GUITARMAKING: Tradition and Technology
A Complete Reference for the Design & Construction of the Steel-String Folk Guitar & the Classical Guitar
by William R. Cumpiano and Jonathan D. Natelson, 392 pages
First hardcover edition published in 1985. First softcover edition published in 1994
||"...perhaps the finest book on
making guitars ever produced."
Guitar Player magazine
" ...finding your book while killing time
in a bookstore back in 1992 completely redirected my life. I fell totally in
love with this craft and made it my full time profession in 1999. Thanks so
much for that."
(Click on book to
order your own copy!)
Updates, changes, feedback,
commentary and clarifications
Numerous other tips, clarifications and additional suggestions can be found in my newsletter archive, and more recently, in my blog. But here are a few important ones to start with:
1- How can I buy this book? Click here!
2- Help! I can't find the truss rod nut your require!
3- All those planning to build a steel string from the book should read this first:
A recommended alternative to the pinned mortise neck joint described in the book.
4- Retrospective thoughts
5- The history behind the writing of GUITARMAKING
6- Headblock dimension typo error in early editions
7- Upper transversal graft dimension error in early editions
8- Swapped captions in the classic bracing sizes diagrams
9- Tips on obtaining rubber rope
10- Workboard shim clarification
11- Workboard clamping shoe caveats
12- Interior fingerboard glueing caul (classic): additional comments
13- "It says to leave only 1/8" for the nut!"
14- "Difference between the classic headblock and tailblock heights only 1/8?"
15- "Where are all the French Polishing details?"
16- Improved truss rod design
17- Dimension error on headblock fixture diagram
18- Clarification and additional caveats on side bending
19- Why don't the heelblock dimensions add up?
20- Is the truss rod spline really necessary?
21- Why does the truss rod nut deflect downward when tightened?
22- NEW Recommended classic workboard upgrade
23- Final width dimension of Side Blanks is not evident. But It should be same as the tailblock height.
24- How do I
tap tune my guitars?
25- How do I compensate a scale other than the one in the book?
26- Another typo error has been discovered. CLASSIC BUILDERS TAKE NOTE.
27- Where are the side blank dimensions?
28- An alternate rosette-making sequence by master luthier Alan Chapman
2- Help! I can't find the truss rod nuts you require!
Folks are having a bear of a time finding the truss rod nut which I recommend--which is actually known as a Hexagonal Brass Spacer Nut. They are also called "standoff" nuts. They are difficult to find because they were used as spacers between circuit boards, and are old-technology artifacts. Here are some alternatives.
1- They are used in the electronics industry so are available in
large electronics hardware supply catalogs. They are often available in large hardware
stores in the fastener section n (but not in brass which is preferable--you want the bolt
to strip first, before the thread in the rod does). Unfortunately the ones in hardware
stores are most often found only in steel. However, brass standoffs/spacer nuts can be
bought in larger quantities from fastener specialty houses such as:
2- Charles Castell writes: "Stewart-MacDonald Guitar Shop Supply offers a brass nut in their catalog; it is item #1018 and is listed as Gibson-style, 5/16" hex with the required 10-32 thread. Offered in a package of 6 at just a tick over a buck a piece. They seem a bit pricey, but I've not had any luck finding the standoffs you mention in the book. So this seems a reasonable alternative to me.
FLASH! Here's a recent testimonial from Bernard
Arnest, who tried option 1:
I am near complete on my first guitar, following your book, and went to look for a truss rod nut. I read the comment that stewart-macdonald's nuts seemed pricey, and so went to hhsmith [noted as #1 above]. They don't sell directly, so I called two of their different local dealers. One got back to me --after over a week-- and quoted me a price of $35 for 100. 100 is more than I need, but ok, I suppose that's not so bad. I ordered them. They had to come from the manufacturer in texas. One week after that, well, today, I call up and they haven't shipped from texas yet, but whould by friday, and then another week to get to me. And just a minute ago, he called back: the original quote was wrong, it's $1.40 apiece, and still the min. quantity of 100!! Stewart-Macdonald doesn't seem so unnattractive after all... and I could have had them a week ago. Or, it seems that the local guitar shop has nuts, whether they're the right size over the phone they didn't seem to be able to tell me.. But just if you want to put this experience over the website, that others might as well go to stewart macdonald from the start. If that's the manufacturer's price, it isn't going to be any cheaper anywhere.
3- Sanaka Thompson of Kona, Hawaii, writes: "I guess the
StewMac ones work. LMI sells one that is round outside with a hex socket, but it's steel,
small O.D. so not much bearing, seems fragile, and is expensive like the rest.
The easiest and best source IMO is McMaster Carr - http://www.mcmaster.com . Catalog page 3048. Or look under "spacers and standoffs" in the "fastening and sealing" section on their homepage.
10-32 thread x 3/4" length, Brass, 5/16" hex - Part # 90308A241 - $3.00
IF YOU DON'T WANT TO MAKE YOUR OWN TRUSS ROD I've tried a commercial design with excellent results: it's a "Double-Acting Truss Rod" which works in both directions, designed, I understand by my old teacher Michael Gurian. It can be purchased from Stewart MacDonald (or more inexpensively, from) Allied Luthierie (www.alliedlutherie.com) and most recently from Grizzly Industrial (visit their Amazon.com website and buy it there). You will have to rout a 1/4 round-bottom slot within 1/8 of the bottom of the shaft, and then file the threaded blocks on each end of the rods into a curve so that it will seat properly into the round bottomed slot. I far prefer to make them accesible through the soundhole, than weakening the headstock/neckshaft joint area ("the hyoid") as they presume you will want to do. However, I consider the best option is to make the tightening nut emerge from the headblock inside the guitar, regardless. Due to their sizes (the Stew-Mac and Allied Luthierie rods are 14" and the Grizzly is 17") you will either have them start under the second fret rather than under the nut for the short ones, or drill a small access hole through the upper transversal face brace for the long ones. Either compromise is acceptable, although not optimum.
3- Hardware-Based Neck Joint
IMPORTANT: A recommended alternative to the book's pinned mortise and tenon neck joint!
A new hardware-based neck joint replaces the book's indicated pinned mortise-and-tenon joint. Easier, better. You can interpolate the information into the book's text, or if you've already have built the neck and body, you can upgrade the design from pinned to bolted with several addition drilled holes. I've been updating the page over the months, to clarify and improve its content.
4- Retrospective Thoughts
Jon and I wrote this book in 1985, and to my great joy and satisfaction, it has become a perennial bestseller, and consequently the standard textbook in the guitarmaking field. It was truly a labor of love, its creation being an enormous challenge, burden and sacrifice (writing the book was like "enduring a protracted illness") but the payoffs have been ample: over the years I have made countless friends and thankful acquaintances as a direct consequence of that effort. Proudly, I have watched the field blossom with a great number of new, fine builders which have "graduated" from the book and from their additional efforts and sacrifices. Together with a burgeoning market for fine handcrafted guitars, these many new builders have entered the field and lifted American guitarmaking into a true Golden Age. I'd like to think that my book has contributed to this resurgence.
6- Headblock dimension error in early editions
The earliest (hardcover) Rosewood Press editions of the book had a glaring dimension error in diagram 9-6 on page 191 captioned "The mortised headblock three-view (steel string)" It showed the over all height of the headblock as 3 5/8" PLUS the 1/2" height of the angled cutoff segment on its bottom. People following this would have ended with a headblock on the guitar which was considerable taller than it should be (some folks informed me that they just trimmed it off). The error was corrected in subsequent editions which properly showed that the overall height of the blank was 3 5/8 BEFORE the 1/2" angled cutoff segment was sawn away--leaving a headblock with its mortised end 3 1/8" in height. For those with the older editions, here's a scan of the corrected version found in the corrected editions and the paperback:
7- Upper Transversal Graft dimension error in early editions
In early editions of GUITARMAKING, diagram 7-7b on page 15 showing the "schedule of final finished brace-blank cross sections" has a dimensional error shown for the steel string UTVGFT (upper transversal graft). It shows the width of the graft as being 1". In later editions, this was corrected to show 7/8", after it was realized that the space alloted for it between the headblock and the upper transversal face brace was 1", and you can't fit a 1" object into a 1" hole!
8- Swapped Captions
A reader discovered swapped captions on the CLASSIC section of diagram 7-7b on page 151. The handwritten legends for "ros.gft (2) 4"L" and "utvgrft (1) 5 1/2" are mistakenly swapped. No profound consequence. These grafts are minor components simply used to "strap" the soundboard fibers together at weak areas, or areas of shear stress, and their precise length or width are not a critical factor in the sound or stability of the guitar.
9- Tips on Rubber Rope
Readers are complaining that automobile inner tubes for making the rubber rope (for roping the back while glueing it onto the guitar shell) are getting harder and harder to get, and while the book text discourages people from using a truck inner tube (because its "too heavy") it seems expedient to recommend these now--because these indeed are far more common to find since they are not obsolete. Check the Yellow Pages for truck repair shops. You may have to slice the rope into a slightly narrower band than what's indicated in the book, but you'll get a far, far longer strip than if you had used an automobile tube.
UPDATE: A clever correspondent has suggested that bicycle inner tubes are plentiful and available from bicycle stores, and can be sliced into long strips and tied together.
UPDATE 2: Correspondent Matt Jacobs writes:
"I found an excellent source of
rubber rope from the Physical Therapy
field. It is marketed under the name Thera-Band, or Cando, it is low powder
latex rubber that comes in rolls 6"X 6yds It also comes in different
thicknesses and tensions. You'll probably want to get the heavy or extra
heavy stuff. It runs from $8-15, a good source is www.wisdomking.com, but I
imagine you could get some through a Physical Therapist."
10- Workboard shim note
The diagram on page 36, figure 3-5b shows a dotted rectangle in the middle of the shim. The text describes this as where a small piece of cork should be glued.The shim supports the transversal face braces right under the workboard shoe. The text does not suggest that in addition, since the lower cross strut is arched, the clearance to the workboard is reduced and thus you should sand that piece of cork until it is thinner by the indicated arch deflection--than the outer cork lining.
11- Clamping shoe caveats
I should add the following update comments: You must be careful when tightening down the clamping shoe over these shims. The thumbscrew or bolt must be snugged down firmly but be aware that if you bully the tightening of the clamping shoe, it is entirely possible to collapse the tiny cork rectangles and break something on the top. It can also tip the headblock back into the shim cavity and spoil the neck angle. My concern over this possibility has driven me in recent years to occasionally use an alternate to the workboard shim, and on my most often-used workboards, dispense with the workboard shim altogether and hollow out the actual workboard with a scraper blade directly under where the arched lower transversal (CL) or the arch upper transversal and x-brace (SS) is expected. It is a choresome step, but a reasonable alternative. Use the same arch template that you used to arch the brace, as you check your progress while scraping the hollow in the workboard. Extend the scraped area, feathering it up towards where the soundhole would be, and down gently towards where the bridge would be. It should be correct in depth right at the main brace locations, but if it is not strictly accurate to the way the top would flex below or above the main brace(s), it is not critical--because the top plate is resilient and will temporarily distort without harm if the precise curvature of the workboard doesn't precisely match the curvature of the top. But it should match right under the main brace(s).
12- Internal fingerboard gluing caul
Confusion has arisen from:diagram 10-16 on page 220 titled "The upper face brace/cross strut caul." it is indeed the upper face brace caul for the steel-string, not the classical, and no distinction is made in the text. A reader building a classic guitar must extrapolate from the text, and use the actual measurements from the instrument to devise the caul for their classical..I will be uploading a classical IFG caul diagram soon at this location.
13- It says to leave only 1/8" for the nut!!
I'm trying a classical. I'm in the process of laying out the head block. It says to measure an 1/8 of an inch from the line where the plane of the head meets the plane of the neck and draw a line. This to indictate where the fingerboard is to begin. But the nut's going to be a 1/4 inch thick. Right? Shouldn't I leave a 1/4 inch space?
Yeah, it sounds funny, I know. But it makes sense after you read ahead, on Page 63 "Applying Headstock Veneer" and then page 310, Step5-- Classical only: Cutting Back the Nut Slot." Jon's technique was to cut back the veneer AFTER the fingerboard is glued down, in case the board ends up wobbled away a bit after the glue dries. It makes sense, since the maximum thickness nut blanks available commercially come 1/4" thick, and if it wobbles to 1/4+1/32, then you're going to have a bad fitting nut. His procedure avoids this possibility.
14- The height difference between the classic headblock and tailblock is only 1/16" !!
I'm to the point in your book where I'm geting ready to arch the back on the classical guitar I'm making. I realize that the amount of the arch is much less in the classical, than on the steel string, but I can't figure out how the arch is supposed to look if the tail block is 3 3/4 inches and the head block is pretty close to that, at 3 11//16. If it's just 1/16 this seems small enough so that it could be disregarded. I'm thinking about adding an additional 1/4 to the tail block to provide for a more significant arch. What do you think?
No, don't add any extra height to the tailblock Remember that the 3
3/4" tailblock is mounted on the top, so the starting difference is not 1/16"
but more like 1/8" to 3/16". The difference will be exacerbated when you trim
the sides down to the proper taper between the two blocks, and then applying the large
sanding board. That will apply a slope to the top of the
headblock, and remove yet another 1/8" or so from the front of the headblock. The arch on the classic AND the slope of the back are indeed subtle.
15- Alas, no French Polishing details...<sob>!!
I have gone through your book very throughly as preparation for
my building of a Martin Guitar with their kit. I found the book very well written and
What I hope to know more about is French Polishing since I don't have the use of spray equipment available to me. Your book has just that one picture on it and it doesn't really describe very much. Can you give me some tips and pointers?
My book never intended to cover French Polishing at length. Because first, I am not a master at it, and second, the directions required are quite lengthy--not that its a long process, but that it takes a lot of space to explain it properly--and the book had reached 400 pages as is. But you have at least two options:
16- Improved truss rod design
Over the years, I've been using this improved truss rod design which features a more massive bearing cap. The cap on the rod design pictured in diagram 4-16 page 55 simply did not grab the top rod adequately, and tended to distort and tip under load. This new design should reduce rod distort ion in the cap area. Wrap the rod with gummed foil as in diagram 4-16 in the book. The new truss rod will interfere with the soundboard when entering the neck tenon into the body mortise, so you must make a minimally adequate slot in the soundboard to clear. NOTE: The diagram shows the bearing cap extending beyond the neck tenon. In shorter scale guitars, this might result in insufficient clearance space from the hex nut to the upper transversal face brace to accommodate the particular nut driver that you may have. You should anticipate this possible dilemma. If in doubt, cut the truss rod wire shorter so that the bearing cap actually sits inside the tenon (you may have to carve a widened and deepened space for it if you do so).
Folks are having a bear of a time finding the truss rod nut which I recommend--which is actually known as a Hexagonal Brass Spacer Nut. Here are some alternatives.
17- Dimension error on headblock fixture diagram
On Page 212, Diagram 10-1a, the distance from the center of the bolt
hole to the bottom of the headblock fixture should be 7/8-inch, not 1 1/8-inch as shown.
People using the fixture to secure the headblock during assembly must have already noted
that the bolt didn't line up with the truss rod hole in the headblock. You will have to
redrill the fixture to line up the holes.
18- Clarification and additional caveats on side bending
>I bought your book a couple weeks ago, and have read it from
cover to cover.
>I don't know, however, how to make the side plates. There weren't any pictures,
>and because you don't mention lengths, as each guitar can be different, it is unclear.
>Could you please add some clarification? Your book is wonderful, thanks for all the help.
All the information you need to prepare the sides for bending is in
the book, but unfortunately it is dispersed in different locations, so I'll provide a map
What are the starting dimensions of the side blanks?
The dimensions are in the Bill of Materials in the back of the book, on page 383:
How do I reduce the blanks to their finished thickness?
You can find that information at the end of Chapter 5, Step 5- Planing the Sides, page 113: The beginning of the chapter on page 104, specifies that the beginner should strive to reduce the side plates to .085" on steel string and .080" on classic. For the beginner, the sides and back are particularly tricky to hand plane to a consistent thickness, so I would highly recommend that you ask the supplier to reduce them, or barring that, locate a small millworking shop in your area that will put your plates through their abrasive planer for you, and reduce them to the precise specified thicknesses. Also, obtain several other side blanks of less precious material and have them reduced also so you can practice before attempting to bend your $500 brazillian rosewood sides!!
There are numerous additional suggestions and comments on side
bending in my newsletter archive.
19- Heel dimensions don't add up?
> I am a first time builder and am having
> the heel dimensions. On pg. 50 of the manual it states
> that the height of the heelblock and shaft taken
> together is 3 1/8". Also my carved heel compares well
> with the heel curve outline full size drawing on pg.
> 82, but the heel seems to be about 7/8" short,
> especially when compared with the side dimensions of
> 4". Is this a typo, and if it is, would it be
> possible, instead of redoing the entire shaft and heel
> again, to simply glue on an extension to the existing
> heel to make up the difference?
Fear not. You're on the right track. No
repairs or additions necessary.
Take a look at the headblock drawing on page 191. The sides of the guitar will taper from a maximum of 4" at the tailblock to about 3 1/8" at the headblock. I say "about" because by the time you include the thickness of the soundboard and the thickness of the back plate; and after all the trimming of the headblock to get everything to fit just right, it may end up a bit more or less. But you'll see that 3 1/8" is in the ballpark for the the heelblock and shaft taken together; as well as for the headblock, and your sides will be trimmed down to meet them both in Chapter Nine.
20- Is the truss rod spline really necessary?
>Tthe truss rod I used is a pre-fab model from
> that seems to sit higher in the slot than your example
> in the manual. Is it absolutely necessary to insert a
> spline over the rod, and do I dare increase the depth
> of the truss slot? Is there a substitute for the
> spline you suggest?
No, don't make the slot deeper. 1/8-inch left of wood under the slot is optimum. Less, and you risk cracking the neck shaft when you tighten the rod and that would be disastrous. If you leave the bottom "web" too thick, the truss rod will lose efficiency. The deeper the rod is in the neck, the more efficiently it works. The spline essentially is there to take up space, but it also performs a minor function of cushioning the upward pressure of the rod on the brittle ebony fingerboard.
21- Why does the truss rod nut deflect downward when tightened?
> Is it normal for the hexagonal nut and spacer block to deflect
when tightening for the fingerboard arching? ( I used your new spacer
No, a small downward deflection is inevitable. But if it
deflects dramatically downward, the problem could be as simple as that the rod hole in the
spacer block is too large, or as problematic as the entire rod having too much slop (room)
in the slot. Everything has to be snug in there. If not, something's going to distort when
Note: Change in truss rod cap recommendation
> Please confirm the measurements of the bearing cap shown as an update.
> Your more massive cap seems to have shrunk some in height and width and
> gotten thicker only. The old way you drilled through the 1/2" side with 3/16
> & 13/64 and now its through the 1/4 " side leaving very thin sides is this
> correct or am I missing something? the old cap was 1/2 x 3/4 x 3/16 and the
> new is 1/4 x 1/2 x 5/8 (.625)
Yes this is a better design, although it requires a bit more precise drilling. However the earlier (vertical) cap tended to rotate under tension and sometimes jam. The longer, lower (horizontal) model holds on to more of the rod's length and stands up to more tension with less distortion of the rod.
The guitar's workboard needs to be attached to the
table by the neck extension, such that the body portion sticks out over the edge
overhanging in cantilever fashion.
22- Recommended classic workboard upgrade
In the steel string guitar, the body is built separately from the neck, so the neck extension can be any convenient length that allows you to clamp it to the table. If the workboard sags a bit when you work on the guitar box as you assemble it, it´s not a worry.
But on the classic workboard, the situation is a bit more complicated (and I´m sorry now I gave it such short shrift in the book. I´ll try to make amends here). Note: CLASSIC ONLY!
When you rope on the guitar back, any considerable force applied to soundbox can cause the body portion of the workboard to sag, and this can eventually result in a bad situation: since the neck and body are integral from the start, the neck remains on the neck extension that is clamped down to the table, and the body portion is allowed to "float". This could potentially cause the neck to emerge from the assembly process permanently set un-straight (relative to the plane of the soundboard). This small imprecision could be exacerbated if your workboard material is especially flimsy.
The Spaniards resolve this problem by attaching a "foot" from the far end of the workboard to the floor, to keep the whole structure from sagging. But since the foot gets in the way of roping, I prefer the cantilevered workboard scheme.
True, if you read further down in the book you will see that there is an allowance for the neck not landing up perfectly level with the rim of the soundboard: you will be required to plane and sand the fingerboard flat and to the proper inclination after it is glued down to the neck shaft. This in effect "trues" the neck after any number of inaccuracies have crept in during the assembly process.
But it would be better if our technique didn't result in a situation that would require another technique to remedy! So I would recommend now that the workboard, as described in the book, be screwed down to a rigid beam placed down the centerline of it's undersurface. You could use half a dozen countersunk flat-head lag-screws or lag-bolts from the hardware store. Any verifiably straight stiffening beam would do, from a construction-grade 2 x 4, or hardwood beam of any cross section (or if you can find one, a length of strudy angle iron), extending from the nut to the bridge location ( you could counterbore it right at the workboard shoe-bolt location, so you don't need an impossibly long workboard shoe-bolt). The thus-stiffened work-board can be flush-bolted to your work table to hold it down, or if you don't want to drill holes in your living room dining table, just extend the beam past the headstock so you can clamp the beam firmly to the table, ahead of the workboard.
23- Final width dimension of Side Blanks is not evident.
A notable lapse. One version of the confusion caused is represented by the following query:
> So far I haven't found how wide or high the sides are to
The Bill of Materials on page 382 list the side blanks as 5" (4" CL) wide--but all the blank sizes listed there are somewhat larger than their finished dimensions.
Indeed, the side blanks need to be narrowed to 4" (SS) or 3 3/4" (CL) -- the same as the lengths of their corresponding tailblocks. The book shows that this may be done after the bent sides are glued to the top (see Step 1, page 213), but it is actually better and considerably easier to reduce the sides to these finished blank dimensions BEFORE bending the sides.
How DO I tap tune my guitars?
Sorry to tantalyze my readers, but my building technique doesn't
include "component tuning." My experience of thirty five years has led me to the
conclusion that "tuning" the sound of a guitar is an illusion and a
chimera, and those who publicly advocate that they can accurately control the
response of a guitar by responding to noises derived from tapping parts of it, are simply seducing
the innocent, or at best, self-deluded.
The short discussion in my book about tapping the soundboard while reducing the thickness of the soundboard was, admittedly, my passing on in an uncritical way, the vague mumbo-jumbo that I was fed by older builders during my years at the guitarmaking factory in the early 70's, when I first learned guitarmaking. The illusion persisted until I concluded, sometime after I wrote the book, that it was, well, an illusion...not to say, a crock.
Now, twenty years after writing the book, I have a different point of view. It's summarized by the following entry into Newsletter 19:
You heard it here first: "Tap-tuning" has been oversold. Let's put it into perspective.
When I was learning guitarmaking, I went through what you're going through: I had HEARD about something called tap tuning and was mystified and tantalized. I presumed that it held THE secret for easily-replicable, world-class results. I actually thought that the alchemy of guitarmaking was somehow locked inside this arcane act of wizardry called "tuning" the top, that it was something that only the select and most sensitive few knew about, and that they weren't going to tell me, so I would somehow have to learn it myself.
Builders, alas, often do things on guitars which they really don't know WHY their doing it, but they do it because they were taught to do what the teacher did, and they're afraid if they stopped doing it, the nice sounds they usually get will go away. So they keep doing it. Then, when you ask them why they do it, they are all too happy to MAKE UP vague, fanciful, jargon-laden accounts which will leave an impression that they know exactly why they're doing it in minute detail, and by implication demonstrate that they can somehow successfully manipulate all these invisible sonic phenomena on the guitar with ease. Luthiers usually won't stop you from believing that they are wizards. This is most painfully obvious in the realm of "tap tones."
Yes, I do "tap" a top to derive some rudimentary information about its anatomy. What you hear does give you some feedback clues which is useful and helpful. The "tuning" part is what is so misleading to beginners. At some point (most often toward the end of the process) many builders attempt to make some final changes in the anatomy of the soundbox, which the builder believes (I've selected this word carefully:) believes, is exercising some control over the final results. Both, skillful experts and deluded fools, equally, scrape here and there, tap, press, reach inside, remove a little on the back braces, on the top braces, and then at some point say THERE. It's just right.
Then there are other builders, masterful experts and deluded fools alike that DON'T. They believe that they're making all the crucial decisions in the INITIAL stages of the construction: materials selection, materials dimensioning, design: that is all they need to achieve the desired results.
My approach today seeks to achieve a balance between a guitar's structure and it's compliance (ability to yield to slight fluctuations in string tension, i.e. vibrations)--factors usually in opposition to each other. I have come to rely largely on visual cues--not auditory ones--displayed by the actual materials at hand and an evolved sense of how to reduce the guitar's structure to an optimum state. Thus, the beauty of the vibrating string can be realized with as little impediment as is structurally possible--given the need for the guitar not to distort under string tension over time. This is a long way of saying that if you want to learn how to "tune" the soundbox by the sounds the individual parts make during assembly, well... please inquire elsewhere.
How do I compensate a scale other than the one in the book?
There is no simple and practical way to calculate the precise figure
for the location of the saddle midpoint for any given scale, because there are
too many factors that add to--or subtract from--the compensation requirement. So
the luthier has little choice other than locating the saddle midpoint according
to a process which is purely empirical, or in other words, relying on a quantity
that has been confirmed by prior experience.
The experience of this luthier is that the .15-inch figure for the saddle midpoint compensation increment is successful for any of the modern guitar scale lengths, from 24.9 to 25.7.
This may be unsatisfying for a technically-minded person to learn, because one would reasonably expect that there must be a single and unique figure that should be calculable for any given value of scale length distance.
This uncertainty has bedeviled luthiers for literally hundreds of years. The reason why no practical way of determining compensation with mathematical precision has ever been developed, is that luthiers are not physicists or mathematicians.
There exists, however, a formula which was proposed to me by a highly respected physicist, before he passed away several years ago. The formula takes into account most, but not all, the variables that affect the compensation requirement for any given string. Unfortunately, the formula requires a great deal of data acquisition in order to plug in the numbers in the right places. So for any given guitar, the formula is pretty much useless to the luthier for practical reasons. But that is not to say that much can't be learned from it.
The formula reads as follows:
C is the compensation, or more accurately--given that the 12th fret divides the scale length precisely in half--it is the amount the saddle-half of the scale is longer than the nut half.
H is the string's height over the twelfth fret
E is the Modulus of Elasticity of the string's core material (a constant which differs for each material, obtainable from a Properties of Materials textbook)
A is the cross-sectional area of the string if it is monofilament; or of the string's core if it is wound.
L is the length of the scale
T is the tension of the string when raised to concert pitch.
I once plugged in the data for an actual string during a experimentation session, and the formula indeed predicted a value that corresponded closely to a midpoint distance of .15", but it took me several days to acquire the necessary information. When I reported this to the physicist, he warned me not to rely on the formula until I had accumulated a voluminous amount of data. Needless to say, I had to drop the project, and return to my empirical figure, .15.
The formula predicts that the compensation requirement:
In discussing this with yet another physicist recently he pointed out that, in
terms of pure precision, the resulting value of C will only be true to obtain
zero pitch distortion when the fret is pressed at the 12th fret. Indeed, the
distance H is not constant over every fret, but increases slowly from the first
to the 20th. So the formula is accurate only at the 12th fret. But then, that is
how us luthiers verify the accuracy of the bridge's placement: by pressing down
on the 12th fret and comparing the given pitch to the 12th fret harmonic, which
is our test standard.
Here is one luthier's brilliant solution to this eternal dilemma:
However, any setting on his guitars appllies only to a given set of
strings, at a given action setting, and has to be readjusted as the strings wear
in. But therein lies madness.
26- Typo error in book, re: classical bracing arch offsets
An Australian correspondent writes:
On page 22 in step 3 you describe the arching of the back braces. You refer to figure 10-20 (b for classical). This figure shows the following arch offsetsbackbrace#1 1/16" , length 12"backbrace#2 3/32" , length 12"backbrace#3 1/8" , length 13"backbrace#4 1/8" , length 15"The lengths come from the material list on page 222You also refer to the arching process of the arched soundboard braces.The arching of some of the soundboard braces is described in chapter 7. On page 150 you describe the proces of making the arch templates (step 4). At the end of this step you describe the arch templates for he classical back bracesbackbrace#1 1/16"offset , span 12"backbrace#2 1/16"offset , span 10"backbrace#3 1/8" offset, span 10"backbrace#4 1/8" offset , span 15"It seems to me that there is a difference between the arch templates (in chapter 7) and the actual braces (in chapter 10).I'm not sure which offsets are correct.
John Natelson, who wrote the classical guitar sequences of the book responded:Mr. Koot is correct. There is a misprint on page 150. The #3 back brace template should have been given a 13" span. I can only assume that the 10" span for the previous template was inadvertently repeated by the typesetter, and we never caught it . I'm surprised it took this long for someone to point out the error. [The book was written in 1984]
The #2 back brace template is OK as written, because the offset will be very close to 3/32" at the 12" length. Mr. Koot's solution is also correct. He should follow the dimensions shown on page 222, as they are right for the guitar being built from the book. Given his attention to detail, I expect he will build a nice guitar.
26- Where are the side blank dimensions?
An embarrassing shortcoming, granted: It is not immediately apparent what the starting dimensions of the side blanks should be. You have to search for it. It's on the Bill of Materials at the END of the book, on page 383, along with all the starting dimensions of all the parts of the two guitars. For the record, they are: CL: 4" x 30" ; SS: 5" x 32" . The blank thickness, 1/8-inch, is the starting thickness. The blanks must be planed to approximately 3/32, or thickness sanded to .085--the difference being the greater accuracy obtainable with machine sanding.