September 22, 2012

Q: In your book you recomend 30′ for a top radius, if one decides to buy a commercialy made disc. On the other hand I saw that James Goodall and Robert Taylor use 50′ and 65′ radiused dishes and Jim Olson & Kevin Ryan make FLAT non-radius tops. Olson said that he feels that these produce more responsive tops. So why, exactly, do you recomend a 30′ one? Wouldn’t a 65′ one be better since it is closer to being FLAT?

A: It’s a question of balancing various factors — very similar to a cook’s gauging how much of this and/or that spice or flavoring to use in making a dinner.

In the guitar (instead of carrots, lamb, and oregano) the ingredients are the string tension, the torque from the bridge, the mass of the braced top (which is arrived at by any combination of top thickness and bracing mass), top bracing and reinforcement (that is, the pattern and layout of the braces, as well as their profiling and height), and desired target sound (the resultant mix of monopole and dipoles, as well as sustain and dynamics).

There is no “correct” way to make a guitar. If there were, they’d all sound the same — just as if there were only one recipe for making French onion soup: then all French onion soup would taste exactly the same. In the biological realm, it would be equivalent to having every wife, husband, boyfriend, girlfrend, son, daughter, etc. be clones. So let’s forget “the one best way” of doing something complex.

Jim Olsen holds that a flat top is the most responsive. Very well. But what, exactly, does that mean? Does this not have something to do with how thick and/or stiff the top is, or how it is braced? And how it responds to the mechanical pull of the strings? I suspect that it does. So if we were to imagine a VERY thin top that is made flat, it would be easy to imagine it buckling or caving in under the pull of the strings… unless the bracing were beefy enough to make up for the weakness of such a flat plate. In other words, you could make that flimsy face hold up by adding more reinforcing.

You could also/instead make that flimsy face hold up better by putting an arch or dome into it. Arched structures are stiffer and more stable than flat ones — just as a pointed or arched roof on a house will hold up to rain and snow better than a flat one. Western architecture took a mighty step forward when structural doming became possible: the materials themselves — rather than supports, trusses, beams, and buttresses — achieve the required structural integrity, Analogously, if you put a dome or arch into the guitar face, you could use less bracing and achieve the same stiffness with fewer materials (i.e. less mass).

Less mass is good; it means that the strings have to strain less to coax sound out of the top. The strings need to work harder to get sound out of a heavy top — exactly as a horse has to pull harder to make a heavily loaded wagon move. You can appreciate that different archings/domings induce different amounts of stiffness into a plate. As can different thicknesses of that plate. And so can different sizes and layouts of braces. These are, in fact, the three main ingredients of top-making, exactly as flour, water, and eggs are three main ingredients in bread making. And in both guitars and bread the ingredients can be mixed or combined in different ways to produce a successful product. In the guitar this goal is: a top that is intelligently constructed and reasonably lightweight (which goes to sound), and also able to hold up to the pull of the strings (which goes to long-term stability of the guitar).

In the guitar, ridiculously small (or small-seeming) amounts of these various ingredients can make a difference you can clearly hear. For instance, a bit more or less top thickness can offset a bit more or less doming. A bit more or less bracing can offset a bit more or less top thickness. And so on. Identifying and using only one of these factors as being “most important”, appealing though that idea is, turns out not to be realistic. If the guitar were a political construct rather than a mechanical one, then it would work best as a democracy in which every component is (not to make a pun) given its proper voice. To make one into a “leader” (in our sociopolitical sense of the word) is not what a guitar is all about.

September 22, 2012

Q: Recently I bought your books & DVD and I found one sentence particularly interesting: you mentioned that if a guitar with a normal flat back had an arched top, its dynamics would be unique. Can you please reveal from your experiences in which direction the sound will change, compared to that of a normal flat/domed top?

A: It’s an interesting question, and to my knowledge no one has yet made a guitar like this. Mario Beauregard of Quebec, on the other hand, has been making something truly new: nylon string guitars with arched backs and flat tops.

The arching of a plate stiffens it: it improves its stiffness-to-weight ratio. And, acoustically, it raises the plate’s pitch: its vibrational behaviors are shifted toward high-frequency signal — such as the violin has. A small highly domed plate is not likely to have a good monopole — that is, a good low end. Also, the greater the arch, the shorter the sustain is likely to be.

Cellos have a low end, and they have violin-like arched plates — but they are huge compared to a guitar. So part of what we are discussing is the SIZE of the plate, in addition to its doming vs. flatness. But it would be difficult to play a cello-sized guitar.

There’s another factor too: what wood the arched plate is made of. Traditionally, all arched-plated instruments (violins, violas, cellos, standing basses, and jazz guitars) have used spruces and maples — spruces for the tops, maples for the backs. Maple does not have much sustain compared with some of the woods used in guitars, especially Brazilian rosewood (although, in my experience of the maples, Eastern rock maple has the most, Western broadleaf maple has the least, and European maple — which is a sycamore — has some). Therefore, if we’re talking about a guitar with an arched spruce top and a flat maple back, it would likely have a sound characterized by a quick attack and a quick decay: bright, brisk, zingy, sharp, and not much sustain.

Sustain is not a factor in arched-plate and bowed instruments. They don’t need natural sustain: they will make sound as long as the player continues to scrape his bow over the strings. In the guitar on the other hand, because it is a plucked rather than a bowed instrument, the sound stops as soon as the strings do — just as happens with the banjo, lute, koto, ukulele, mandolin, dulcimer, harp, or harpsichord. [NOTE: the harp and the harpsichord are both excited by plucking action; the piano is excited by hammering action.]

It’s not likely that these traditions had such acoustic considerations behind them. The science of acoustics didn’t yet exist, and early European makers would of course have used the woods available to them — in this case the European alpine spruces and maples. They were a long way from having access to imported exotics from the New World. Also, in those days, the cost of labor was cheap and the cost of materials was high, so a thick plate of an imported exotic wood (that you’d carve down into an arched surface, and in the process wasting much of the wood) would have been quite expensive, compared to a thin plate of the same wood such as would eventually be used on guitars.

December 18, 2011

Q: In my limited experience with classical guitars there seems to be a need to have a more flexible top on the bass side and a more stiff top on the treble side, giving the warm and low sound on the bass and more sustain on the treble, as well as preventing the percussive anvil sort of trebles . The flamencos seem to have a somewhat less flexible bass and a flexible treble side, which gives them a somewhat percussive sound with rapid decay in sound. This is at least consistent in the examples I have and have had the opportunity to play. 

If one were to attempt to make a classical guitar as stated above would a change in top thickness to accommodate the additional stiffness required on the treble side be in order, or would changing or stiffening the bracing on that side be a better option than making a guitar with a top that is not of uniform thickness? 

So…essentially, bracing or thickness or both?

– – – – – – – – – – – – – – – – – – – – – – – – – –

A: Your question addresses the primary issue of whether it is desirable to design a plate that relies on symmetrical design, or an asymmetrical one, for optimal functioning. The matter is clouded by the fact that some remarkably good guitars of both types have been made. Also, some pretty unimpressive guitars of both designs have been made. So there’s no clear winner, and I’m not convinced that the mere fact of symmetrical or asymmetrical construction is the most important consideration. I mean, if it were, one of these designs would produce consistently better results than the other.

As I read your question, I am translating it (to myself) into language that I am most comfortable with. So my answer might put a different spin on things than you’re used to. Let’s see if this makes sense to you.

To my knowledge, flamenco guitars achieve their characteristic sound by having looser, more flexible tops in general than classic guitars. Specifically, they are so loose that the monopole (the base) is discharged quickly — thus giving those instruments the characteristic traditional “dry” sound without a lot of sustained presence; their sound is more percussive. Usually, if one believes in asymmetrical construction, the treble side is made a bit stiffer than the bass side. In any event, compared to classic guitars, flamenco guitars have a more prominent cross-dipole. They are “looser” in that mode, in which the top moves side-to-side across the centerline and the bridge teeter-totters with one wing going up as the other goes down, and back. You can get a sense of this looseness by simply pressing down on one of those tops with your thumb: you will probably feel them “give’ fairly easily. You can also test for cross-dipole compliance by lightly putting one or two fingers of one hand on one wing of the bridge and lightly tap on the other wing with your other hand. You should feel a definite and instantaneous displacement as the bridge see-saws relatively freely. Classic guitars will be built less loosely, and will accordingly have less mechanical “give”. Neither one of these designs is “good” or “bad”, by the way; they are simply different — and different for a reason.

To my knowledge, these characteristics of flexibility and movement in a guitar top are best achieved by careful BUT SYMMETRICAL calibration of the plate — and I do not try to build any mechanical or dimensional asymmetry into my own guitar tops. But I acknowledge that other makers of successful guitars use dimensionally asymmetrical faces, so I’m inclined to believe that the motions of the cross-dipole (or any other mode) apply to a variety of architectures. At that point, the issue becomes that of basic calibrating so that one is “in the ballpark” and doesn’t overbuild or underbuild. I go into several detailed discussions of this concern in various chapters of my book The Responsive Guitar. You’ll undoubtedly find some of my ideas from there worth at least thinking about.

Your question addresses the question of whether it’s appropriate to add or subtract bracing in order to accommodate to changees in top thickness — as a matter of making material stiffnesses — and hence vibrational action — be consistent. The matter of achieving a balance between stiffness and/or looseness through top thickness vs. bracing mass is central to lutherie. And in this balancing act, there are two factors that inform my thinking.

First, in the traditional approach, if one were to make part of the top thinner, one would indeed want to “compensate” for it by making the bracing a little stiffer. This is assuming that the maker’s goal is to have AN EVEN GRADIENT OF MECHANICAL AND VIBRATIONAL STIFFNESS FROM THE BRIDGE TO THE PERIMETER, IN ALL DIRECTIONS. This is certainly my goal. In purely practical terms, this is surprisingly tricky to do until you begin to understand what you’re doing and have some practice at it; after that, it’s surprisingly easy. So an experienced hand and eye are really useful to have. I should add that, incidentally and technically, the gradient that I visualize in my work is only even in the sense that there are no lumps or irregularities of localized stiffness between the center and the perimeter; but it is not the same slope on all axes, in the sense of being identical. The longitudinal gradient is stiffer than the lateral gradient.

Be all that as it may, the second factor is, I think, just as important. It’s also interesting, subtle, and elegant — and obvious. So much so that it took me years to see it. It’s the “water running downhill” principle; you know: that water will find a way to run downhill regardless of trees, rocks, or irregularities of slope or terrain — because that’s the nature of water. In fact, such downhill movement of water cannot be prevented short of putting up a barrier or obstacle that exceeds the power of gravity over water.

Interestingly, sound energy is the same, except that instead of running downhill it wants to radiate off a guitar top — with all the freedom of water running downhill. It’s the nature of sound energy to dissipate into its surrounding medium, be it air or water. If we think of sound energy as seeking its easiest path “out”, as water wants to find the easiest path “down”, then it’s a short step to seeing physical unevenness in a guitar top as being analogous to unevenness in downhill terrain. And unless any of this unevenness is significantly huge, both water and sound will continue to flow and radiate. Tweaking any of the minor irregularities of slope, terrain, or structure will by no means stop any of the flow or radiation; they’ll find a way to get from here to there.

Let’s take a look at how this might work in a guitar. Let’s assume that you have a dimensionally asymmetrical plate, as you described above. And let’s further assume that the structure — irregularities and all — is “in the ballpark” as far as not undermining the monopole, cross-dipole, and long dipole. Or, saying the same thing with different words, that the irregularities are such that they allow the capacity of the plate to engage in these modes without messing any of them up). The top will flex and bend and seesaw and vibrate just as all the theories, diagrams and Chladni patterns suggest. The question then becomes: what makes you assume that such a top and its vibrational modes have to function symmetrically around the guitar’s centerline? Or that the various vibrating quadrants and subsection of the face will map out as being active with elegant evenness, symmetry, and consistency? I mean, no one expects water to flow downhill over a natural terrain in a straight, even, consistent, and predictably regular line, do they?

Your question cites differential side-to-side construction. This is the axis of the cross-dipole, which is (in theory) a see-sawing action around the centerline. If we were to imagine two kids on a playground see-sawing up and down on a teeter-totter, that device will be pivoting on its center point as a matter of the manufacturer’s design. But, suppose one of the kids is heavier than the other one? That would introduce an irregularity into the flow of their play. The manufacturer of the teeter-totter wouldn’t care about that, of ocurse; only the kids would. And to anyone to whom the kids’ fun was important, they could compensate for this disparity in mass (in the “playing field” or “gradient”) of that teeter-totter by simply adjusting the fulcrum point to a somewhat off-center position. Then, being better balanced, the kids could see-saw happily and without strain: same mass, same energy, same frequency, easier and more harmonious oscill.ation.

This is close to my sense of how the guitar works. To repeat, using other words: if there is sufficient unevenness in the top because of any design idea of the maker, and the design variable isn’t so huge as to throw a monkey wrench into the natural functioning of the guitar, then that “uneven” top will accommodate to the needs of the energy flow of that irregular structure all by itself. It’ll adjust, within some limits (of the natural capacity for flexibility of its woods), and perhaps wind up fulcruming, say, 1/16″ off the centerline. It can do that because the guitar lacks a fixed fulcrum in the way that a teeter-totter has one. Therefore, the vibrating quadrants of the top may be a little bit lopsided or asymmetrical in actual movement, etc.. But, as far as dissipation of strings’ energy is concerned, nothing has been prevented; it’s simply found its way out via an alternate path from what “the manufacturer’s blueprint” might have suggested.

This doesn’t fully answer your question yet, but my answer required me to have sketched in this background before addressing it specifically. This background, to repeat once more, is that tonewood that has been worked to more or less optimal dimensions has a certain innate flexibility of vibratory potential. There are no fixed fulcrums or vibrational nodes. And it may not matter that you’ve made an irregular plate — as long as you have not made it so uneven that you’ve pushed the plate past some limit of being able to perform its principal vibrational tasks.

So, to sum up: my answer to you is in four parts:

First, that yes, if you make a top thinner on one wing, which necessarily weakens it, then there’s a logic to adding bracing stiffness to it to make up for that weakening. I believe that one should aim toward at least some standard of evenness of physical/mechanical/tonal gradient if one’s goal is to make better and more reliable guitars.

Second: I believe that these maneuvers work most effectively if the top and braces are “in the ballpark” as far as optimal mass and stiffness are concerned — rather than the system being overbuilt as is often the case. Or underbuilt, if you’ve gone too far in thinning. If you’re overbuilding, then the thinning and bracing work that you are considering might be nullified or overshadowed by the fact that the structure is still too stiff. Or maybe the part that you’ve thinned will work fine, but the part that y ou haven’t thinned will be inhibited. But you won’t get 100% cooperation from such a top.

Third, you will probably produce minimally uneven tops no matter what you do; guitars in the real world always have something or other that’s not optimal.

Fourth, as with the example of water cited above, and if your gradient is not too unevenly made to begin with, then what you’ve constructed or misconstructed probably won’t matter. Or at least it won’t matter very much within the context of the flexibility of vibrational potential that the top has. The top will bend itself (sorry about the pun) to work in any way it can, to release its load of sound energy. It’ll modulate itself physically and vibrationally. As I said above, the vibrationally active areas of the top may functionally be a little bit asymmetrical, things may be a bit off-center, vibrational patterns might not be quite mirror-image, etc. But this is no big deal: the top plate has the capacity to function at least a little bit like that in order for the system — as it is physically constructed, with its unevenness and irregularities — to engage in an adequate monopole, cross-dipole, and long dipole. Finally, the sound will get out, sometimes because of, and sometimes in spite of, and sometimes without being much bothered one way or the other by, the work you’ve done. And I think this is where a bit of the magic comes in.

I hope this makes some sense.

December 18, 2011

Q: If the soundhole is not in the traditional location at the end of the fretboard, is there a better bracing pattern than the X-brace, in your experience? 

A: The soundhole is where it is, as a matter of tradition rather than critical thought: it’s always been put there. One might put this in terms of history trumping dynamics. History and tradition notwithstanding, the guitar soundhole has a tonal role to play, and I devote an entire chapter of The Responsive Guitar to the mechanical and dynamic functions of the soundhole with respect to brace location.

As far as the mechanical dynamics go, the soundhole in the Spanish guitar is outside of the main vibrating area of the face; it’s isolated from it by a massive brace that acts like a dam, and the comparatively delicate fan bracing on the other side of it does its work without being affected by exactly where, above that dam, the soundhole is. In the steel string guitar, instead, the soundhole is inside the main vibrating area of the face. It represents a mechanical perforation of that plate — and it necessarily weakens it. Imagine a drum head (a vibrating diaphragm) with a great big hole in it, and you’ll be able to grasp one of the principal bad dynamic ideas in the steel string guitar.

As far as bracing placement is concerned, my opinion is that the acoustical work of the bracing is more important than the specific location of the soundhole, and that these shouldn’t be in conflict with one another; therefore, I think there’s more to be said for moving the soundhole “out of the way” than moving the bracing around. Those kinds of judgments depend, of course, on understanding the functions and possibilities of various bracing systems. You don’t just want to move stuff around randomly.

Speaking of tradition vs. critical thought, the Kasha guitars (with the innovative Kasha bracing) were the first ones to focus on the bracing layout first and the soundhole placement second — in spite of how oddball those guitars looked. I give the Kasha people credit for understanding about putting the soundhole in a place where it helps rather than hinders. The soundhole’s dynamic function is to act as a port (as per the discoveries of 18th century Dutch scientist Christian Huygens, which I go into in my book), and as such doesn’t HAVE to be in any particular location. I recommend reading my book if you haven’t already.

Whether or not one moves the soundhole, it’s useful to have an idea of what each bracing layout can do, in terms of its mechanical and vibrational possibilities. Or impossibilities. There’s a logic to each bracing pattern and each one can be tweaked and altered in many ways — some subtly, some radically. And, as I said, part of the challenge is to not put the soundhole where it’ll create a problem. Either way, we’d have to understand how these factors interact before going on to talk about “better” or “worse”… because there are many ways to spoil the efficacy of any blueprint pattern and there are many ways to “get it right”.

But, let’s get back to your question about “X” bracing and soundhole location. The virtue of “X” bracing is that it ties the face together so as to create the possibility of a dominant monopole motion. Now, it won’t work nearly optimally well if the bracing/top are overbuilt and too stiff, or if the plate isn’t properly or consistently tapered, etc., and your job is to learn to do an INFORMED balancing act. Plus, the soundhole is right in the middle of this, sort of like interrupted ceiling beams that are holding up a roof that itself has a great big hole in it.

If you can get comfortable with the idea of relocating the soundhole to somewhere else then you do have to think about what to do with its area of topwood that is newly available as vibrating diaphragm. I mean, you’re creating an empty space bigger than any other empty space on that braced top. You could close the “X” brace up a bit… but that would necessarily open up the bass and treble quadrants, and you’d have to figure out if you were comfortable with that. As I said, it’s all a balancing act. If you didn’t want to mess with the balancing act then you might think about installing one or more finger braces into that space, to tie it into the rest of the bracing. I don’t have a better specific answer for you than this.

My unspecific answer is to think of what your changes might signify in terms of the main modal movements of the top: the monopole, the cross-dipole, and the long-dipole. Mainly, “X” bracing is a recipe for bringing out the monopole; it ties everything together. Fan bracing is a recipe for facilitating cross-dipole; there’s nothing there to prevent or inhibit that mode. Ladder bracing is a recipe for emphasizing long-dipole; it destroys the monopole and the cross-dipole.

So, if you were thinking of closing in the angle of the “X”, you would be justified in suspecting that this will facilitate more cross-dipole: the legs of the “X” would be stiffening the plate in a different way, as a function of their new orientation. So, the equation might look like: (Take away soundhole) + (closing in the “X”) = (more cross dipole). A second equation might be: (remove soundhole and add a bit more topwood) + (leave “X” the same) = (maybe a bit more monopole). Another equation might be: (remove soundhole) + (enlarge the space by spreading the “X” legs out) + (make new bracing accommodations to reinforce this larger space) = (?).

My point is that if you can accept that there’s some actually useful information contained in technical jargon such as “monopole”, “cross-dipole”, and “long-dipole” (which are simply formal words for some basic concepts of top vibration, and hence sound) then I think you can begin to have really interesting ideas about how to problem-solve your next guitar project, and make it better.

November 3, 2011

Q: Assuming you’re looking for a back to work in tandem with the top, as opposed to a reflective back, should the back also be thinned till it “relaxes”, as you do on your guitars?

A: Ummmmm… this is a really interesting topic that very few people have done any thinking about — and most of the ones that have are classic guitar makers, not steel string guitar makers.

The matter is too complicated for me to write fully about in this format, especially as I have written about exactly this kind of thing in my book. Have you read my book’s chapter on the functions of the guitar back? If you haven’t, it’ll be useful for you to do so. Mainly, my answer is based in the proposition that the job of the guitar top is to generate an optimal mix of monopole, cross dipole, and long dipole signal… which gets converted into sound a bit further on down the line. The back has a different function — although, frankly, almost no one that I know of has ever considered making a back that might have a purposely dominant monopole, cross dipole, long dipole, or whatever.\

The back has not been studied like that. And one indicator of this circumstance is that while guitar tops have been made with all kinds of variants of “X” bracing, double-X bracing, fan bracing, lattice bracing, ladder bracing, Kasha bracing, radial bracing, and even the most oddball experimental bracing, over the years… 99.99% of all guitar backs have been made with three of four parallel braces since the back was invented. Period. So our information about the possibilities of the back is limited to one model of bracing that has been done over and over and over and over again. I show some experimental back-bracing ideas on page 91 of my book The Responsive Guitar; take a look at them.

Also, consider that it doesn’t matter how the back is constructed if it is not allowed to be active. For instance, Bluegrass guitars are played with the guitar’s back resting against the player’s body. These backs are significantly damped out. That is, they are prevented from participating in the dances of the frequencies. Would it matter to that kind of guitar that the back has been thinned to the relaxation point? Not at all. That back isn’t expected to do anything. The technique of playing the typical bluegrass guitar (standing up, strap around shoulder, guitar resting against player’s body) does not concern itself with the back’s doing anything in particular except maybe acting as a reflecting surface and otherwise keeping the dust out. And, as I say in my book, (at the risk of becoming unpopular): the use of a highly resonant and expensive wood on the back of a guitar that has no use for a functioning back is to waste the wood.

But aside from all this, to get back to your question, the short answer is “yes”. My prejudice is to make the back more flexible than other makers typically do. The reason for making both the top and the back flexible to begin with is that everything else you do to them does nothing but stiffen them up. You brace them, dome and stress them, and attach the perimeters to the guitar rims. Pretty soon, you’ve got something that you’ve (perhaps inadvertently) made really too stiff.

But too stiff for whom? For you? Maybe; or maybe not. For me? No, I don’t really care. For the strings and their work? Yes: they care.

I first got onto this idea, years ago, from an interview with David Rubio in [long-since disappeared] Guitar And Lute Magazine. Rubio recommended thinning the free (unclamped and unbraced) top until it had no tap tone of its own. If it still had an identifiable tap tone, it would be introduced into the guitar’s structure and responsiveness. But if one introduces a “tone-neutral” top (or back) into the system one could then build an appropriate tap tone back into it by bracing it, attaching it to the guitar, and bridging and stringing it. The basic equation is: if you start out with this, and then add that and something else, you wind up with this + that + something else = something greater than what you might think you have..

November 3, 2011

Q: Do you use the same X amount of flexibility for all your guitar tops? Is there any reason to have a different, Z, level of flexibility when you use woods of different species? 

A: I certainly try to for the same level of stiffness in every guitar top I make, regardless of species of wood used, for reasons of consistency of sound and musical responsiveness.

However, it’s not quite a simple yes-no. The thing is, if you’re going to build a guitar that’s slightly bigger or smaller than the last one you made, then you’ll need to factor some accommodations into your measurements.

A bigger guitar top is weaker than a small one of the same absolute mechanical stiffness (i.e., the same mechanical stiffness is asked to cover a larger span or area), and will have to be left thicker to compensate for that weakening. And vice-versa. For example, imagine standing on a plank that serves as a bridge to cross a 5-foot wide creek, and a longer but otherwise identical plank spanning a 10-foot wide creek. The latter will sag more when you stand on it. Your weight is the same, just as the guitar’s string tensions are the same. The resistance over the span needs to be adjusted, however, if you want the sag to be the same amount.

That “sag”, in the guitar, goes to vibrating-plate motion, which has everything to do with sound. You probably don’t care how much sag there is in a simple footbridge, but in the guitar the ‘sag amount’ corresponds to how much or how little the guitar face can move and flex in order to produce sound. There’s a direct correlation, as sound is nothing but excited air molecules. Finally, we’re (you’re?) trying to build guitars that are optimally permeable and receptive to the strings’ energy level and budget. Assuming the use of standard strings of a standard scale — which goes to the energy budget — this implies the same (or at least comparable) optimal amount of structure.

October 16, 2011

Q: Obviously, your method [of working tops to stiffness than to target dimension] is going to lead to different thicknesses for every piece of wood of a certain species to get the same flexibility. I am curious, though, if you find that different species have to be worked to a different degree of flexibility? For example, say you thin your steel string Sitka tops to have X amount of flexibility with a Y weight on them. Do you use the same X amount of flexibility when you are using Engelmann or Cedar, as well, or do you find that you need to develop a Z amount of flexibility for a different species? Thanks.

A: You’re correct that in theory no two pieces of topwood will wind up being exactly the same thickness if one follows my method. That is, we’re looking to achieve a consistent level of RESISTANCE, and different woods will have different proportions and densities of xylene, cellulose, and fiber with which to achieve that level of resistance.

This level of resistance isn’t some theoretical number that’s gotten by formula — although it can be gotten that way. The level of resistance is organic to the guitar: it is set by the top’s need to work with the strings’ pull, modulated by the kind of sound (character, sustain, overtones, etc.) that you might be after. And that’s all. Various gauges of strings, of various scale lengths, exert a certain amount of pull which, when excited, provide the motive force and energy budget. This is, of course, affected by things like how hard the player plays, bridge height and torque, etc. I don’t think any of this is exactly new information to anyone who’s been paying attention.

If the top is too resistant to the strings’ pull, then the mechanical response of the guitar is hampered. It is compressed into (i.e., limited to) regions of high-frequency/low amplitude activity/signal. You might or might not like that sound, but it will be a limited sound. If the top is too wimpy and flexible then it MIGHT have to rely on the bracing to restore its dynamic balance to a higher level of stiffness and hence response. The bracing will reinforce, or undermine, or overpower, what the top itself is able to do. It’s a partnership.

Steel strings on a guitar exert a pull of around 180 pounds. Nylon strings exert a pull of nearly 100 pounds. Let’s say that the strings on your guitar exert 125 pounds of pull and torque when tuned to pitch. I’m just grabbing a number here. Now consider: it really doesn’t matter whether your guitar has a Sitka spruce top, an Engelmann top, a redwood top, a European or Lutz spruce top, a cedar top, a koa top, a mahogany top, or a plywood top. That top is, in every case, going to be driven by 125 pounds of string pull/drive/torque. We’re assuming everything else being equal here: guitar size, soundhole size, bridge height, etc.

The question is: why would you put a top with any different stiffness (than that needed to deal with a 125 pound pull and torque) on your guitar? Put it another way: if string gauge were like octane in gasoline (i.e., a measure of its ‘oomph’) and top stiffness were like tire pressure (a certain ease or hardness in car maneuverability), then regardless of what octane gasoline you fill your car’s tank with, why would you change the tire pressure every time you gassed up?

Now, there are different things than mere stiffness going on. There’s also internal damping and mass. Different woods WILL behave a bit differently, at identical stiffnesses, when excited by strings, because of these other factors. Some woods will suck the strings’ energies up pretty quickly and damp their motions. Some will be vitreous and live and allow the strings to remain excited for longer. Some will be internally brittle. Some will be internally tough and ropey. Some will be very dense; others will be like Styrofoam, etc. You get the idea. So there’s a lot to be said for familiarizing one’s self with the average tonal potential of different woods, as well as which woods tend to be more consistent in qualities and which species have a wider, less consistent, range of qualities depending on which plank or log you’re working with. The main thing is to work with woods that have the least energy loss possible. You want the energy to go into the air (sound) and not into the woods and materials of the guitar.

If you’ve ever been to a lumber yard you’ll have noticed that some planks of a given wood are dense and heavy while other planks right next to them are not. Such things affect a guitar’s behaviors, and need to be factored into your calculations — if only to the extent of your using the same selections of woods on the guitars that you make. You may or may not have a clue as to what difference any characteristic that you’re aware of might make, but it’s smart to not throw uncontrolled variables into your work if you can help it.

Having said that, EVERY guitar will produce a monopole, a cross-dipole, a long-dipole, and whatever other mode of motion you think is important enough to consider. If you don’t know about these, please stop reading this right now and read up on these fundamental vibrational modes of a guitar top: they’re critical. Every guitar has SOME mix of these modes, and every guitar has a fixed energy budget with which to excite these — depending on how the maker has knowingly or ignorantly designed his system to ALLOW, FACILITATE, INHIBIT, SUPPORT or PREVENT certain movements of the top.

August 8, 2011

I do as much writing for website guitar discussion forums as I can, in addition to answering questions that people email me personally. I can’t really keep up with this demand very well, especially as so many of the questions are duplicates and I wind up giving the same answers over and over again. So I thought that I could eliminate a lot of this repetition by posting some of the questions I’ve gotten, along with my answers. Here’s one such:

Q: In The Responsive Guitar book you go to great lengths to discuss importance of and your method for top “stiffness testing”. I realize you would not want to divulge the optimum number you look to achieve for your guitars. Could you give us a range of numbers that you see from you experience that a new builder could use as a starting point?

A: Yours is a good question. To my mind it’s not so much a question of there being a “right number” or “right quantity”, as finding a method that delivers that information in a way that the brain can take meaningfully. In our culture, weights and measurements and statistics are how such information is most easily taken in and digested.

In other times and other places, however, the same information was transmitted differently, using different language and different tools. But it was the same information. One alternative method that I learned (from master luthier Jose Romanillos, who is certainly a traditionalist in the school of Spanish guitar making) is the following:

Take your joined top plate and start to thin it. It doesn’t matter if you do this with a plane or with a power sander. It is only necessary that you thin evenly, and not leave the plate full of lumps and low spots. Flex it from time to time to get a sense of its stiffness along the grain. Stiffness along the grain is considered the most critical indicator of where you want to wind up, as opposed to strength across the grain or diagonally to it.

You’ll notice that the plate is stiff, of course. How stiff? Well, stiff enough so that when you are pressing your thumbs against one side while holding onto other side with your fingers, and are bending the plate by pushing it with your thumbs, you will find that the spot that your thumbs rest against will be resistant. It will be resistant to the extent that if you keep on pushing so as to bend the plate, it will crimp at the points where your thumbs are. That is, you will induce a bend at those points that is different from the bend that the wood will take between those points.

That tells you that the wood isn’t ready to bend evenly yet. Keep on removing wood. By the way, if you’ve read my chapter on the Cube Rule you’ll understand that removing seemingly small amounts of wood will make a huge difference in the wood’s measured stiffness. So don’t hack a lot of wood off too quickly: go slowly and methodically.

Keep flexing the wood and removing wood. There will come a point at which the wood will “relax” in your hands and, when you press on the long axis with your thumbs, the board will begin to make an even arc along its entire length. It’s not fighting you.

That’s your starting point. No fancy equipment other than your hands and fingers, and a bit of sense of the wood, is needed.

You can of course keep on removing wood, and you can do so until you’ve reached a threshold on the other side and rendered the wood too wimpy to be useful on a guitar. (At the extreme, you can imagine how relaxed and unresistant a paper-thin slice of wood would be, right?) Your next twenty years can be happily spent exploring the range between these two extremes — which define a range of thickness that’s probably on the order of 1/16 of an inch. It’s pretty amazing what a few thousandths of an inch can do — and that’s not even considering the possibilities of selective tapering, bracing, and thinning!