Official Luthiers Forum!

OK, I get what Chladni patterns are showing...as I understand it,
basically the glitter accumulates at the nodes of the vibrating top,
and those nodes can change with the frequency of vibrations. 
Great.



Now...I'm at a loss as to what to do with this information.  The
books I've read (Cumpiano, Benedetto) don't even mention it. 



Are there any links or resources out there that provide some guidance
about how to use the info that Chladni patterns are providing? 
How to alter the bracing, top shaving, etc...??

Erik,

Search here and elsewhere on the web for Alan Carruth. There are loads of his posts chock full of great info.

_________________
now known around here as Pat Foster
_________________
http://www.patfosterguitars.com

Check out the GAL Big Red Books.

There are a number of posts here that you can access in the archives by searching for Chladni. The best source of information are the GAL articles about the subject as they have the pictures of what the patterns look like for different kinds of guitars at different frequencies.

The most knowledgeable person here on the subject is Alan Carruth so when you search the archives for chladni, his posts are the ones that will help the most.

I wish we had a video of one of the workshops that Alan has done at ASIA symposiums on the subject as they are the best lectures I have ever seen on the subject.

As Alan will say, it is not the specific frequency but rather the shape of the pattern that will help to show how well all of the elements of the guitar are coming together to produce a sound at a given frequency. If the pattern is unfocused or does not converge or close, there are things that can be done to refine the shape by shaving braces and such.

Start by searching the archives here, the GAL archives, check out Alan's website http://www.alcarruthluthier.com/, and wait for Alan to respond to this post.

I will have to check out the GAL archives and the Big Red Books...thanks!



Yep, if there's anything I'm "in tune" with, it's Alan's experience in
this area (being a scientist myself, he's one of my closet luthery
heros).  I've already read the ~200 or
so posts that have the word "chladni" in them, and found the UNSW site
on this.  But what I still find lacking are practical guidelines
on how to alter the top if flaws are revealed in the Chladni patters...



....and even what "bad" patterns look like compared with "good" ones (maybe tight nodal lines vs big broad nodes...not sure).



I was under the impression that there exist some general rules about
how to tweek the braced (but not yet glued to the rims) top based on
what the Chladni patterns look like at various freqs, but maybe I'm
wrong about this...???



An example...how to close the ring mode depending on where it's open...stuff like that.



Another example...to avoid wolf notes, how best to shift in frequency
the main resonance (this corresponds to the most ring-like ring mode,
right?) if it corresponds to a musical note...stuff like that.

If the ring is open, then the top is not naturally bending tangent to the ring at the point it's open. The solution is to make the top more flexible on that axis, so you thin the bracing and/or top plate in that area and the ring should close.

Every time you thin the bracing, you lower the frequency of the ring mode and so you have to balance closing the ring with keeping the fundamental of the top where you want it in terms of pitch.

Al can be a lot more specific, but that's the quick and dirty version of what I took away from my time studying with him.

_________________
Bob Garrish
Former Canonized Purveyor of Fine CNC Luthier Services

All of this Chladni stuff is a work in progress, and quite empirical at the moment. Obviously the way the top and back vibrate before you glue them down should have something to do with the way the guitar works when it's done. Just as obviously, there are so many things that are changed by the act of gluing those plates down that tracing out the cause and effect of it all is far from trivial. We're chipping away at it, slowly, but it's not as if this was Homeland Security: funding's a problem.

FWIW, my own experience suggests that it's the mode shapes that matter most. In particular, you'd like to see some of the 'ring' type modes 'closed'. On the top the one that seems to have the greatest effect is the 'ring-and-a-half' or 'ring+' mode; the one that has a (you hope) closed ring in the lower bout, and a curved line across the upper bout roughly along the shoulder brace. There actually seem to be two or more different modes that can become the 'ring+', deepending on how the bracing is done and so forth, but whichever one it is, getting it to 'close' seems to be the key.

By 'closed' I mean that the ring is a complete one, with no gaps. It may not be round, or even very smooth, although, again, smoother rings seem to work better. In some cases the ring in the lower bout will have one or more 'parenthesis' with it: you can see )O( or )O, with the parenthesis joined to the ring, and that's OK.

Once you get that, you can look higher in frequency to see if there are more modes with 'closed' rings. There will be 2-ring and 3-ring modes as you go up. The best guitars I've made, at least according to the folks that have tried them and heard them, tend to have a lot of these modes, and they are well formed (symmetric, and smooth), and very active (narrow node lines) with high Q values (narrow bandwidth).   

Many moons ago Oliver Rogers, who just passed on, did a long computer modeling study of the 'free' plate modes of the violin that enabled him to publish 'mode influence surface maps'. These told you where to take off wood to alter specific mode frequencies. The problem for us is that guitar plates are much less standardized than violins; even small changes in shape, wood, or bracing patterns alter the way things work. The only way to get really good at it is to make a lot of one design, and then you're only really good at that one design.

Again, FWIW, the braces below the bridge on the top seem to have the greatest effect. You can reduce the lower tone bar to almost nothing at the end. On my tops making the finger braces too low tends to get me into trouble. Other than that, about all you can do is try something and see what happens.

In the beginning you may find that nothing happens, no matter what you try. That's probably because the braces are too tall: they so dominate the stiffness that the top can't influence the modes at all. Once you get to a better balance between top and brace stiffness things will start to happen; sometimes 'way too fast. Take a few shavings at a time, and be patient. Once you find something that works, do some more of that. When it stops working try something else.

Sometimes the guitar will be balanced on the head of a pin, so to speak; one or two shavings in the wrong place can really alter the tone. Again, take it easy.

Many thanks Alan!



So if I'm reading the tea leaves correctly on this example from UNSW:

http://www.phys.unsw.edu.au/~jw/guitar/patterns_engl.html



The ring mode at 221 Hz might be closed by shaving slightly the tone bar underneath the open part of the ring...right?



So would the "ring+" mode be the one at 438 Hz or maybe 511 Hz?



Do I see a hint of a 2-ring mode at 530 Hz (lower bout and soundhole)?

So how would gluing on the bridge affect all of this? After all it is the heaviest brace your going to put on the top, apart that is from the rims/back which are also 'bracing' the top. I do my final tuning after the guitar is closed, bridge on and strung up, by ear not pattern.

Colin

_________________
I don't believe in anything, I simply make use of a set of reasonable working hypotheses.

[QUOTE=Colin S]I do my final tuning after the guitar is closed, bridge on and strung up, by ear not pattern.
[/QUOTE]



By strategic thinning of the soundboard?  Or do you actually reach in and work on the braces some more?

erikbojerik asked:
"The ring mode at 221 Hz might be closed by shaving slightly the tone bar underneath the open part of the ring...right?"

That's the 'ring-and-a-half' or 'ring+' mode; see the line across the upper bout? My feeling is that the braces are already too low on that top, and you might not be able to get the ring to close fully at this point.

"So would the "ring+" mode be the one at 438 Hz or maybe 511 Hz?"

No: again, 'ring-and-a-half' or 'ring+' are just two names for the same thing.

It looks to me as though that top is asymmetrically braced.

I used to test 'free' tops with the bridge on. It was something of a pain: sadly, there's no sort of tape that will allow you to stick one on temporarilly that will really replicate a good glue joint. Even the worst tape tended to peel the top wood up when I removed the bridge.

I finally did some tests that indicated that the mass and stiffness of the bridge tend to have opposite effects on the mode shapes and pitches, so that they seem to cancel out to some extent. I know this is a cop-out, but at that point I gave up testing tops with the bridges on. I find that gluing on the bridge in the end normally drops the 'main top' tap tone pitch by about 1/2 semitone, and try to take that into account as I work, but otherwise I've stopped worrying about the effect of the bridge in tuning.

I've gotten to where I very seldom feel the need to do any tuning after the box is assembled. Once in a while I'll shave back braces a little, but they're usually easy to get at. That's the real point of 'free' plate tuning: it's a lot easier to work on things before you close the box!

If I may be so bold as to speak for Al here for a moment, one of the major contributions he's made to this science is that he's been able to correlate free plate Chladni patterns to final results with the top glued at the edges, with a bridge glued on, with all this attached to a box, and with finish and string tension on it. So, yes, the guitar top winds up with more stuff on it, "clamped" around the plate and with a Helmholz resonator behind it, and yet the free plate patterns are predictive enough.   That requires a certain leap of faith, to be sure, but that is what Al has done...made that leap...and now he's been generous enough to share the techniques.

Yes I know all of that Rick.

Thanks Al for the additional info, I guess if you do enough of this, then you know how the final result will work out. Just interested to know how the patterns are effected once the box is closed and the bridge in place.

I've spent all of my professional life (35 years or so) as an academic scientist working on and devising methods of analysis of seismic data and more recently 3D tomography in deep crustal structures. So when I get in the workshop I like to leave the science behind and just become an artisan, but I appreciate that it is a very valid approach to the business of guitar acoustics. I know the guys at Loughborough used it when designing and building the all-polymer guitar, still one of the best I've ever played.

Colin

_________________
I don't believe in anything, I simply make use of a set of reasonable working hypotheses.

Well, Colin, when I'm in an earthquake, I like to forget that there are seismographs and scientists staring at squiggles on paper and just simply do whatever it takes to save my ass.

To a guy who has lived in California earthquake zones, the work of seismologists looks as irrelevant as Chladni patterns look to you.   Many years of expensive study...no appreciable success at predicting "the big one." At this point we'd best look to Nostradamus...

That doesn't mean that I don't appreciate your work analyzing quakes or think that maybe just...maybe just someday, you guys will be able to get ahead of studying what did happen when the earth shook and let us know ahead of time that we need to do something to keep from having a building fall on us. But for now, I think that the study of top motion is perhaps better understood than the movement of the earth...

[QUOTE=Colin S]

I've spent all of my professional life (35 years or so) as an academic scientist working on and devising methods of analysis of seismic data and more recently 3D tomography in deep crustal structures. [/QUOTE]

A Geophysicist??? You don't wear a bow tie and gaudy braces do you?? Seems to be the trade mark of many Geophysicists I deal with in my job in the oil patch. Walking into the Geophysics department of some of my clients is like walking into MacDonalds...all thats missing is the big floppy shoes

[QUOTE=kiwigeo] [QUOTE=Colin S]

I've spent all of my professional life (35 years or so) as an academic scientist working on and devising methods of analysis of seismic data and more recently 3D tomography in deep crustal structures. [/QUOTE]

A Geophysicist??? You don't wear a bow tie and gaudy braces do you?? Seems to be the trade mark of many Geophysicists I deal with in my job in the oil patch. Walking into the Geophysics department of some of my clients is like walking into MacDonalds...all thats missing is the big floppy shoes

[/QUOTE]

Mud logger eh? I'm more the grey bearded outdoor type geophysicist, with just a hint of eccentricity in the sock department. Just love setting off monster explosions or towing a big array. Although I usually leave the actual getting sea sick bit to those lower down the foodchain. Though I did spend a long time on a boat in the tsunami aftermath for the UN.

Colin

_________________
I don't believe in anything, I simply make use of a set of reasonable working hypotheses.

Did the mudlogging thing many years ago. I was about to quit the job (if you can call it that) after 2 years when some Wellsite Geology work came up. Dont know if taking the work was a good thing or not. Financially very lucrative but after 20 years in the oil patch I'm ready for life outside the patch and more time spent in the workshop and generally having a more normal life.

[QUOTE=Alan Carruth]erikbojerik asked:

"The ring mode at 221 Hz might be closed by shaving slightly the tone bar underneath the open part of the ring...right?"



That's the 'ring-and-a-half' or 'ring+' mode; see the line across the upper bout?  [/QUOTE]



OK great, thanks!  This helps.



So which of those photos shows something closest to the "ring" mode? 

The 221 hz one, if the left side were closed like the right, would be a ring mode. Back to Al, but does the ring ever happen without the line in the upper bout?

_________________
Bob Garrish
Former Canonized Purveyor of Fine CNC Luthier Services

I would think that if you dispensed with the transverse brace in the upper bout (which I do), you'd at least lessen that node.

[QUOTE=Bob Garrish]Back to Al, but does the ring ever happen without the line in the upper bout?[/QUOTE]



That's what I thought the ring mode was...and that the ring-and-a-half (or ring+) was just a higher-order (higher Hz) mode.



All those photos have action around the UTB.

The 'ring' mode is the one just below the 'ring+' in pitch: I don't remember the frequency offhand. On most guitar tops the 'ring' is almost never well closed: too much difference in cross grain stiffness between the two ends. Rick's guitars might be different, and, if so, I might try to close the 'ring' instread of the 'ring+'.

I made one guitar that had two patterns that looked like 'ring+' modes. It was a large bodied classical, and if I had the fans trimmed a certain way the 'ring' would close across the middle where it normally just necks in a bit, to form another 'ring+'. I'd never seen that before, so I finished the guitar up to find out how it would work. Sadly, when I got it together, one of my hearing aids blew, and I couldn't hear it very well. On top of that, of course, these things change a little.

When the customer got it, he complained about how 'uneven' the sound was, with the high E string being much different from the others. In the end, I had him ship it back and refunded his money. With two ears I could hear that it was, indeed, badly uneven.

I kept careful track on the next couple, and figured out that a little shaving of the outer two fans below the bridge would bring the 'ring' mode back to the 'correct' shape. I got in with a finger plane and took about two light shavings each off those four braces, and the sound was fine. By this time the customer had found something else he needed the money for, so I ended up selling it to another buyer, who has been quite happy.

A word to the wise....

I hope at some point to see a little more 'science' behind this Chladni free plate testing stuff. Right now, as Rick says, it's mostly just empirical; trying to find corellations like the one above(but, hopefully, without the pain!). There are some logical inferences that suggest why one sort of pattern might be better than another, but no 'proof' yet, and there may never be. I've got one experiment going, comparing 'free' and 'bound' modes on round plates, in which the preliminary results are about what I'm expecting. This has me suspicious. I'm trying to get some time in a laser lab to do some holography, as that might shed some light, so to speak, but until then I don't want to get all 'het up' and fool myself.

But Al, if you can't fool yourself, then what fun is it?