X brace vs. tone bars ?

[michaelcj]

I'm getting ready to start an A5 build.

I have done a search here for the topic [without any clear results] and am sure it has been discussed, so sorry if this is a repeat.

I would like to hear what differences, tone, "personality", projection, folks have experienced with these bracing systems in otherwise similar A5, or for that mater F5 style mandolins.

Built will be maple/sitka.

[Marty Jacobson]

You're right, there are not any clear results. There are too many factors to pin the tone of an instrument to the bracing pattern alone. Dr. Dave Cohen has done some of the best experiments with this, but many others have, too.
Bracing patterns are one of the many variables you can use to control the final tone of an instrument, but in the end it basically comes down to overall plate stiffness, mass, and longitudinal (neck to tail) and transverse (side to side) stiffness.
Your best bet is to copy an instrument that is known to be good as closely as you can, and it will be a good start.

[michaelcj]

Thanks Marty, I was able to find some of the old threads by googling Cohen [couldn't find them searching here].... Think you hit the nail on the head... not enough difference, all else being "equal" to deviate from the "norm".

Moderators can close/or delete the thread if you want.... I got the directions I wanted.

Thanks

Mike J

[Big Joe]

All else being equal, there is a noticeable difference from the two types of bracing. However, the size, shape, and placement of the bars will also affect the outcome. The material they are made from acan also affect outcome. It is not the most important issue in building, but if you wish to aim for a certain tonal output then bracing is certainly a consideration. Again, it is only one aspect of the whole, but it is still significant.

[Fretbear]

Listening is believing;
You can listen to older Youtubes and recordings of Ronny McCoury playing his X-braced Gil and then later his Loar, both good, but different.


http://www.youtube.com/watch?v=zStc3FjJ-T8

[fatt-dad]

the inside of my Cohen. . . I call them, "Wacky Braces." I know he uses some geometric formula to set the top and back arch. Is it like Gibson, et. al.? No clue. I'd guess the nature of the carved arches has some effect on the bracing - just like the material properties of the chosen wood.

I think beyond those other factors, whether they are parallel, X, or wacky is down the list of variables. Then again, I'm no builder. . .

106239

I do love this mandolin!

f-d

[Willie Poole]

Then again I hear that The Loar 700 doesn`t have any tone bars or braces and they are said to sound great, I don`t know because I have never seen or heard one myself....I have heard that tone bars were used to "tweak" the final sound of a mandolin and it must have been a lot of trial and error to get them just right...I guess that is why a lot of builders tap tune tops when they build until they get them like they want and then build them all to the same specs...Good luck with you trials, it will probably take a while to get what you like but some times people get lucky on the first go around...

Willie

[Eric Michael Pfeiffer]

I can share wih you my experience and personally what I think. After owning a Weber Yellowstone and a Summit F-100 I can tell you that the ound of "x" bracing isnt for me. They say x braced mandolins project the sound a bit more and create a more balanced sound across the ranges or registers. I cant say personally with regards to projecting the sound more....not really sure, but I did notice my mandolins seemed to not have stronger bass or stronger mid-range or highs but def kinda consistent in that regard. In other words, not one range seemed to stand out from the other....so mor balanced I'd say " yes"
However in my opinion, personally, they don't have that, what I call "monroe sound" or "blugrass kinda sound" that parallel ton bar bracing mandolins seem to have more consistently.
Not that x bar bracing is a bad thing, I just dont think it lends to a traditional bluegrass mandolin one if thats what one is after...like I was. I also felt that both the x braced mandolins I had, didn't have quite as much volume or bark...maybe its because the sound was being projected more and thrown away from the mandolin

[Dave Cohen]

Every account I have ever seen describing the "difference" between "tone bar" braced mandolins and X-braced mandolins alluded to comparisons of one "tone bar" braced mandolin to another X-braced mandolin. Needless to say, there are numerous variables involved in whatever the difference is between one mandolin and any other mandolin. The only way to make a valid comparison between the sonic characteristics of X-braces and of "tone bars" is to eliminate all of the variables except the braces themselves. It is not enough to make two mandolins which are "very similar" (whatever that means). So, what I did was make a single mandolin with a back plate attached to the ribs, lining, tailblock, and heelblock by about 60 small (#2x3/8") screws.

I did the experiments in 2004 in Rossing's lab at NIU. The mandolin started out with completely conventional "tone bars". I did holographic modal analysis, spectra, air resonance measurements, etc., on the mandolin with the "tone bar" bracing. Then I removed the back plate and installed a Virzi thingy made according to the GAL drawings. I repeated the experiments the next day with the Virzi thingy installed and the back plate re-attached, then removed the back plate again, carved out the braces carefully, and installed an X-brace made as similar as possible to Gilchrist's X-brace. I repeated the experiments the next day with the X-brace installed and the back plate re-attached. Then I removed the back plate, carefully carved out the X-brace, and installed my own "wacky braces" (Butch Baldassari was the first to call them that). Finally, I repeated the experiments with the wacky braces installed and the back re-attached the next day. I gave an invited paper on it at the Spring, 2005 ASA meeting in Vancouver, B.C. The results are summarized in my chapter in the 2010 Rossing book.

All of that stuff involved measurements which could be repeated. What I didn't say in either the talk or the book chapter was that I also played the mandolin and listened to it (to the best of my ability) with each iteration. And here is the part that I couldn't put in a peer-reviewed publication: The mandolin sounded substantially the same to me with every iteration. I really couldn't tell any difference from my auditions of each iteration. Iow, the mandolin sounded the same with each of the different bracing patterns. Now, there are obvious problems with my subjective observations. One is that human audial memory is not very good. There was a ~24-hr interval between each of the observations. So if there were nuanced differences between the iterations, it is certainly possible that I would have missed them. Still, the listening observations were consistent with the small differences in modal frequencies, & etc., observed in the physical measurements. Btw, the mandolin is a very good sounding mandolin. Fatt-dad, as well as others, wanted to buy it from me (Fatt-dad referred to that mandolin as "Opus I"). I didn't sell it because at the time, there was more physics yet to be done with it. It is currently on loan to the Allred Assoc.company.

http://www.Cohenmando.com

[michaelcj]

Thanks Dave for the detailed "from the horses mouth" reply. One of the things I love about this forum is the "sages" dive right in without talking down to those of us at "lower pay grades".

Although I find your 'whacky braces' really interesting [same with kasha's on guitars] I'm a bit of a traditionalist. I have built several guitars and a couple of "flat top mandos" this will be my first carved top. I can see now that there is no compelling reason [at this stage] to deviate from traditional. After all that's what we have you guys for!!

Mike J

[peter.coombe]

Every account I have ever seen describing the "difference" between "tone bar" braced mandolins and X-braced mandolins alluded to comparisons of one "tone bar" braced mandolin to another X-braced mandolin.

Yes, and often those "differences" are between mandolins made by different makers. Those sorts of comparisons should be confined to the bin. Dave's approach is valid and the results are certainly interesting. One other valid approach is to make a large number of mandolins, identical in every respect (including the maker), except half are tone bar braced, half are X braced. One can then look at averages and work out if there is a statistically significant difference. Unfortunately it is not very practical unless you have a big research grant.

[Dave Cohen]

Michael, there is absolutely nothing special about my "wacky braces". At the time of those experiments, I really hoped that there would be something different or (even) better about them. Alas, that was not to be. The results are a clear victory of empirical science over hopes and desires, etc. The only reason that I still use the wacky braces is that I have gotten used to working with them. Now, there is some utility in varying bracing patterns to accomodate differences in the ratio of longitudinal to lateral stiffness of your plate. I discussed that in a bit more detail in an answer I gave in the Questions column in American Lutherie #95 (Fall, 2008).

http://www.Cohenmando.com

[michaelcj]

Dave I've been a member since way back in the black and white days, and I've saved most of the issues.... I'll dig and have a look thanks again.

MikeJ

[CeeCee_C]

Yes, and often those "differences" are between mandolins made by different makers. Those sorts of comparisons should be confined to the bin. Dave's approach is valid and the results are certainly interesting. One other valid approach is to make a large number of mandolins, identical in every respect (including the maker), except half are tone bar braced, half are X braced. One can then look at averages and work out if there is a statistically significant difference. Unfortunately it is not very practical unless you have a big research grant.

I know that this is an old thread, but I found it to be very interesting. I wonder how many mandolins you'd have to build and test to have sufficient power to detect a difference.

[Dave Cohen]

The answer to that is based on the empirical nature of science. Scientific experiments have to be both verifiable and falsifiable. To be verifiable, someone else should be able to repeat them with the same methods and equipment and get the same results (within the limits of measurement). To be falsifiable does not mean that they should ultimately be found wrong (or false),but rather that if there is anything wrong with the results, someone will be able to do an experiment in the future that will show the (previous) experimental results to be wrong. In other words, nothing in science is ever proven. We can disprove, but never prove. There is, in principle, always the possibility that an experiment will come along and show our results to be wrong. That invalidation doesn't necessarily have to happen, but the possibility for invalidation has to be there.

What does all that have to do with your question? It says that in principle, you would need an infinite number of X-braced and tone-bar braced mandolins to be sure. For the practical reasons of time and money, we can't do an infinite number of experiments, so we settle on some practical compromise and do enough experiments to be able to gain some statistical confidence. If one can control all of the other variables and conditions well enough, a practical number of experients may be a small number, like say, three experiments. If the number of other variables is large, the necessary minimum number of experiments may be much larger. In the case of stringed instruments, the number of variables is large to the point of not even knowing what they all are. That means that to do the kind of experiment(S) which Peter described would take a very large number of trials and/or samples, and even then might not be definitive. That is why I chose the one reconfigurable instrument approach. Based on what I saw with that approach, my bet would be that the other approach would not show a measurable difference between X-braced and tone-bar braces mandolins.

http://www.Cohenmando.com

[CeeCee_C]

Good answer. I should have added "with an x% probability of detecting the difference.".

[Bernie Daniel]

I'm not completely sure what Dave means by measurable difference - but I expect he means a difference that can be shown via the deflection on some electronic instrument designed to measure sound quality/quantity?

That of course is different than whether some human being can (in a blind test) actually hear a difference between two instruments with a success significantly better than chance.

This very discussion has happened many time on this forum - leading mostly to the debaters agreeing to disagree (at best).

What makes it seem so futile to me is this. Suppose you did set up a perfect double blind test and you actually found a person who's hearing is so precise and certain that he could pick out a great mandolin from a mediocre mandolin with high probability. After all that what would you have?

One person's opinion and that's all.

[Bernie Daniel]

Edit -- the last portion of my post (#17) kind of went adrift -- it should have read:

What makes it seem so futile to me is this. Suppose you did set up a perfect double blind test and you actually found a person who's hearing is so accurate and precise that he could pick out difference between a tone bar braced mandolin from an X-braced one with some reasonable level of probability. Then that person could render an opinion on which sounds better. After all that what would you have?

One person's opinion and that's all.

My guess is most individuals would not be successful in picking out the bracing. So how many angels can dance on the head of a pin anyway?

[sunburst]

I'm not completely sure what Dave means by measurable difference - but I expect he means a difference that can be shown via the deflection on some electronic instrument designed to measure sound quality/quantity?

That of course is different than whether some human being can (in a blind test) actually hear a difference between two instruments with a success significantly better than chance.

Seems to me that a double blind test is a way of "measuring". If you show a statistically significant preference for one over another, or even a statistically significant detection of a difference, you've measured something.
...And I should add that if you don't show a statistical difference you've still measured something, though the measurement is near 0.

[Pete Jenner]

Is anyone aware of any carved top mandolins with no bracing? Sometimes I wonder how necessary bracing is given the arch.

[Bernie Daniel]

Seems to me that a double blind test is a way of "measuring". If you show a statistically significant preference for one over another, or even a statistically significant detection of a difference, you've measured something.
...And I should add that if you don't show a statistical difference you've still measured something, though the measurement is near 0.

Well you are correct if you have a statistically significant result you have measured the fact that one person can distinguish between the two mandolins -- yes for that individual you have indeed shown that there is some audible difference. So my earlier post was off the mark.

My point, which I did not make even after two tries was even if one (or a few individuals can detect a difference) it doesn't mean much if most others cannot detect it.

My belief is that in a properly conducted experiment perhaps only a very few, or even no one would be able to pick it up. But my opinion is no more valid then those who think it could be detected of course.

The one thing I have on my side in the debate would be that when these tests have been conducted in a properly controlled manner ---for the most part at least -- the participants fail to detect the better instrument -- e.g., Strads versus modern violins at more than chance.

[Bernie Daniel]

Is anyone aware of any carved top mandolins with no bracing? Sometimes I wonder how necessary bracing is given the arch.

There is one manufacturer who does sell both braced and unbraced instruments -- maybe "The Loar"?

This kind of supports the idea that the bracing is not for support but for tone.

OTOH if that is true tone bars or cross braces are not structural can anyone explain why the tops of so many Gibson F-2's and F-4's have sagged when the transverse brace fails?

[Pete Jenner]

There is one manufacturer who does sell both braced and unbraced instruments -- maybe "The Loar"?

This kind of supports the idea that the bracing is not for support but for tone.

OTOH if that is true tone bars or cross braces are not structural can anyone explain why the tops of so many Gibson F-2's and F-4's have sagged when the transverse brace fails?

I don't think any conclusions can be drawn or ideas supported by one example from one factory.

Surely, the sole function of braces and so called 'tone' bars is to increase stiffness. If that's the case then a slightly thicker top should produce the same result. It would be interesting to see how a thicker top with no bracing affects Q also.

[CeeCee_C]

So how many angels can dance on the head of a pin anyway?

I don't know, but the Sunis and the Shi'ites have been at war over this very issue for hundreds of years. So it must be important.

[mandobassman]

Surely, the sole function of braces and so called 'tone' bars is to increase stiffness. If that's the case then a slightly thicker top should produce the same result.

I remember reading a article written by Roger Siminoff, who knows about as much as anyone about Gibson Lloyd Loar mandolins. He stated with absolute certainty that the primary purpose of the tone bars is to shape the tone and not for support. The fact that they are there allows the top to be made thinner than would otherwise be the case, but their main function is tone.

[Bernie Daniel]

I don't think any conclusions can be drawn or ideas supported by one example from one factory.

Surely, the sole function of braces and so called 'tone' bars is to increase stiffness. If that's the case then a slightly thicker top should produce the same result. It would be interesting to see how a thicker top with no bracing affects Q also.

Peter the truth be I really don't have any experimental data to tell me what the role of tone bars or x-braces (i.e., just structural, just for tone, or both).

Nor have I even built a mandolin so everything I say is my opinion -- which is not worth too much.

When you say "Surely, the sole function of braces and so called 'tone' bars is to increase stiffness." I have no data to either refute or support that view (and I assume it is your opinion also?)

But if you read post #9 in this string it is obvious that Dave Cohen does have knowledge, and he has built mandolins, and he has also actually done experiments using different bracing on the same mandolin top. Now I submit that you can't get more relevant information to the question at hand than that.

One mandolin or not it is still real data and I don't see any other actual experiments refuting Dave's work.

Bottomline, I'm a scientist and I go with the data that's all. :)

[Bernie Daniel]

I don't know, but the Sunis and the Shi'ites have been at war over this very issue for hundreds of years. So it must be important.

Too bad they did not ask me -- I could have given them a SWAG! :)

[sunburst]

Braces are intended to (and do) add stiffness. The thing is, "tone bars", being roughly parallel and in line with the grain of the top add almost all of their stiffness in the fore and aft direction. Wood (especially softwoods like spruce) is considerably stiffer in that direction already, so tone bars add to the difference in stiffness lengthwise and crosswise the grain. X-braces cross more wood fibers of the top (because their arms are diagonal), so they can add stiffness in both directions. With that in mind, however, it is up to the luthier how stiff he/she leaves the braces after carving, and generally the top itself is by far the stiffer part of the system (as opposed to the braces contributing most of the stiffness), so contribution from the braces is secondary to the top itself. In other words, the stiffness added by the braces, whatever configuration, has only so much significance in the total stiffness of the top, and the stiffness of the top has only so much influence on the sound of the mandolin, so... how much difference in sound should we expect from bracing systems? To me, it seems that we can't expect very much difference attributable to brace configuration alone.
Also, as I often do, I look for consensus of opinion in situations like this. When there is debate and little consensus, that says to me that any difference is small if it exists at all, because an obvious difference would result in more consensus of opinion. (Sort of back door science...)

[Pete Jenner]

But if you read post #9 in this string it is obvious that Dave Cohen does have knowledge, and he has built mandolins, and he has also actually done experiments using different bracing on the same mandolin top. Now I submit that you can't get more relevant information to the question at hand than that.

One mandolin or not it is still real data and I don't see any other actual experiments refuting Dave's work.



Yes Bernie, I have seen and played Dave's mandolin with the removable plate/plates (can't remember if the top and back were removable or not.) and discussed some of these topics with him and John at Dave's place in Virginia.

[Mike Bunting]

Jesse McReynolds played a mandolin with no tone bars, but I can't remember the builder.

[Bernie Daniel]

Jesse McReynolds played a mandolin with no tone bars, but I can't remember the builder.

He also played a fiddle with a 20-penny nail for a sound post....although that might be OT.

[pfox14]

A master luthier once made the point to me that the least number of braces you can get away with, the better. Braces are needed for structural strength to prevent the top from caving in or distorting, but in theory a top with no braces would vibrate the most and probably sound the best. I've always just copied what has been time-tested to be the right number, size, and shaped braces for a given design.

[Dave Cohen]

On the other hand, the top plates of Greg Smallman's classical guitars owe all of their structure and vibratory properties to the CF/balsa lattice bracing. The relatively thin (ca 0.5 mm) top plates are really just membranes for pushing air. Do they 'vibrate less' for being predominantly bracing? Apparently not. They are generally acknowledged to be louder than conventional classical guitars. Meaning they push more air, and to do that, the plates' motions have to have greater amplitudes.

So what does it mean when someone says something 'vibrates more'? Faster/ That would mean higher frquency. Sometimes better, sometimes worse, but not ncessarily 'better'. Does it mean that more of the plate vibrates? If so, that is false. All of the normal modes of motion of plates are global, i.e., they occupy the entire area of the plate. That leaves greater amplitude.

A plate with no braces will not necessarily vibrate with greater amplitude. As Sunburst already pointed out, braces allow a given stiffness with less mass. Remove braces altogether, and you have to make a heavier plate to get the same overall stiffness. The UK physicist Bernard Richardson has shown that a plate with lower effective mass will vibrate with greater amplitude for a given driving force, in turn moving more air. And a plate with lower actual mass will have lower effective mass.

http://www.Cohenmando.com

[Pete Jenner]

Remove braces altogether, and you have to make a heavier plate to get the same overall stiffness.


Yes I thunk that too ;)
I am going to build a mandolin with no braces in the same batch (to be commenced in the next few weeks) as two other braced mandolins and a mandola. All I can hope is that it makes acceptable mandolin noises and has good volume. I will report back.

[Perry Babasin]

"The Loar" 700 and "The Loar" 400 have no tone-bars...