Sunday, 10 November 2019

Building a 13 Course Lute, Start to Finish, 7: Bracing the Belly

Hello again. Today, another adventure in lute making: shaping and gluing the braces on the backside of the belly. Perhaps not the most glamorous job in the lute maker's resume, but essential to the sound and structure, and eventually the action, of the finished lute.

At the beginning of this blog series, somebody on social media asked me a question something like this: is building a 13 course lute all that different from building a renaissance lute? My answer could only be superficial, but I said yes--and that's mainly down to the fact that whereas a renaissance lute has a relatively flat (though not completely flat) fingerboard, a 13 course lute has one with a very pronounced arch. Practically every operation bends--often, quite literally--toward accommodating that feature. Today, we'll really start to see that principle in action.

[Small editorial comment here: throughout this post I will refer to 'braces' and 'bracing,' rather than 'bars' and 'barring.' I think the former is more a British usage, the latter more commonly a North American one. Writing in Canada, and in Canadian as I do, I'm generally more likely to go with British spellings, though in this case the Britishism seems a bit foreign to me. However, I've decided to use it to alleviate confusion when I start talking about go-bars and the go-bar deck.]

In my last post, I carved the rose. Upon finishing that job, I hung up the belly one more time in the light box for 24 hours, in a low-humidity environment, for another round of humidity cycling (described in more detail here, in the 5th instalment of this series.) After bracing the belly I won't be doing any more humidity cycling, since the braces are glued across the grain (and to put the assembly in a low-humidity environment at this point would be disastrous.)

So no more humidity cycling with the belly, but I will nevertheless continue to carefully control the relative humidity in my workshop, in two ways: by using a small oil-filled heater to warm my work room, and, when the relative humidity is a little too high for that to handle, by turning on my dehumidifier. These two little appliances can make working conditions uncomfortably warm at times, but together they enable me to keep the r.h. in my shop in the low-40% range, which is optimal for crucial jobs like gluing on braces, and, a little further along, gluing the bridge.

Here's my brace material: alpine spruce, from the same source as my soundboards. It's sold in split billets, which guarantees that the pieces I cut will be practically free of runout. (Watch out for that knot!)

After cutting braces to generous length and height, I take them to a quite exact and uniform thickness with a bench plane.
I have a planing box for the purpose--just a small plank with some edging--which fits into my bench vice.



All three sets of braces are 'roughed in' for all three lutes. The brace dimensions and locations are based on historical examples, and on my own years of experience building a lot of lutes with the same or very similar material throughout. (One does get to know one's material quite intimately after a while.)

I bend thicknessed sheets of material for j-braces using the bending iron and a bit of water from the spritzer. Later, when they are cooled and dry, I will cut individual j-braces from them in the bandsaw.
Once the braces are prepared, I'm ready to go ahead with gluing them on the belly--and I will complete all the work to follow in today's post in about 3 hours. It's essential for me to do all this work at one go, to ensure the freshest gluing surfaces and the most consistent shop humidity. These precautions--these principles--will do much to ensure the finished instrument is as durable and as sound as it can possibly be.

I scrape the belly. Not to remove much material, only to refresh the surface of the wood by removing the slightly grimy, gluey and oxidized outer layer. 


I mark the centreline, and the positions of all my braces.


I locate the body of the lute very carefully on the backside of the belly, and trace the outline. I use wedges to lift up the middle of the belly (remember the belly scoop from episode 2?


I then shoot the bottom of each brace, to give a fresh, perpendicular gluing surface--and, on quite a few of them, to create a carefully controlled amount of curvature on the underside of the brace.




Take, for instance, brace #7--the highest on the belly, the one closest to the body-neck joint and therefore to the fingerboard. The fingerboard, at the body-neck joint, will be significantly arched--somewhere around 3-3.5mm across approximately 100mm of width. That fingerboard arch cannot simply stop at a dead-end where it joins up with the belly; there must be some sort of transition between the fingerboard curve and the (relatively) flat surface of the belly. To achieve this, I plane curves into the top three braces, gradually easing the amount as the braces approach the rose.

This is the way I will support the belly to glue on this curved brace: with carefully-shaped curved cauls. The cauls are made of strips of cardstock, cut to length, stacked, and taped together.

As you see here, it's not only the top three braces that are curved--there is also a small amount of curve in the first two braces in front of the bridge. (I'll explain the reason for the curves in those braces when the time comes, in a future post.) I also place a thickness of cardstock under the rose, to support it, and a few small wedges of card to support the uncurved braces below the bridge.

What can I say? The three-dimesional geometry of a 13 course lute belly is very complex indeed. Stick with me; all will become clear. 

I locate the belly very precisely on top of these curved cauls, on the bottom floor of the go-bar deck. Almost ready to begin gluing... but first, back to the braces for a minute or two.

A little more shaping to do: I use my trusty low-angle block plane (with a very sharp, and slightly curved, blade) to slightly taper the braces, so they are narrower at the top. A few strokes with the plane on each side of the brace is enough, and I am careful not to reduce the already-established nominal thickness of the brace at the bottom.

This is the kind of profile I'm looking for.

Here's how I shoot the bottom of the j-brace: by running it over the upturned bench plane in the vice.

Almost ready to glue, but first I ink black the bottom surface of the three bars that cross the rose (note the authentic 16th century Sharpie.) There are three bars across the rose, and I black only the portion of the bottom surface that crosses it. 
Here's the best shot I have of the whole go-bar deck and glue pot rig. I'm reaching up on the top of the deck, where the oak go-bars are kept (and, may I say, also rocking quite a saucy pose.) I'll be reaching overhead for go-bars throughout the gluing process.
Here's how I glue. Starting at the bottom of the belly, I take a bar, hold it above the rim of the glue jar and brush glue upward onto the bottom surface. I keep dipping the brush in the jar, then brushing hot glue upward, ever higher, until I've got good, even coverage with glue that is a pretty uniform temperature.
And then, I get the bar down quickly in place on the belly, giving it a slight rub as I do so, to ensure good initial contact. Holding the brace in spot with the palm of the left hand, I reach up with the right to grab a go-bar. Still holding the brace securely with the lefthand palm, I use the left fingertips to locate the tip of the go-bar on the top of the brace, while the right hand swings the upper end of the go-bar into place directly above, against the ceiling of the go-bar deck.

Huh?

Yeah. I just read that over again, and it might seem a bit confusing. But really, that's how it goes, and short of actually doing a video of me working (which I'm not interested in doing), I can't make it much clearer. You'll have to just take my word for it--or, better yet, try it yourself.

Still holding the brace with the left, I place two, three, four, or even five more go-bars, as needed, in order to secure the brace. Then onto the next.

Here's the first brace. Seems like a good start.

Oh yes, and after I glue each brace, I clear away excess glue with a long-handled brush whose ferrule I've bent for the purpose.
About half-done. It's getting to be quite a forest of go-bars, and I must be careful as I glue each successive brace not to nudge the go-bars of previous braces. (A domino-style disaster is thus averted.)

Full shot of the go-bar deck. This one was made by and belongs to Grant Tomlinson, whose shop is right next door to mine, and who kindly lends me it whenever I need. (I have my own set of go-bars which I use with the deck, however.) 

We see the top bar here, the one with the most curve, with the belly beneath supported by the curved cauls.

One more shot of the go-bar deck? Why not. It's a simple piece of machinery, but it works perfectly. I love that kind of technology.

And next morning, after carefully taking down each go-bar, and removing the belly from the deck, I see my result. All the braces are securely glued, and they look reasonably orderly.

However, most of these braces, I will confess, are not placed with an excess of precision. I get them in basically the right spot, but, as you can see, there is some randomness to the angles of most of them (the top two, especially, seem a little askew--or is it the third one?) But that's actually what I'm looking for, a little randomness. One sees it all the time looking at old lute bellies. As with their rose carving, one gets the sense that the old makers actually distrusted too much symmetry, proportion and control, and actively as a principle courted a certain amount of chance and randomness in their work. Perfection is a kind of death, both visually and sonically; imperfection is energy and life.

 Next time: shaping the braces, and fitting up to the body. Have a wonderful week.







Saturday, 2 November 2019

Building a 13 Course Lute, Start to Finish, 6: the Rose

Hi again. Today I shall talk about the 13 course lute's rose--preparing the belly for the pattern, gluing the pattern on, and carving the rose. There's much ground to cover, so let's get started.

As I mentioned at the end of the last episode, after thicknessing the belly I placed it in the low-humidity light box for 24 hours, then gradually allowed the humidity to rise over a couple of days until the belly stabilized in the ambient humidity of my workshop (about 40-45% relative humidity.) I'm now ready to thin the rose area, to prepare for gluing the pattern.

The belly centreline and rose position are marked. The central area of my belly is around 1.5mm thick; I'll want to thin the rose area to just under 1mm.

I use a small, curved fingerplane to remove most of the material. It seems to work best (to minimize tear-out) to plane diagonally across the grain. I can only plane up to the belly centreline, since the grain direction will reverse on the other side of centre. I'll need to approach from the opposite direction.

I work in a crosshatch pattern, over an area that is slightly larger than the diameter of the rose pattern I'll be using.

I do the fine thinning of the rose area with a small, sharp hand scraper. I don't want too abrupt a transition from thin to thick, so I also use it to blend these areas.

The rose pattern--photocopied onto 60 lb. watercolour paper--is cut out with about 5mm extra width around the perimeter. I locate it very carefully, and hold it in place with one hand...

While with the other I spread hot hide glue under one side of it, then the other, and then on the top of the pattern. The paper more or less becomes saturated with glue, and the glue on the outside acts as a lubricant for my finger to rub the pattern securely in place, and press out any air bubbles.

Once the pattern's on, I set the belly aside overnight to dry. I can begin rose carving the next morning.

I'll mention one thing about the glue itself. When I glue on a rose pattern, I mix a little alum into the glue, by dissolving a bit in a dish with hot water, then adding it to my glue jar. Adding alum tends to counteract (a little) hide glue's tendency to contract as it dries--a sometimes-useful, sometimes-maddening property. In so doing, I hope to minimize shrinking and distortion of the rose pattern.

My rose carving station--cutting mat, reading glasses, a clean, soft brush, a block of wax, and a photo of the original pattern, which is from an early-17th century archlute from the Sellas workshop.

This rose is going to eventually have a chip-carved border, so before I can begin punching out the pattern I need to do some preliminary layout work with my compass. (One foot has a sharp point, and the other a small, sharp blade.) The first step is to find the true centre of the pattern--which, owing to distortion from the gluing process, may or may not be as marked. I need to adjust the compass and work by trial and error to find the true centre point.

Once I've found the centre, I can transfer it to the topside of the belly by pressing a fine pin all the way through. If I'm careful and don't push too hard, I can make a very tiny mark that will be visible on the top of the belly but won't mar the pattern.

Once I've got that point, I place directly over it a small piece of plexiglass with a pin-mark in it. This is my centre point, and I can use it to lay out all the circles I need for the various rings of the chip-carved border.

Here's a good look at the circles that I've scribed. I don't push too hard with the compass to deepen the cuts--I've found that doing that can distort them pretty badly. Instead, I scribe them fairly lightly, and at a later stage I'll deepen them with my carving knife. 
These are my two main rose carving knives, both made with lengths of 2mm graver steel set in pear handles. The upper one is shaped as a small chisel, while the lower one has a fine, slightly curved knife edge.


Quick note on the carving station. I have three roses to carve, for three lutes; each will take about 24 hours, or 3 good, long days at the bench, making 9 days in total. I made it only a couple of days before... umm... certain parts of my anatomy rebelled, and I had to reconsider my method of work. For the first time, instead of sitting, I decided to try a stand-up carving station instead. I'm happy with the result: the roses turned out nicely, and various parts of my physical person were thankful as well.

I 'punch' the pattern out in two stages, using both of my knives. First, I make a relief cut about halfway through the wood, along all the lines...

Then cut at an angle into those outline cuts, to remove a chip of wood.

I go over the entire pattern--except for the outer edge--making these 'relief' cuts.

Then I use my tiny chisel to punch through the remaining thickness of wood.
Why do it in two steps? Well, I used to use just the chisel to try to cut all the way through, but found that it's impossible for me to shape and sharpen a chisel that will cut well (and I mean almost flawlessly) all the way through the 1mm thickness of wood. I almost always found that instead of a pure cut, the chisel would compress some fibres, and I'd end up crumbling or even breaking out pieces--especially pieces cut across the grain--that I'd then have to glue back in place. It was a pain. I'd spend time fixing, and the pattern would end up weak in crucial areas. This two-part method takes a bit more time to do in the initial stages, but I end up with a physically stronger rose pattern that allows me to be more bold with the top carving. I actually end up saving time, and get a better result.

The majority of the pattern is punched through from the back...

But the outer circumference is punched from the front. The chisel seats nicely in the scribed ring, and I get a very round and orderly-looking result.


I keep my photo of the original rose close by as I begin work on the top carving, because I don't want to confuse the over-under pattern that 'weaves' the organic and geometric parts together.

I us my knife to 'peck out' all the points where the elements cross; all the 'crotches' among the vines and leaves; and all the terminus points of what will become parallel v-channels in the geometric elements.

I'm always amazed by how making single, sweeping cuts to relieve the edges of the vines makes them instantly come to life.

Three cuts are enough to finish these leaves--one down the middle, then one small scooping cut at an angle on each side of it. (It should go without saying that one needs to be conscious of grain direction when doing this; it usually works best to follow the grain 'downhill.')

For the smaller leaves, there is no actual cut, just a pin-prick which I then widen by twisting a small, sharpened point of my own making.

There's the profile. It's a salvaged broken drill bit set into a piece of dowel for a handle. The angle of the tip is pretty flat, and I've ground some (rather random) facets into the sides. It seems to work well.

I make the parallel v-channels with three cuts (all of them free-hand, without any kind of straightedge, just in case you were wondering.) The first cut is the most crucial. I mark out the spacing between them first, by making a couple of parallel knife-pricks.

Then I set my little curved blade in one of the marks, and push it away, rather lightly, in a slightly rocking motion, toward my destination, the terminus point that I've already marked.

Then I pull the knife back toward me, hopefully in the very same line I've just made, confirming and deepening the cut. (It's hard to say exactly how deep the cut is, but I suspect it's at least halfway through the thickness of the wood.)


Then I make an angled cut alongside the central line, first on one side, then the other.

The result should be one nice, crisp, v-shaped channel, sharing the  width of wood in an amicable way with another nice, crisp, v-shaped channel.
And this is where we're at so far: all the vines, lines, and leaves as crisply cut as I can make them. All that's left to do is define the outer scribed rings, and execute the chip-carved border.

First: deepen the scribed rings. Again, I'm going about halfway through the material.

Then: lay out the border. I make these cuts by eye, estimating the angle of the cuts and the width between them. I use the geometric elements of the rose pattern to help orient the cuts.

The border's all laid out--not with perfect uniformity, perhaps, but hopefully with some assurance and energy.

Now, the chip-carving. With the cuts you see here, I'm working with the grain and cutting away from me, but as I get round to the top of the circle I'll be going against it, and will need to reverse my knife blade to cut toward me.
Another v-channel. Always cutting with the grain, I'll need to change my cutting direction four times on this outer side of the ring, then four more on the inner side. (On the very outside ring, I won't make a v-channel, only a bevel cut on the inner side.)

That's about it for this rose. I'll go over it once more with a very sharp knife to relieve little shaggy edges that catch my eye, and perhaps tidy up some cuts that don't quite make the grade. In general, though, I try not to revise my work too much. My sense of how the ancient makers cut their roses is that they worked quickly and surely, and didn't look too much in the rearview mirror, but pressed on to the next job. This gives their work a liveliness and energy that I find very appealing, and I strive for it in my own rose carving.
So there we have it: nine days' work at the rose carving bench. Seems like a lot of time to spend on a purely decorative item, but then again the rose is where, more than practically any other place on the lute, the maker asserts his/ her/ their craft. It's really an emblem of beauty for its own sake.

But it is a long time to stare intently at very small things. I'm going to go outside now, and look at the horizon for a while. I'll be back with the next instalment of the blog in a few days; until then, please have a look at the two other roses I carved alongside the 13 course rose, during my nine-day stand.

For the 8 course lute: the rose pattern from the 1592 Venere, in the Accademia Filarmonica, in Bologna.

For the 7 course lute: a variation of the Gerle/ Bellini pattern, drawn by Ray Nurse.