Just had a thought ...
 

It just struck me ...on the body of an acoustic guitar, there are three, and only three, areas where absolute flatness is required ... on all other areas , curves rule OK ..

A: Most importantly, the surface on the upper bout where the neck heel shoulders contact the ribs needs to be absolutely flat (as of course do the neck heel shoulders) ... unless ... the heel has a parallel contour rather than a "V" contour.

B: The area underneath the glued fretboard extension would preferably be dead flat.

C : The base of the saddle slot should be dead flat.

On the neck, things are slightly different, naturally ..I am only referencing the body of the guitar.

As long as you can cope one suface to another you can have absolutely no flat surfaces! Theoretcally....

Not necessarily.

As long as the heel and sides meet without gaps, they can be any shape one wants. Flat is easiest to fit, but not the only option.

I often make the underside of fingerboards very slightly concave. Doing so assists with a gap-free visible seam at the edges. Even a slight convexity will produce a gap at the visible edges.

People like to believe that if there isn't perfect contact along the entire surfaces of the bottom of the saddle and the bottom of the saddle slot, "sound" is lost. It sounds nice, but I know of no objective experiments that prove that to be true. One could effectively argue, I suggest, that with the portion of string pressure pushing down on the top of the saddle, some mismatch in surface contact doesn't make an audible difference. Again, without objective evidence, it's just arm-chair theorizing. In short, we don't really know, but people still put in the effort to make the surfaces "perfectly" contact, regardless.

While the mechanics are somewhat different, the argument that one needs full, complete matching of surfaces suggests that arch top and violin family instruments would sound better if they had the entire length of the bridge contacting the top. I think most accept that single-piece archtop guitar bridges sound better than screw-mechanism two piece versions, but, regardless, there isn't full contact between the bottom of the bridge and the top.

http://www.stewmac.com/product/image...tar_Bridge.jpg

http://www.koentoppguitars.com/guita...es/full/12.jpg

I never use a flat section at the head block, it's always curved. I don't really like the look of a flat section there, a curve gives it a nice touch. I have arched sanding blocks to rough out the shape of the neck heal. Basically it's the opposite or congruent curve used for the head block itself.

One thing I do strive to do, in light of the flat saddle slot argument, is make it so that the face of the heal is fully flush to the ribs at the headblock. I use a bolt on butt joint and like to think of it more as a traditional Spanish heal only bolted together. Most builders will rout out the middle area of the heal face leaving the cheek edges proud for only about 2mm along the edge to make flush contact with the sides. I am also armchair hypothesizing on that too but I like to make it a flush contact and it's not that hard to do by just flossing the heal after roughing it out.

On most of my guitars too I build the lower bout flat. When I do put a slight arch in the top I just glue the bridge on flat too.

It always seemed to me that the arch top bridge design was wasting energy some how. It is what it is though.

True , Louie, ... but as far as heel mating surfaces are concerned, why would any rational human being even contemplate anything other than two plane surface contacts ? Assuming a tapered heel, that is ... a parallel heel is a different matter entirely.

I get Charles' observation about concavity on the underside of the fretboard extension ... eminently sensible ... but would nonetheless point out that violin techs take the most extreme care to ensure that the feet of a violin bridge match perfectly with the contours of the soundboard.

Some builders want the look of the rounded upper bout, thus you have to cope the heel to the upper bout, which you'd end up doing even if the heel area were flat. Doesn't matter the shape of the heel.

As to the violin bridge, it is fit because the top is arched. This is also done with acoustic steel strings built with a domed top.

Redir, can I just ask, do you make your heels (sic), tapered, like Martin do, or are they parallel ?

Thinking out loud, what IS rational about guitar building? The majority of us in this subforum are not going to do this as a primary source of income. Building a guitar is a blend of art and science. Thus we pursue what is rationally impractical for the sake of the art. I'm sure the pride or satisfaction of "pulling it off" a driving factor. Even if it's the simplest task like fitting a saddle to a bridge. Is t rational to spend a few hundred dollars on a piece of wood that was dumped in the river some 300 years ago? Perfectly rational... if you build guitars!

I prefer the look of the curved upper bout, whether with parallel or "V" contoured necks. Flats can give a sort of sunken look. If one uses a procedure that involves "flossing" the joint, the difference is about two strokes (the way I do it), so not much of an imposition. Flat front to back makes referencing everything a lot easier, but once you get your eye in, so to speak, a very slight convexity in the sides front to back will give an almost perfect fit on a "dished V" (AKA standard, heel shape) with straight, slightly undercut shoulders, even before you start flossing. Lots of 3D geometry to get your head around, there!

I'll add another "flat"... the bottoms of binding channels. Very hard to achieve with most of the typical router tooling and not totally necessary to bother, but it sure makes gapless binding a lot easier.

I'm another fan of continuing the curve of the upper bout through the neck joint. I made a pivoting jig that allows me to sweep the neck through my bandsaw at the desired radius to fit the end of the body. My heel sides are parallel so I haven't tried it on a tapered heel design, but I would think it would work for that too.

I would think it should... If you make the cuts on your jig with the parallel heel, fit it to the neck block area, then shaped the heel to be tapered, it should be fine as well.

Tapered. I like to make them very much like Martin does except of course for the bolts.

https://lh3.googleusercontent.com/9-...=w1280-h720-no

Good point! And I think your idea would give a bit cleaner result.

I've built a few guitars that had individual saddles for each string, and I did find the response and volume of each string to be surprisingly "even", so I ponder whether this is because the individual saddles seat better against their respective sockets as opposed to a long saddle against the bottom of a long slot?

That's one possible variable that might have contributed to the even response.

In the late 1970's Charles Fox made a guitar that had 6 individual bridges, each with its own saddle. He found that it lacked sympathetic resonance between strings, which he attributed to the disjoint bridge pieces.