Light weight fiddle

[violins88]

Thanks guys for your comments. I can use all the help I can get in my quest.

John Schmidt: my unfinished back plate currently weighs 82 grams and has an M5 of 382 hertz. This results in a kback of 1.2. Did I do the math right? My unfinished top plate currently weighs 79 grams and has an M5 of 370 hertz. I am measuring M5 using Audacity software and the peaks are sharp and repeatable to within a few hertz. My back wood has an SG of .37 and the top has an SG of .35. The arching for both plates is very high and the radii are as smooth as I can get them. I am mindful of Strobel's graduation suggestions.

Don Noon: I stopped carving the top a few weeks ago in order to work on the back and see where that took me, but the back has been stubborn. I am proceeding carefully, trying to keep everything in balance. I am keeping my data in a spreadsheet and will share it when I finish. So far, my results with light tops have also disappointed but that could change.

I am continuing to measure the harmonics with my bowing apparatus and have been applying the technique to my back plate. Same trend, the harmonic content of a bowed string fades as the plate is thinned. I have read everything I can find about string harmonics but can't find any references about what affects their relative amplitudes. Was that the class I dropped in school? This is really fascinating.

The top is much too heavy and the frequency is too high. The back weight is very, very low. I will be very surprised if this violin works out. But I would like to see it finished.

[Don Noon]

The back weight is low for a normal violin, but who knows what's going to work with .37 density back wood. 

The top, not finished, naturally is heavy with a high taptone... but the taptone looks on the low side for that weight and .35 density of wood.  Perhaps the stiffness is on the low side.

With the choice of woods and ultra-light focus, I would be surprised if this violin sounds "normal".  But it might hang together, be playable, and have an acceptable voice for fiddling.

[violins88]

The back weight is low for a normal violin, but who knows what's going to work with .37 density back wood. 

The top, not finished, naturally is heavy with a high taptone... but the taptone looks on the low side for that weight and .35 density of wood.  Perhaps the stiffness is on the low side.

With the choice of woods and ultra-light focus, I would be surprised if this violin sounds "normal".  But it might hang together, be playable, and have an acceptable voice for fiddling.

Don,

I agree that it might have an acceptable voice for fiddling. I hope we get to hear about that eventuality. Recently a good violin maker told me that someone will buy a "too thin" fiddle. People like different things.

[Roger Frankland]

Well, I am rapidly approaching that category of those who tried ...

Top plate ( SG=.35) is now 68 grams and M5=334 hertz.

My goal was 60 grams and 350 hertz and I didn't even come close. I still have some clean up to do around the edges so the numbers will fall some more. It is fascinating to me that it doesn't seem to matter where I remove wood at this point, the tap tone drops. The grain on this piece of spruce seems pretty open to me. It ranges from 10 to 20 per inch. Might that account for the low tap tone (stiffness?)?

Bottom plate (SG=.37) is now 78 grams and M5=372 hertz.

I have read that the back plate should be within a semitone of the front plate which means I need to take it down some more. Is that right? I have become rather expert at removing backs, so if it doesn't work ....

[Don Noon]

My goal was 60 grams and 350 hertz and I didn't even come close

 

You can have a goal of mass, or a goal of taptone, but the wood will tell you what combination is attainable.  Arch height has some influence, but it's mostly the wood.  Here is your progress (in bright green) plotted against a bunch of mine (these are all without the bass bar).  The black diamonds are Cremonese datapoints.  The low-lier is the only one of mine I plotted that has unprocessed wood, and the arching was quite low. 

 

I personally don't put too much stock in taptone ratios, especially since you're using abnormal wood combinations.  You could probably go to 60 g on your top, but I'd quit thinning of M5 got much below 320 Hz.  This is from someone who tends to build relatively light.

[Roger Frankland]

You can have a goal of mass, or a goal of taptone, but the wood will tell you what combination is attainable.  Arch height has some influence, but it's mostly the wood.  Here is your progress (in bright green) plotted against a bunch of mine (these are all without the bass bar).  The black diamonds are Cremonese datapoints.  The low-lier is the only one of mine I plotted that has unprocessed wood, and the arching was quite low. 

taptone vs mass.jpg

 

I personally don't put too much stock in taptone ratios, especially since you're using abnormal wood combinations.  You could probably go to 60 g on your top, but I'd quit thinning of M5 got much below 320 Hz.  This is from someone who tends to build relatively light.

That is really interesting, thank you. And good advice. I'll keep forging ahead.

[Roger Frankland]

You can have a goal of mass, or a goal of taptone, but the wood will tell you what combination is attainable.  Arch height has some influence, but it's mostly the wood.  Here is your progress (in bright green) plotted against a bunch of mine (these are all without the bass bar).  The black diamonds are Cremonese datapoints.  The low-lier is the only one of mine I plotted that has unprocessed wood, and the arching was quite low. 

taptone vs mass.jpg

 

I personally don't put too much stock in taptone ratios, especially since you're using abnormal wood combinations.  You could probably go to 60 g on your top, but I'd quit thinning of M5 got much below 320 Hz.  This is from someone who tends to build relatively light.

I have another data point on my top plate. Weight 64 grams, M5=331 hertz. I'm not satisfied, but it is time to stop. Regarding graduations, Strobel says the center should be 4.5 mm thick. Others say less. My limited experience suggests that Strobel is right. What do you guys do? This sure is complicated.

[curious1]

I have another data point on my top plate. Weight 64 grams, M5=331 hertz. I'm not satisfied, but it is time to stop. Regarding graduations, Strobel says the center should be 4.5 mm thick. Others say less. My limited experience suggests that Strobel is right. What do you guys do? This sure is complicated.

The top should be 4.5mm thick? That is much too thick.

2.5-3.5 is normal.

[Don Noon]

I checked my Strobel book, and he gives 4.5 as the upper limit for the center of the back.  For the top, he gives a nominal 3mm for the center.  Even the Cannone isn't 4.5 thick on the top (more like 3.5).

 

Here's the updated chart showing your path in green:

You appear to be just at the higher end of the Cremonese zone, and heavier than anything I've built (except for my first one, which isn't on the chart).

It seems as if you are backing off of your original goal of going for super-duper light and everything else being secondary.  However, my guess is that you could still go quite a bit lighter on the top without horrible consequences.  If you were building for a soloist with the right arm of a gorilla, maybe I'd stop thinning now.  If this is for a casual fiddler, thinner should be fine.

 

[Roger Frankland]

I checked my Strobel book, and he gives 4.5 as the upper limit for the center of the back. For the top, he gives a nominal 3mm for the center. Even the Cannone isn't 4.5 thick on the top (more like 3.5).

Here's the updated chart showing your path in green:

taptone vs mass 2.jpg

You appear to be just at the higher end of the Cremonese zone, and heavier than anything I've built (except for my first one, which isn't on the chart).

It seems as if you are backing off of your original goal of going for super-duper light and everything else being secondary. However, my guess is that you could still go quite a bit lighter on the top without horrible consequences. If you were building for a soloist with the right arm of a gorilla, maybe I'd stop thinning now. If this is for a casual fiddler, thinner should be fine.

I checked my Strobel book, and he gives 4.5 as the upper limit for the center of the back. For the top, he gives a nominal 3mm for the center. Even the Cannone isn't 4.5 thick on the top (more like 3.5).

Here's the updated chart showing your path in green:

taptone vs mass 2.jpg

You appear to be just at the higher end of the Cremonese zone, and heavier than anything I've built (except for my first one, which isn't on the chart).

It seems as if you are backing off of your original goal of going for super-duper light and everything else being secondary. However, my guess is that you could still go quite a bit lighter on the top without horrible consequences. If you were building for a soloist with the right arm of a gorilla, maybe I'd stop thinning now. If this is for a casual fiddler, thinner should be fine.

Well, you guys come through again. I find Strobel a little vague so I really appreciate your knowledge. I am still OK because the center of my top is currently 3.4mm. And, so far, the top is ringing nicely so that is a help. I am not dropping my goal of super light. Still reaching for 300 grams overall.  But yesterday, I was listening to an expert violinist playing scales. That sound is just so incredible.

[Roger Frankland]

here are my final data points for my top plate - log 83 with SG of .35.  This is before cutting f holes and adding bass bar.

 

thin outside perimeter:

weight = 61 grams

M5 tap tone = 331 hertz

 

thin bouts to 2.4 and 2.5 mm respectively:

weight = 58.5 grams

M5 tap tone = 314 hertz.

 

my instinct is the sound might have been a little more balanced at 61 grams

[Don Noon]

oops... I thought your numbers were with the F-holes cut when I plotted them (all mine are with the F-holes).  Big difference, but still might not be too bad.

[Roger Frankland]

oops... I thought your numbers were with the F-holes cut when I plotted them (all mine are with the F-holes).  Big difference, but still might not be too bad.

Well that clinches it. Next step: cut F-holes and remeasure.

[Roger Frankland]

Well cutting the f hole dropped me to 261 hertz which is a disappointment and perhaps the consequence of trying to go light with minimally aged, untreated wood. It is beginning to seem impossible. I am encouraged that the plate still sings nicely in my bowing test however! My plan is to attempt to tune the plate back to 350 hertz via the bass bar, but that may be overly ambitious, even with a high bar. This is going to take a few days and some experimentation. Any advice would be appreciated, as I think I am at a crucial junction. It is also important to me is to achieve a fiddle with good color and balance In the tone at least through the first position. My experience with my previous fiddles like this one is that they improved after a year, so there is another consideration. But the immediate problem still is where to tune the bass bar to.

[Roger Frankland]

Well, here's the results after barring the top plate.

Prior to cutting f holes: 59 grams. M5 = 314 hertz

Cut f holes: 58 grams. M5 = 261 hertz

Add high bar: 65 grams.. M5 = 353 hertz

M5 was more difficult to measure after barring ...theres a lot more going on.

Although not the normal assembly sequence, I am inclined to attach the sides next and watch what happens. I've been tracking M5 since the wood was billets. To me, this is all quite fascinating although maybe not so profound.

[curious1]

Well, here's the results after barring the top plate.

Prior to cutting f holes: 59 grams. M5 = 314 hertz

Cut f holes: 58 grams. M5 = 261 hertz

Add high bar: 65 grams.. M5 = 353 hertz

M5 was more difficult to measure after barring ...theres a lot more going on.

Although not the normal assembly sequence, I am inclined to attach the sides next and watch what happens. I've been tracking M5 since the wood was billets. To me, this is all quite fascinating although maybe not so profound.

How tall is your bass bar? And the general shape?

[Don Noon]

7g for a bass bar seems like way too much.  I have generally found that trying to compensate for a floppy plate by using a very stiff bar doesn't work.  Maybe just slightly more than normal... ~4.5 - 5g is the most I'd use.  But height and shape are important, too.

[curious1]

7g for a bass bar seems like way too much. I have generally found that trying to compensate for a floppy plate by using a very stiff bar doesn't work. Maybe just slightly more than normal... ~4.5 - 5g is the most I'd use. But height and shape are important, too.

That was my thought too. He's trying too hard. I don't think I'd go much above 16mm (with top) for highest point of bass bar and much above 5gms.

[Roger Frankland]

Thanks for the comments, guys.

Well. I cut the bar to the length and profile suggested by Strobel, thicknessed it to 5mm, and cut the height plus plus with the intention of trimming it down as I tuned M5. So it is high, ranging from 5 to 16 mm. But then my first measurement came in at 353 hertz, which was my original target. So I stopped, looking for your helpful comments. Sounds like I should start by lowering the tails and thinning it. 5 grams would make me happier, 4 grams and I'd be almost ecstatic. My data base is growing.

Perhaps thinking about using the bar to tune M5 is a mistake. What I have read is that the bar was originally added to prevent violins from breaking due to the introduction of steel strings. Actually, I think a real value of the bar is to help a violin hold it's tuning... Ie not deflect.

I was just reading Curtin's paper that you suggested in your forum, Don. Interesting. The Kreutzer Stradivari top plate (unbarred) apparently weighs 55 grams and rings at 276 hertz. That's pretty close to where I am at! Hmmm, my plate stiffness can't be all that bad, but I do understand that this is a game of tenths. That last tenth I removed from the bouts probably did it to me. On the other hand, my dentist told me that tenths is more precise than he works to!

[Don Noon]

I trimmed a bass bar to obtain a specific taptone on the free plate only one time... on my first violin.  I had to trim it far beyond what I thought looked right, in order to get 360Hz, and the result was poor.  I had to put in a more normal shaped one, and it was much better.  I do keep records of the taptones (just in case, and it's easy), but that's as far as it goes for me.

 

Roger, is that 16mm max height excluding the top thickness?  If so, that would put the total dimension (which I use) at 19mm or so, which I think would be excessive.  16mm including the top would be the maximum I'd try.  I think any more is just dead weight.

[Roger Frankland]

I was measuring just the bar. Thanks for the info.

[curious1]

I was measuring just the bar. Thanks for the info.

The reason you are having difficulty reading M5 is because the bar is too stiff in relation to the plate. I would take it to 16mm including the thickness of the top (14mm + top), as both Don and I have suggested. This is about the maximum.

[Roger Frankland]

Just an update. I just trimmed my bass bar to 14 mm and 4.6 grams, because I am actually pretty good at following suggestions. Following is my data. I am including deflection numbers as I measure that too.... 20 pound load applied to center of the plate via an apparatus including a dial readout in mils.

Finished plate prior to cutting F-holes:

Wt 58.5 grams

M5 314 hertz

Deflection 92 mils

Plate after cutting f holes

Wt 57.5 grams

M5 261 hertz

Deflection 130 mils

Plate after attaching oversized bar

Wt 64.5 grams

M5 353 hertz

Deflection 75 mils

Plate after trimming bar

Wt 62.1 grams

M5 332 hertz

Deflection 80 mils

My prior experience is this deflection will go to almost zero when the sides are attached.

[Smith]

 .

[Roger Frankland]

Its probably to late to say this, but don't go below the generally accepted thicknesses to meet your goal of a light weight fiddle. There is an obsession with light weight instruments lately, perhaps always has been, but it is better to make an instrument more stout, and perhaps miss a customer or two, than make a light and thin one, only to have a customer bring it back a year later complaining of cracks, a sinking top, and wolf notes. One of my teachers used to say "a light instrument is like a light camera-it just feels cheap!"

Probably good advice and I should heed it, but my sense is that so many violins are unresponsive because they were over designed to prevent returns. Also there are also so many unsubstantiated "facts" about how various components of a violin affect the final result....

I am beginning to understand that there is an acceptable weight region for the top plate dependent upon wood and arching ... thanks to the wisdom of members of this forum. You are correct that If I create an exceptional violin I won't be able to sell it due to uncertainties regarding its durability. But I don't care. I just finished a great article in the latest New Yorker magazine about fusion reaction and ITER. Now there is an almost insurmountable challenge with a dubious outlook. But they soldier on.