Official Luthiers Forum!

I fired up the audio generator/amp today and got the cat covered in glitter- what's next?

Seriously, I have some Chladni patterns but I really don't know how to interpret them, or what I should do in the way of more shaving braces, etc. .
I've bought/watched Al Carruth's DVD (excellent, but I'm a bit thick) and read Al's article in BRB#3, but of course the 'real thing' doesn't always look a lot like the pictures.

The soundboard is spruce, bracing pattern is Rodriguez, and I've done some shaping on the braces. Without the detour into the world of Chladni patterns I would be checking on the thicknesses and probably thinning down the bracing some more before gluing the top to the body. (BTW, I do know that the UTB/LTB aka 'harmonic bars' are pared down at the ends 'steel-string style, not in the classical way.)

I actually have two soundboards with identical bracing- a 2nd SB, cedar with the braces totally rectangular (as-glued,unshaped) gave the 'same' pattern at 285Hz as the spruce SB (shaped braces) gave at 258 Hz. So that would fit with the idea that paring down the braces lowers the freq of the mode?

Anyway, any opinions/input will be welcomed. (I'm pretty thick-skinned- really!). I don't want to go crazy shaving away braces in a futile quest for 'better patterns' here.....
I'll try to post some pics- the file names have the Hz #s.
Thanks
John

Well John,

reading and understanding chladni patterns make abut as much sense to me as fish entrails do. Although I hear tell the medicine man over in the Alan Carruth camp is real good at it.


blessings

Padma

_________________
.

Audiences and dispensations on Thursdays ~ by appointment only.


.

The resonances are quite low enough already, I would definitely not shave more if it was my build. What is the top thickness and how did it feel (light/heavy/stiff etc)

_________________
Build log

The first shape at 174Hz approximates reasonably close the main resonance on the completed guitar, at least with the Torres and Bouchet patterns I have a little Chladni experience with. I would like to see it somewhere in between 185 and 210 Hz. I agree the bracing looks substantial (except the harmonic bar scallops ) but we don't know the weight...Perhaps the low points are a feature of the Rodriguez bracing? You can't really tell until you build at least a couple and start seeing the trend

_________________
Build log

Thanks for the comments; I appreciate them.

Some more info-
Weight of the braced top (about 1/4" oversize all around) is 175g. Englemann spruce- nice light stuff.
Top thicknesses are in the 2.1-2.3 mm range
I deflection-tested the unbraced top and it deflected .300 over 18" span with a 1kg mass.
The Rodriguez plan (Tom Blackshear) shows the ends of the transverse braces taken down to 6.5mm; I've gone farther than that (though not started the scoop any closer to mid-brace) because I stabilize the neck block with CF struts so I'm not so concerned about the structural role of the transverse braces. The 'treble bar' aka diagonal brace is 'scalloped' where it passes under the lower transverse brace- this is as per the plan.
The bracing looks heavier in the photo than it is, I think- I had an angled light aimed to cast some shadows. The fans are 5x5 (central), 4x4 and 3hx5w (outers) as per the plan. It still seems pretty substantial to me...but this is the first time I've built from the Rodriguez plan.
My inclination would be to do some more gentle sanding, perhaps 'flatten' the under-bridge brace a bit and feather the edges a bit more...nothing too dramatic.
Unlike some, I do believe that it's possible to build 'too light' even if the guitar doesn't explode or fail structurally...this is a bit of a recent revelation to me, so I'm trying not to overdo the brace shaving!.


So, is the 258Hz pattern the 'ring mode' ?
Cheers
John

While it is OK for the UTB, not sure if the CF help against the collapsing between hole and bridge? We shall see 2 years from now

What I can say is that the low resonances, coupled with that huge box Rodriguez plan, could potentially give you a huge bass. As for the general tone quality, and the treble, which is the real tricky part, I guess that depends on the quality of the wood and the quality of your little decisions.

My gut feeling, even as I don't have experience with the pattern, remains that you have thinned and carved enough. The plate weight is OK, resonances definitively do not seem to call for any further scalloping, I see no reason to go at it more. You can find a bunch of Chladni photos on Gary Lee's gallery. It should help.

Btw, is that plan calling for solid laminated linings? Even if it doesn't, I suggest using them.

Before gluing it to the neck you could trim the overhang completely and see how the frequencies change. They might go up a little.

_________________
Build log

Update: I had a visit on Tom Blackshear's site and looked around. The design does have solid linings, definitely use them too. They help with such a large contraption.
Tom mentions he expects a F-F# top resonance when the guitar is done, which is 174 to 185 Hz. You have that mode at 174. There you go.

_________________
Build log

John,

I wish I had some more helpful advice for you. I only have experience with steel strings so I am not sure if I understand exactly what I am seeing but if I saw these for one of my steel strings I would say it might be a little high. Specifically, the mode at 258Hz is what looks like the ring+ mode. I recently strung up a 28.5 scale baritone where that mode was 270Hz, not that different. This seems strange because your top is so much thinner and the bracing, while maybe a bit larger than a normal classical, doesn't seem large for a steel string. The mode at 208 seems to be being affected by the diagonal bar, I usually see that one travel up and around the top of the soundhole, like a horseshoe. Again, for fan braced instruments that may never happen, I am not sure.

If I gave any advice it would be to continue shaving and pay most attention to the mode at 258. If you can manipulate the circle to be wider and encompass the two outer parenthesis I would see that as a good thing. I regularly string up steel strings in the 220 range, sometimes even a bit lower for a small guitar so I would assume you have some room to work here.

For what it is worth, I don't see any correlation I can trust between any of the free plate resonances and a the resonances of the finished guitar body(although this is something I hope to be able to do one day with much more data). The data Alexandru puts forth concerning the resonances of the finished body are great to have but doesn't have much meaning against your mode at 174 I don't think.

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

Are you shaving a certain brace area because of a certain pattern, or are you shaving in general til the pattern improves?

That is a pretty looking round mode but it is only specific to that bracing pattern. It is not related to the ring SS mode. I wish Gary would chime in as he has a lot of experience with Chladni on Rodriguez and I'm sure he would agree this particular top is already stretched.

_________________
Build log

I did just look at some of Gary's data from his archives on his site and it does appear that his finished (I assume) modes are very similar to John's current ones. I am surprised by that, but that is nothing new!

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

It seems for some reason classical free plates do not have a proper looking ring mode. Maybe it corresponds to John's 174 Hz mode, or the one below it which looks like 3 vertical lines at about 150 Hz - not pictured here but expected. There are a couple low Hz shapes that are common to all bracing schemes (probably dictated by the actual plate), but the higher frequencies (wherebracing starts to play a bigger role) start to look very different. I get one thing with Torres, another with Bouchet, and Rodriguez again is different.

_________________
Build log

Thanks for the helpful comments, everybody!

I'm really just getting started on making some measurements (Chladni, deflection) so that if I 'hit it lucky' with a good-sounding guitar I'll have some data to be able to repeat the process.
So, I've basically just shaped/sanded to what I think would be 'OK' based on looks and feel, and then done the Chladni pics.
The bracing & thicknesses are about where I'd guess they should be, based on the plans and pics I've seen of similar guitars by other more-accomplished builders.
With classicals, the tricky thing is to get decent treble response, so I don't want to get things too 'bassy'.
Alex: thanks for the mention of (Garrett?) Lee - some interesting stuff at his website.
http://www.leeguitarworks.com/

I like laminated linings, and use them on all my guitars lately. I'm also using laminated (IRW+yellow cedar aka Alaskan cypress) rims, so plenty of stiffness there. It will be interesting to see what sort of Chladni patterns I get once the box is closed up.

Cheers
John

Good call! I got the 'three vertical lines' mode at 145Hz. Based on Alan Carruth's DVD and the steel-string examples with ring mode, I think I was focused more on the 200-300Hz region, so didn't take a pic of the 145Hz pattern, though I did sketch it on my worksheet.

John

The stiff rims and linings are just great for a large and not too stiff plate. At the other etreme if you look at the famous Hauser 37, the plate might seem stiff but look at the linings, just a couple dentellones every inch or so. There are other variables that compensate: low doming, extended distance between harmonic bar and tail etc.

A bassy guitar is not bad at all if matched by decent trebles. Matter of taste in the end. I think for great tone and trebles I think you need a refined process (build more and more) and great wood. I like to use Chladni and deflection just to make sure I am positioned where I want with the type of tone and balance, a sort of double-check for the gut feeling. In any case it's bewildering how much maneuver space there is between the accepted "voicings" for a classical guitar.

_________________
Build log

Thanks everybody- esp Burton & Alex- lots to think about.

Just for reference, I'll attach a pic of Rodriguez-style bracing from Garrett Lee. Seems to be in the same 'ballpark' with mine, which is reassuring.

Cheers
John

I've been using symmetric bracing on classicals for several years now, so I've got to think back a ways on this. FWIW...

Those modes do seem a bit low in pitch to me. It could be in part because the top is over sized, which lowers things more than you might think. If I'm reading your text correctly, though, the top seems a bit thin to me, and those braces are not over large. I think the mode you're seeing at 174 is the 'ring+', and 208 is the 'ring'. If I'm right, that could be a problem.

I've been looking more at mode shapes than frequencies. The pitch tells you something about the stiffness to mass ratio, and that, in turn, tells you something about the actual stiffness _assuming_ the mass is in the ball park. But I've seen far more direct correspondance between mode shapes and the final sound than the mode frequencies provide.

A few years ago I made up one of my 'Grand Concert' classicals, just about the same size as a Rod, and saw mode shapes similar to what you're seeing there. The 'ring' mode normally looks like a sort of open X, with arcs down from the top and up from the bottom, and inward from either side from below the hips to above the waist. On that top the side arcs curved inward, to meet below the soundhole, and there was the usual arc downward from the shoulders and above the soundhole as well. Since there was a 'normal' 'ring+' mode at a bit higher frequency, this top in effect had two ring+ modes. I was curious, and left it like that. Big mistake.

It sounded OK when I strung it up, but when the cutomer received it, it was not OK at all. It was very uneven sounding, ad the high E, in particular, had a much diferent character than the others. He tried some different strings, but, in the end decided to ship it back. Since, at that point, he needed the money more than the guitar, I sent him back his deposit, and tried to figure out what to do. When I got it back I could hear the problem clearly.

The obvious first step, since I had one anomoly that I know about, was to figure out how to get the 'ring' mode back to it's proper shape and see if that might help. I tried a few things on my next couple of classicals, and found that shaving the lower ends of the outer fans, and possibly the outer ends of the bottom cutoffs, would switch the side arcs of the ring mode to the proper configuration. It did not take the removal of much material to do this.

I reached in with a small finger plane, and took two shavings, one for each side of the 'roof', off the ends of each of the two outer fans on each side (I use nine). The shavings were about 2" long. It worked.

In your case, the 'ring' mode is higher in frequency than the 'ring+', and I've observed since that this does seem to go along with this sort of shape. I'd try reducing the height of the braces in the lower corners of the top, and see if you could get the shapes and frequencies to a more 'normal' relationship.

I will note that I'm making this recommendation on the basis of one bad experience. That's not much data. OTOH, given that the cost of that single datum was pretty high, and I really don't want to have that happen again, I figure I'll play it safe.

Alan-
Thanks for those comments! Definitely some things to 'mull over'.
I will trim the top to the correct size and do some work on the braces and see what changes with the Chladni patterns.

I'll report back here.
Back to the shop!

John

I trimmed the soundboard to just a fraction (1/16") over the outside profile of the rims, and took down the height of the closing braces by about 30%, and planed/sanded a longer 'ramp' on the ends of the fan braces.
I didn't see much change in the patterns, but I don't really know what I'm looking for, either. I think the top is thin enough (and everybody seems to agree on that) but shaving braces (moderately) doesn't seem to have much effect - so what's in control here?

'After trimming and shaving'....
The 145 Hz pattern is there at 142Hz (same shape)
The 174 Hz pattern is there at 178Hz (slight shape change )
The 209 Hz pattern is there at 203 Hz (close to same)
The 258 Hz pattern is there at 258 Hz (more broken shape)

There were some other 'patterns' that formed- 235 Hz, 282, 312, 362 and a bunch more confused patterns at higher frequencies. Some of the patterns seemed to be 'partials' - only one area of the top was vibrating with clear nodes/antinodes, the rest just had a scattering of glitter. I don't know if these had much meaning - perhaps I was just being more observant this time around??

I've attached some pics. (Perhaps I need to consult Padma's fish entrails at this point???)

Cheers
John

John, I think you are in the right ballpark. I have not done Chladni analysis on the true Rodriquez-patterned guitars I built in the past. Your numbers are 10 Hz or so higher than what I usually see these days, however, I now use 7 fans, a light treble bar, and a very thin bridge patch instead of a strap. I suspect the strap has a large influence in raising the pitches of the lowest modes. I would go with what you have.

My next reply is a more general discussion regarding your question of "what next" with Chladni?

John's question of, "what's next?" is an excellent one, and is a question I have been trying to put into context since I was inspired by Mr. Carruth's and others' writings two years ago to use Chladni analysis on my classical guitars. As a former research scientist, I'm extremely aware that obtaining data simply "because you can" is a waste of time. On the other hand, when I began down this road, I knew I would first need to build an ample data set (on many guitars) before it would be useful, i.e., draw correlations between physical observations (peak resonant frequencies and modal patterns) and the tonal qualities of the finished guitars. My view of Chladni analysis is that it is one of a number of tools or senses that I can use to understand the dynamics of my tops. Chladni analysis is interesting in that it yields both quantitative and spatial information. The technique is not necessarily better than tapping, flexing, and listening, but rather, a complementary technique.

Four things I have found useful about Chladni analysis:

1. Because it is a simplified system (that employs one frequency at a time), I now have a much better mental model of how a guitar top moves across the frequency range, and how I need to (attempt) to control the resonances across all the pitches we play on guitar. Only by doing this in a hands on way did I gain a much better understanding of the big picture. I may not completely understand the patterns, but embedded in them is essentially the tonal character of each guitar. I started to think of each vibrational mode like a gear in a car. The more gears, the smoother the shifts. If a gear is broken or the wrong shape, the car can't run efficiently at a given range of speed. This is my analogy for a guitar's balance, smoothness of note-to-note transitions, projection, and the basis of wolf notes.

2. Free plate analysis of the lower modes tells me whether the top assembly is flexible enough to support a good low end once it couples with the air column in the assembled box. I use this in conjunction with top deflection to arrive at a target which is based on experience from previous guitars that were tonally superior. Graduating the unbraced top according to deflection in both axes is by far the most important factor for me to get the sound I'm after. Slightly different bracing schemes and brace carving have much less influence, and Chladni analysis suggests to me this is true.

3. Chladni analysis of the assembled box in the 400-1000 Hz range tells me if I will have a good first string (which is everything in the classical world). If the patterns are numerous and highly resolved, the first string will sing with good focus and clarity. This has been especially true for double tops. Because this is in the assembled box, this step is more or less confirmatory for me and is good for the confidence (but not much else right now). I have yet to do much tweaking of top thickness and braces in assembled boxes.

4. Chladni analysis has been the most useful in developing a double top model that sounds very close to a traditional guitar. Without Chladni, it would have taken me three years to do so empirically, instead of 1 1/2. I reasoned that a double top must behave similarly in Chladni analysis to a conventional top if it is expected to have a traditional sound. I found this was the case. Using the data I had from my conventional guitars, I had resonance/pattern targets to shoot for in free plates and assembled boxes. While there are some expected frequency shifts (because of the lighter double top plates), higher amplitudes, and better resolved patterns, the general idea is to at least match the lower modes of vibrations. In the free plates, this can be done by altering the thickness and stiffness of the Nomex, thickness of the wood skins, doming, and bracing. In the assembled box, this can be altered by the width of the linings and weight of the bridge. To a point, the above concepts would also be more or less true if you wanted your cedar guitars to sound as close as possible to your spruce guitars, and vice versa.

Two things that Chladni analysis can't do for me:

1. Tell me how the first guitar I analyze with Chladni will sound. With a data set of n=1, you are still in the dark, but keep your chin up, things start to look bright and really interesting at only n=2!

2. Tell me how to carve braces. The method has not been sensitive enough for me to detect these changes (at least on classical guitars).

Gary-
Thanks for those thoughts/pointers!
One idea that you mentioned reinforces an observation I made when I was playing around with the Chladni patterns. The top actually was vibrating at a LOT of different frequencies, so that it was hard to 'pick' the ones that deserved drawing or taking a pic. I'm assuming that the 'more different ways' the top can vibrate, the better, so I took that as a good sign.

All additions to the 'knowledge bank' here are appreciated! (I'm also starting to get a bit more disciplined with 'build books' and filing of info here. About as far as I can imagine from the 'paperless office', but it's all just wood, right??)

Cheers
John

Gary, those are interesting observations. I'm curious, on the free plate analysis of the lower modes, when graduating the top thickness do you do this on the back side of the plate or on the front of the plate? I've considered something similar but wondered how the braces would fit if I graduated the back side and wondered how it would look if the front is graduated.

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

Alan,

This thread reminds me that you had said you might do a Chladni/top tuning DVD for classical guitars at some point in the future. Any plans?

Thanks,
Max

_________________
Max Bishop
Brighton, Michigan

I'm not a Chladni guy so maybe someone can help me out, I don't really understand how the free plate frequencies relate to a completed guitar. I would expect the mode shapes and frequencies to be very different on a fixed plate vs. free plate, much higher on a fixed plate.
Do you figure that the fixed plate frequencies will jump "x" Hz or something when you glue it to the rim. Has anyone botherd to figure out how much it jumps, I know it'll depend on body size, particular mode, etc..?
I suspect it'll jump up a lot.
Thanks,
JIm

_________________
Jim Watts
http://jameswattsguitars.com