Vionola/Rotola: The Pinblock

In constructing an instrument, one must choose some starting point to which the rest of the body is oriented, or registered.  For the vionola, I chose the pinblock as this starting point because it is a) probably the most solid and stable component, b) an interior part (although not interior-most), c) would have precise faces to which other parts can register, and d) itself needs to be constructed with precision (i.e., would be easily thrown out of symmetry if parts it registered to were imprecisely made).  

For reference:

The instrument in overview; the pinblock is at the left end, beneath the soundboards, running between the dark colored endcap to just left of the bridges.

I collected the materials for construction:  walnut for the endcaps, nut, bow, bridges, crank handle, and base; Baltic birch ply for the pinblock (more on that below); Sitka spruce for the soundboards; mahogany dowel for the axle; zither pins and strings; red piano felt for bearing surfaces.  

Some detail on the pinblock: 

Location of the pinblock beneath the soundboards (blue highlight)

End view

As you can see, the block is constructed from eight trapezoids with 22.5° sides; for scale, the base (outward face) of the trapezoid is 1.9" W x 2" L and the distance from the outward face to the center of the axle is ~4".  The material for the pinblock is two layers of 17mm Baltic birch plywood, with the plane of the plies running parallel to the longitudinal axis (coplanar to the outward face).  The reason why it is constructed out of eight trapezoids rather than simply cutting an octagon out of a single block of wood -- whether laminate or solid -- is that, in order for the pins to have a) a stable pinhole and b) maximum exposure to endgrain against which to bind, they are best driven perpendicularly through laminations with alternating grain direction.  Therefore, building the pinblock in such a way as to have the grain and plies oriented optimally for all pins means each pin essentially gets its own mini-pinblock.  (In fact, one could understand the vionola as basically eight single-string zithers arranged in a rough cylinder.)  

A well-equipped shop would make short work of this; it's exactly the sort of thing a table saw with a good mitre sled could toss off literally in minutes.  Unfortunately, my shop is very much in the fledgling stages and bang-for-the-buck calculus led me to choose a bandsaw for my first floor machine (I need to be able to cut curves).  It's a good model, but bandsaws are by nature less precise than table saws and I'm a novice, so this was going to take me a while, even just to figure out how best to execute the cuts.  Between initial set up of the saw and experimentation, it ended up taking a few weeks of off-hour messing about.  

In my first attempt (all test cuts used 2x4 scraps, resawn, planed flat, and squared by hand), I tried a T-slot mitre that I bought as an accessory:

The degree scale on T-slot mitre turned out to be not only crude, but out of calibration, despite my best attempts to compensate. You'll note that the wedges were each just a little too big, so the last one couldn't quite squeeze in.  My next attempt used the bandsaw's table tilt:  

Although the table's tilt is easier to measure by making use of a digital gauge, I had the opposite problem this time:  the wedges were each just a tad too small, so the octagon couldn't close up.  

Another requirement I had was replicability:  I needed to be able to cut wedges, do something else, and then cut wedges again and be able to depend on that the wedges in each set would be exactly 22.5° and come together to make a perfect octagon every time.  The best way to do that with the equipment I have is to build a jig; so, that's what I did.  Again, after quite a bit of trial-and-error and bandsawing and planing, I ended up with this:  

With which I was able to produce this:

Ta-da!  And there was much rejoicing.  I was able to repeat this again, as well, so I took a deep breath and cut into my expensive Baltic birch ply:  

After sanding the mitered faces flat, the narrow dimension of the large face (which is the wider of the parallel sides of the trapezoid) measured almost exactly 1.9" on all of them (only 1/1000" off), so I was pleased with that.  A test fit yielded good results:

Next was the glue-up:

Once the glue had set and I could look at the pinblock as a unit, I saw that there was one joint that didn't quite fully seat; I believe the thickness of the glue, which I had not accounted for, produced that.  However, the gap was very small and I was able to fill it with glue and sawdust.  

Here is a shot of the completed pinblock in context with the uncut spruce for the soundboards, walnut for the endcaps and nut, and mahogany, all placed next to a ruler for scale; the overall length between the outsides of the endcaps is ~22" (the axle will extend a bit beyond that at each end).  

Next step is to make a plug for the center of the octagon, for which I'll use walnut, and through which a hole for the axle will be cut; that will also need to be as precise as I can make it.  My next installment will focus on that process, plus making the nut-end endcap (right hand side per overview pic) and setting the whole thing on the axle.