Wednesday, July 10, 2013

Oooops. I over-thinned my back plates!

I live all the way over here in Ireland. I'm in the process of building the steel string guitar from your book and have just noticed that I have planed the back too thin (0.085"- the dimension as the sides)
I am using Amazaque for the back and wondering am I wasting my time completing the build with this back and just bite the bullet and purchase a new set of plates. I don't need my first build to be perfect but still don't want to be disappointed when the build is finished due to this error of mine

Hope you have the time to reply to this. I just want to say that your book is brilliant and has given me a new addiction and already planning future build. 


I seem to be getting a lot of mail from Ireland lately!

I don't know if you've bent or assembled the sides yet.  Or what size the guitar is to be. Or how coarse the surface grit pattern is, that you'll have to remove with sanding. Also, I've never faced the precise dilemma that you're facing. So I can't vouch for the quality of my advice. But i can think of some potential disadvantages to making a guitar with a back that is no thicker than .085".

1- If it's the large guitar precisely like the one in my book that you're building, the back is likely to sag somewhat between the back braces and look somewhat lumpy and peculiar.

2- If it's a very dense wood, it is likely to crack easily, like a thin sheet of glass--either spontaneously if you let it dry out, or after a gentle impact.

3- Acoustically, the effect of thin plates on a large guitar is to drop the resonant frequency of the sound box, likely lending the guitar a somewhat tubby sound.

If I was given a set with a thin back plate to make a guitar from it, I would choose to make a 13-14-inch parlor guitar. A back of that thickness would not be disproportionate. But I don't know how far along the rest of your project is, so I don't know if that's very useful advice. Sorry.

Will a 3-piece neck solve my action instability problems?

I live not far from you in New England... therefore.. yes I am experiencing neck warping and bowing issues. I have sent the guitar to Fender, who says they have kept the original Guild Custom Shop, and they bleeped me royally. They agreed that the neck was a warranty issue, but claimed that because the TOP had a little wear on it they wouldn't do the warranty work unless they replaced the top at a cost of $500.00 to me. I fought and fought, but they held fast, and because I love the guitar, I agreed to let them do it. They did and rebuilt the neck, replaced the top and bridge and saddle. I got it back and 3 weeks later had to send it back to Nashville for a neck reset. Horrible craftsmanship. Anyway they fixed it and I played it for 6 months and here I am again with the neck problem. I have made three custom saddles varying in height to be able to adjust the action, and it seems to work okay. 
     What I am looking for from you is, (I am a carpenter by trade), [is a solution] to remove the neck and fret board, take 1/2" out of the middle of the neck, replace the cutout piece with a piece of blond maple, lose the truss rod, and laminate the three pieces together, mount the fret board and remount the neck to the soundbox. In my 35 years experience with wood, I feel there is a 70% chance the problem will be permanently resolved. And if I have to go back to using the different saddles then so be it, but I would like to try my idea. Can you hint to me in yes or no fashion if my idea of a three piece laminated neck may solve this issue?


Can I hint whether inserting a 1/2 inch maple spline in the neck--in replacement of a steel truss rod--is going to resolve "neck warping and bowing issues"? Well, to have an opinion I would have to know for certain if your action-height instability problems are in fact the result of a neck that is insufficiently stiff, and whether removing the steel truss rod and replacing it with a maple spline would in fact make it more stiff.

Well, maple is stiffer than mahogany but not as stiff as steel. Mahogany is only "moderately" stiff while maple is of "high" stiffness. On the other hand, mahogany is far more dimensionally stable than maple (as regards to changes in environmental humidity fluctuations). Your plan to replace steel with maple to make the neck both stiffer and more stable seems a bit iffy. I don't know how massive or how flexible your steel truss rod is--I don't keep track of the nature of every truss rod on every guitar model in the industry. But I've seen some truss rods that are enormously stiff while I've seen others that seem as stiff as a limp noodle. So I can't opine in that regard without knowing what's being replaced.

But consider this: your problems with action-height instability can happen even if the neck itself remains perfectly straight and un-bowed. For example, a perfectly straight neck can rotate upwards at the heel. That is, the entire neck AND headblock can rotate upwards in the soundbox--with everything stretching around it. The soundboard itself can rise and drop in a variable climate, independent of what the rest of the guitar is doing. The entire guitar from tip to tail can stretch like a long bow--every element between tip and tail stretching just a tiny bit, to add up to a noticeable action-height change. So what is the likelihood of a 1/2-inch maple spline inserted laboriously into your neck, in order to resolve action-height instability? Not very high, I would guess.

What is indeed seriously troubling is that Fender convinced you to replace the entire top to resolve action instability problems because it "had a little wear" on it--at your expense. Either you're not telling me the entire story, or they really have taken you for an innocent rube on this. If they knew that your action instability was the result of your top collapsing--this would warrant a new top--it is they who should have picked up the tab. They may not have told you the real reason WHY the top needed replacement (a "little wear" is ridiculous) but if they noted serious distortion on the soundboard they should have owned up to the situation--instead of telling you that a little wear had to be resolved before they could solve your action problem. Very troubling. Shame on them.

Saturday, January 26, 2013

Which glue to use? A conversation with an expert.

This was the original question posed to Bee Miller, the technical service rep of Franklin International, makers of, among other things, the Titebond line of adhesives:

I'm a professional guitar maker in Northampton Massachusetts.

I've used only Titebond ("classic" I think you call it now) for most glue ups since I started making guitars 40 years ago.

The only drawback was it's relatively short open time. Some of my glue ups required extra open time, because it takes a long time to apply the glue with precision to all the surfaces that need to be covered during a complicated assembly procedure. I found the glue begin to start to thicken up before I had completed its application.

I've tried the several new titebonds that have recently become available, and ended up preferring Titebond II because it remains wet and creamy for a longer time, which gives me plenty of time to spread it with precision before closing the seam. The others, classic, Titebond III and Extend all seem to be extremely thick and quick to skin over. Extend especially seemed extremely thick and hard to spread--relative to the others.

But recently I've heard that titebond II is not the best choice for other reasons, not the least of which the fact that it is less creep-resistant than the others. Creep resistance is important, because guitars are under constant string stress and any distortion of the structure over time is prejudicial to its performance.

Given all these factors, which particular Titebond product would you recommend? I need good creep resistance and easy spreading.

Miller's response:

Titebond Liquid Hide Glue and Titebond Extend Wood Glue have the greatest creep resistance and longest open and working time of any glues we carry. Both however have a limited one year shelf life and I am afraid you may have sampled some Titebond Extend that had gotten too old. If you would send me your delivery address and phone # (for delivery purposes) I could send you a fresh bottle of each to try for free. Many luthiers are big fans of the Liquid Hide Glue because jointsbonded with this glue can easily be steamed apart for repairs...The Titebond Extend has large particles added that plug some of the pores of the wood to slow down the absorption of water/moisture. Over time these
large particles want to settle to the bottom so this glue begs to be given a good stirring every 60 days or so, and can actually be used longer than the 1 year shelf life if kept stirred and still maintains a
good pourable, spreadable consistency. The Titebond Liquid Hide Glue can get a little thick if temps are below 65F or so, and it can be overcome by giving the bottle a warm water bath. It also can last longer than the one year shelf life. However, the animal proteins begin to break down and the glue can lose bond strength. There is a simple test to tell if it is still usable. Simply smear a thin film on a piece of paper and allow it to dry (this can take longer in a high humidity environment.) If it dries crisp and brittle and can be broken, it is still good to use. If it dries flexible or gummy, bond strength will be compromised and you should discard it.

I would welcome your feedback after experimenting for a few weeks.

I wrote back:

Let me take advantage of this opportunity to ask you something further about Liquid Hide Glue. I've been making guitars for over 40 years and from the earliest years of my apprenticeship--and during the years of my career--I persistently received the following lore-as-truth from other practitioners:

1- Hide glue was the very best glue for making musical instruments because the resulting film separating the parts were crystalline and brittle, which served as the best sound conductor between them.

2- But that advantage applies only to hide glue made in a pot from flakes. Commercially pre-prepared room-temperature liquid hide glue was to be avoided because it was very bad, however. No one ever specified exactly how it was bad, just that it was to be avoided.

3- Because of the many headaches of actually using traditional flake hide glue, most instrument builders opted to put up second-best synthetic evaporative glues for their wide temperature range and convenience. It was said that the resulting film between the parts dampened the sound more than the "crystalline" hide glue film. But the convenience factor was the determining one. Still, the pre-prepared, convenient form of hide glue was nonetheless to be mysteriously avoided.

Are you familiar with this history of poor reputation of liquid hide glues within some sectors of the industry, in comparison to the traditional hot-pot form?

Indeed, what precisely are the differences between the resulting joint made by hide glue from flakes in a hot pot vs. commercially pre-prepared liquid hide glue?

Miller's response:

We have not done any testing with traditional hot flake hide glues for many years. But we have always found that if our Liquid Hide Glue is used before the proteins break down (remember the paper smear test? brittle is good) it forms a strong, crisp, brittle film that is as good as provided by the flake hide glues. When we do shear failure tests on Hard maple, Titebond Original fails on average at 3,600 psi with 77% of the bonding surface area showing wood fiber failure. Titebond Liquid Hide Glue fails on average at 3,591 psi with 72% wood fiber failure.

[Ed: This is the standard test for adhesives. Two wood blocks are glued together with the test adhesive and then are pulled apart by a machine. The two numbers represent a) the pressure in pounds per square inch at which the two blocks finally pull apart and b) the percentage of wood at the joint from one block that remained steadfastly adhered to the other block (fiber failure). The remaining fraction where the glue failed and the joint surface remained clean and clear of wood fibers from its neighbor  remaining clear area where the glue failed at the joint surface. So the higher percentage of wood fiber failure (i.e., glue success), relative to the amount of glue failure shows how successful the performance of the adhesive.

We have always assumed that any rumors of poor performance with our Liquid Hide Glue was primarily an issue of using glue that was too old and did not form as brittle of a film because the proteins had begun to break down.

Our favorite phrase around here is "One test is worth a thousand expert opinions." I encourage you to test and convince yourself one way or another - hopefully we both win!

This conversation has evoked skepticism of the traditional lore in this regard--that commercially pre-prepared hide glue is inferior to hot-pot flake glue. All it takes is for several users to ignorantly assume that commercial pre-prepared hide glue need never be tested or agitated--and then as a result experienced several costly failures--in conjunction with the familiar older-is-always-better bias--for a false glue meme to be created. Stay tuned.

Hats off to Alan Chapman



Alan Chapman is a brilliant New England luthier. Indeed--a luthier's luthier. In my mind he is a distinguished model for other builders--an indefatigable researcher as well as a master craftsman. He is an eternal tinkerer of the guitar's form, each iteration revealing to him a yet deeper "hidden truth" of the instrument. His finished guitars are correspondingly all the wiser: loud, robust and impressive. His modus operandi is to create test guitars, for example, one with a replaceable belly--that is, one that allows him to remove the entire lower belly of the soundboard, starting from half of the lower transversal face brace, and to replace it with new lower sections with different bracing schemes, noting the difference in each case. For him, this goes a long way to satisfy such queries--as closely as anyone can imagine-- "all things remaining the same, what is the difference between a five-bar and a seven bar fan?" Or what effect do the bottom cutoff bars make? or any number of other queries. In the photo above, he has made a test guitar with the back glued on with the braces on the outside--to actually hear what changes as you remove material from them. He also glues on toothpicks to the bridges of his test guitars, to ascertain the effect of changes in mass. Brilliant? Sadly, I cannot convey what the fruits of his labors are--how could I? The changes can only be perceived--not described in text form. But his experiments speak only to him and make him all the wiser for it.
     Alan writes:
     I appreciate the sound of traditional, fan-braced guitars.  For the last twenty-five years, I designed and built traditional fan-braced guitars attempting to capture and define the warmth and beauty of that sound found in 60ís and 70's Bream recordings.  Who can resist the nuance of a vintage Hauser or Romanillos guitar or the warmth of a 60s Ramirez?  Even so, I am increasingly drawn to the power, clarity, sustain and greater dynamic range of some newer designs.
     New designs with lattice and core-box tops play louder with less effort.  A player can play longer with less fatigue and include more technically demanding material in his or her repertoire.  Most professionals agree the recital material of the 90s is much more demanding on average than a decade or more ago.  Players are acquiring more new pieces and playing more chamber works with less and less time to polish and settle.  Lattice and core-box guitars allow the player to bring new pieces to a higher level quicker while conserving physical resources. 
     Australian builder Greg Smallman, perhaps inspired by Martin x-braced guitars, created his first successful lattice-braced guitar in 1981.  The lattice (picture a diagonal porch lattice) consists of eighteen interlocking braces.  It is an effective and strong bracing technique which allows a guitarís top to be about one-half the thickness of a traditional top.
     In my building, I recently ran a series of eighty test tops to arrive at a stable design.  I now offer a lattice guitar with a traditional back (rather than a Smallman-type molded, braceless back) which gives the player a warmer experience.  Any lost power is restored by reducing inertia structurally throughout.
     The lattice is now a widely accepted bracing pattern which I believe will rival the popularity of traditional fan bracing.  Simply building a lattice guitar is no guarantee of making a fine guitar, of course.  Using the lattice is as subtle an art as any luthierie but well worth the effort.

from an article written for the Boston Classical Guitar Society newsletter