Theoretically I can raise Mode 5 by removing wood where?

[fscotte]

2 hours ago, Marty Kasprzyk said:

"Who are you going to believe, me or your own eyes?   Groucho Marx

Oliver Rogers used theoretical computer finite element analysis to show how the various vibration mode frequencies changed with wood removal in different areas of the top and back free plates. He found the plate edge areas are quite sensitive to wood removal.  

Three of his papers are in the CAT journals which are attached.

I found similar results in my plate thinning efforts.  But rather than believing others it might be more believable to do your own experiments.  Remove wood in a small area and see if there is a change in frequency for the various modes.  Try different areas and make a map of the results. Keep good records and enjoy the resulting arguments.

Newton's third law:  "For every physicist there is an equal and opposite physicist"

May 1990.pdf 23.7 MB · 0 downloads Nov1991 pdf.pdf 28.3 MB · 0 downloads May 1993pdf.pdf 19.59 MB · 0 downloads

Thanks for those studies.  I'll definitely look into those.  I have been keeping detailed records of my own, removing pieces of wood and receording Mode 2 and 5.

 

Other than removing wood from the edges, I've yet to see a Mode go up.

[Peter K-G]

Dampen the plate and warm it, then re-shape under pressure to get a higher arch.

Leave it to dry for a week.

You can rise M5 up to 30 Hz this way (M2 will go down). But the question is, why didn’t you make it higher in the first place 

 

[Anders Buen]

19 hours ago, Peter K-G said:

Dampen the plate and warm it, then re-shape under pressure to get a higher arch.

Leave it to dry for a week.

You can rise M5 up to 30 Hz this way (M2 will go down). But the question is, why didn’t you make it higher in the first place 

 

Interesting information! I have used some water on the inside of a varnished plate for rising the wood. The plate warped and the modes moved somewhat. 
Now I assume from own experiments with violins that the arch heights changes with the humidity conent and thus the relative humidity cycling with the climate. Have you seen a similar effect on the assembled bodies, e.g. before and after entering an UV box dryingg the wood?

I would suppose that the arches goes down when the wood dries in a violin body. Does this fit with makers experiences?

Edited February 14 by Anders Buen
Spelling corrections

[David Burgess]

2 hours ago, Anders Buen said:

I would suppose that the arches goes down when the wood dries in a violin body. Does this fit with makers experiences?

Yup.

 

2 hours ago, Anders Buen said:

Interesting information! I have used some water on the inside of a varnished plate for rising the wood. The plate warped and the modes moved somewhat. 

Yes, that happens. When wood originally dries, the outside dries in an expanded state, held that way by the wood inside which has not dried as much yet. Then the inside dries while held in that expanded state by the outside wood which has already dried.

Stress relieving by applying moisture to a finished or nearly-finished plate, and then letting it dry can move shapes around quite a bit.

Some makers try to minimize this by moisture-cycling their rough billets a lot before starting to work on them.

[Andreas Preuss]

3 hours ago, David Burgess said:

Some makers try to minimize this by moisture-cycling their rough billets a lot before starting to work on them.

Does this really help? 
 

Often I think we are trying too much to control the material. So I was rather working WITH the warping effect. I found it interesting that warping of the top plate only occurs when the plate is sufficiently thin. So I would partly use this effect to see if my plate is thin enough. Once it is warped it is stable so basically it is not a bad thing. Problematic is only that if you start with a flat underside it will end up to be deformed. From there I got the idea to calculate the underside surface of my top in the opposite direction so that warping would flatten it out. 

[Don Noon]

7 hours ago, David Burgess said:

Some makers try to minimize this by moisture-cycling their rough billets a lot before starting to work on them.

Or one seriously intense cycle.  Wood often warps during the torrefying process, but not after that.

[Evan Smith]

On 2/12/2024 at 8:38 AM, Don Noon said:

But I thought the idea was to make a good sounding violin, which I suspect is inversely related to manipulating free plate mode frequencies in bizarre manners. 

I thought that graduating a plate always manipulated the free plate mode frequencies, I do not understand the reference to it as being "in bizarre manners."

[Evan Smith]

On 2/11/2024 at 6:08 PM, fscotte said:

So the pencil marks is where tea leaves showed mode 5 at 347 hz,

So the theory is that if I remove wood on the outside of the modal lines, top and bottom, mode 5 will actually go up?

 

Having the tea puddled around the plate shows a bad speaker setup, misplacement of the plate supports or a low responsive plate. Ideally the tea should be blown out of the middle and two clean lines left. Not knowing which confuses the issues.

As far as raising frequencies by removing wood outside of the lines, in a very general sense it is theoretically true, but basically nonsense. There are too many variables to come to that Rule of thumb. Removing weigh at the end of a vibrating member will certainly raise the frequency, but since the vibrating member will be glued down and not allowed to vibrate, what was accomplished was to make that area more flexible by removing stiffness=wood =weight which allows the frequency to go up, instead of having any effect on the actual stiffness as is measured by the mode numbers.

I have seen mode 5 go up when removing wood from the middle of the top plate many times. The central area  between the corners, it only means that I still have more weight there than the relative amount of stiffness that is being used to maintain the frequency as it exists.

Plate modes are a general measurement used to determine whether a piece has adequate stiffness vs weight to build an instrument to a certain requirement as per your goal for "That" particular instrument. Nothing about plate mode numbers top or back determine the goodness of, or the desirability of any instrument.

So if the size of the instrument is X, and the edges are X, and the general thickness is X, with some experience, you learn that the mode 5 should be between about K and M if it is not, then you decide to start a new top or accept the fact that the instrument will be this, and not possibly that. But specific numbers are not magical, there are lots of variables.

Hope that made some sense

[Evan Smith]

On 2/12/2024 at 12:28 AM, Andreas Preuss said:

But you could also ask what is the purpose? I always recommend to make experiments on the assembled playable instrument in the white. If tap tones would be significant, wood removal in tap tone influencing areas on the assembled instrument should make clearly audible changes. It doesn’t. 

It does make audible changes in sound and in play ability if you remove wood from the right places at the right time. Just randomly removing wood just because you feel like it won't produce the results that you hoped for. The chances of hitting the right places are similar to playing the lottery. The areas to remove wood that will produce results are often 2 cm in diameter and no larger than 5cm and could be anywhere on the fiddle. It is an algorithm that is in constant flux, it changes everytime you mess with it. Some times you will hear someone say that removing wood from a certain place will brighten the G string, and at one point it will, but only when everything is in place for that to happen. At another time an entirely different place will accomplish the same thing. Quite some time can be spent between 2 spots then,, add a third and fourth up to dozens. It is an extremely complicated algorithm, and in many ways it almost appears quantum, but I'm certain that it all has some physical explanation, I sort of get it on an intuitive level, but I don't understand it. It's like speaking a language, you never question it, or think about it you just do it.

On 2/13/2024 at 9:40 AM, David Burgess said:

My current opinion is that "good" violin sound and playing properties are contingent upon getting stiffness-to-mass ratios "right" everywhere on a violin, and not just those affecting the lower plate modes.

This is exactly what I am referring to, this is what gives a refined sophisticated sound that is desirable. The famous "Italian Sound". When a violin has been perfected to that point, it is funny in that you can patch it, crack it and repair it, the arch can deform and be reshaped, and most of the time it will retain it's basic character. The violin still wants to behave as a whole, not just a bunch of parts trying to vibrate together but not quite there. There appears not a clue of understanding as to how violins even work.

[Don Noon]

4 hours ago, Evan Smith said:

I thought that graduating a plate always manipulated the free plate mode frequencies, I do not understand the reference to it as being "in bizarre manners."

Gradual thinning of a plate to maintain a graduation pattern is one thing, but local digging out of wood from certain areas just to obtain a specific free plate mode frequency is something else… and that is the part that to me is bizarre.

[fscotte]

11 hours ago, Evan Smith said:

Having the tea puddled around the plate shows a bad speaker setup, misplacement of the plate supports or a low responsive plate. Ideally the tea should be blown out of the middle and two clean lines left. Not knowing which confuses the issues.

As far as raising frequencies by removing wood outside of the lines, in a very general sense it is theoretically true, but basically nonsense. There are too many variables to come to that Rule of thumb. Removing weigh at the end of a vibrating member will certainly raise the frequency, but since the vibrating member will be glued down and not allowed to vibrate, what was accomplished was to make that area more flexible by removing stiffness=wood =weight which allows the frequency to go up, instead of having any effect on the actual stiffness as is measured by the mode numbers.

I have seen mode 5 go up when removing wood from the middle of the top plate many times. The central area  between the corners, it only means that I still have more weight there than the relative amount of stiffness that is being used to maintain the frequency as it exists.

Plate modes are a general measurement used to determine whether a piece has adequate stiffness vs weight to build an instrument to a certain requirement as per your goal for "That" particular instrument. Nothing about plate mode numbers top or back determine the goodness of, or the desirability of any instrument.

So if the size of the instrument is X, and the edges are X, and the general thickness is X, with some experience, you learn that the mode 5 should be between about K and M if it is not, then you decide to start a new top or accept the fact that the instrument will be this, and not possibly that. But specific numbers are not magical, there are lots of variables.

Hope that made some sense

Correct on all accounts!  Bad speaker setup is true.  I try not to worry about the definition of the chladni lines, I just need to know the what I'm hearing is actually Mode 5, so double check with crappy speaker setup. 

And the idea of raising Mode 5, yup, I have always been suspect of this.  That's why I posted this question.  I've yet to see clear evidence of this happening. I'd like to see a video of some removing wood "just outside" of those Mode 5 lines and watch the Mode 5 freq go up.

Your post goes hand in hand with the other thread referring to the "musical" part of a violin.  The edges of a plate are not "musical" just as you state here, they don't vibrate so much when attached to the rim.

[Marty Kasprzyk]

12 hours ago, Don Noon said:

Gradual thinning of a plate to maintain a graduation pattern is one thing, but local digging out of wood from certain areas just to obtain a specific free plate mode frequency is something else… and that is the part that to me is bizarre.

If you look at Jeffrey Loen's book "Thickness Graduation Maps" you will see many highly uneven graduation thicknesses of violins made by the famous old makers.  But his book shows modern violins are often much more uniform which to me is bizarre.

[Marty Kasprzyk]

4 hours ago, fscotte said:

>

Your post goes hand in hand with the other thread referring to the "musical" part of a violin.  The edges of a plate are not "musical" just as you state here, they don't vibrate so much when attached to the rim.

The antinode of the violin plates become smaller and smaller as the vibration frequency increases. These anitnodes have alternating +or - phases and the sound waves comming off of them often cancel in the interior area of the plate's surface. But at the plate edges some of the out of phase antinodes don't exist and the remaining ones aren't canceled and are able to produce some sound.

The attached drawing of a flat plate illustrates this effect. The pictures from the attached report show the actual sound pattern coming off of a violin top at various frequencies.  Notice that the relative amount of comming off of the edges becomes greater and greater as the frequency is increased.  

So the edges may not effect the sound of the lower range of a violin much but they do at the higher end.

macine and violim sond fields.pdf

[David Burgess]

3 hours ago, Marty Kasprzyk said:

If you look at Jeffrey Loen's book "Thickness Graduation Maps" you will see many highly uneven graduation thicknesses of violins made by the famous old makers.  But his book shows modern violins are often much more uniform which to me is bizarre.

We modern makers do it partly because we can, without too much extra time and effort.
The hope is that by doing so, we can target sound and playing properties better than the makers of old, who produced sound and playing properties which ranged all over the map, from horrible to divine.

[LCF]

7 hours ago, Marty Kasprzyk said:

The antinode of the violin plates become smaller and smaller as the vibration frequency increases. ...

...The pictures from the attached report show the actual sound pattern coming off of a violin top at various frequencies.  Notice that the relative amount of comming off of the edges becomes greater and greater as the frequency is increased.  

 

Thanks Marty Kasprzyk!

 

Interesting to see that for those notes illustrated there isn't very much happening outside of the ff holes except where the whole garland is moving up and down in phase.

 

That was a long report to wade through (phew!) but when I finally drilled down to fig 6 and fig 7 of the bridge motion at various points in paper D --- it is quite interesting to contemplate. 

[fscotte]

Yes very interesting to see little movement outside of the f-holes.  I wonder if removing large amounts of wood here would reduce weight without affecting stiffness.

[LCF]

16 minutes ago, fscotte said:

Yes very interesting to see little movement outside of the f-holes.  I wonder if removing large amounts of wood here would reduce weight without affecting stiffness.

That's only for that one bowed note and its harmonic so  hard to generalise from that.

 

It's a strong area of the garland and strongly conserved like that over time  for some reason(s). I think that altho the area outside the ff is not directly connected to the island it is still strongly connected to the back plate and it bridges the upper and lower bouts of the top.  

I would try the added weight before I ever tried to thin, or you could bang a few accelerometers on there. 

[Dr. Mark]

There seems to be agreement with other work, which is good.  We can compare e.g. Gren et al's Figure 9 at 285 Hz (reproduced by Mr. Kasprczyk above) with the calculated A0 mode at 294.3 Hz in Figure 5(a) on page 6 of http://tinyurl.com/3zc37wv5 to obtain (or lose!) confidence in ostensible similar data.  I think I can hear Mr. Zuger in the far, fardistance ('read this dissertaaatioooonnnnnnn.......')

12 hours ago, Marty Kasprzyk said:

The antinode of the violin plates become smaller and smaller as the vibration frequency increases. These antinodes have alternating +or - phases and the sound waves coming off of them often cancel in the interior area of the plate's surface. But at the plate edges some of the out of phase antinodes don't exist and the remaining ones aren't canceled and are able to produce some sound

Now my neck hurts again.  I must be losing my language acuity....

[Don Noon]

5 hours ago, fscotte said:

Yes very interesting to see little movement outside of the f-holes.  I wonder if removing large amounts of wood here would reduce weight without affecting stiffness.

4 hours ago, LCF said:

That's only for that one bowed note and its harmonic so  hard to generalise from that.

The figures show mode shapes for A0 and then two octaves above that.  Skipped over are the signature modes of CBR, B1-, and B1+, where there is significant movement of the C-bout ribs.

12 hours ago, Marty Kasprzyk said:

The antinode of the violin plates become smaller and smaller as the vibration frequency increases. These anitnodes have alternating +or - phases and the sound waves comming off of them often cancel in the interior area of the plate's surface. But at the plate edges some of the out of phase antinodes don't exist and the remaining ones aren't canceled and are able to produce some sound.

I would argue that the arching of a violin increases the area of the antinodes toward the middle of the plate, and attachment to the ribs takes the edges of the plates out of play.

Experimentally, I am totally convinced that the upper bout of the top is where the higher frequencies are most efficiently radiated, and whatever theoretical things you want to propose have to deal with that.

[Anders Buen]

18 hours ago, Marty Kasprzyk said:

The antinode of the violin plates become smaller and smaller as the vibration frequency increases. These anitnodes have alternating +or - phases and the sound waves comming off of them often cancel in the interior area of the plate's surface. But at the plate edges some of the out of phase antinodes don't exist and the remaining ones aren't canceled and are able to produce some sound.

The attached drawing of a flat plate illustrates this effect. The pictures from the attached report show the actual sound pattern coming off of a violin top at various frequencies.  Notice that the relative amount of comming off of the edges becomes greater and greater as the frequency is increased.  

So the edges may not effect the sound of the lower range of a violin much but they do at the higher end.

macine and violim sond fields.pdf 2.77 MB · 11 downloads

Sound radiation from a "baffled" window or a wall, are like that. Most of the mid and high frequency energy comes from the edges. Around the coincidence frequency the plate starts to radiate more from "all over". That is when the bending wave speed becomes similar to the speed of sound in air. At lower frequencies the air particles just speeds past the plate vibration and short circuits the sound radiation. Above the coincidence (it is a rather wide frequency region, as it depends on the angle we look at the plate), the bending wave speed "catches up" with the airborne part and, eventually, passes faster. Then more sound radiation is produced for a given vibration amplitude.

A violin box is kind of a "self baffled" 3D plate. 

I think the maximum radiation direction out of a plate can be calculated from the sine to the soundspeed in air over the bending wave speed in the plate. It behaves much like the refraction of light in glass - air or water air contact which also can be predicted by the speed of light in the two media.

I think the cross grain sound-speed and bending wave speed is the most important one for a violin. But that is still a hypothesis.

Edited February 17 by Anders Buen
Spelling corrections and some added info

[LCF]

1 hour ago, Anders Buen said:

Sound radiation from a "baffled" window or a wall, are like that. Most of the mid and high frequency energy comes from the edges. Around the coincidence frequency the plate starts to radiate more from "all over". That is when the bending wave speed becomes similar to the speed of sound in air. At lower frequencies the air particles just speeds past the plate vibration and short circuits the sound radiation. Above the coincidence (it is a rather wide frequency region, as it depends on the angle we look at the plate), the bending wave speed "catches up" with the airborne part and, eventually, passes faster. Then more sound radiation is produced for a given vibration amplitude.

A violin box is kind of a "self baffled" 3D plate. 

I think the maximum radiation direction out of a plate can be calculated from the sine to the soundspeed in air over the bending wave speed in the plate. It behaves much like the refraction of light in glass - air or water air contact which also can be predicted by the speed of light in the two media.

I think the cross grain sound-speed and bending wave speed is the most important one for a violin. But that is still a hypothesis.

There are some good animations and discussion of what you mention here:

"Evanescent Waves (sound radiation from a plate)"

https://www.acs.psu.edu/drussell/demos/evanescentwaves/evanescentwaves.html#:~:text=Flexural bending waves with higher,of sound in the fluid.

[LCF]

Of course one aspect of standing waves in any medium is that they are comprised of travelling waves in contrary directions which add and subtract in such a way as to give a stationary pattern. And something else which usually sneaks by without much attention is that the antinodes we're mostly  discussing, the so-called 'musical zones'(!?!) are velocity antinodes but the velocity nodes are bending antinodes. There's always a two phase system in a standing wave between potential energy (eg bending) and kinetic energy (tea leaf bouncing). If nothing appears to be happening vis a vis the leaves there is nonetheless just as much energy in the plate, in the bending zones. Those bending antinodes are not generally the flexural boundaries either. You have to look at the geometric features of a plate ie edge conditions, load points, drive points, cutouts, to get a handle on the various boundaries which the travelling waves are subject to. 

 

 

Edited February 17 by LCF
Nodes/antinodes!

[LCF]

7 hours ago, Don Noon said:

... Experimentally, I am totally convinced that the upper bout of the top is where the higher frequencies are most efficiently radiated, and whatever theoretical things you want to propose have to deal with that.

If we made a crazy fiddle where we could swap endpin and neck from one block to another do you think this outcome would follow whichever end the neck is attached to, no matter how wide the bout is??

[Don Noon]

58 minutes ago, LCF said:

If we made a crazy fiddle where we could swap endpin and neck from one block to another do you think this outcome would follow whichever end the neck is attached to, no matter how wide the bout is??

As long as the bridge location and F holes are kept in the same relation to the body geometry, and somehow you kept the scale length, I think it would sound the same except perhaps for some of the lower modes, where the neck and endblock have significant movement.

The F's and tight arching in the waist and upper bout I think play a large role in determining the higher frequency mode shapes, radiating efficiency, and therefore sound,

[Marty Kasprzyk]

4 hours ago, Anders Buen said:

>

I think the cross grain sound-speed and bending wave speed is the most important one for a violin. But that is still a hypothesis.

I very much agree. 

I use 0.8 mm plywood for my plates and I can independently vary their longitudinal and cross grain stiffnesses (therefore their speed of sounds) with interior braces in various combination of directions.  Attached is a photo of viola No. 44 plate with three major cross braces.

The next viola, No 45 plate (attached photo) was made without these cross braces and the resonance frequencies in the FRC curve were lowered considerably which made the viola much deeper sounding than No. 44.

Three 13mm high cross braces were then added to the plates's outside surface of No 45 (like a giant Sam Z. gluey experiment) and the FRC curve was now similar to No. 44 as was expected.

These outside cross braces were then cut down in steps to 7, 3, 2, 1mm and a photo is attached showing the final 1mm step.  FRC and playing tests were done after each step.  The A0 decreased 221 down to 196 Hz as seen in the attached plot.

With low or no cross braces the viola's sound seemed overly deep and "muted" which players didn't like.    I attribute this to a big output in the low range 130-400Hz with little midrange ~500-1000Hz output. 

So cross grain stiffness can have big effect in experiments with actual instruments and not just in theory.