Wood treatments by Stradivari and Guarneri

[Michael_Molnar]

On 7/1/2021 at 11:57 PM, Bruce Tai said:

Unfortunately we still don't have publicly available CT data that tell us about the weight, density, thickness, and weight at the same time. It would be important to know if the very thin Strad plates were built wit denser wood. More importantly, CT cannot reveal chemical composition. So we don't know if Stradivari also adjusted thickness according to wood treatment method. 

Do you have a bullet list of those wood treatment methods?

[David Beard]

2 hours ago, Evan Smith said:

I have noticed from experience that thinning between the upper eyes leads to rapid arch rise and neck dropping as the stresses in the arch equalize themselves. There is a certain sound advantage to this but I would rather have longevity than instant gratification. Early on I would end up thinning there to get what I wanted out of it. In knowing what I'm looking for now, I've been able to accomplish it in other ways that allows for a much stronger box.

I've never seen this sudden arch rise and neck drop at all on a DG type arch with slightly thicker grads in the center.

That's interesting.  Thanks for sharing those observations.   

For me, so far I make that area on the thin side of normal.  Definitely not thick.  I want good ability to twist there.

[Bruce Tai]

On 7/4/2021 at 1:30 AM, Michael_Molnar said:

Do you have a bullet list of those wood treatment methods?

In a nutshell, the relevant ones are:

Minor additives (tens of ppm): copper sulfate, iron sulfate

Major additives (thousands of ppm):

1. Stradivari: table salt, potash

2. Del Gesu: alum, lime
 

We don't know the order and application conditions.

We don't know about washing or potential baking/boiling steps. 

[Bruce Tai]

On 7/3/2021 at 8:19 AM, Marty Kasprzyk said:

The most complete set of data I've seen came from:

 M.A. Pyrkosz, "Reverse Engineering the Structural and Acoustic Behavior of a Stradivari Violin", Dissertation, Michigan Technological University, 2013. http://digitalcommons.mtu.edu/etds/634

Pyrkosz took the CT scans of the "Titian" 1715 Strad to show in his table 4.23 that the top plate had a density of 0.35 g/cc with a weight of 48.1g and the back plate had a 0.570g/cc density and a weight of 88.9g.  The plate thickness maps are shown in figures 4.69 and 4.70. 

With modal analysis of the various violin mode frequencies he was able to estimate the top and back wood's elastic modulus.

I think it would have been easier to cut up the violin into wood strips and test the pieces. 

Thanks a lot. I have read this thesis before but missed the data. 

The top is 48.1 g with bass bar with wood density of 0.35. So the average thickness may be 0.24-0.25 mm. 

That's a very thin top made of low density spruce. And yet it survived well and sounds great. 

And yet Joseph Curtin thought a 54 g top without bass bar (for a Strad) is already very light. 

 

[Bruce Tai]

Sorry, website glitch. Double post. 

 

[Bruce Tai]

Sorry, website glitch. Multiple post. 

 

[Bruce Tai]

Sorry, website glitch, multiple post. 

 

[Evan Smith]

You ok Bruce, ya keep repeating yourself

[Bruce Tai]

Sorry, website glitch. Multiple posts. 

 

[David Burgess]

5 minutes ago, Bruce Tai said:

Thanks a lot. I have read this thesis before but missed the data. 

The top is 48.1 g with bass bar with wood density of 0.35. So the average thickness may be 0.24-0.25 mm. 

That's a very thin top made of low density spruce. And yet it survived well and sounds great. 

And yet Joseph Curtin thought a 54 g top without bass bar (for a Strad) is already very light. 

 

I think your conclusion about an average thickness of 0.24-0.25 mm might be a little off.

[Bruce Tai]

So David, what is the area of the violin plate? Is it 500 square centimeter? I am trying to calculate average thickness based on the area. But the area number is difficult to find on the net. 

What is the average thickness of the Titian top? 

[Philip Perret]

1 hour ago, David Burgess said:

I think your conclusion about an average thickness of 0.24-0.25 mm might be a little off.

Indeed.  Case of a misplaced decimal point.  And from practical experience I have a very hard time believing the weight of the top with a bar being 48.1 grams.  

[HoGo]

4 hours ago, Bruce Tai said:

Major additives (thousands of ppm):

1. Stradivari: table salt, potash

2. Del Gesu: alum, lime

Doesn't potash and alum generally desctribe the same thing? Potassium aluminium sulfate?

[Don Noon]

4 hours ago, Bruce Tai said:

So David, what is the area of the violin plate? Is it 500 square centimeter? I am trying to calculate average thickness based on the area. But the area number is difficult to find on the net. 

What is the average thickness of the Titian top? 

3 hours ago, Philip Perret said:

Indeed.  Case of a misplaced decimal point.  And from practical experience I have a very hard time believing the weight of the top with a bar being 48.1 grams.  

At first glance, the 48.1g weight seems hard to believe, and that was my initial reaction.

However:

1) My 352 back length model (closest to Titian)  area is 545 cm^2.  The Titian grads go as low as 2.2mm; if you figure 2.5 mm average thickness including the edge, and .35g/cc then you get 47.7 grams. (the 545 cm^2 value is a 2D outline... includes a button, with no subtraction for F-holes, and no addition for arching.)

2) My "snakefiddle" is 355 long x 208 lower bout, a bit bigger than the Titian, and currently has a top that just happens to weigh 48.1g including varnish and internal casein coating (but no bar).  It's also slightly thicker than the Titian, probably ~2.6 mm or slightly over on average, but also lower density wood, at .31.

So 48.1g for the Titian top seems perfectly reasonable, WITHOUT the bass bar.  The Titian bass bar is relatively bulky, and I would guess it's well over 4g.  It comes down to details of varnish, exact area of the plate, graduations, etc. which could push the calculation a few grams either way.  But it's pretty close.

 

[Don Noon]

Yikes... double post!

[David Beard]

6 hours ago, Bruce Tai said:

In a nutshell, the relevant ones are:

Minor additives (tens of ppm): copper sulfate, iron sulfate

Major additives (thousands of ppm):

1. Stradivari: table salt, potash

2. Del Gesu: alum, lime
 

We don't know the order and application conditions.

We don't know about washing or potential baking/boiling steps. 

So, those sulfates seem like good candidates for the wood preservatives you always suspected.

But the others seem like deliberate parts of the maker's finishing work.

Didn't you say that Stradivari's use of salt and lye was less consistent?

[Rothwein]

Would they tend to darken the wood over time?

[Philip Perret]

2 hours ago, Don Noon said:

At first glance, the 48.1g weight seems hard to believe, and that was my initial reaction.

However:

1) My 352 back length model (closest to Titian)  area is 545 cm^2.  The Titian grads go as low as 2.2mm; if you figure 2.5 mm average thickness including the edge, and .35g/cc then you get 47.7 grams. (the 545 cm^2 value is a 2D outline... includes a button, with no subtraction for F-holes, and no addition for arching.)

2) My "snakefiddle" is 355 long x 208 lower bout, a bit bigger than the Titian, and currently has a top that just happens to weigh 48.1g including varnish and internal casein coating (but no bar).  It's also slightly thicker than the Titian, probably ~2.6 mm or slightly over on average, but also lower density wood, at .31.

So 48.1g for the Titian top seems perfectly reasonable, WITHOUT the bass bar.  The Titian bass bar is relatively bulky, and I would guess it's well over 4g.  It comes down to details of varnish, exact area of the plate, graduations, etc. which could push the calculation a few grams either way.  But it's pretty close.

 

Thanks for your explanation.  

[Bruce Tai]

2 hours ago, Rothwein said:

Would they tend to darken the wood over time?

As you can see from our published photos that have been color-calibrated, the answer is no. 

[Bruce Tai]

14 hours ago, Don Noon said:

At first glance, the 48.1g weight seems hard to believe, and that was my initial reaction.

However:

1) My 352 back length model (closest to Titian)  area is 545 cm^2.  The Titian grads go as low as 2.2mm; if you figure 2.5 mm average thickness including the edge, and .35g/cc then you get 47.7 grams. (the 545 cm^2 value is a 2D outline... includes a button, with no subtraction for F-holes, and no addition for arching.)

2) My "snakefiddle" is 355 long x 208 lower bout, a bit bigger than the Titian, and currently has a top that just happens to weigh 48.1g including varnish and internal casein coating (but no bar).  It's also slightly thicker than the Titian, probably ~2.6 mm or slightly over on average, but also lower density wood, at .31.

So 48.1g for the Titian top seems perfectly reasonable, WITHOUT the bass bar.  The Titian bass bar is relatively bulky, and I would guess it's well over 4g.  It comes down to details of varnish, exact area of the plate, graduations, etc. which could push the calculation a few grams either way.  But it's pretty close.

 

 

Assuming 550 cm^2 (includes arching), 2.3 mm thick, 0.35 density, that's 44.3 g. Bass bar 4 g -> 48.3 g.

Very close to the Titian measurements. 

Am I more correct this time?     

 

[Strad O Various Jr.]

no, you mean 0.23cm thick not mm

[Don Noon]

2.3 mm average thickness is too low, since 2.2 is the minimum in some places, and there are some areas up to 3mm, as well as the block areas and edges, which would be thicker still.  Just eyeballing the graduation map, I think 2.5mm might be a good average.  Get the Titian poster from TheStrad.

But still, the point is that rough calculations get you in the right ballpark, and there are several sources of uncertainty that can move the estimate around by a few grams... but still within the ballpark.  It's silly to sharpen your pencil when the numbers you start with are fuzzy.

[Andreas Preuss]

9 hours ago, Philip Perret said:

Indeed.  Case of a misplaced decimal point.  And from practical experience I have a very hard time believing the weight of the top with a bar being 48.1 grams.  

Since the plate weight was calculated from whatever scans (was it CT?) there is certainly some error possible.

However I know from a reliable source that a contemporary maker with a reputation for good sound arrived at 45g (w/o bb) as the lightest plate in his production. Apparently this is possible knowing how to balance material properties and arching in order to thin down the top plate to the extreme.

[David Beard]

I thought I'd read the Titian has spots that are only 1.9mm??

[Advocatus Diaboli]

The titian has a back weight of 98.5ish grams, and a top weight of 66.6 grams with a bar according to some old notes of mine.  I'm not sure how anyone calculated the CT weight at 48.1??