a cylinder or compound radius neck when leveling a fretboard in order to keep the neck as designed. With that in mind is it safe to say in general that a compound fretboard is leveled by removing wood in line with the taper of the strings from the bridge to the nut? A cylindrical fretboard is leveled by keeping the removal of wood parallel to the center line of the cylinder? What I am not sure about is how much feathering in is there on a compound neck to across the fretboard.

I read something by someone that indicated the removal of wood by following the taper of the neck and strings will accomplish the feathering. Once the fretboard has been leveled properly then my assumption is the same fundamental method is used to level the frets. In both one should remove as little material as one can and still accomplish the desired objectives.

Cody

[Pierre-Antoine Roiron]

You are almost right : the idea is to sand along the straight lines that generates the surface of the fretboard. For a cylinder, these lines are parallel to the neck axis. For a compound fretboard, these line are not parallel to the neck axis, nor are they parallel to the taper : most of the compound fretboards have radii that evolve from 12" to 16", and they are cone shaped, but the ratio between the radius and the taper is not constant along the neck. In fact the fretboard edges are not part of the lines that generates the compound fretboard's conical surface. They are closer than the ones that generates a cylinder shaped fretboard, but not part of them.
In the real life, I use a straight caul and sand the fretboard (or the frets) along the strings'axis for a compound radiused fretboard, and a radiused caul for the cylindrical fretboards, sanding along the neck axis. As you do.

The ideal fretboard would have a radius/taper ratio that would be constant. I tried to compute it some years ago, but it lead to very small radii at the nut, and that was not comfortable. Maybe I made a mistake?