Theoretical question about increasing cross stiffness on a violin top plate (with f holes) by adding bars.

[Andreas Preuss]

If I want to increase cross stiffness on a loose violin top with the help of a wooden reinforcement MOST EFFICIENTLY on which lines on the surface would this be?

This question also implies that I am not necessarily thinking of straight lines. So to illustrate it better I made a couple of drawings of patterns which I considered.

So in practice this means adding minimum mass in form of bars to achieve the maximum increase in stiffness. In an experimental setup I would glue the crossbars trim their height to match a preset weight and measure the tap tone of the x mode before and after. 

 

So which of the patterns would most efficiently increase the cross stiffness of the loose plate?

1. one bar across the center above the f hole: 

 

2. two bars connecting the corner blocks 

3. two bars across the upper and lower bout at their widest distance

4. following the nodal lines of mode 5

[David Burgess]

In which vibrational modes would you most like stiffness to increase?

I've run across more situations in which less cross stiffness was an improvement, than I have where increased cross stiffness was an improvement. What is the outcome that you are trying to accomplish?

[Ronald Moretto]

I come from guitar building, where we use bracing to control stiffness and shape tonal response. In violins, though, as we know, cross-grain stiffness is mainly governed by arching rather than internal supports. That said, if your goal is to increase cross stiffness with minimal added mass, placement is crucial.

A single bar across the center might stiffen the plate, but it could also overly restrict movement, potentially killing responsiveness. A more efficient approach might be a curved or diagonal reinforcement that follows the natural stress distribution of the plate (following the arching flow?). This would allow for added stiffness without overly dampening resonance.

I’d be curious to see how different patterns shift the X-mode frequency. 

[Marty Kasprzyk]

1 hour ago, Andreas Preuss said:

If I want to increase cross stiffness on a loose violin top with the help of a wooden reinforcement MOST EFFICIENTLY on which lines on the surface would this be?

This question also implies that I am not necessarily thinking of straight lines. So to illustrate it better I made a couple of drawings of patterns which I considered.

So in practice this means adding minimum mass in form of bars to achieve the maximum increase in stiffness. In an experimental setup I would glue the crossbars trim their height to match a preset weight and measure the tap tone of the x mode before and after. 

 

So which of the patterns would most efficiently increase the cross stiffness of the loose plate?

1. one bar across the center above the f hole: 

 

2. two bars connecting the corner blocks 

3. two bars across the upper and lower bout at their widest distance

4. following the nodal lines of mode 5

I use all of them.

[Andreas Preuss]

57 minutes ago, David Burgess said:

In which vibrational modes would you most like stiffness to increase?

I don’t quite understand what you mean with vibrational modes. Supposedly this is for the assembled instrument. Right now I just need information on the loose plate.

 

59 minutes ago, David Burgess said:

What is the outcome that you are trying to accomplish?

As usual I am dealing with problems which are outside classical violin making. I am making top plates which are in a traditional sense by far too thin making a reinforcement grid necessary. I simply noticed consistently that adding cross stiffness on such a top plate adds a desirable soprano character. 

[Andreas Preuss]

1 minute ago, Marty Kasprzyk said:

I use all of them.

Your work was a great inspiration. 
 

What I am trying to figure out is the absolute minimum of a reinforcement grid.

[Davide Sora]

It looks like guitar makers have evolved to lattice bracing (a sort of grid). Best of luck with the fit of all that braces on a curved board.

http://www.stevenwalterguitars.com/uploads/5/9/7/1/59716739/6671244_orig.jpg

[Marty Kasprzyk]

58 minutes ago, Davide Sora said:

It looks like guitar makers have evolved to lattice bracing (a sort of grid). Best of luck with the fit of all that braces on a curved board.

http://www.stevenwalterguitars.com/uploads/5/9/7/1/59716739/6671244_orig.jpg

 I'm an archenemy so I make my plates flat.  This makes it easy to fit all the braces. 

[Marty Kasprzyk]

1 hour ago, Andreas Preuss said:

Your work was a great inspiration. 
 

What I am trying to figure out is the absolute minimum of a reinforcement grid.

You have it backwards.  You should figure out the absolute minimum of the plate's weight.  The function of the plate's surface is to just move air which doesn't require much strength or stiffness at all.  Paper, like used in speaker cones, might work fine if you couldn't accidently poke a finger through it. 

On the other hand, the reinforcement grid has to be stiff enough to support the downward string force.

The goal is to achieve the minimum total weight of the plate plus the stiffening grid to maximize sound output while still having a desirable frequency response.

 

[Don Noon]

The question lacks the details necessary for a coherent engineering answer.  Generally, for efficient bending stiffness you want tall and thin beams.  For something that sounds like a violin, the answer is going to be completely different and not based much on basic engineering principles.

[Andreas Preuss]

7 hours ago, Davide Sora said:

It looks like guitar makers have evolved to lattice bracing (a sort of grid). Best of luck with the fit of all that braces on a curved board.

http://www.stevenwalterguitars.com/uploads/5/9/7/1/59716739/6671244_orig.jpg

It’s naturally more time consuming than fitting a normal bass bar and looked at the beginning quiet complicated. Now I developed a method which is reasonably fast. 

[Andreas Preuss]

5 hours ago, Marty Kasprzyk said:

The goal is to achieve the minimum total weight of the plate plus the stiffening grid to maximize sound output while still having a desirable frequency response.

That’s what I am working on. 
 

I have just seen in the response curves that augmenting the number of bars starts to worsen the result and it seems that the location of grid bars makes the difference between a better or worse response curve. So my assumption is that if I place bars where they can work with highest efficiency will be best to achieve an optimal grid pattern. Right now the weight of my plates is down to under 60g with the grid. I am not really trying to minimise the weight of the grid.

[Don Noon]

11 hours ago, Marty Kasprzyk said:

The goal is to achieve the minimum total weight of the plate plus the stiffening grid to maximize sound output while still having a desirable frequency response.

5 hours ago, Andreas Preuss said:

Right now the weight of my plates is down to under 60g with the grid. I am not really trying to minimise the weight of the grid.

Sound output and frequency response are important for sure, but so are the dynamics and controllability.  Minimizing weight and maximizing sound is lowering input impedance, which means that the player has to provide the energy for that increased sound... it doesn't come for free.  There are also issues with playing softly, and wolf notes.

Just in the last couple if weeks, I have been futzing around with 2 of my shop fiddles, both of them have tops that weigh 53 grams with the bass bar and varnish.  They are functional for me, in fiddle jam sessions, and they can put out a lot sound.  But due to the issues mentioned, they are not what a violinist would want.  I have tried grid bracing years ago, and that was much worse (there are other issues with a grid).

In my experience, 60 g with the bass bar (without varnish) seems to be the practical limit for a violin, and even that is limiting its use to a player wanting the dB output and is OK with the other characteristics.  I find that a few more grams is a good compromise for a wider range of players.

 

[Andreas Preuss]

Let’s start where I can give best answers: 

5 hours ago, Don Noon said:

Just in the last couple if weeks, I have been futzing around with 2 of my shop fiddles, both of them have tops that weigh 53 grams with the bass bar and varnish.  They are functional for me, in fiddle jam sessions, and they can put out a lot sound.  But due to the issues mentioned, they are not what a violinist would want. 

I can’t deliver any proofs in physics for what I am saying but I got by now a (completely?) different view on violins. If you want to go down with the weight of the top alone it doesn’t really work with or without bracing. Here I totally agree with you .
 

But I am convinced that you can lower the top weight under the conditions that

1.there is a sturdier less elastic rib construction

and

2. the rib garland forms a lengthwise arch in the top side. 
 

5 hours ago, Don Noon said:

In my experience, 60 g with the bass bar (without varnish) seems to be the practical limit for a violin, and even that is limiting its use to a player wanting the dB output and is OK with the other characteristics.  I find that a few more grams is a good compromise for a wider range of players.

On classical built violins a top plate of 60g with bass bar is dammed light!
(I was usually happy to end up with 65g with the bass bar (without varnish))

Somehow I am just trying to figure out the same ideal minimum weight for my type of construction. I am convinced that the different rib construction paired with an ‘intelligent’ bracing allows to go below 60g (with the bracing). I am not really hunting for the lowest figures but take the weight of the different parts into consideration. 
useless to say that if I have to sacrifice plus 5g I am going to do it.

5 hours ago, Don Noon said:

Sound output and frequency response are important for sure, but so are the dynamics and controllability.  Minimizing weight and maximizing sound is lowering input impedance, which means that the player has to provide the energy for that increased sound... it doesn't come for free.  There are also issues with playing softly, and wolf notes.

I can’t really give a convincing answer on this other than: so far I didn’t have problems with wolf notes, and from my perspective of playing violin the dynamics between soft and loud were good to control. The ultimate test with professionals will come up soon. Stay tuned. 
 

PS: I’d be interested to know how you constructed your grid and first of all how thick the bars of your grid were. I am using spruce from bass bar material (with upright year rings) and cut them in thin strips of 1.8mm. Secondly I try to avoid any grid pattern that squares up with the year rings direction of the top. I have no other reasons for doing so than pure instinct. 

[christian bayon]

Yes I did it on all my instruments during 3 years, but not glued, carved.

[Don Noon]

1 hour ago, Andreas Preuss said:

PS: I’d be interested to know how you constructed your grid and first of all how thick the bars of your grid were.

I don't know.  That was close to 50 years ago.

For my lightweight fiddles, my "grid" is solid... i.e. the top is low-density torrified spruce, normal construction and bar.  Free plate taptones are sky-high, but assembled the mode frequencies come out at the normal places.  I just don't think you can get away from the different playing character of ultra light weight.  Assuming you want normal playing character.

[Mike_Danielson]

Just some thought:  if you want to increase the stiffness, then increase the arch height.  Mechanical engineers learn this when looking at plates and what happens when you arch the plate.

Lattice bracing is now the thing in classical guitars.  I think lattice bracing strongly inhibits the teeter-totter motion of the bridge which causes a lose of acoustic power because the pressure waves are out of phase with each other on each side of the top center-line (the pressure waves leak across the bridge and cancel each other out).  There was a French guitar maker in the 1950s, Bouchet, that built guitars with a large brace that went under the bridge (almost the entire width of the top) that would also inhibit this motion.  His guitars are well thought of.  Perhaps his design should be brought back.

[Don Noon]

36 minutes ago, Mike_Danielson said:

Just some thought:  if you want to increase the stiffness, then increase the arch height.  Mechanical engineers learn this when looking at plates and what happens when you arch the plate.

For spruce, the crossgrain stiffness actually drops for an arched (carved) vs. flat plate due to the extreme weakness of the off-quarter direction.  

[Marty Kasprzyk]

4 minutes ago, Don Noon said:

For spruce, the crossgrain stiffness actually drops for an arched (carved) vs. flat plate due to the extreme weakness of the off-quarter direction.  

Flat plates with high braces avoids this off-quarter problem.

[Marty Kasprzyk]

4 hours ago, Andreas Preuss said:

Let’s start where I can give best answers: 

I can’t deliver any proofs in physics for what I am saying but I got by now a (completely?) different view on violins. If you want to go down with the weight of the top alone it doesn’t really work with or without bracing. Here I totally agree with you .
 

But I am convinced that you can lower the top weight under the conditions that

1.there is a sturdier less elastic rib construction

and

2. the rib garland forms a lengthwise arch in the top side. 
 

On classical built violins a top plate of 60g with bass bar is dammed light!
(I was usually happy to end up with 65g with the bass bar (without varnish))

Somehow I am just trying to figure out the same ideal minimum weight for my type of construction. I am convinced that the different rib construction paired with an ‘intelligent’ bracing allows to go below 60g (with the bracing). I am not really hunting for the lowest figures but take the weight of the different parts into consideration. 
useless to say that if I have to sacrifice plus 5g I am going to do it.

I can’t really give a convincing answer on this other than: so far I didn’t have problems with wolf notes, and from my perspective of playing violin the dynamics between soft and loud were good to control. The ultimate test with professionals will come up soon. Stay tuned. 
 

PS: I’d be interested to know how you constructed your grid and first of all how thick the bars of your grid were. I am using spruce from bass bar material (with upright year rings) and cut them in thin strips of 1.8mm. Secondly I try to avoid any grid pattern that squares up with the year rings direction of the top. I have no other reasons for doing so than pure instinct. 

Back in 2003 George Bissinger wrote a paper "Relating Normal Mode Properties of Vioolins to Overall Quality: Part II: Modal Averages/Trends" see https://stacks.stanford.edu/file/nw043ph6421/CAS_nw043ph6421.pdf  to see the entire paper.

He found that "good" violins had better sound radiation efficiencies than "bad" violins which is shown in his below attached Figure 4.  

He also found that the radiation efficiency of violins in the lower frequency ranges up to about 2000Hz in general was  quite low as shown in his Figure 3 which is also attached below.

Bissinger's measure of sound production efficiency as the ratio of the violin's actual sound output divided by the theoretical amount of sound produced by a rigid baffled piston of the same surface area.  His description of this efficiency measurement is also attached.

So it be concluded that a traditional violin with an arched top plate is a poor good sound producer compared to a rigid piston.

Therefore if you wanted to make a much better projecting (higher sound producing with the same bowing effort) violin it should have a top plate  built like a baffled rigid piston. But it should also be light weight.  There is nothing new here-Bissinger's efficiency definition is described in his no. 2 reference by Fahy in 1987.

One way of increasing the plate's rigidity while minimizing its weight is to use a thin plate with very high ribs as I had shown in my earlier photograph of my most recent violin plate which is again attached.

[Don Noon]

1 hour ago, Marty Kasprzyk said:

So it be concluded that a traditional violin with an arched top plate is a poor good sound producer compared to a rigid piston.

Therefore if you wanted to make a much better projecting (higher sound producing with the same bowing effort) violin it should have a top plate  built like a baffled rigid piston.

It not be concluded, the way I see it.

If you're only interested in the lower frequencies, yes... the woofer speaker model applies, and getting a large area moving uniformly is ideal. 

But even in the chart you show, the difference between good and bad violins doesn't show up strongly until you get to 2 kHz and above, where woofers are crap.  Even a midrange speaker runs out of steam around 2 kHz.  We're asking a violin to cover a huge frequency range, and it does that by going into a variety of modes with multiple antinodes which a rigid piston won't do.  Ignore those higher modes and you get a "bad" violin.

[LCF]

5 hours ago, Mike_Danielson said:

Just some thought:  if you want to increase the stiffness, then increase the arch height.  Mechanical engineers learn this when looking at plates and what happens when you arch the plate.

Lattice bracing is now the thing in classical guitars.  I think lattice bracing strongly inhibits the teeter-totter motion of the bridge which causes a lose of acoustic power because the pressure waves are out of phase with each other on each side of the top center-line (the pressure waves leak across the bridge and cancel each other out).  There was a French guitar maker in the 1950s, Bouchet, that built guitars with a large brace that went under the bridge (almost the entire width of the top) that would also inhibit this motion.  His guitars are well thought of.  Perhaps his design should be brought back.

Greg Smallman's lattice braced guitars are (can be) very loud. Very thin top, lightweight lattice, strong ribs.  Principle action of a guitar bridge is pumping up and down, drum-like. 

 

 

[Marty Kasprzyk]

4 hours ago, LCF said:

Greg Smallman's lattice braced guitars are (can be) very loud. Very thin top, lightweight lattice, strong ribs.  Principle action of a guitar bridge is pumping up and down, drum-like. 

 

 

Thanks, Here's another interview which shows Smallman's lattice bracing.

 

[Andreas Preuss]

7 hours ago, Marty Kasprzyk said:

thin plate with very high ribs

I suppose you meant high grid reinforcement beams.

9 hours ago, Don Noon said:

For spruce, the crossgrain stiffness actually drops for an arched (carved) vs. flat plate due to the extreme weakness of the off-quarter direction.  

One of the reason why my plates are bent. 

6 hours ago, Don Noon said:

and it does that by going into a variety of modes with multiple antinodes which a rigid piston won't do.

From my experimental results I was concluding that the sum of cross bar stiffness on my reinforcement grid audibly changed the soprano quality. But also that the position of cross bars can have eventually negative side effects notably in the overall loudness. This was my motivation to look isolated on the placements of cross bars. 

While alterations on the reinforcement grid couldn’t alter the basic frequency pattern above 2kHz it has a massive influence on the average frequency amplitudes below 1kHzwhich seems to confirm Bissingers findings. (We should also keep in mind that 1975hz corresponds to B6 which is way up on the E string) 

To me the soprano quality (or what is typically expected from violinists) comes from an ‘ideal’ balance of the sum of low frequencies below 1 kHz and the sum of very high frequencies between 2kHz and 4kHz where on a graph the average of low frequencies should be slightly below the average of high frequencies. If reversed the violin sounds too dark too diffuse and has a low range in dynamics. 
 

I was able to trigger the amplitudes between 2 kHz and 3kHz with the height of the long arch of the top plate but couldn’t figure out any means to control/increase frequencies between 3kHz and 4kHz. Sometimes the overall adjustment would bring a little more or less response but unfortunately not consistently and understandably enough to have a receipe in hands. 
 

Overall, the ‘piston mechanism’ seems to have only negligible influence on the high frequencies when using my grid structure. My current top plate was bent from wood not precisely cut in split direction and I am very curious to hear what result the next top shaved down to the fiber direction will bring. Ideally it would boost the frequencies between 3kHz and 4kHz. 

 

[Andreas Preuss]

13 hours ago, Mike_Danielson said:

Just some thought:  if you want to increase the stiffness, then increase the arch height.  Mechanical engineers learn this when looking at plates and what happens when you arch the plate.

Lattice bracing is now the thing in classical guitars.  I think lattice bracing strongly inhibits the teeter-totter motion of the bridge which causes a lose of acoustic power because the pressure waves are out of phase with each other on each side of the top center-line (the pressure waves leak across the bridge and cancel each other out).  There was a French guitar maker in the 1950s, Bouchet, that built guitars with a large brace that went under the bridge (almost the entire width of the top) that would also inhibit this motion.  His guitars are well thought of.  Perhaps his design should be brought back.

I have taken a few things from guitar making as an inspiration to develop new ideas in violin making.