Starting Work on a New Cello

Well, today’s the day to start work on a new cello. The first step is to choose the model. I have three, all of my own design, varying about a centimeter in length between them. The first cello I ever made (outside of school, that is), was a Strad copy, and the next a Goffriller copy; but I quickly discovered that using an established model was – well, boring. You kind of have to do it with a violin; Strad or del Gesu, it’s what’s expected. Either you make a faithful copy, or a variation, which ends up looking like a misconstrued copy. Violas and cellos, however, are different – the models vary widely, so you can design something original that doesn’t look like a bad attempt at a copy.

So where do we start? As the Red Queen said, begin at the beginning, and when you reach the end, stop. The beginning is the string. We tend to think of a violin as an instrument with strings attached; but the truth is just the opposite: it’s a string with an amplifier attached. The sole purpose of that varnished box is to make the sound of the string audible. The violin is the ultimate green machine; a perfectly self-contained sound system. Nothing to plug in; like the Beatles said, all you need is love, sweet love (and a bow).

Violin design has always been driven by string technology: makers since the very first, Andrea Amati in Cremona back in the early 1500’s, have basically taken whatever strings were available and made a box to go with them. Without getting bogged down in string theory, here’s the basic fact we have to deal with: in order to produce a sound, a string has to have a certain tension. And that tension depends on the mass of the string and its length. The greater the mass, the shorter the string needs to be. Old cellos are huge. Why? The strings had to be immensely long – all that was available was gut, and even then the C string, often braided, still had to be as thick as a rope to get enough tension to make it respond.

But add some metal, and the thickness goes down considerably. Heavy Metal meets Amati; the results, especially in the past few decades, have been strings that give you all the sound you want or need at a much shorter length – so you can design and build a smaller cello.

So: we start with a string length. Oh, no! Not more math! This is supposed to be making a cello, not beginning algebra! Hang on; this is the last of it. The neck is in proportion to the F-stop, which is the distance from the top edge to the notches of the F-hole. Why? So that when you shift into fourth position, you hit the note. So you choose a string length you want, do the math, and come up with an F-stop, and then go from there (and that’s that; if you want to know more, check the photos at the end of this posting. In fact, that’s the way we’re going to do this: basics up front, more detail for those who want it at the end).

We have a pattern. We need a form, to build the cello around. Here’s where we need to stop and send a huge thanks to a colleague, Bill Monical. He did an exhibit of instruments and then published a catalog called Shapes of the Baroque; and it opened our eyes (he’s the one who pointed out that instrument design has always been driven by advances in string technology). We had in mind a certain way that these things must be made; he showed us that there were all kinds of methods – inside form, outside form, no form at all. Everyone I know uses an inside form, as do I; I once made a viola without a form, and while it was exhilarating, I wouldn’t want to do it on a regular basis – kind of like riding a bicycle without a helmet.

Thoreau said to beware of any venture that requires new clothes. Yeah, well; just imagine that guy as a weekend guest, he’d drive you nuts. While you’re making the pesto he’d be right there at the counter telling you if you use a food processor instead of a mortar and pestle it’s not the real thing and you’d better toast the pine nuts. Sometimes a new beginning needs celebrating: in this case not with new clothes, but a new form. The old one, truth be told, was getting kind of ratty; I’d changed it here and there after each successive cello, and every time you use it, you lose a bit of wood at the block areas. So a new form: go back to your hut in the woods, little Henry, sometimes new clothes are just the thing. Like getting behind the wheel of a brand new car – a convertible, on a spring day, with the top down, with the prospect of a long drive ahead through gorgeous country, with a map and a place to end up but no need to get there at any particular time.

So I went up to the local lumberyard on Friday and told them what I needed, and when they heard what it was for I got the usual look of curiosity mixed with pity, and when the guy was cutting the mahogany plank to fit in my car I saw high up on a rack on the second floor what I really needed; too late, he had already cut, so now I had a stack of absolutely gorgeous mahogany. Luckily my son Jack needs some wood for a box to go with a chessboard he’s making. Over the weekend I joined up the pieces. Why not just use one large piece? Movement. Around here, where the humidity can go from forty to eighty in the space of a few days, a single board moves too much; and it has to be as true and flat as possible. So I join up several smaller pieces, like this:

In order to further reduce movement, you want to reverse the individual pieces, so that the grain runs in alternate directions (which reminds me of a sign I once saw in the window of a store in New York: Closed For Alternations):

I draw these lines on the assembled pieces, and number each one, so that I don't have to fumble around for the right piece once I start gluing:

There are too many pieces to glue all at once, so I do it in three steps; each side, and then the middle parts. The two pieces closest to you are scrap wood to protect the joining surface from compression from the clamp:

Once joined, the form is planed to the proper thickness:

And then after the pattern is traced on the form, it's cut on the bandsaw, as close to the pencil line as possible -- leaving just enough wood to file off the kerf marks of the blade (the kerf is the angling of the teeth; like a snowplow, it widens the cut just enough to allow the dust to escape so that the blade doesn't get stuck):

Here’s the form after being cut out on the bandsaw:

I file the form, smoothing the edges and making sure they're exactly square.



Then I lay out the spaces where the blocks will go. There are six -- one at each corner, to reinforce the rib joints, and one at each end:

The next step is to cut the places for the end and corner blocks (as you can see, some of the block areas already look cut out. Since I knew it was going to be cut, I didn't always bother to extend the piece all the way there in the first place -- but I did leave enough so that I could finish the curves beyond the cut line). Once cut, they must be filed to be absolutely square, or else the blocks will be out of true:

And then the form is varnished, which will also help to reduce any effects of weather:

The sides of the form are also varnished, to reduce the chance that any errant glue might attach the ribs to the form. I use very fine sandpaper (400 grit) to make the sides as smooth as possible, and also to break the sharp edges of the form so that they won't crack the ribs when taking the form out:

The very last step is to glue size the places where the blocks will be set. A tree is essentially a hydraulic machine, designed to lift water from the ground several stories into the air. Even cut wood is like a sponge, and even more so when it's been drying for years. You'd be amazed at how much glue it will drink up -- especially at the ends, where it's head wood. You can brush the glue on and literally watch it vanish, stroke after stroke. It takes two or three applications until it stops soaking in; if you don't do this, you get what's called a starved joint when you try to glue something to it. Rather than staying on the surface, bonding the two pieces together, the glue vanishes like rain in the desert sands, and when you unclamp it, the two fall apart. So you have to size the gluing surfaces of the form to prevent this happening.

And that's it. Now we have a pattern and a form. Tomorrow we set the blocks, and off we go.

Designing the cello:

No system of design has survived from the era of the great classical makers, and so any attempt to replicate one involves a high measure of conjecture. There is no lack of theories, however – the subject has engaged mathematicians for centuries, and there are enough ‘systems of design’ to fill several volumes. The problem is that none of them, when used, can seem to generate the models of the great makers. Part of that is because of the way they’re built (as we shall see) – the results stray quite far from the form. But there are some extant forms and drawings that are pretty reliably attributed to Strad (although with the ubiquity of reliquaries in the environs containing pieces of the true cross and leftover body parts, and the fact that the drawings and forms were sold by Stradivari’s grandson, a known rascal, a healthy dose of skepticism is perhaps not out of the question). The drawings show the locations of the f-holes, using compass arcs; but that just creates another problem. So far no one seems to be able to say exactly where the measurements for those compass arcs come from. The drawings do show one thing, however: the lower corner is keyed to the lower eye of the F-hole. And this does seem to be sort of standard for the great makers (sort of being the operating qualifier here). So where does the rest of the outline come from? So glad you asked. I put together a whole bunch of measurements from great cellos; figured which seemed to work the best, and then laid out curves to connect them. Oh, I’m sorry: I meant to say I used a system of design ----

Anyway, here’s how it works:
I lay out the desired F-stop, keeping in mind that the real one will be longer than on the drawing, since it will be over the arch. The Fstop and the neck will give me the string length I want:

Another thing that seems to be pretty consistent is the length of the F-hole. In placing the F-hole, I decide how far apart they are going to be, and how much cut wood they will span (both are critical in determining the flexibility of the top):

Then I can extend a line and that will show me where the lower corner goes:

Interestingly, it seems that the distance from the outside of the eye to the edge is the same for the upper and lower; so depending on how broad the chest is, you know where to put the outside of the lower corner. So far so good, right? As for the rest – a few outside dimensions (upper bout etc), and then – as the Italians say, per occulo. Bene? I guess it’ll have to do until the real thing comes along.