Official Luthiers Forum!

Could someone either explain or point me to a source that explains how Chladni patterns are used to voice a guitar top?

I understand how the patterns are produced and how the material lines up along the nodes which aren't vibrating. What I would like to learn is how they are "read"......or how one knows this pattern isn't exactly right and needs adjusted. Does one find the shapes on a guitar that produces the sound you want then try to reproduce these shapes on subsequent guitars?........or are you trying maximize or minimize the nodal lines?......or what?

Finally, do you usually do this before or after applying finish?

_________________
Formerly known as Adaboy.......

Darryl,

In a big nutshell I think that you had it right with "Does one find the shapes on a guitar that produces the sound you want then try to reproduce these shapes on subsequent guitars?" The problem is that you have to make a lot of the same guitar style to get enough data to find where the important parts lie. In terms of when to test, it seems you can glean the most information from the braced but unjoined top plate. Here are a few archived discussions that go into much better detail:

viewtopic.php?f=10101&t=17955

viewtopic.php?f=10101&t=16268

viewtopic.php?f=10101&t=15626

viewtopic.php?f=10101&t=15381

_________________
Burton
http://www.legeytinstruments.com
Brookline, MA.

I feel that voicing the 'free' top, before it's glued to the sides, is helpful in getting a consistent sound. This is the approach I picked up from Fred Dickens twenty-five or so years ago. I started out using his recommendations, and have been trying to refine the process since then.

I don't think you learn as much about the top by looking at the patterns once it's glued down. The fact is that the constraints on the edges mean that you get pretty similar patterns on all sorts of assembled guitars, good, bad, and indifferent. The frequency relationships of the assembled patterns are useful to know, however.

The objection is raised that the 'free' plate patterns are not good predictors of what the lower-order patterns will look like on the assembled instrument, nor of their frequencies. This is true to some extent: if you make 'matched pairs', with the same wood to the same design and so on, you'll end up with the same patterns in the lower order resonances, but any change, in wood or bracing for example, will alter the assembled modes, even if the 'free' plate modes are very similar. I'll note that even matched pairs tend to sound different, probably due to small differences in mas and stiffness distribution that might not show up in the lower-order assembled modes, but that, at least in some cases, can be seen in the free plate modes. These small differences effect the way the instruments work in the higher frequency ranges, where you can't use mode patterns for various reasons.

As I get more experience, and do more data mining, I'm coming to the conclusion that the mode shapes are more important than the exact pitches. In particular, I feel that free plate mode shapes that have a lot of 'kinks' on the node lines (which denote 'lumps' of either mass or stiffness) end up making guitars that don't sound as good (to me, at least) as ones with 'smoother' patterns. This is ongoing work, of course.

I published a series of articles on free plate tuning in 'American Lutherie' back in '91-'92 (!), and they are in one of the 'Big Red Books', of course. We've learned a bit since then, but the basic info is still good. Hopefully that can be suplimented soon. Aside fro my own work, Mark Blanchard has done some interesting and useful stuff that has helped me, and, I'm sure, others to get better and more consistent results.

Burton, thanks for the info and the links were great! I read all of them.

Al, thanks for the your post......I always learn from them.....but they raise even more questions <smile>.

After reading the links Burton posted above and Al's response here, I have these questions.

Anyone using Chladni patterns to test asymmetrical, X braced tops? Does anyone have pictures or drawings of Chladni patterns produced from X braced tops? Will non-symmetrical bracing produce the ring in the lower bout and does it help to get the ring closed on this bracing style?

When the back is tuned a semitone above the top, this contributes to a strong bass response. This would be true even on asymmetrical, X braces guitars wouldn't it?

Is this an example of the ring and a half mode where the ring needs closed?


Thanks!

_________________
Formerly known as Adaboy.......

Non-symmetric bracing tends to yeild non-symmetric modes. You can work the bracing in such a way as to get the modes to be symmetric, but why not just go the easy way and use symmetric bracing?

One possible reason to avoid the easy way out (aside from the fact that it's less manly ) is that you might prefer the sound of an assymmetricly braced top. When I did my 'matched pair' experiment, players and listeners at the ASIA Symposium commented that there was a slight difference in timbre, with many calling the asymetric top a little more 'traditional' sounding, and the symmetric (double-X) top being more 'modern'. Whatever 'traditional' and 'modern' mean.... There was about a 2:1 preference for the symmetrically braced guitar, but it was slight (people liked it a little better).

I did write that one up in 'Guitarmaker', sometime in '05, iirc, and there are drawings of the 'free' top and back modes in the article. I'll note that I was trying to match them as much as possible, and got lucky on that pair.

Once you get the top and back glued to the sides you normally can't tell whether the bracing is symmetrical or not, and the semitone tuning works equally well for both cases. I'll note that on some asymmetrically braced tops I have seen diagonal dipoles on the assembled guitar: instead of the usual 'cross dipole' and 'long dipole' you get two modes with the node ling running diagonally across the top, one 'dexter' and one 'sinister'. Once in a while you get those _and_ a more or less normal looking long dipole as well; I suspect that's a function of particularly strong coupling with the A-1 air resonance, but since it's rare I haven't had much chance to study it.

The pic you have there is a good example of a decent mode shape. I'm putting less and less emphasis on 'closed' ring modes, and more on the sort of smooth node lines you have there. Often it's very hard to get the ring to close much better than that, and going to great lengths to try makes more trouble than it solves. I'd bet that guitar will sound fine.

BTW, I tend to see the 'ring+' at between 230-250 Hz, and would worry a bit if it got below 215 or so. Remember that the frequency is an indicator of the ratio of stiffness to mass of the top, and unless it's pretty heavy, a low frequency means that the stiffness might not be enough to hold up over the long term. If the top is thin, or floppy, you might have to reduice the bracing quite a lot to get the sort of balance between brace and top stiffness that results in a nice, smooth mode like that. This leaves the whole assembly weak.

So much to think about!

Also, and you probably got a bit of this from the links, it can be beneficial to test the top at every stage as you get it ready for the guitar. I generally test the top at 3 different places before gluing it in the rims. First, the top glued up and cut to rough shape at .125. Then I test again with the rosette installed and the top thicknessed to the point I will start bracing it. And then, finally, all along the bracing process. Most of this data isn't telling me too much right now but the hope is that in the long run it will be revelatory.

_________________
Burton
http://www.legeytinstruments.com
Brookline, MA.

Howdy Al C.,

I've chatted with Mark B. about this but I'd be interested in your take on mode line thickness as relates to both projection and clarity in the final product.

Alan D.

Burton:
You're following the same path that Mark Blanchard did: a useful one, but why re-invent the wheel? He was nice enough to send me a copy of his Healdsburg talk, and if you missed it at the NEL meeting, I'll be happy to go over it with you at some point. Maybe we can get Mark to put his findings up here, or publish them in AL or Guitarmaker.

Mode line thickness is a measure of the mobility of the plate. You have to get an acceleration delta of greater than 1g to throw off the glitter, and that implies a certain amplitude if the frequency is known. As the amplitude goes up you get that level of acceleration change closer to the node line, and the line gets narrower.

For a given power input, you'll get more amplitude at a given frequency if the losses are lower. Thus narrow line widths are a handy indicator of low losses, which usually favors high frequencies.

You ear is more sensitive to high freequencies than low ones, so the more high end there is in the sound, the better the instrument is likely to project.

Also, if the Q-values of the modes are generally high (the losses are low) the spectrum will tend to be more 'peaky', and this seems to correlate well with both clarity and projection. A peaky spectrum will tend to give different note somewhat different proportions of overtones, and these proportions will change more with small changes in pitch, as when you use vibrato. Since your senses are set up to detect changes (on the theory that stuff that doesn't change much won't eat you, and you can't eat it) this gives you something the think about as you listen, so you tend to notice it more.

That's my take, anyway.

Al,

You amaze me! You must have answered the same question one hundred different times. You are always polite, patient and understanding. I want to thank you for your selflessness and dedication to teaching. I know my guitars are much better for the help you have provided, and we have never even met!

Thanks,

Hi Al,

Actually, it was Mark's talk that got me interested in testing all of these points. I have 10 or so cosmetically downgraded tops I have been prepping to test in the rough and take through some stage of completion with the goal of using this data to correlate with Mark's data on matching the top to the optimal sized instrument. I have been meaning to email Mark, or ask you, for the specific modes and frequencies he was looking at from his lecture to compare them to what I find. I remember reading that you and Mark both favor a stiffer plate too and I am interested in trying to find some relationships in the modes for a thinner top and larger braces as compared to your data from stiffer plates/symmetric braced tops. I have been building lately with the symmetrical bracing and want to compare it to a more standard style for some of my guitar sizes. I would be grateful if you could check over my lists of tests before I start to point me in a few directions or cut off some dead ends before I get too far into them. I will hopefully have it close by our June NEL meeting.

_________________
Burton
http://www.legeytinstruments.com
Brookline, MA.

Al, You use a symmetrical bracing pattern. Do you still get fish shaped patterns or is that an artifact of the asymmetrical pattern? Closing the tail of the fish can be a real challenge and often the best you can do is an amorphous blotch. From your comments above, it seems that you now feel that closing the pattern is not as important as we once thought.

It would seem to me, however, that well defined pattern lines AND a closed ring+ would be the best of both worlds. I still feel that the closed pattern is a good thing because the free plate vibrations will match the mounted plate vibrations. That implies that the bracing is not fighting vibrations that the rim is forcing the top to make.

If a symmetrical bracing pattern produces closed Chladni patterns than symmetrical would seem to be the way to go.

I am interested too if it has been easier to close a symmetrical bracing pattern but for the discussion I have been able to close an asymmetrical (standard Martin style) one also.

_________________
Burton
http://www.legeytinstruments.com
Brookline, MA.

I, too, think it's probably better in general to close the 'ring' type mode patterns, but not if you have to do violence to the bracing and end up with paterns that have 'dead' spots or 'kinks' in the lines. This whole thing is evolving (slowly, to be sure), which is as it should be. Athena was the only good idea to spring full grown from anybody's head.

I do think it's easier to close the patterns on symmetrically braced tops, which is one reason I keep making them. But, as Mark points out, the bracing is only a 'fine tuning' control on the top stiffness distribution, and most of the sound is in the top itself. If the stiffness ratio of thw wood, or the aspect ratio of the pattern, don't allow the 'ring+' to come reasonably close to closure on the unbraced top, then it will be difficult to get it to close with bracing, no matter what you do.

Burton:
I'll be happy to look over your experiments.

I do think that one of the things that makes the plates work well is a certain 'balance' between the top and bracing. I tend to use a slightly thicker top than many people, and somwaht lighter bracing, which yeilds a top that's a little heavier than a thin plate/heavy brace top would be, but gives better 'free' plate mode patterns. If I wanted to use a thinner top I might use more braces, to distribute the stiffness a bit and make it less 'lumpy'. Again, this is just my opinion, with very little hard data to back it up. OTOH, hard data is scarce in this business, and I do like the way my guitars sound.