|
I did a bit of testing on this some time ago. The idea behind it is that it works the way a harp does, and harps typically do sound darn nice. The questions are then; what is it that the harp is doing that makes it sound that way, and does angling the neck up on a guitar do the same thing?
Harp strings typically pull upward on the soundboard at an angle between 30 and 45 degress. Thusit seems reasonable that the tension change as the string vibrates could pull directly on the soundboard, and cause it to move up and down, producing sound. To test this I use a 'wire pull' * to pluck a harp string exactly in the center of it's vibrating length. In theory this would produce only odd-numbered partials in the transverse force signal, and even-numbered ones in the tension change signal. I recorded the sound of the pluck on my computer and used an FFT program to look at the overtone structure. I saw both odd- and even-order partials, and it was interesting that the odd-order partials decayed at a different rate from the even-order ones. This suggests that they are coupled somewhat differently to the soundboard. I'll note that classical harpists learn to pluck the strings in the center, which maximises the tension change signal and minimises a high frequency sound that also acts through the tension vector.
The Humphrey 'Millenium' model guitar uses a five degree upward angle on between the strings and the soundboard. This is pretty small, compared with what you see on a harp, and one would expect the tension change signal to be correspondingly weaker. I took advantage of the detatchable neck on my 'test mule' guitar to see if that was the case. I cranked the neck up until the strings were at five degrees to the top, and tried the same sort of 'wire pull' pluck in the center of the string. In this case, there was not very much energy in the even-order partials, suggesting that the tension change signal was not strongly coupled to the top, and didn't produce much sound, as one would expect.
I'll note that archtop guitars, with the necks back-angled at 3-5 degrees, would be expected to work about the same way, and, again, when I tried the experiment there was not a lot of even-order partial in the tone.
The 5-degree upward angle of the strings on the Millenium model does seem to put some stress on the top, though. Humphrey used a couple of different top bracing schemes to fight this, I believe, mostly involving cross bracing in front of the bridge to keep the top from dishing too much. When I saw Sharon Isbin playing one I noted quite a bit of dishing for an inch or so in front of the bridge, and was surprised to find out after the performance that the guitar was only three months old at the time. He had gone to a sort of 'lattice' brace system, and the top plate itself was very thin. I can't speak to the probable longevity of these tops: I'd have worried if it were one of mine, but he had been building them that way for years, and seemed to have a handle on the thing. The other major concern I'd have would be for trhe peeling stress along the back edge of the bridge, but, again, it may not be the problem I'd expect it to be.
There is another way to get more of the tension change signal into a guitar top; raise the height of the strings off the top with a taller saddle. This is not as effective as what a harp does, since the bridge-rocking mode of vibration causes part of the to to move upward while another part moves down, giving some 'phase cancellation'. Obviously you do need to think about the torque on the top if you do this, and the stress on the back edge of the bridge, too, so it may not be much of an advantage.
I have to make a point here: we can often hear things that are difficult to measure. In particular, even a small enhancement of a normally weak or absent partial can alter the timbre of a note quite noticably. My test mule is impossible to play with the neck up at a five degree angle, so I could not do playing tests. This is another of the long list of things I'd like to get around to doing a better job on at some point.
* Loop a length of fine magnet wire behind the string and pull until it breaks. The force is always the same, and you can 'pluck' at a known point and in a known direction. Neat.
|
Login
Register
FAQ
).