Monday, October 29, 2007

Interior framing

With the horse now glued to the baseboard, the interior framing continues with the gluing of the knees. I traced out the places where I wanted them, making sure that they were at right angles to the edge of the baseboard and exactly flush with the edge. Getting a right angle was easy on the spine and tail, and as for the bentside, I basically eyeballed the knee placement there. There are 8 knees on the bentside, spaced a bit closer together in the treble where the curvature is tighter and more widely towards the bass. The spine is supported by three knees, and the tail has one in the middle.

Before gluing the knees, a pair of small holes was drilled through the baseboard to mark the correct location for dowel holes drilled from beneath, as mentioned earlier.

The knees were glued with polyurethane glue. The back corner was fastened with a single brad, while the front was clamped:


Using a pneumatic nailer was very helpful, as I could hold the knee in place and have it instantly nailed down at the back without things shifting around.

Note the two corner blocks at the tail: these will be the endpoints of the spine and bentside liners.

When the glue dried, lots and lots of holes were drilled and dowels fitted: 3/8" dowels to hold the underside of the wrestplank support blocks, and 1/4" dowels to hold the knees, corner blocks and lower belly rail edge.

Next, the spine liner was glued in place:


This is joined to the bass wrestplank support with a pair of dowels and is cut short of the tip of the instrument by the width of the tail liner (seen resting loosely on the opposite corner block). The tail liner will overlap both the spine and bentside liners to keep it from shifting under string tension.

Having prepared the tail liner to its final dimensions, it gets glued on next:


Behind the tail you can see my length of door stop composite lying on the spine liner. It will help me in getting a preliminary idea of the bentside liner's curvature.

Making the bentside liner will be the tricky part: it will be kerfed (i.e. a bunch of saw cuts will be made most of the way through its thickness) to make it flexible, and is likely to require some trial and error to determine how to best do it. I have a spare piece of poplar lying around to practice on before I try this for real.

Friday, October 26, 2007

Horse and baseboard united

Nothing earth-shaking, perhaps, but today the horse was at last glued onto the baseboard:



Before doing so, the joints between the wrestplank support blocks and the lower belly rail were dowelled. Since I intend to dowel into the lower edge of the belly rail and the two knees supporting it, I traced out the position of these components with pencil, and just before gluing, I drilled a series of tiny holes through the baseboard, positioning them at the midpoint of the thickness of the knees and belly rail. When the glue is dry, I'll drill larger dowel holes from beneath, using these little ones as markers, without having to guess where the components are on the other side of the baseboard.

I've also laid out the positions of the knees in pencil, and once again I'll drill small locator holes to help me dowel them after they are glued down.

Thursday, October 25, 2007

Soundboard wood

Now the single most expensive part of the project is at hand: my soundboard wood, which I ordered earlier this month from Germany, finally arrived the other day. I had to go out to the airport, get the paperwork for clearance through Canada Customs myself (no brokers were used, but fortunately the transaction was quick and painless), then take the clearance forms to the shipping company's warehouse, have them release the package to me and get it home in my van. The package was an 8-foot long sandwich of spruce boards between two half-sheets of plywood. It just barely made it into the van.

I told the wood suppliers what I was aiming to make, so they provided boards in various lengths to suit my needs. I stacked and stickered everything to help it dry out, as the boards had been steamed for a while to remove stresses after they were resawn. Here are the two piles in the basement workshop:




This is alpine spruce from Europe. Trees that grow at high altitude have tightly-spaced growth rings because the growing season is short and cold. For use in soundboards, the wood should ideally be quarter-sawn. I received it about 1/4" thick; it will be planed down to about 1/8", glued up and then selectively thinned by hand-planing.

Having such expensive materials to deal with—the total cost, including shipping, duty and taxes must be at least $750 Canadian—makes me a bit nervous. In addition, there is no substitute for prior experience in making a soundboard, which I admit is a liability for me. Thinning down a soundboard is not a mathematical exercise: since no two pieces of wood are alike, the degree of thinning suitable for one soundboard is unlikely to match any other soundboard. You must work your materials in a way that accounts for the idiosyncrasies of what you have at hand, and that will vary from board to board. Still, I have to get my feet wet with this eventually, and I will give it my very best shot.

For now, everything will have to dry out for a while. Cold, dry weather is approaching, and I have the boards stacked up in the furnace room, which is warm and dry, to help things along.

Sunday, October 14, 2007

Horse: Gluing together

Once again, my current professional schedule leaves me with little time to work on the project, but there are some little things I've been able to do that didn't take too long.

I have now glued the component parts of the horse together. First the wrestplank was glued to its two support blocks, using the same polyurethane glue that helped assemble the component parts of the pinblock:


Next, the upper belly rail was glued in place:


Once the glue dried, I drilled holes for dowels to reinforce the joint between the wrestplank and support blocks, and between the upper belly rail and support blocks:


These dowels compensate for the fact that some adjoining surfaces in the horse involve end grain, which does not form the best glue joints. I sized end-grain edges with a thin coat of glue to try to counteract their tendency to suck up glue and thereby deprive the joint of it, but the dowels provide additional security.

I did a little work on the lower belly rail as well, gluing its two knees to the back and nailing them in place while the glue dried:


The silver-coloured metal angles are right-angle assembly guides, which hold the knees square until the glue dries.

Although I had earlier thought about first gluing the lower belly rail to the back of the horse, and then gluing the knees against the lower belly rail, some dry-assembly experiments convinced me that the lower belly rail was too tippy to glue against the upper belly rail and still guarantee a good right angle with the baseboard. That's why I decided to glue the knees to the lower belly rail now, so that it will be held at a right angle once it is glued to the bottom.

Since I need to drill dowel holes to reinforce the joint between the lower belly rail and the support blocks—another end-grain gluing situation—I decided to glue the lower belly rail to the horse now, and make the horse a complete unit that will be attached all at once to the baseboard:


Drilling the dowel holes will be much easier while the horse is still separate from the baseboard. Once the joints are secure, the horse will be glued down at last.

If you look closely, you can see two dowel holes drilled into the right-hand wrestplank support block. These will hold the spine liner in place. I also took the opportunity to chamfer all the inside edges of the horse that will come into contact with the soundboard, just to free its edges a little bit once it is glued down.

Wednesday, October 10, 2007

Baseboard: Cutting out the bentside and tail

I've been busy with various professional commitments since my last post, but I managed to find a few spare moments to work on the baseboard.

In describing how the curvature of the bentside was laid out, I mentioned that the bentside shape had to be corrected at the tip of the instrument. I managed to do so by nudging the far end of my plastic strip (shown in an earlier photo below) about an inch to the right, and then I traced out the final position of the bentside parallel to the strip and 95 mm to the right. I also drew a line parallel to the left side of the instrument at a distance of 185 mm. At the point where my bentside curve encountered this straight line, the bentside ended and a straight line angling leftwards at 40 degrees defined the tail of the harpsichord. Where this line met the left edge of the baseboard denoted the final length of the instrument, which worked out to 2245 mm (about 88.5"), just a whisker off from the original 2242 mm.

Then it was simply a matter of taking the baseboard outside and sawing out the curve with a jigsaw. I smoothed the edge with a file and let my mother, who has a good touch for sanding, clean up the edge even further:


The finished product: