Eric Builds a Banjola, Part I: Planning and Design
I’ve been interested in building musical instruments for most of my adult life. Over the years I’ve built (and, in most cases, subsequently discarded) numerous cigar-box guitars, bucket basses, and copper-pipe woodwinds. But over the past few years I’ve started getting better about finishing projects, and about paying more attention to detail. I’ve also entered a phase in my life where I have the money and space to acquire some real tools.
Before the build, though, a bit of history:
A few months ago I binged on NPR’s Tiny Desk Concerts, and a few days into that binge I came across this video:
I’d been listening to Wovenhand for years by this point, but mostly the rockier full-band stuff. David Eugene Edwards all alone is haunting, and I found the instrument he was playing fascinating.
In the video he claims that it was made by the Pullman company (which also made railroad cars), but I believe he’s incorrect — in fact, I’m pretty sure his instrument was built by August Pollman, who produced a bunch of “mandoline-banjos” in the late 19th century (as well as a number of brass instruments). Basically, it’s a banjo neck (5 strings, with 1 string 5 frets shorter) on a mandolin-style body. Edwards’ instrument seems to have a much deeper body than most mandolins I’ve seen.
Much more recently — sometime in the 80’s, I believe — Edward Victor Dick began producing a similar instrument, terming it a “banjola”. I’ve been using his name for the simple fact that it’s got fewer syllables, although my instrument will probably have a shape closer to the Pollman models of a hundred years before.
I decided that I wanted to build one. Specifically, I want to build an instrument that emulates the sound of the one that Edwards has. This would be a fun goal because I had something to shoot for, but could allow myself some freedom in how I manage it.
This goal dictated several important aspects of the design: I will use a five-string setup, it will be strung with nylon strings (I’ve ordered Aquila Super Nylgut), the body will be a bit on the small side and shaped more on the “almond” end of the spectrum than the “pear”, and will be significantly deeper than a mandolin.
This is also a bit of an experiment. I have been unable to find good information on how the sides and back of a chordophone affect the sound. People care a lot about soundboards, and with good reason — the soundboard is directly coupled to the strings by the bridge. The question is whether the back and sides need to be free to vibrate in order to contribute to the sound.
My theory is that the back and sides can contribute to the sound, but that they act primarily as acoustic reflectors rather than resonators. In order to test this, I’m going to use cheap (though pretty!) lauan for the back, stiffened slightly by epoxying the board after cutting. (If I had a vacuum-bagging setup, I’d vacuum-bag this. I don’t have such a setup, though.)
The sides will be even more unusual than the back. Typically, on most instruments whose construction is anything like this, the sides are thin wood that is steam-bent to form the desired contours. Interior “kerfing” is used to provide a broader gluing surface — it’s basically the 3-dimensional equivalent of an L-bracket. This allows for very thin sides indeed: in some cases, slight over-sanding can mean that an instrument’s sides are translucent.
Instead of bending thin wood, I’m going to cut the sides out of plywood. I’ll stack lots of 1/2" birch plywood, glue it all together (the plan is to use 7 layers, for a total body height of 3.5"), and then sand it on a table-mounted belt sander until it’s thin before gluing on the neck, soundboard, and back.
This might mean that I end up with a very dead-sounding instrument. If I hate it, I’ll just un-glue the sides from the soundboard and neck and re-build the sides and back with more traditional methods. If it sounds fine, then I’ll be able to contribute some meaningful info about how the back and sides affect the sound of a perpendicularly-coupled chordophone.
I was originally going to build the neck myself. My father recently gave me a rock maple 4x4 that he’d bought in the 70’s, so it’d be nice and dry and neat in a “family heirloom” sort of way. But the shaping required to build a neck is significantly beyond anything else required by this project, and I’d rather not mess up that particular piece of wood. So I ended up ordering a pre-built neck off of Amazon, which cost about $65.
This meant that lots of my design could comfortably be based off of calculations from that neck. Instrument-making is neat in that there are lots of measurements that don’t have to be precise at all, and lots of measurements that have to be precise but only relative to the other things on the instrument. In this case, the main design consideration is “scale length” — the length of the vibrating portion of the string.
The scale length is very important. The twelfth fret must occur halfway between the nut (the end of the neck near the tuners) and the bridge (the thing the strings pass over on the body, which transmits the vibrations of the strings to the soundboard). This is because of how the Western musical scale is constructed: there are 12 half-steps to an octave, each fret raises the pitch by a half-step, and an octave consists of a doubling of frequency (one octave up from the A whose frequency is 440 hertz is an A with a frequency of 880 hertz). To double the frequency of a string, you halve its length. Thus, raising the pitch of a string by an octave can be done by halving its length, and you want that to happen at the 12th fret. (Technically, the twelfth fret shouldn’t be exactly halfway, because when you fret a string, the tension of the string increases slightly, which raises the pitch a touch more than simply changing the length would have. However, a nice thing about a movable-bridge design — which banjos are! — is that you don’t have to account for this when building, because you can slide the bridge around when you first string the instrument up to adjust the intonation.)
So as a baseline, I traced the heel of the neck onto some foam-core board and marked the edge of the board as the 12th fret. I then used the neck to measure that same distance again from the edge, to get the bridge position. I could then place whatever shape I wanted using the heel location and bridge position as my anchor points.
I played around with different outlines, ultimately deciding that I liked a simple teardrop shape. The tail end is 7 and a half inches from the bridge line, and the greatest width is a foot (so, 6 inches in each direction from the center line at the bridge). I didn’t use any complicated drawing hardware or formulae for this — I just freehanded the curves.
Now that I’ve got an outline, it’s time to start tracing and cutting wood. I ordered spruce from Stewart-Macdonald for the soundboard, which comes in two halves; I need to plane a common edge and glue the halves together before I cut the soundboard out. The back can be cut directly from the lauan. I’ll copy the curve from the foam core onto some cardboard and trace that onto the wood. The pieces for the sides will be a bit time-consuming and will need to be cut a bit larger than the back and soundboard, to account for loss as I sand them down.
I’ve also started brainstorming “branding” stuff — I want to woodburn the head stock and glue down a label to the soundboard, but I need to figure out what those pieces will look like.