My classical challenge

I took a class from the great mandolin builder James Condino on how he does hand applied stains and sunbursts. His stains are water based because he does use FP over the top and says if he used DNA for the solvent the alcohol in the FP would pull the color back. The seminar was about the coloring, not the finish itself. I haven't tried with FP - I do a lot of staining with DNA as a solvent followed by lacquer.

The traditional Gibson bursts were all done with hand applied stains and then finish over the top - as always, experiment on scrap.

 

I have been playing the heck out of this thing and really enjoying it. But I have also been trying to understand it better. If your remember back a month or so I said that I had some software that would let me analyze the frequency content of a continuous waveform. The process is called a Fast Fourier Transform, FFT, and is based on a theorem that says that any continuous wave can be described as the sum of a whole lot of sinusoids of different frequencies. The software does that math and displays each component frequency and the relative amount of it.

I used this back when I was building the top - the process is called "voicing" or "tap tuning" and it amounts to tapping on the top and listening to the sound that comes out. For years luthiers have been using their ears (or maybe a strobe tuner) to listen to these notes, the software tells us what they are.

What the software actually did for me was give me confidence that the top I was building, and later the box when it was closed, had notes and frequencies that were in the realm of a guitar, that it wanted to vibrate more or less as a guitar wants to vibrate.

OK, now that it is finally a guitar, what does that software tell us? Actually lots. We know that when we pluck a string it vibrates at some frequency determined by its physical properties - its mass, length, and tension. If we change one of those, say tension, the frequency changes. However it also tries to vibrate at some other frequencies that are integer multiples of the basic frequency - a string tuned to vibrate at 110 hz (A2, the open 5th string) will simultaneously vibrate at 220 and 330 and 440 hz ad infinitum. It will also vibrate at a few other frequencies corresponding to "partials". The software can show all of this.

Presto, here is the frequency spectrum of the open 5th and 1st strings



The top part of the top graph is the time domain response - you can't tell what the note is but you can tell how long it rings out, well over one second (1000 msec). That is your holy grail, sustain.

The bottom part of the upper graph is the frequencies included in that A note as it rings out. The general feelings is that lots of lower partials makes a note "warm" or "rich", lots of high partials makes it "bright". Having only a few partials makes it "simple" where lots of different partials makes it "complex".

The second graph is the high E string, it should not have any frequencies below 330, which is E4 itself.

You can see that both of these notes have some other note frequencies buried in them - that is what makes a guitar string sound like a guitar string. These are most often the fourth and fifth of the fundamental note.

What does all of this mean? Frankly I am not sure. One of the fascinating things to me is to compare the A string of one guitar to the A string of another - what is the mix of frequencies in each? Can I somehow correlate what the software shows me with what I am hearing?

This is only the start. I currently have a stack of printouts from several different guitars. I see patterns, some things make sense, some don't. I can see myself falling down a rabbit hole

 

Before I answer PhredE's questions (and I will) two more comments about the spectrum analyzing software. First, it is absolutely killer for comparing changes to a guitar. Lets say you want to try some new strings or fancy expensive bridge pins made out of petrified narwal tusk, you can record it and take a few FFT's before the change and then after and compare - do you really hear the additional sustain or whatever else you think it going on.

An example. A local musician brought a guitar to me to change the expensive boutique pickups for some even more expensive boutique pickups. He said he had recorded it with the old ones. I took waveforms of both the new and old pups and predicted that the new bridge pup would be brighter and a little louder than the old one, the neck was going to be mellower and darker. The next day I got a text - "bingo, you nailed it".

Second thing I want to empathize is that what I have done so far is no where even close to a good high school physics demonstration. The experiment is full of inconsistancies and holes, I'm doing this to gather information, not to prove anything. If I wanted to be more rigorous I would build a support for the guitars that isolates them from the workbench (or my belly). I would make the room more soundproof, I would devise a calibrated "plucker", insure that I was always plucking in the same place, and most importantly, make the tests double blind. I haven't done any of this so the data I am gathering is indicative but not conclusive.

Farther down the rabbit hole....

 

^^^^ I was in engineering school in the late 1960's and remember FFT's all too well. When I think of the infinite sum of sinusoids I realize just how cool this software really is.

The other fun part of the software is the way it confirms the basic string theory that says that the partials will be proportional to n/2 times the characteristics of the string, where n is an integer. And lastly, if you notice that the partials in the graphs are growing by one or two hz each time that is the inharmonicity of the strings - its much worse with steel but nylon shows it too.

But I have to throw out a little anecdote here. I was working in my shop one day, writing on a piece of paper when my wife walked in. "Wat'cha doing?" says she. "Just calculating a compensation factor" says I. "You're using math?" asks my incredulous wife. "Ah, yes, that's what it is for".....

 

Probably time to put the finishing touches on this thread and let it pass of into where ever old threads go. Its been a lot of fun and I really want to thank everyone who has followed along with me - I appreciate all the nice comments and encouragement, its been a fun trip.

Think back to the original challenge. I decided to build a slightly unfamiliar guitar out of materials on hand using as simple tools and methods as I could. I limited my self on the adhesive and finish, and limited my methods to those I had used in my first guitar over 15 years ago. Along the way I bumped into a very modern tool, that also had links to my past some fifty years back. Let me briefly hit on each of these

Guitar. The guitar is a fan braced classical built on the Torres/Hauser format. It turned out well. Several people have played it by now and they all give it passing marks. I took it down to my local music store and it holds its own when played against the mid priced Yamaha and Cordoba classicals hanging on the wall.

It has good balance across the strings and up the neck. Notes are clean and true, it plays in tune with itself. The action is reasonable (seems high to me LOL), the neck is comfortable (seems wide to me LOL), it sounds pretty good playing classical songs like Hesitation Blues or House of the Rising Sun or Angi (LOL). I can even muddle my way thru Classical Gas.

I'm going to call the guitar a success.

Material challenge. Everything that went into this guitar was in my possession before I started with the exception of the tuners ($65), a couple of sticks of fretwire ($10) and a pint of Everclear (I don't remember, lets say 10). The other stuff wasn't free, obviously, I had paid for it some time in the past, but my point is, I didn't go out with a shopping list to build this guitar.

Fir. I don't know whether I am surprised or not by the tone wood. I wanted to believe that the engineering properties of the wood made it a candidate and that the bonk tones were promising. But right up to the very end I was prepared for this to be a disaster. Its not. It has a very good nylon string guitar sound. Charlie thinks so, my wife thinks so, the folks at the music store thinks so. What does that say about all our sacred cows about tone woods.

The woods worked well - planed easily, bent without any issues. I think the top looks good - it is slightly more amber than a spruce top but the grain lines are even and straight and close - it might be mistaken for a nice piece of spruce. The back and sides aren't stunning but they are growing on me. I'm reminded of the Leonardo Project where people preferred the sound of dark guitars (rain forest woods) when they could see them, but blind folded they couldn't tell one wood from another. If you ever play this guitar wear a blind fold.

Tool challenge. My original idea of building with only hand tools would not have worked. I'm glad I realized that. However limiting myself to the tools in Guitarbuilding turned out to be not as bad as I thought. I do much of my work with hand tools as it was - sharp cutting tools are just a pleasure to use. However it was a new experience to not automatically turn to the band saw or the belt sander, to stop and think "where is my gents saw" or " I need a block and piece of sand paper". By far the hardest part was thicknessing the sides and plates with hand planes and scraper, but I got it done. I'm glad that C&N allowed the router for binding channels and the drill press for tuner holes, and I certainly think bending with the Fox machine is better than a hot pipe. But, hey, the tools all worked.

Methods. The two parts of the building methods that were new to me were the Spanish heel and building on a solera. I had done one Spanish heel before, the geometry worked out mostly by luck, I'll admit I really didn't understand at the time. I had never built on a work board before. The combination does work well and once you understand that the geometry will be built in from the beginning it makes sense. I would change my solera if I ever did it again, but for this guitar I am very satisfied. I'll add that if (and when) I build steel strings in the future they will be separate body and neck and built in molds as before.

Hot hide glue. I had used HHG in the past for those joints where I thought it appropriate - necks and bridges that I knew might need to be taken apart and vintage instruments where it was "correct". My challenge here was to use it for the entire guitar. It worked and worked well, but certainly changed my thinking about each glue up. You can't just run a bead of something out of a bottle and then say "gosh, where is that orange clamp..." Everything had to be ready and in its place, glue hot, clamps open, cauls in a nice little row where I could pick them up. Any of the complex joints needed to be rehearsed so I knew exactly the steps I was going to take. All of my joints are sound (the bridge hasn't flown off yet), clean up is easy, if feel kind of snobby saying "yeah, its all built with hide glue..."

French polish finish We have beat this one bloody. I am not satisfied with the finish, but that is an operator problem, not the material or method. The finish is fine for this guitar, I'm glad I did it. I know I can do better on the next one.

Luthier Spectrum Analyzer Hot damn, batman, isn't this fun. All I can say is that I don't have a clue what this thing is telling me or how to use it to build a better guitar but I sure had fun and that won't stop. When I built my first guitar all those many years ago it wasn't just to own another guitar, I wanted to understand how a little box of wood makes all those pretty sounds and whether I could do anything to influence it. Its a left brain / right brain thing, I like the wood and craftsmanship and all of that, but I'm still an engineer and I keep asking Why? Maybe the Analyzer can help, maybe it will just confuse me.

In conclusion... I set out to challenge myself. To make something a little different, to limit myself in several ways. I wanted something to look forward to when I went out to the shop each day. I wanted a little frustration, a little satisfaction. I wanted to question some things, to learn some others. I think it was a success.