Positioning of tone bars inside soundboard of an A5 or an F5

[Nick Gellie]

I came across a nice photo on Northfield mandolins Facebook page which showed where they positioned the tonebars on an F5. I did not copy it across to respect their copyright.

Where do you put tone bars inside the soundboard of an A-5 or an F-5 relative to the F-holes and why? Does it affect the tone on the instrument as much as the shaping or graduation of the top?

[sunburst]

I put the bars pretty close to the position of the bars in the original F5 mandolins. Why? Because when I started building mandolins, I assumed that they had to be there to be correctly placed. Now that I know that exact positioning is not important, I put them there because I always have.
Overall stiffness and mass are the most important aspects of the top that affect the sound. Top thickness, arching, graduation, bracing and other factors figure into that in various interrelated ways, so there is no good way to consider the effect of "shaping and graduation" separate from tone bar placement, or any factor separate from any other. It is all related.

[HoGo]

I came across a nice photo on Northfield mandolins Facebook page which showed where they positioned the tonebars on an F5. I did not copy it across to respect their copyright.

Where do you put tone bars inside the soundboard of an A-5 or an F-5 relative to the F-holes and why? Does it affect the tone on the instrument as much as the shaping or graduation of the top?

What they do is typical placement as seen on vintage (20's) Gibson F-5 mandolins. There is pic on their site showing their mandolins with well worn copy of my F-5 drawings. You don't have to worry about copyright.
I use that same placement just because I always did, just like John. No one really knows if there is audible difference in this asymmetrical placement and something more symmetrical.
There is inherent problem with this placement that it creates large area of unsupported arch between bridge and tailpiece right on the axis of string tension. In case the arch is weak (insufficient thickness or weak shape of arch) this area will bulge and recurve will sink under tailpiece. Eventually, if the top is really too weak, the whole top may collapse (top joint opens wide and recurve can break across grain). Placing tonebars closer may help keep the weak arch from bulging (I just did it on repaired collapsed top on Schoffield F-5 that lasted 40 years till it collapsed).
I once built F-5 out of extremely soft piece of Engelmann and put the bars closer together and right under bridge adjustment wheels and crossed them some 2" in front of bridge to support the arch (of course with graduations to help as well). Without that, I believe, the top would collapse within decade or so... (but could probably sound fabulous in the meantime)

Building mandolins is always about getting as close as possible to the "line of failure" but without crossing it. Many notable instruments show such deformations and their graduations show that they may be just "on the line" (Grismans "Crusher", Loar A-5), time will tell the rest.

[fscotte]

From a deflection perspective, tone bars merely replace the stiffness that is lost when you cut out the f-holes. Thus one reason why the back plate doesn't need tone bars. Some mandolins like the Loar brand don't even have bars.

Like when you cut out a hole in your floor for stairs, or a hole in the wall for windows, now you gotta build headers and reinforce the opening, otherwise it will sag and eventually fail. Tone bars, they do the same thing. I don't believe tone bars have anything to do with the tone, or shaping the tone, or any other magical thing. They should be called stiffness bars.

[HoGo]

From a deflection perspective, tone bars merely replace the stiffness that is lost when you cut out the f-holes. Thus one reason why the back plate doesn't need tone bars. Some mandolins like the Loar brand don't even have bars.

Like when you cut out a hole in your floor for stairs, or a hole in the wall for windows, now you gotta build headers and reinforce the opening, otherwise it will sag and eventually fail. Tone bars, they do the same thing. I don't believe tone bars have anything to do with the tone, or shaping the tone, or any other magical thing. They should be called stiffness bars.

I think you contradicted yourself in the first and third sentences. If you build it strong enough you don't need anything... TheLoars are built thicker on top. I played Fender A some 15 years ago that had no tonebars but the top was 1/4" thick near soundholes, just the very edges were thinned so it looked normal from outside.
Guitar backs don't need braces from structural reasons but they all have them.
The best approach to this topic I have seen so far is based on how the whole structure is loaded. There is completely different load on top and back plate. When you tension the strings the top is compressed along grain while back is mostly in bending/tension (there is no soundpost so it has to hold quite a load of bending - on violins it is pretty much just tension on the back) along grain. This load on top together with bridge force create outward forces in areas between bridge and tailpiece and bridge and neck block. Since the area under fingerboard extension is quite thick and straight and tonebars are close this area shows minimal to no deformation - I've seen mandolins that had thin grads under extension and deep recurve they showed some sinking. On the other side of bridge the round arch of thin wood will show some tendency to bulge immediately after stringing and placement of tonebars will affect this tendency. SInce the whole mandolin is being bent by tension of strings, the body gets wider on top at the perimeter and this lifts outer edges of f holes slightly. Typical back is quite thick in the half close to neck so it bends mostly in the half closer to tailpiece. The recurve cen get shallower near tailpiece and arch slightly more pronounced next to it.
The aim is to biuld as light as you dare. Think how stiff is your wood and how much it can handle at given thickness.
When I build mandolin I want it to stand all these forces without breaking apart and get good tone... if you build it clost to Loar specs it WILL sound close to Loar, you can adjust the tone by leaving this thicker and remove wood elsewhere etc. so the whole thing is still in balance and will hold together.

[fscotte]

I don't see a contradiction, I see the Loar brand as doing their own thing, perhaps making the top arch and top thickness a little different so cutting the f-holes doesn't reduce stiffness to a dangerous point. I think we would both agree that if a Loar brand mandolin had identical top thickness and identical arch and identical wood to Grisman's Crusher, then the Loar brand mandolin would have probably broken apart soon after it was built, or maybe ten years later... who knows.

But if you're going to build a top that is thin enough and deflects enough, you will need tone bars. I pointed out that from a deflection standpoint, the tonebars when properly shaped make up for the loss of stiffness when you cut two big f-holes into the sides.

I've seen this myself, testing the plates before f-holes, after f-holes, and then when tonebars are added. The deflections are similar before f-holes are cut, and then after tonebars are shaped to proper Lloar size.

So as example, I shape the plate to near proper thickness before f-holes are cut and tone bars are added. I put the plate on my deflection jig, get a reading of between .025" and .030". I then cut the f-holes out and put the plate on the jig, I get .035" - .045". I then add tonebars and shape them down to proper Lloar size and walah.. I'm back to the target of .025" to .030"

What else am I to conclude from this? If you don't need f-holes, then you don't need tonebars.. But ya gotta have a hole somewhere..

[siminoff]

The specific asymmetrical positioning of tone bars on the original F5 mandolins has to do with tuning the soundboard. As tone bars are moved closer to the center of the soundboard, the frequency of that portion (bout) of the soundboard increases (gets higher). As tone bars are moved closer to the outer portion of the soundboard, the frequency of that portion of the soundboard decreases (gets lower). The treble tone bar is closer to the center of the soundboard to raise the pitch of the treble side or bout of the soundboard. The bass bar is closer to the outside of the soundboard (closer to the bass f-hole) to lower the pitch of the bass side of the soundboard. The tone bars are not there for strength but instead serve as stiffness (tuning) adjusters. As wood is removed from a tone bar, the pitch or frequency of that portion of the soundboard is decreased. As wood is removed from the center of the tone bars, the pitch changes rapidly. As wood is removed from the ends of the tone bars, the pitch changes slowly. As a result the two tone bars can be tuned (tap tuned) differently, by removing wood until a certain pitch is achieved. An arched and graduated soundboard like that of the F5 mandolin does not need tone bars for strength - it will withstand the 44 pounds of string load* of the strings just fine without tone bars. …R

*A set of J74 strings at a 16° string break angle exerts a downward load on the soundboard of 44 pounds.

[Nick Gellie]

I forgot say thanks to all those that contributed to this thread. I am much better informed as a result.

[Dave Cohen]

As in another recent thread, I have attached some interferometry work I did in 2004 on a 1924 Gibson F5. These are the lowest frequency normal modes of a mandolin top plate. I have done these on numerous different mandolins, both vintage, my own, and those of other makers. They all look more or less like this. In different mandolins, the main differences are modest differences in peak frequencies. You can see that they are all overall symmetrical, they occupy the entire plate, and the nodal lines are in principle infinitesimally narrow. Each of the modes is characteristic of the entire plate at their respective frequencies. These were deliberately excited with a single frequency. The applied frequency to the driving coil was scanned until a resonance appeared, then the image was acquired. When the plate is excited by a string, there are many frequency components in the excitation. The result is that most or all of the modes are excited. The resultant motion is what is called a fourier superposition of all of the individual modes. It is an algebraic sum of all of the amplitudes at each point, and is overall symmetrical. There is no such thing as a frequency for one side of the plate or the other. A given normal mode will be excited the same regardless of where on the plate the excitation is located, and will have the same frequency. The only exception is if the excitation happens to be on a nodal line for a particular mode, in which case the mode will be weakly excited, or possibly not excited.

These modes are for an assembled instrument. For a free plate, which is what Siminoff is describing, the modes are different, but the principles are the same. You can see a discussion, complete with some chladni patterns, for rectangular free plates in the Fletcher & Rossing text, "The Physics of Musical Instruments" (2nd Ed.) in chapter 3, pp 81-3. When Siminoff bonks his free plates, he is exciting many modes at once. He thinks he is doing something different when he bonks different sides of the plate. It doesn't really matter where on the plate it is bonked. What is actually happening is that he has to hold the plate differently when he bonks the different "tone bars". If he hears different frequencies, it is likely because in one of the bonks, he is holding the plate at a node for one of the modes, and holding the plate at an antinode for that particular mode in the other bonk. So, he is not exciting completely different motions. The motions are always the same, regardless of where on the plate it is bonked. He is just accidentally extinguishing or greatly attenuating one or a few of the modes by holding the plate at their antinode(s). Iow, by holding the plate at a different location, he is accidentally selecting out one or a few of the modes that would otherwise be present. Ergo, he observes a different frequency. This is really basic stuff about plate motions; it goes back at least ~100 yrs to Rayleigh's work, maybe even Helmholtz' work.

[brunello97]

Separate the thumps from the chaff.

Mick

[murrmac]

Guitar backs don't need braces from structural reasons

I am not sure that there would be universal agreement on that ...

[Ron Cox]

thanks for this. Any and all info on tone bar placement helps us (me) learn instead of the trial and error the original builders had to go thru. Tone bar placement could be random and so would your individual results.

I personally am very thankful to all of you who post, as I learn something from each and every post.