The Great Bracing Experiment - Judge For Yourself
Oh no, is this another radical bracing thread?! Not really. I conducted an experiment in bracing. I invite you to read what I did, listen to the sound samples, and judge for yourself.
I set out on this project not to prove one thing or another, but as a sort of pure science-type experiment. Although the idea was simple enough, a lot of work went into this - I hope all will at least find the results interesting.
The problem as I saw it: there are many opinions out there on ukulele bracing, and really no evidence to back up the many claims. Even two ukuleles built side-by-side, as identical as possible but with different types of bracing, is not really great proof of anything. And many believe that there is not much of a difference to be made.
The solution: build a ukulele with a given type of bracing, then take it apart and change the bracing and put it back together. Sound samples recorded from the SAME ukulele with different types of bracing could be compared and maybe something definitive could possibly go on record. This had nothing to do with "what's better" and everything to do with "what's different". The answer could be nothing worth mentioning, but I wanted to see (hear) what I could, and share my results with UU.
Since I am not a professional luthier, I really had nothing but some time to lose, as irrespective of the comparison results I was still going to walk away with a new ukulele. I decided on a tenor since that is where my experience lies.
The top, back, sides, and neck are solid mahogany. I thicknessed the top to just over 1/32" and the back and sides to about 1/8". The endblock and heel block are both mahogany.
The fretboard and bridge are rosewood, and the saddle and nut are bone. I strung it with Worth clear strings - since they come double length I was able to use one half of the set for each version of the build. It has Waverly open gear nickel plated tuners.
The initial build went well and without too many surprises. I did crack one of the sides on the bending pipe, but luckily had another piece of matching mahogany to replace it. This is the first time I tried Pete's barrel bolt connection, and I like it a lot better than trying to set a threaded insert into the end grain of the neck (which is certainly do-able, but no fun).
I braced the back with two cross braces; the top got a cross brace on either side of the sound hole, a bridge plate, and three thin braces - one on either side and one notched to fit across the center of the bridge plate - to form the fan. The material for the bridge plate and all bracing is Engleman spruce.
After completing the initial build I recorded a sound sample to serve as a control for the experiment. Then I used my wife's hair dryer and a thinned-down putty knife to separate the fretboard from the soundboard, and also to disassemble the sound box.
All through the initial build I looked into alternative bracing. I considered lattice, kasha, and minimal radial designs. And then I ran across the bracing theory of E. Palazzo and realized that his concepts would dovetail perfectly with my experiment. Palazzo first gained international attention in 1988. I will say that his concepts are by no means revolutionary, but they are interesting , and following them would lead (perhaps) to a final bracing design that would contrast with traditional fan bracing. Also, trying out this theory of bracing allowed me to leave the fan bracing in place since it would just be modified through the voicing process.
In a nutshell, Palazzo claimed that since no two pieces of wood are exactly alike, the bracing for any soundboard should develop organically through the voicing of that individual soundboard. This voicing is akin to tap tuning, except that (like Ervin Somogyi) the goal is not a specific tone, but rather the best tone for that specific soundboard. Palazzo believed that deflection testing was an extra step since the goal is responsiveness and tone, and numerical measurement can only put you in the ballpark. Furthermore, he said that the soundboard was most accurately voiced while under the stress if simulated string tension. More on how I managed this in a moment...
In order to voice the instrument before the sound box is glued up, it was positioned on the mold. Cauls are generally made to match the future glue joint around the perimeter, and the cauls are uniformly clamped against the mold. I used the solid spruce lining, which I had to remove anyway register the soundboard against the mold. Another caul sized like the fretboard was clamped where the fretboard belonged.
In order to voice the soundboard in this way, it was necessary to make a fixture to hold the mold. This allowed me to clamp down the soundboard without modifying the mold itself. As for string tension, there are various methods of attaching a stand-in bridge to which strings are actually attached and tensioned. Luckily, I already had the actual bridge in place from the fan bracing phase of the experiment. The fixture that held the mold also had tuners mounted on it so that I could basically string it up while the soundboard is not attached to the instrument and the bracing is accessible for modification.
Once the soundboard was mounted on the mold and the appropriate string tension was applied, the fun part of following Palazzo's method began. I tapped all around the board looking for the spot with the best ring. This turned out to be between the bridge and the edge of the soundboard. I was surprised at this because I had assumed that the bridge plate and central brace if the fan was going to dampen this area. I still don't know why it rang so well right over that brace.
The next step was to identify the area with the "worst" ring - i.e. muffled or dead tone. This spot turned out to be on either side of the sound hole, where there was no bracing. WHAT?! This made no sense to me, so I had to go back to Palazzo's body of work to try to understand.
He claimed that the responsiveness of any given point on the soundboard could only be very generally predicted by that spot's proximity to a brace. I was beginning to get skeptical. But having gone this far already, I decided to keep going.
I fixed a small piece of spruce directly under the dead spots. When the hide glue was holding firmly I again tapped all over the top to see what had changed. To my surprise, there was an improvement in the area where I added material. I know this is counter-intuitive, but it was working. I added another small piece of bracing butted against the other piece I had just added.... And that area rang like a bell! Even better, the original "best tapped" area near the bridge didn't seem to have changed at all. Now the worst sounding spot had moved toward the narrow bout approaching the fretboard. Seriously? Yes.
My mission became clear. By cautiously adding small bits of bracing, I was eventually able to chase the dead spot under the fretboard. (along the way I did have to either remove or pare down some of the bracing that I had added - not every addition was an improvement). At that point everywhere else on the soundboard responded similarly (and beautifully) to tapping. The added bracing did not seem to deaden the soundboard at all. The sound was much different than it had been with the simple fan bracing, so I decided to quit while I was ahead and reassemble the instrument.
This turned out to be a little tougher than the initial glue up. I definitely have a new respect for those of you who do repair and deal with this frequently! But I got it done. After stringing it up and getting the strings settled I checked the intonation and it was spot on. Again, this was a great surprise since I assumed that changing the properties of the sound board would affect the relationship between the saddle and the nut. I believe that this is why Palazzo's tapping is done under string tension. The final step was to record a sound sample to compare with the first. I recorded in the same room, the same distance from the mic.
Well, enough of my yakking. Below are the sound samples with pictures of each type of bracing. Please judge for yourself. Following the sound samples there is a link to a short video about Enrico Palazzo.
Fan Bracing Sound Sample:
(Never used photobucket for video before, so here's a YouTube link to the same video just in case):
http://www.youtube.com/watch?v=HWBCrvL6wjo
Palazzo Bracing Sound Sample:
(YouTube link to the same video):
http://www.youtube.com/watch?v=Oub5h7V8y4U
Recognition of Enrico Palazzo:
http://www.youtube.com/watch?v=NCpGCpHfaSs
Oh no, is this another radical bracing thread?! Not really. I conducted an experiment in bracing. I invite you to read what I did, listen to the sound samples, and judge for yourself.
I set out on this project not to prove one thing or another, but as a sort of pure science-type experiment. Although the idea was simple enough, a lot of work went into this - I hope all will at least find the results interesting.
The problem as I saw it: there are many opinions out there on ukulele bracing, and really no evidence to back up the many claims. Even two ukuleles built side-by-side, as identical as possible but with different types of bracing, is not really great proof of anything. And many believe that there is not much of a difference to be made.
The solution: build a ukulele with a given type of bracing, then take it apart and change the bracing and put it back together. Sound samples recorded from the SAME ukulele with different types of bracing could be compared and maybe something definitive could possibly go on record. This had nothing to do with "what's better" and everything to do with "what's different". The answer could be nothing worth mentioning, but I wanted to see (hear) what I could, and share my results with UU.
Since I am not a professional luthier, I really had nothing but some time to lose, as irrespective of the comparison results I was still going to walk away with a new ukulele. I decided on a tenor since that is where my experience lies.
The top, back, sides, and neck are solid mahogany. I thicknessed the top to just over 1/32" and the back and sides to about 1/8". The endblock and heel block are both mahogany.
The fretboard and bridge are rosewood, and the saddle and nut are bone. I strung it with Worth clear strings - since they come double length I was able to use one half of the set for each version of the build. It has Waverly open gear nickel plated tuners.
The initial build went well and without too many surprises. I did crack one of the sides on the bending pipe, but luckily had another piece of matching mahogany to replace it. This is the first time I tried Pete's barrel bolt connection, and I like it a lot better than trying to set a threaded insert into the end grain of the neck (which is certainly do-able, but no fun).
I braced the back with two cross braces; the top got a cross brace on either side of the sound hole, a bridge plate, and three thin braces - one on either side and one notched to fit across the center of the bridge plate - to form the fan. The material for the bridge plate and all bracing is Engleman spruce.
After completing the initial build I recorded a sound sample to serve as a control for the experiment. Then I used my wife's hair dryer and a thinned-down putty knife to separate the fretboard from the soundboard, and also to disassemble the sound box.
All through the initial build I looked into alternative bracing. I considered lattice, kasha, and minimal radial designs. And then I ran across the bracing theory of E. Palazzo and realized that his concepts would dovetail perfectly with my experiment. Palazzo first gained international attention in 1988. I will say that his concepts are by no means revolutionary, but they are interesting , and following them would lead (perhaps) to a final bracing design that would contrast with traditional fan bracing. Also, trying out this theory of bracing allowed me to leave the fan bracing in place since it would just be modified through the voicing process.
In a nutshell, Palazzo claimed that since no two pieces of wood are exactly alike, the bracing for any soundboard should develop organically through the voicing of that individual soundboard. This voicing is akin to tap tuning, except that (like Ervin Somogyi) the goal is not a specific tone, but rather the best tone for that specific soundboard. Palazzo believed that deflection testing was an extra step since the goal is responsiveness and tone, and numerical measurement can only put you in the ballpark. Furthermore, he said that the soundboard was most accurately voiced while under the stress if simulated string tension. More on how I managed this in a moment...
In order to voice the instrument before the sound box is glued up, it was positioned on the mold. Cauls are generally made to match the future glue joint around the perimeter, and the cauls are uniformly clamped against the mold. I used the solid spruce lining, which I had to remove anyway register the soundboard against the mold. Another caul sized like the fretboard was clamped where the fretboard belonged.
In order to voice the soundboard in this way, it was necessary to make a fixture to hold the mold. This allowed me to clamp down the soundboard without modifying the mold itself. As for string tension, there are various methods of attaching a stand-in bridge to which strings are actually attached and tensioned. Luckily, I already had the actual bridge in place from the fan bracing phase of the experiment. The fixture that held the mold also had tuners mounted on it so that I could basically string it up while the soundboard is not attached to the instrument and the bracing is accessible for modification.
Once the soundboard was mounted on the mold and the appropriate string tension was applied, the fun part of following Palazzo's method began. I tapped all around the board looking for the spot with the best ring. This turned out to be between the bridge and the edge of the soundboard. I was surprised at this because I had assumed that the bridge plate and central brace if the fan was going to dampen this area. I still don't know why it rang so well right over that brace.
The next step was to identify the area with the "worst" ring - i.e. muffled or dead tone. This spot turned out to be on either side of the sound hole, where there was no bracing. WHAT?! This made no sense to me, so I had to go back to Palazzo's body of work to try to understand.
He claimed that the responsiveness of any given point on the soundboard could only be very generally predicted by that spot's proximity to a brace. I was beginning to get skeptical. But having gone this far already, I decided to keep going.
I fixed a small piece of spruce directly under the dead spots. When the hide glue was holding firmly I again tapped all over the top to see what had changed. To my surprise, there was an improvement in the area where I added material. I know this is counter-intuitive, but it was working. I added another small piece of bracing butted against the other piece I had just added.... And that area rang like a bell! Even better, the original "best tapped" area near the bridge didn't seem to have changed at all. Now the worst sounding spot had moved toward the narrow bout approaching the fretboard. Seriously? Yes.
My mission became clear. By cautiously adding small bits of bracing, I was eventually able to chase the dead spot under the fretboard. (along the way I did have to either remove or pare down some of the bracing that I had added - not every addition was an improvement). At that point everywhere else on the soundboard responded similarly (and beautifully) to tapping. The added bracing did not seem to deaden the soundboard at all. The sound was much different than it had been with the simple fan bracing, so I decided to quit while I was ahead and reassemble the instrument.
This turned out to be a little tougher than the initial glue up. I definitely have a new respect for those of you who do repair and deal with this frequently! But I got it done. After stringing it up and getting the strings settled I checked the intonation and it was spot on. Again, this was a great surprise since I assumed that changing the properties of the sound board would affect the relationship between the saddle and the nut. I believe that this is why Palazzo's tapping is done under string tension. The final step was to record a sound sample to compare with the first. I recorded in the same room, the same distance from the mic.
Well, enough of my yakking. Below are the sound samples with pictures of each type of bracing. Please judge for yourself. Following the sound samples there is a link to a short video about Enrico Palazzo.
Fan Bracing Sound Sample:
(Never used photobucket for video before, so here's a YouTube link to the same video just in case):
http://www.youtube.com/watch?v=HWBCrvL6wjo
Palazzo Bracing Sound Sample:
(YouTube link to the same video):
http://www.youtube.com/watch?v=Oub5h7V8y4U
Recognition of Enrico Palazzo:
http://www.youtube.com/watch?v=NCpGCpHfaSs
Last edited:
