Official Luthiers Forum!

I've been seeing more and more people referring to the glitter technique to help voice the top. Can somebody here provide a more detailed explanation of 1.) how to do it, and 2.) what the results tell you and how they can be applied to future builds.

Thank you!

John

If you sprinkle glitter on to the top that is flat and excite it or use something to make it vibrate , the glitter will migrate to show areas of the top that move.

John, I think there is a comprehensive discussion in either Big red 1 or 2, I will try to find out for ya.
from what I understand, you use a speaker to generate a constant tone that will cause the top that is suspended some how to show nodes of movement. With experience you can adjust the bracing to control where these hot spots are. Thus creating a consistent sound from guitar to guitar. That is my very limited knowledge.
Maybe others will chime in, I'm sure that Al knows much about it, as does John Osthoff.LanceK38896.9567708333

_________________
Support the OLF! Bookmark our STEWMAC link Today!
Lance@LuthiersForum.com

That is really interesting, i'd like to hear about that too! Thanks!

John,

This might get you started:

http://www.phys.unsw.edu.au/~jw/guitar/guitarchladni_engl.ht ml

I've only barely worked in this area after some reading of Alan Carruth's posts on various forums and at the urging of my father-in-law who taught acoustics of musical instruments as a physicist and was acquaintence of Carleen Hutchins, who did early work on the technique with violins.

I believe you need to excite the top with a pure tone to eliminate overtones.

From what I understand, the top of a guitar with a desireable sound will tend to show certain patterns of vibration that become visible by where the glitter accumulates as the top is vibrated at varied frequencies. As the top vibrates, it tends to sort of "flap", in waves, if you can imagine a potato chip going from curvy to flat to "opposite curvy" and back again at the frequency at which it's being drive by the speaker. The parts that move bounce the glitter towards the parts that don't move so much. It's sort of like the glitter gathers in the "troughs" of the waves. Thus the glitter patterns, which change as the frequency of the speaker driver changes. Certain patterns are good to have happen at certain frequencies.

Some people vibrated the tops of good-sounding guitars and vibrated them to see what patterns are generated and at which frequencies they happen.

It seems the magic part is that the bracing can be tweaked to influence the patterns to enhance your chances of getting a nice-sounding guitar by making the patterns resemble those of known-good-sounding guitars.

That's all I know. So far, I've only briefly made patterns on my one and only, an OM. Really interesting, but at this point, I don't know what to do with it. The next step for me would be to learn how to interpret the patterns and make appropriate changes to a top.

This is just a start from a novice. Perhaps Alan will show up.burbank38897.0599768519

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

Others have written extensively about this but here are the main points for guitars. Carved top instruments are a little differnt:

You work on your top after you have glued the braces to it but before you have shaped them to their final dimensions.

You will need a source of sine waves with a method of controlling the frequency. I signal generator is good but some guys use a generator that runs on their pc. Google on wingen. You should have a small (< 6") speaker that is capable of handling 25 watts. You'll need an amplifier to get to about 20 watts.

You will need four foam pyramids that are about 1-1/2" on the base and about 1-1/2 inches high.

Place the top on the the four pyramids such that the pyramids are near the edge of the top. braces down.

Sprinkle craft glitter on to the top.

Put on your hearing protection. Sine wave can be very damaging and don't sound as loud as they actually are.

Turn on the tone and move the speaker very close to the top. Start at a frequency of about 60hz. Slowly raise the frequency until you notice the glitter hopping about on the top. You should also notice that the glitter is hopping about more in some places than in others. Move the foam pyramids around so that they are beneath parts o the plate that aren't moving very much. Fine tune the frequency to get maximum glitter movement. Move the speaker around the top to excite different parts of the top.

You should start to see the glitter take on a pattern of lines and empty areas. Write down the frequency and draw a sketch of the pattern.

Redistribute the glitter until it covers the top evenly. Slowly raise the frequency of the tone until the glitter starts to hop at this new frequency. Move the pyramids around to get the greates movement of the glitter. Note this new frequency and draw the new pattern.

Repeat the process with ever increasing frequencies until you get between 6 and 8 different patterns.

This is the first step in the process. The next step is to examine the patters and try to determine what to do to the top based on this information. I'll post on this after lunch.

I will not pretend to be any sort of an expert on glitter testing, but I do use it while building my guitars. It is a type if measurement that I use to while carving my braces on my tops. I do some testing on backs too. Many builders do not put any stock into this type of measurement. That is ok. But they probably use some other feedback mechanism to take “measurements.”

Let me take a second to talk about what “I” mean by “feedback mechanisms.” When I am bending my sides I use a spot thermometer (the ones stewmac sell) to measure temperature of my bender. I can get an accurate measurement on the thing and since I am using a heating blanket I can just adjust my rheostat to dial in my temps. The feedback of the rheostat adjustment is the temperature I am reading off the themometer . Recently I saw a builder do a demo on side bending over a hot pipe. He was using a propane torch to heat his bending pipe and he used a few different feedbacks to tell when the pipe was at the right temperature. He would spritz water on the pipe and watch and listen. He was looking to see the water “dance” a little, but not bounce right off the pipe. This was his first indication. He would also smell, and even did smell the mahogany burn a little. (A little too hot!) I am sure some other things were coming into play including the size of the flame, just a general feel of the heat etc. Anyway his way relies on experience, senses and makes things a little more sexy and traditional. Ya know StewMac wasn’t around a 100 years ago.


Back to the Glitter Testing. The idea is that when we “tune” or “voice” a top, builders use different tools for feedback. I don’t have a great tonal memory, so when I tap something and listen, I am not good at writing in a journal or remembering what it sounded like. I still do lots of taping and I listen more for “dead” spots than a particular tone or overtones. I also flex the tops a lot throughout the process. I have done some deflection testing in the past (a type of stiffness test) but I don’t do too much of that at the moment. Most of that stiffness testing is done by “feel” now. Hey see I am becoming maybe I am becoming more of a traditionalist.


Basically the idea of glitter testing as mentioned is to excite the top with different frequencies. I use a sine wave generator and a speaker to do this. I support the “free” top on some foam pyramids. Then I sprinkle poppy seeds on the top. Running through the frequency will cause the poppy seeds to dance and certain frequencies. The link provided here and in a previous post will explain it much better:

More on Glitter Testing

I am using this glitter stuff at the moment as more a database collection, but I have learned a few things that seem to changes the shapes of modes while I carving. It is nothing that I will profess any conclusions at the moment.

Another thing I do is spectrum analysis. That is pretty cool too, but I don’t have any conclusions on that yet either. Basically the spectrum analysis correlates to the glitter patterns but gives a graphical printout of the spectrum. (I am currently looking at about 30 to 1000 hz. I measure my free tops, my backs, my rims with the top attached the box, and the guitar when strung up. It still is information gathering, but I am managing to get some consistencies in my measurement from guitar to guitar within a particular model.

I hope to study with some of the greats at some point Ervin Somogyi’s for instance. I am sure I would modify some of my processes but I think I would still take my measurements. For me the idea of having the measurements, make things repeatable and more predictable. So in a nutshell, I will continue to use my glitter testing and spectrum analysis to add to my database. I will continue to tap, flex, scratch and listen too. Maybe over time I will develop the taping skills that the “scientific measurements” won’t be necessary, but I suspect I will still keep that data for reference.

Al Carruth has been mentioned here too. Do some research and you will see lots of information that Al has shared on this stuff. He knows way more about this stuff than I ever will

In my previous post I described how I generate glitter patterns. John Osthoff described something similar. John brings up an interresting point. I'd like to discuss this.

Currently, John is in the "Naturalist" phase, as are most of us persuing a scientific approach to guitar building. I call this the naturalist phase because it is similar to the work of the naturalists of the late 18th and early 19th century. These naturalists spanned out around the globe colletcing specimens of plants and animals. They spent their time drawing pictures and noting where the various species were the same and where they were different. They did this for a hundred years or so before anybody took this data and started developing theories on the nature of species. (Darwin) So, when trying to figure out the natural world, you have to collect a lot of data before you can use that data for anything.

For the most part we are in this phase when it comes to measurements of musical instruments. However, some trends seem to be emerging and now it is becomming possible to take the information that has been gathered and use that to guide us when we build a new instrument. Here's where things get a bit more contraversial. Unlike Darwin, using glitter patterns to tune your top won't get you burned in effagy. ( I'm not so sure about the violin builders on this). Here are a few things that we think we have learned:

One of the patterns that appears on a guitar top is called the "ring and a half mode". This pattern is the closest to forming a complete circle in the lower bout. Most likely, however, the circle will not be completely closed. The objective is to shave the braces until the pattern closes into a ring. You most likely will not succeed with this but the closer you get to fully closed, the better.

Why would making a closed ring be better? For now the answer seems to be that we don't know. I personally have a hypothesis. The basis of this hypothesis is that once a top is attached to the sides of the instrument the ring will be closed. The reason that this is so is that you can't have an anti-node (a place where the top is vibrating alot) at the edge if the edge is not allowed to move which is what happens when the top is attached.

If you trim your braces such that the top doesn't have any anti-nodes at the edge than attaching the top to the rims won't be changing the behavior of the top (much). If there is a big anti-node at the edge than that means that there is a piece of bracing pattern that is allowing it. Now when this is attached to the rims, the braces and the rim are working against each other. The braces want the top to vibrate one way and the rims want the top to vibrate a different way.

If the ring is closed than there is no anti-node at the edge and the braces and the rim both want the top to vibrate in the same way. The top is a simpler sound radiator. It is still very complex but it is simpler than the unclosed top is. I believe that a simpler sound radiator is going to have better behavior in that the harmonics that is effecently readiated are going to be closer an even multiple than a more complex sound radiator. This should give you a tone that is more musically pleasent.

I can't emphesize strongly enough that why closing the ring makes for a better sounding guitar is still pretty much unknown. My idle speculations not withstanding. In fact, many builders claim that it doesn't help the sound of the instrument at all.Mike Mahar38897.5873958333

I by no means have any expertise in this area. But my common sence tells me the glitter gathers into the areas with the least sympathetic vibrations. i.e. transitional nodes along a wave lenth or terminal nodes. By closing the loop the area inside the pattern is acting as its own diaphram suspended in the major diaphram (the area outside the pattern).

What this means to the overall resonance of the intire top will take some thought. but off the top of my pea brain this would as Mike explains eliminate termial nodes and leave only trasitional nodes.MichaelP38897.6020949074

Thanks Mike: saves me a lot of typing!

One way to get at this is to think about a string. When you pluck a string it vibrates in a lot of different ways. You know that if you touch it right in the middle you hear the octave of the tone that the string is tuned to. What you've done is to 'kill' all of the vibration modes that are moving [have an 'antinode'] at that point. The center of the string is a 'node', a non-moving point, for all of the even-order partials of the string; the vibrations that are at multiples of two times the fundamantal frequency. The 7th fret location divides the string into thirds, and is a node for all the third-order partials , and so forth. If you put a little paper ring around the string and touch it at a node point, like the seventh fret, it will tend to migrate to another node, in this case, to the 19th fret.

The thing is that strings are really simple vibrating systems, and that's why they have such nice regular patterns and frequency relationships. The patterns on a guitar top are more complicated; for one thing, because the top is more or less two-dimensional, the moving sections are areas and the nodes are lines in between. But,just like the little paper ring, some dust on the surface will tend to 'dance' over to the places that aren't moving. The important thing to remember is that it's not just the 'hot spots' that are determining things; the vibration involves the whole plate, just as you need the whole string on the guitar. Making changes in one area, even one that's not 'doing' much, can change things in other places.

Ernst Chladni found this plate vibration stuff out back in the 1830's iirc. He used a violin bow on the edge to drive to the plates, and sand to make the patterns visible. The big problem he, and everybody else, had until recently, was in figuring out the pitch of the tone. Now we've got digital tuners, and computers, and all sorts of electronics that can chop and dice the signal seven or eight ways while you sip your coffee. It's hard to overemphasize how cool this is to us old farts, who remember when you needed about $10,000 worth of lab equipment to do the same thing, and you couldn't play Solitaire on it.

The series I wrote on plate tuning is in the Big Red Book #3, iirc. We've learned a bit since that came out, but it will get you started. One of the main things we've learned is that there are a lot of ways to do this stuff, and what I wrote is _not_ gospel. I generally go for somewhat lower frequencies these days than what are in that article.

As Mike said, the theory is the weak part right now: we really don't know just what the 'free' plate patterns 'mean' when you glue the thing together. OTOH, I've got data that indicates that those patterns do tell you something useful: people tend to prefer guitars that have better 'free' plate patterns, for one thing. Again, Mike's analysis agrees with my current thinking, and the most important thing to get is 'closed' and smooth 'ring-type' mode patterns. The frequencies tell you something about the ratio of stiffness to mass, but what they indicate beyond that is hard to say.

The only way to learn this stuff is to experiment, and keep records. There's so much we don't know that enybody can discover something new with just a little persistance. If you do see something interesting, _please_ talk about it! Togther we can learn a heck of a lot more than any one of us will have the time to find out.

[QUOTE=Alan Carruth] Thanks Mike: saves me a lot of typing!
[/QUOTE]
It didn't semm to save you all that much

I've used the Chladni plates in teaching physics and we used a violin bow to create vibrations. Could you bow the edge of the guitar top (before it's attached, of course) and learn anything?   Hmmmm. Would be cheaper and easier than a signal generator, I would think. On the other hand, spruce may not be dense enough to transmit the vibrations.

Ron

_________________
OLD MAN formerly (and formally) known as:

Ron Wisdom

Somewhere in the middle of Arkansas......

[QUOTE=old man] I've used the Chladni plates in teaching physics and we used a violin bow to create vibrations. Could you bow the edge of the guitar top (before it's attached, of course) and learn anything?   Hmmmm. Would be cheaper and easier than a signal generator, I would think. On the other hand, spruce may not be dense enough to transmit the vibrations.

Ron[/QUOTE]

Ron -

My understanding is that by using a signal generator, you can excite a single mode at a time, so you can clearly see the shape of the mode by observing the glitter. You also know the frequency of that particular mode.

Bowing the edge would only excite those modes for which the edge was not a node - and, probably several at once, so the glitter pattern wouldn't tell you much. And, it wouldn't give you any information about frequency, unless the top vibrated clearly enough to allow you to identify a note by ear.

Also, you would have to somehow hold the top in order to bow it ( moreso than by setting it on foam triangles, which is enough support to excite it with a speaker). That would change the response of the top.

Phil

[QUOTE=Alan Carruth]

As Mike said, the theory is the weak part right now: we really don't know just what the 'free' plate patterns 'mean' when you glue the thing together. ....   [/QUOTE]

Alan -

it seems like this type of testing (and, tuning) is generally done to a braced top plate, before the plate is assembled to the sides. This would certainly make it difficult to predict the top's response when you glue it to the sides and constrain its edges.

But, if the top plate were glued to the sides (no back, yet), it could still be excited, and observed, and also still be tuned, either by reducing plate thickness (locally, or globally, as required), or reducing bracing stiffness.
It would also be easier to support the assembly (for example, top up, with the lower edge of the sides resting on three foam wedges) without affecting the response as much.

This frequencies and mode shapes would still change when the box was closed, but you'd be much closer to the final response, wouldn't you?



The whole idea of vibration testing the top (and/or top and sides) really intrigues me. I already have an audio-range signal generator, and I will certainly try this. It would be a tremendous help to have real data on the components of each build.


And, I should have the top for my first guitar braced soon, so the timing of this thread is great for me.


Phil

But it would give it a nice "woody" tone

I've seen one or two 'how to's' on using a bow to drive plates. The supports have to be fairly firm; rods stuck in blocks in one case, with big stick-on pencil erasers to contact the plates. You have to get them in just the right places when you drive the modes, but they do damp out all but one or two in the process, so it works pretty well, or so I'm told. I've never tried it. There are a number of other ways to drive plates that avoid some of the noise of using a speaker, such as using stick-on magnets with a drive coil, but most of them introduce other problems and require that you already know what the modes look like, at least to some extent. You might get someplace by Googling on 'Fuhr's method', or maybe not.

The problem with testing plates after they're glued to the sides is that they all have the same series of modes, more or less. I've built pairs of guitars that looked the same once they were together, with the same mode shapes at the same pitches, but they still sounded noticably different. So far the best correlations I've found have been with the configurations of the 'free' plate modes.

I _think_ that what we're actually doing when we tune the lower order free plate modes so that the shapes are 'right' is setting up the mass and stiffness distributions that will give 'better' high frequency performance. Some initial experiments seem to confirm this, so, of course, I'm suspicious. Any time I see the result I expected to see there's the chance that I'm fooling myself, and a better experiment is called for. Anyway, the point is that once you get the box together there's so much coupling going on that above about 700 Hz you're in the 'resonance continuum', where it's hard to seperate things out. They just overlap too much. 'Free' plate tuning seems to be a way to 'get at' the things that make this work properly in a more controlled way. Maybe. ;)

[QUOTE=Phil Marino]

This frequencies and mode shapes would still change when the box was closed, but you'd be much closer to the final response, wouldn't you?[/QUOTE]

In the past couple of years I have lurked in building forums and read many articles regarding using glitter techniques. I will state upfront that I have no true knowledge of its use or effects on the completed instrument. One problem that always bothers me is all the work tuning a top, either using glitter or not, and then attaching it too the rim. My thoughts have always been that it would make the most sense tuning after the top is secure in its final resting place so that it can react in its true environment. I also know that many builders attach the back first to help glue cleanup in the visible area of the inside. Of course this wouldn’t allow any great access to the continued tuning of the top. I couldn’t find the quote I wanted from David Hurd’s web site but from what I remember the thought was:
Luthiers can make good instruments using knowledge, experience, touch, and feel without the use of science. However, he feels that the ability to use science allows luthiers the ability to repeat the building of high quality instruments. Great info on David Hurd's web site


I have no real knowledge yet and won’t be able to start my build till the end of July. My focus is currently on work, vacation, and a Million dollar Poker Tournament that I have qualified for. This type of discussion is always interesting and I will continue to read in preparation of my first guitar. Thanks to all for your sharing of thoughts and experiences that I will soon put to good use.

Philip
Great info on David Hurd's web site

_________________
aka konacat

If you think my playing is bad you should hear me sing!
Practice breeds confidence and confidence breeds competence. Unfortunately, I'm stuck in practice.

[QUOTE=Mike Mahar] using glitter patterns to tune your top won't get you burned in effagy.[/QUOTE]

Maybe not, but it does get you some grief, mostly by the people who have not taken any time to try to understand this stuff. Can you build great guitars without this type of testing? Absolutely!

[QUOTE=Mike Mahar]
One of the patterns that appears on a guitar top is called the "ring and a half mode". This pattern is the closest to forming a complete circle in the lower bout. Most likely, however, the circle will not be completely closed. The objective is to shave the braces until the pattern closes into a ring. You most likely will not succeed with this but the closer you get to fully closed, the better.
[/QUOTE]

I can get this mode to be a very nice circle before I put my tonebars on. Once I add the tone bars the bracing pattern becomes non-symmetrical and the mode will open off an edge. I understand why this happens. I don't get too shook up about it and tend to leave the ring open over on the lower left.

[QUOTE=Mike Mahar]
Why would making a closed ring be better?
[/QUOTE]

Mike goes into the reasons why this may be true. It sure could be. I feel that the best way to achieve the nicest closed circle in the mode is by using symmetrical bracing. Al did a survey a few years ago where he built 2 guitars as close to the same as possible except one used a symmetrical bracing scheme vs. a "traditional" asymmetrical scheme. In his survey, if I remember correctly, 2/3 liked the symmetrical bracing vs. 1/3 liking the asymmetrical. Regardless, I go against the grain of Al's survey. I still use asymmetrical bracing. I believe asymmetrical bracing is not as efficient in producing the sound from the guitar, but I feel it was used on so many great guitars that what "we" have grown to believe a guitar should sound like this. I remember hearing some folks say the kasha style guitars (which I guess are very efficient) sounded “sterile.”

[QUOTE=Mike Mahar]
I can't emphesize strongly enough that why closing the ring makes for a better sounding guitar is still pretty much unknown. My idle speculations not withstanding. In fact, many builders claim that it doesn't help the sound of the instrument at all [/QUOTE]

I think Mike is correct in saying that closing the ring “helps the sound,” by making the guitar more efficient, but does helping the sound mean it sounds better? It depends on who is listening. I was listening to this ol’ blues guy talk about guitar sounds and he would hate my guitars. There are way too many overtones, too much sustain. He would prefer a $100 cheapy that goes “prank … prank.”
johno38899.7299074074