Soundboard: Intermediate glue-up stages

I'm still gluing the soundboard together, as my glue has a long clamping time.

After making up 4 pairs of boards, the next step is to glue 2 consecutive pairs together. A full-sized sheet of plywood now acts as the work surface:


Note the second glue-up of short pieces at the back of the work surface, as well as the extra shims needed to keep pressure against the full length of the glue joint: these compensate for the staircase-like profile of each pair of boards.

Before gluing, I spent the morning levelling and cleaning up the glue joint of each pair with a cabinet scraper. The process was tiring, as there is quite a bit of friction in pulling the scraper along the surface, but the tool did a great job of taking off thin little shavings.

Tomorrow, after scraping the product of this intermediate stage, each set of 4 boards will be glued together to complete the soundboard.

Soundboard: Glue-up

Today I completed the process of planing my boards down to about 3.5 mm. After laying the boards edge-to-edge, I chose to rejoint two edges just to clean them up. Then I trimmed a bit from the front of each board, as the planer had sniped each one a bit at the end of each planing pass.

Next I brought out the instrument, which has been resting in a corner of the workroom since the framing was finished in early November, and started laying out the soundboard wood in sequence on top of the liners. I clamped the leftmost board in place along the spine, keeping the closest edge about 1/4" in front of the upper belly rail, and marked the final length of the board about 1" beyond the far end of the liners. Each board received the same treatment, and as each was trimmed it was replaced back onto the instrument in sequence. Here's a picture of the process:


I'll use the off-cuts to make up a cosmetic veneer for the wrestplank surface.

Originally my plan was to make the soundboard long enough that the front portion would be cut off and made into the wrestplank veneer. Unfortunately, several of my boards have defects close to the far end, and if I had pulled them all forward so that they were flush with the front end of the wrestplank, these defects would have ended up on the finished soundboard. I'd rather not have knots and splits on my soundboard. By keeping the front edge of the boards almost flush with the upper belly rail, the defects end up beyond the perimeter of the instrument, where they can safely be cut out.

At last the time came to actually start gluing the boards edge-to-edge. I made up a gluing platform consisting of an 8-foot section of plywood, two 8-foot long pine boards, clamps, wedges and a bunch of bricks.


The bricks got wrapped in paper because they like to shed red dust all over the place.

One pine board was clamped parallel to the long edge of the plywood. A sheet of wax paper was laid down and two soundboard planks were placed on top. A long skinny off-cut of soundboard wood was placed along the edge of the second plank, and the second pine board was clamped about an inch away from this. Numerous pairs of wedges were inserted between the pine board and the off-cut.

The actual glue-up is quite simple. Two soundboard planks are folded together face-to-face to expose the mating edges and a bead of fish glue is squeezed on and brushed out. After replacing these planks on the plywood sheet, the wedges are squeezed together, which presses the two planks together, the off-cut serving to protect the outside edge from being indented by the wedges. Bricks are placed on top to keep everything from buckling upwards as the pressure increases. The set-up looks like this:


The "gift-wrapped" bricks ended up that way because they were still shedding dust.

With this set-up, I can glue up several pairs of boards each day. There are 8 boards in total, and I'll soon be able to glue 2 pairs of boards together, and so on until the entire soundboard is complete.

This method is similar to what harpsichord makers did centuries ago in gluing up their soundboards, according to surviving accounts.

Soundboard: Edge jointing

Last weekend, after jointing one edge of my soundboard material at the router table, I took each board to the table saw and sawed defects from the edges. This shrank the widths of each board somewhat, but it was necessary to get rid of large knots and pitch pockets, which might have weakened the board and could have been torn out during planing or jointing.

Next I set up the planer and started planing each board face until roughness and oxidized wood vanished. I removed only enough to clean up each face. When this was done the material was from 6-7.5 mm thick, so I continued planing all the boards in sequence on the same planer setting to get their thicknesses to match. Since the material is quite thin, I rode each soundboard piece on top of an 8-foot long pine backer board: this is safer than trying to send something very thin through a planer on its own.

The boards are now about 5 mm thick, and I'm aiming for a starting thickness of 3.5 mm. My friendly harpsichord maker advised a thickness of a "heavy eighth inch", an even eighth being about 3.2 mm.

I jointed the sawn edges of each board to clean them up, and as an experiment I tried laying boards edge-to-edge to see how well they fit. Even after several light jointing passes, the boards wouldn't quite come together. It appeared that the length of the jointer fence was insufficient to remove a very subtle bow from the edges.

I thought I might have to invest in a good-quality large hand plane such as a jack plane, which together with a shooting board is usually what is used for jointing soundboard planks. But, after consultation with my father, we decided to adapt the shooting board principle to the router table and save the expense of the plane, which would easily be over $200.

A traditional shooting board is a long wide board to which a second narrower board is fastened. The step between the boards receives a long hand plane laid on its side, and the material to be jointed is placed atop the narrower board with the edge slightly overhanging it. The plane is pushed forwards and, guided by the edge of the narrower board, planes the edge of the desired board straight. Using a long plane ensures that any minor defects on the narrow board's edge are bridged by the plane's long sole.

We came up with a shooting board made of a 12" wide piece of medium-density fiberboard (MDF). MDF is flat, straight, stable, heavy and cheap. Atop this went a narrower piece of MDF with five slots at right angles to the long edge. Five carriage bolts were inserted through the large piece from below, with the heads countersunk, and hand knobs were screwed onto the opposite ends to hold the narrow piece down.

The slotted narrow piece acts as an adjustable fence. A piece of soundboard wood is laid against it and the fence is adjusted until the soundboard material just slightly overhangs the edge of the bottom board. Then the knobs are tightened to hold the fence in place. The entire package is pushed through the router table with the bottom edge running against a piloted router bit. This shaves the soundboard plank flush with the MDF edge.

In action the shooting board looks like this:


It's "incredibly ugly", as my father put it, but it cost $20 to make and it works. The bricks weigh down the soundboard plank and keep it from shifting around as the assembly is pushed forward.

This router bit


takes care of the trimming. You can see the pilot bearing running against the MDF edge.

After jointing each edge once and trying planks side-by-side, the fit is pretty much perfect.

Soundboard: Selecting boards

At long last I am ready to begin work on the soundboard.

The wood, as you will recall, was purchased back in October and has been sitting around for about two months. Vogel sent me something like 2 square metres, which is almost twice as much as I need, but the material only comes in increments of 1 square metre. It's good to have some extra in case something goes wrong; perhaps I'll build a smaller instrument someday with the remainder.

I spent several hours recently going through all the wood again, thinking about how to put it to best use. There are a number of considerations to balance in selecting the planks:

• It would obviously be wasteful to cut a long board into short pieces for use in the treble, especially since Vogel specifically sent me a variety of lengths. So the longest boards should ideally remain in the bass.
• The boards are sounded out by rapping them with one's knuckle. The idea is that "ringier" boards are better suited for the bass, while duller-sounding ones should go higher up. This is complicated somewhat by the fact that the boards are mostly of different lengths and sound different by virtue of that alone. I gave it my best shot by aiming to have some sort of gradation of "ringiness" as I arranged my boards from left to right.
• I've heard suggestions that coarser-looking grain should go in the bass and finer grain in the treble. One of my books says that there shouldn't be wild changes in grain density from one board to the next. My boards seem pretty good in this respect, except for one that has a patch of very wide rings right in the middle.
• My friendly harpsichord maker says that if boards want to bow, they should be arranged so they all bow upwards. The soundboard in a finished harpsichord is slightly crowned anyway, so any natural tendency of this sort should be noted and taken advantage of.
• Grain runout must be accounted for. Wood fibers don't run perfectly horizontal in sawn lumber but come up to the surface at an angle. On rough lumber, these little whiskers can be stroked with the fingertips: one direction feels rough, while the other is smooth (like a cat's fur). This must be noted when planing, otherwise the blade can catch the whiskers and start ripping fibers out of the surface like a loose thread being pulled on a garment. If the boards are properly oriented, however, the blade simply trims the fibers harmlessly.

After examining my collection of boards with all these factors in mind, I chose the requisite quantity, arranging them suitably from left to right, decided which face might end up as the visible upper surface—though it's a bit early to be completely sure, as the boards are still rough—marked down the grain runout on each face and made a note of defects such as chips and knots. Although the wood is of good quality, most boards have some knots or pitch pockets (near the edges, fortunately) that will have to be cut out. The exception is the left edge of the lowest board, which will be glued to the spine liner, so any modest defects there are unlikely to have an impact on the sound.

Today I did some preliminary straightening of one edge of each board. Since the wood is rough-sawn, all four surfaces will need attention, but everything must start with one straight edge to place against the table saw fence. The two longest boards were not that straight—one of them was out by about half an inch in the middle—so I had to attach a long straight board with double-sided tape to each of these in turn and run the whole package through the router table to get a half-decent straight edge. I was impressed by the gigantic splinters that flew all over the place as I did so: the wood is quarter-sawn, so the router bit was basically splitting the edge off along a growth ring. Still, no damage was done, and at least I can now take all this lumber to the table saw. Here it is, waiting for tomorrow:

Key panel, key frame and miscellaneous details

I've been busy this past month with various concerts, so I've collected smaller bits of work I've done over the past little while into this single post.

Back in October I glued up the key panel, out of which the keyboard will be made. This consists of several 1/2" thick basswood boards 6.5" wide (the width of one octave) and of varying lengths to suit the 8-degree taper that the keyboard must have in order to fit inside the instrument. The key panel isn't a big deal at this early stage, but here's a picture anyway:


It still needs to be cut off straight (but at an 8-degree angle) along the back.

This will sit on the key frame, which is made of soft maple:


The key panel, when cut to size, will be almost flush with the front of the key frame and somewhat short of the back edge. At the back, the rack—that strip of maple sitting behind the key frame—will be glued on edge. This will have a series of vertical slots which will accept a metal pin driven into the end of each key. When the key is depressed, the pin and slot will ensure it moves straight up and and down.

The balance rail—the chamfered rail in the middle of the key frame—will be the pivot point for the keyboard. Each key will have a mortise punched most of the way through it from above, and a pin will pass through this mortise into the balance rail. This second guide point together with the rack will keep the keys moving vertically, without skewing sideways and touching their neighbours.

Back in mid-September, I visited A&M Wood Specialty again and selected a large plank of Alaska yellow cedar to use for the exterior casework of the harpsichord. I ended up with a board about 9 feet long, 10 inches wide and 1" thick (all rough dimensions). Since the moisture level of this board was a bit high, I purchased it and arranged to have A&M keep it for me until it dried out somewhat. By late November the board was ready to go, so I had A&M resaw and surface all the resulting pieces to my required final dimensions of 7.5" wide, 3/16" thick. They did an excellent job for an eminently reasonable fee of about $25. I ended up with 4 long resawn boards plus one thicker full-length piece that came from the original board when it was ripped to width. This will be used for mouldings, of which there are a great many on an Italian instrument.

Alaska cedar is a lovely wood of pale yellow colour and very subtle straight grain: it's quartered so the grain is quite even. It has an oddly herbal smell when freshly-cut, quite unlike the usual resinous evergreen fragrance. In fact it smells as if someone has been smoking pot somewhere nearby!

The case wood is thin, so the bentside can be made by simply shoving a board against the glue-smeared knees and clamping until the glue dries. Other sides are treated similarly, except that they need no bending. The numerous mouldings that go along the top and bottom edges of these boards will also strengthen them and keep them from being too floppy. A nice example—one of many in the Italian harpsichord tradition—of form and function being combined.

During the course of October and November, various items needed for the harpsichord arrived from various destinations: cloth padding of various types, boxwood plates which will make up the natural key covers, strips of ebony from which the sharps will come, maple arcades to decorate the front edge of each key, tuning pins, bridge pins, hitch pins, and so forth. I was lucky to have purchased all this stuff at a time when the Canadian dollar was skyrocketing to all-time highs against the US dollar. It also gained against the Euro, which helped offset the cost of buying things from Europe.

My sources are:

The Instrument Workshop: cloths, key covers and sharps, some specialty tools, plectra and strings.
Hubbard Harpsichords: maple arcades and specialty drill bits.
MicroMark: several small tools.
Marc Vogel oHG: brass and iron pins of all sorts, soundboard rose, soundboard wood.