William Cumpiano's 
String Instrument
Newsletter #19

JUST HOW STIFF?

WHERE TO TAP?

CALLING HARRY LIME...

SIDE SPRING

SOUNDHOLE/ BINDING
ROUTER

OAK GUITARS

HEADSTOCK-ACCESS
TRUSS RODS

CROOKED NECKS

CLASSICAL TRUSS RODS?

PADAUK GUITARS

SADDLE COMPENSATION
CONSIDERATIONS

 

Tap tones, and so forth
William R. Cumpiano 2000, All Rights Reserved


Just HOW stiff?

I was wondering how the inexperienced guitar maker can determine the thickness of the top and size of braces when wood varies so much in stiffness. I suppose one could follow someone else's measurements, but what to do when creating an original bracing pattern or changing the size of the instrument? How stiff must a guitar top be to retain stability?

The top's inherent characteristics are telegraphed by the peculiar sounds it makes when the luthier taps it. The experienced luthier grows familiar with the sound the top should make when it's ready to go on the guitar. Others just flex the top, reducing its thickness with a plane or scraper until it feels right. Others just pick the stiffest material they can find, and reduce it to a minimal thickness that past failures have shown them not to go beyond. Yet others do deflection tests of strip samples and then make decisions based on that input. I can't give you a formula or recipe. Sorry.

Generally the braces simply should be as stiff as "can be" so they can be reduced in size to an optimum. That happens when it is sawn from stock which is split in two directions. With little behind him/her to go by, the beginner has little choice but to begin by copying the work of experienced forebears, until the medium becomes more familiar to all the senses...and then work to develop a style and technique which is original.


Where to tap?

I am working on another classical guitar (this  instrument making thing is addicting) and have a question about "quenching the fundamental". I hope you don't mind answering another question.

I am currently thinning the top plate and I think I understand when to stop. The problem is where to tap and where to listen. There is a photo in your book showing you tapping the top. The photo shows you holding the plate with two fingers at the upper bout (the plate hangs at an angle) and tapping a spot that appears to be off-center on the lower bout (almost straight down from where you are holding the plate). Is this the only place that you tap?

The reason I ask is that I get different results depending on where I tap (indeed, where my ears are relative to the plate). If I tap in the same position as in the photo in your book, I hear no fundamental, rather a jumble of tones. If I lightly tap in the middle of lower bout and have one ear directly opposite of the tap position a couple inches from the board, I hear a faint fundamental which dies quickly and lots of low tones (like the fluttering of a birds wing). If my ear is off-center or the plate is not perpendicular to my ear canal, I hear a fundamental tone.


I expect to get very general information from tapping, nothing specific. I tap around the bridge area, either holding the top where the picture shows, or just stick a finger through the soundhole and let the board hang from my finger. Put simply, I'm just picking up auditory cues at several stages during the brace-thinning process. As the "anatomy" of the top changes by my chiseling material off the braces, so does its' response to my tapping. I'm listening for any clear, persisting bell-like, or musical, tones which are predominant when the top is too stiff. Those tones are the ones that I hope to "quench" by paring material off, bit by bit. The clearer sounds change to dispersed, mixed low sounds, (the "fluttering" sounds which I think you describe) without any significant or outstanding fundamentals predominant; is where I want the top be, after paring the braces progressively down. The fans usually end up pretty low and light, if I've selected a particularly stiff top (which I ordinarily do on classics--usually the well-cut boards with a lot of medullary rays prominent on the surface). The braces are slit from extra-ordinarily stiff spruce which I've selected out and reserved from the general pile.

Obviously I start to tap right after I've glued the last full-height brace on the top--to establish a start point. Usually the taps at that point, before put a chisel to the top, produce woody tones that are similar, fairly high in pitch, fairly clear and musical--focused you might say. At the end of the process all I hear is predominantly low and flappy, unfocused sounds. I know it sounds like a paradox--that I should be trying to extinguish the musical sounds from the top--but remember that at that stage the top is a "free membrane" rather than a "clamped membrane" (the other pertinent acoustical model is a "stretched membrane" like a banjo head). When the edges of the top are eventually secured to the rigid sides of the guitar. The musicality returns, believe me.

 

Calling Harry Lime...
The attached photo is of an instrument that I inherited from my Grandmother.

zither.jpg (176363 bytes)

It had been referred to in our family as a mandolin. I
suspect it may be something else as I have not found any descriptions of a mandolin of this type. It is about 24 inches long and 16-18 inches wide. There are 5 strings across the fret bar with an additional 27strings arranged like a harp. There is a songbook in the wooden carrying case that dates to 1873. The sad part is that the glue failed allowing the face to curl up from the botttom as a result of the spring tension. Any information you could provide would be greatly appreciated.


What you have there is German concert zither.


Over the years I have restored several. If you've ever seen the old Orson Welles movie, "The Third Man" you can hear it playing the theme and background music throughout the film. It is actually a concert instrument,
used with symphonic accompaniment for waltzes and 19th century light classical music. It is placed on a table with the fretboard nearest the seated player. The player frets the fretboard with fingers of the left hand while plucking the melody notes with the left, usually the thumb strikes the strings that are being fretted and the rest of the fingers on the right hand reach over the "harp" to stroke the chords and arpeggios on the open strings.

The distortion you describe is typical. The old hide glue usually lets go in time. Restoration usually involves removing all the strings and the unglued parts, flattening the warped pieces with heat, cleaning all the old glue off
and reglueing, and purchasing a new set of strings which are available at a German zither store in New York City with a German name that starts with an M. A web search would probably yield up the store name. That's MEISEL. I just remembered.




Side Spring
I'm writing now is to thank you for the help I've gotten from your book. The Martin instructions are a bit of a joke and without your book I'd never have put the guitar together. Mind you, I didn't follow everything in your book. For one, your book speak of building without a jig, but I'm absolutely sure now that for kit guitars you must have a jig - the reason, in your book teaches sides bending and assumably the sides are fitted straight after. Not possible with the kits - when I received they were already opening up. And they continued to straighten up by the day! I thought this little bit of info would be good for you to know, in case, you get questions from other beginners like myself.




Well, that's not unusual for sides to "open up" (presumably you mean that the ends drift apart). They always do unless you keep the ends restrained. Also you usually have to touch up sides dry against the iron just before assembling the guitar because they always "drift" to some degree or other. They also don't have to match the template perfectly, since they are naturally flexible and can be coaxed a little to conform with the template line without any consequences.


Soundhole/Binding Router
First, congratulations on your nicely done guitarmaking book. As you can guess I am an aficionado guitar builder (I a biologist, actually) trying to do my first guitar. I have a question: Is the base of the router that you show in your book custom-made to have pin holes to fix the router and cut the circular rosette channel? If not, what brand/model is it? Also, for doing the binding rabbets you show a stop in the base of the router, again, what brand is this?


I simply drilled a careful series of little holes in my router base at 1/16" intervals. A small headless brad through the top and into a workboard below serves as pivot. What you saw is called a laminate trimmer base with an adjustable follower wheel. These are fairly common, the one in the picture is a Rockwell.


Oak guitars
Congratulations on the fine book that you wrote. I really enjoyed reading it and it has helped my greatly. I have one question? Why can't you use woods like "Oak" with a nice grain pattern that is common on the market. Flat saw or Quarter saw?



You can use any bendable hardwood (oak is imminently bendable) for guitar soundboxes and make a good-sounding guitar with it. Oak is never used commercially, however, presumably because it has such massively huge pores that it is hard to fill.  Reputedly, it is also rather brittle/ breakable. You may have other troubles with it: I have never used it, staying away because of its reputation. But I have seen people make guitars from oak. Dick Boak from Martin Guitars showed me an oak prototype once. They all sounded just fine.


Headstock-access truss rods
I am beginning my second instrument. I purchased the box set of materials from LMI and they included a 14" truss rod (Part #TRS) that is essentially the same in design to the one I manufactured per your book on my first instrument. Laying the rod up in the slot, it comes up a good 2" short of filling the slot from nut to heel (measurements are from memory). I could excavate the heel end of the shaft a little deeper to accommodate the adjustment mechanism and that would fill the slot up a little more leaving only the nub of the adjustment screw exposed once all assembled. My question: Is it ok if the rod does not extend all the way up to the nut end? If not, I have to figure out how to install the rod with the adjustment screw at the head end, which is apparently the the preferred method for which it is designed (although the catalog does indicate either way is ok). If I leave the adjustment at the heel end, I could get all but an inch to an inch and a half of the slot filled by leaving just a nub exposed for adjustment. Is THAT ok?


It's too bad the catalog at LMI is misleading. Its a headstock-adjustable rod that simply is too short to have it turned around to access the nut at the head block. Period. My experience has taught me that headstock- adjustable rods are not preferable to headblock- adjustable rods.


The problem with leaving the adjuster under the nut, with access to it from a hole in the headstock,  is that you seriously weaken the neckshaft/headstock joint when you cut the truss rod slot clean across it. There is precious little cross-sectional area at that location to start with: about the area of a large postage stamp.

Correspondingly, I've repair dozens of Gibsons with fractured headstocks: fractures waiting to happen.  A combination of noodle-thin necks and oversize access holes right at the critical joint the inevitable culprit. Is it wrong? As wrong as hundreds of thousands of guitars that are built that way. Will it break for sure? Yes, unless you're prepared to baby the instrument and get a perfect form-fitting case that doesn't let the guitar slide around at all inside. And try not to bump the headstock against anything while you're using the guitar!

As far as putting the adjuster of the too-short truss rod in the headblock, you'd have to be a structural engineer to predict how short it can be before you stop getting a fair curve and start getting a lump in the wrong place in the neck when you tighten it against the string tension. Wish I could help you on that one. If you have to use that LMI rod, place the adjuster in the headstock and try not to shape too much of the wood away from the region under the access hole. And treat the guitar gingerly...



Crooked necks...
Having read much of your writing, I've concluded that you would be the ideal person to seek advice from. Recently I purchased a Guild Custom Shop F30R-LS (handmade in Nashville, TN) and in my post purchase obsessive-compulsive paranoia I've noticed several imperfections involving the alignment of the neck with the body of the guitar, 1) the neck is not aligned with the center line of the body, i.e., when measuring the position of the neck with the center line, the neckis positioned about 3/16'' to the right. 2) it appears as though the neck is not fit at a perfect right angle to the body, i.e., looking at the back of the guitar, the neck appears to angled slightly to left. It should be noted that the bridge appears to be correctly aligned with the neck. I guess my question(s) is a) Having not previously owned a hand-made instrument (and therefore having never examined one for such a length of time) howcommon/uncommon are the aforementioned traits? b) is there any cause for concern regarding the structural integrity of this guitar?

Not having x ray eyes good for a thousand miles away, I can't diagnose or predict the longevity of your instrument. I can venture, however, that if the action is comfortable and there are no marked deficiencies in intonation or playability, then the fact that the neck doesn't quite align with the centerseam of the top is in itself trivial.

In a hand made guitar the bridge, and thus the string array, is set relative to the placement of the neck, and not to the centerseam of the guitar. So its not uncommon to see the neck not strictly aligned with the centerseam of the top. Fear not on that regard.




Classical truss rods?
I have recently seen classical guitars with truss rods in them (such as La Patrie, Jim redgate, etc.). I have not seen many luthiers using truss rods on their classicals. I understand the reason for them on steel-string acoustics, but was not aware that the tension of the nylon-strung guitars was that much to make it susceptible to collapsing. Does the use of a truss rod in classical guitars reduce the value, sound quality or any other characteristic? The functionality of the rod is quite evident for structural soundness, but what about for value when it ages? I doubt a Stradivarius violin with a truss rod is anywhere close to the value of one without. Your thoughts?

You can rest assured that if someone ever put a truss rod in a Stradivarius, he would be put on death row soon after.

You're right about being puzzled about the role an adjustable rod would play on a classic guitar. You'd put an adjustable rod in order reduce the effect of string tension on the neck shaft. But a classic guitar's strings pull barely 80 pounds: not likely to deflect a relatively short and massive beam of mahoganyy and ebony. So to me the idea of putting an adjustable rod in a classic seems kind of silly.

A non-adjustable rod is another matter. Ramirezes have been putting a heavy, ebony centerpiece through the neck of many of its guitars since the 60s. I put a 3/8" square steel tube in my classical necks. As many luthiers have discovered, placing more mass in the neck can increase volume and enhance the trebles. Some luthiers put additional lead weights imbedded in the headstock and balance it off with lead weights in tailblock. I think it is favorable to do so.



Padauk guitars
I have a guitar building business in Australia. I have had a request to build an acoustic steel string from padauk, and was wondering if you have any advise unique to this timber. I have in the past made a classical, with I think great results, but am not sure about its suitability for a steel string guitar. I found it to have a very loud, almost bell-like tap tone, with a lot of overtones, which does seem to have imposed itself on the tone of the instrument. However does this present a problem with the nature of a steel string? Any advise would be much appreciated

The temptation to use a bright red wood for guitars is huge. I've succumbed several times. Use it as you would rosewood. However, the pores are huge, so you need to fill it several times. The other drawback is that it's bright red on the guitar for several months, and then turns a muddy brown. I haven't tried it yet, but I'm told that the solution is to use a Ultra Violet UV blocker to the final
coats of lacquer. Its a powder that is mixed in with the thinner. Stewart MacDonald Guitar Shop Supply sells a lacquer with UV right in it.


Saddle compensation considerations
I'll try to keep this quick and short, but I first want to thank you for being such a great teacher to so many of us! Regarding what I have read in your book and on dejanews, you say that the compensation you add is .15" and that over a three inch slot, the inclination being 1/8", you only require a straight saddle for correct intonation? I was wondering if I was correct in my understanding of that. I also would like to know if .15" added to scale length only works on guitars with a 25.4" scale and if not how do you adjust for different scale lengths.

Yes, .15 additional length added to the scale as measured down the centerline of the guitar to a straight saddle sloped 1/8" in 3". It's not MY invention. Martin has been doing it for almost a hundred years. They measure it differently, but it amounts to the same thing.

There are simply too many variables that come into play that determine the appropriate compensation for each string in each scale. The other fly in the ointment is that the fretted fingerboard is set up to create an equally-tempered musical scale, where all intervals, save octaves and unisons are shaved or stretched.

Formulas have been proposed but never tested and offered to the public. One was proposed by no less credible a source than Dr. Daniel Haines, the great researcher. It needs some data-testing before being let out in the wild but it predicts that the amount of compensation (C) required by an individual string is:

  • Directly proportional to the square of the height of the strings over the frets (meaning that C is extremely sensitive to changes in string height).
  • Directly proportional to the modulus of elasticity of the string core material (the smaller the modulus, the more elastic the material--so the more elastic the material, the less compensation required)
  • Directly proportional to the cross sectional area of the string core (the greater the string gauge, the greater the compensation requirement).
  • Inversely proportional to the open string length (the shorter the open string length, the greater the compensation requirement).
  • Inversely proportional to the string's at-pitch tension (the looser the string the greater the compensation requirement.)

So you can see that it is an array of factors that affect the compensation requirement in opposing direction by different amounts. So, believe me, just use the empirical compensations found in the book, derived by trial and error over the years by many expert luthiers.

Some people charge big $$$ to let you in on dubious "patented tuning systems" that can be justified to work in theory but must go screwy (just like the rest) when tension distortion changes the variables. So save your money and apply the .15 compensation empirically, a compensation which will work to have you guitar's 12 fret octave satisfactorily accurate when the guitar is well set up. And .15 will work satisfactorily, if not perfectly, for the range of scales from 25 inches to 25 3/4... and .10" compensation works for the full range of classical string scales.



Waist notch

I am making my way through Guitarmaking slowly and on page 34 of the paperback version on the design part it says "step 4 Notching the waist Cut or file a tiny v notch precisely at the deepest point on the waist of the template." I don't get it. How deep is the notch supposed to be? What is the purpose because I don't see it in the drawings in the book? It really has me stumped. Please write back when you can.

As my book describes, when you draw the template on the soundboard, a little bump in the line will be caused by the notch in the template. When you assemble the sides to the top, you will line up a mark on the sides to the bump in the template line to help the sides line up properly to the them up properly. See page 206 Step One> final fitting, where it says "The side that will be glued down is then aligned at the waist marks and clamped along the outline." and Step three on the same page: "Align the waist marks..."


Figuring for alternate scale lengths

Having constructed a lovely guitar from your book, I was going to do this summers guitar in a shorter scale length, say 25 or 24.9 inches... Now this is going to move the bridge a little closer to the sound hole if I use the same body measurements and soundhole placement I used for the guitar in your book... how do luthiers handle this? Do you move the soundhole back towards the neck a little, or do you build a slightly smaller cabinet, or do you just not worry about it at all? I was worried about the aesthetics of bridge placement, as well as insuring that the saddle's downward pressure through the bridge into the cabinet top is in a good central location for dispersing vibrations. - a simpler way to put the question: Do you use different scale length fingerboards on the same model of guitar body?


You might recall that the layout drawing in the book's early chapter specifies the layout procedure for a 25.4 guitar, but is actually quite generic, allowing you to plug in the desired measurements and create a layout for any guitar that you have a scale for.

Thus the soundhole for the 25 inch guitar (just as it does on a 25.4 guitar) will still start 1/8" beyond the end of fingerboard (actually the 21st fret cutoff point), and the string length will still be the scale length plus .15" , i.e., will lie at 25.15" from the nut. As far as the template, when you're only varying the scale by 1/2" you don't have to make up a new body shape template. Just look at the Martin 000 and the Martin OM: both use the exact same 000 template, but the former has a 24.9 and the latter a 25.4 scale.