Deflection testing on Adirondack braces... slab vs quartered

[j. Kinnaird]

I like the setup very well. I think I will go there with my next text

[Alan Carruth]

There were studies I've heard about from the 40's. They still used lots of wood in aircraft, particularly at the beginning of the war, and had issues with glue. One study from the Forest Products Lab showed that you get better glue joints when the wood surfaces are freshly worked. Another found that wood propeller blanks where the laminations were sanded to thickness were more likely to come apart in use than ones that were planed to thickness.

I did see something an a publication from Experimental Aircraft Association on cold creep, which, iirc, was citing an FAA rule. They said that the initial deflection under the maximum design load for a wood structure should never exceed 1/3 of the deflection that could be tolerated in use. This suggest that 'cold creep' which tends to slow down over time, could amount to about twice as much as the initial deflection.

Trevor Gore says in his book that he tries for a 2 degree short-term deflection of the bridge under string tension. Any more than that signals a top that is too weak, and any less a top that is too heavily built to sound good (in his estimation). If you put that along side the FAA rule it makes some sense. A two degree initial deflection implies 6 degrees in the long term, which would be enough to significantly increase the peeling load along the back edge of the bridge.

[BrunoBlack]

This is a very clever test John. I will add my experience to the conversation. As a wood scientist I’ve done quite a bit of research on wood and wood-based materials. However, I have no experience with guitars or elements as small as guitar bracing loaded as bending members. Most of my wood research was related to structural applications and wood/moisture relations. In that context grain direction (radial or tangential) is disregarded when assigning design values in bending because the difference between the 2 aspects is considered to be insignificant. That’s for beams and structural elements used in building. Even when conducting 2”x2” ASTM bending tests using clear wood samples, grain orientation is disregarded when assigning design values.

However, if you test very small bending members like guitar braces, each band of earlywood and latewood occupies a relatively large percentage of the cross sectional area. If there are differences in the material properties for each of the latewood or earlywood portions of the growth rings I can imagine it might possibly have an impact on bending.

Consider a brace with the load acting perpendicular to the tangential surface. The annual rings in the brace resemble laminations like a glue-laminated beam. The properties of the extreme fibers might be different from the next layer as you approach the neutral axis. So there could be differences in stiffness from 1 layer to the next. Now - Flip the test specimen so the rings are positioned vertically while loading the top surface perpendicular to the radial surface. Now there would be an “averaging” across growth ring “laminations” because there would be no horizontal bands of structural differences.

Having said all that, this should have more influence on strength than stiffness. Reduction of grade in structural lumber (for example) influences strength more than stiffness.

Just thoughts off the top of my head. I know nothing about guitar bracing.

[j. Kinnaird]

Originally Posted by Haasome

This is a very clever test John. I will add my experience to the conversation. As a wood scientist I’ve done quite a bit of research on wood and wood-based materials. However, I have no experience with guitars or elements as small as guitar bracing loaded as bending members. Most of my wood research was related to structural applications and wood/moisture relations. In that context grain direction (radial or tangential) is disregarded when assigning design values in bending because the difference between the 2 aspects is considered to be insignificant. That’s for beams and structural elements used in building. Even when conducting 2”x2” ASTM bending tests using clear wood samples, grain orientation is disregarded when assigning design values.

However, if you test very small bending members like guitar braces, each band of earlywood and latewood occupies a relatively large percentage of the cross sectional area. If there are differences in the material properties for each of the latewood or earlywood portions of the growth rings I can imagine it might possibly have an impact on bending.

Consider a brace with the load acting perpendicular to the tangential surface. The annual rings in the brace resemble laminations like a glue-laminated beam. The properties of the extreme fibers might be different from the next layer as you approach the neutral axis. So there could be differences in stiffness from 1 layer to the next. Now - Flip the test specimen so the rings are positioned vertically while loading the top surface perpendicular to the radial surface. Now there would be an “averaging” across growth ring “laminations” because there would be no horizontal bands of structural differences.

Having said all that, this should have more influence on strength than stiffness. Reduction of grade in structural lumber (for example) influences strength more than stiffness.

Just thoughts off the top of my head. I know nothing about guitar bracing.

Very thoughtful. I like the averaging vs laminated idea. If this is how it actually is then carving on the braces ought to have different effects, a variable worth looking to

[Alan Carruth]

Haasome:
Would flat cut brace stock with only a few ring lines in it be more likely to fail in shear? How about cold creep?

[BrunoBlack]

Originally Posted by Alan Carruth

Haasome:
Would flat cut brace stock with only a few ring lines in it be more likely to fail in shear? How about cold creep?

Not to be cute, but it depends on load, duration and description of failure. If you mean simply as compared to radial, possibly. Testing data that I’ve seen doesn’t support that idea when testing large scale testing samples. But I could imagine that being the case in these very small samples. I am curious on how the differences in growth ring properties might (and I mean might) influence performance of flat sawn braces - moving values upward and downward depending on properties at extreme fibers. However, intuitively, I wouldn’t think these braces are subjected to high enough loads under normal use.

BTW- horizontal shear is highest at ends of “beam” at the neutral axis; and zero at mid span of “beam.” ( center-loaded beam) Is this the scenario you are discussing? Horizontal shear can be critical in composite products having relatively weak internal layers — like weak interior plies in thick plywood where rolling shear develops. However, I can’t imagine the loads on guitar braces are that high. Again I don’t know anything about guitar braces because I’ve never even thought about them as structural elements.

What do you think?

From a practical standpoint: if you had a stiffness or deflection target under known load, I could imagine setting up a bending jig that supports the ends of known brace lengths, subject them to the appropriate loads, and compare the brace deflection to desirable target. This would be a type of proof testing for each brace where the specific brace would be sized according to what you are working with. (Of course you can purchase proof testing devices too)

Afterthought— as a guitar building know-nothing - it seems that since there is potentially such a wide variability in strength and stiffness from one brace to the next (based on material properties holding sizes equal) it seems that the difference in the native mechanical values for each sample would dwarf the difference you might see in using radial vs tangential. It seems to me that having a proof-loading device that allows you to test every brace would have more practical benefit.

[Guitars44me]

Holy Smokes!

There are some SMART Folks Hereabouts.... this place is so cool!

And John is about to carve my braces on the BIG BRAZ GUN. I thought they were done already. We spoke today.
I always like to talk to him when it slows down progress on MY build! Hahahaha

There is always something to learn, just as there is always musical progress to be made.

An interviewer asked Segovia why he practiced so much when he was already the best at what he did. He was about 90 at the time I think I read. Segovia replied, "I hope to be much better before I die."

There it is!

Carry on gang and stay healthy

Paul

[j. Kinnaird]

Originally Posted by Haasome

Afterthought— as a guitar building know-nothing - it seems that since there is potentially such a wide variability in strength and stiffness from one brace to the next (based on material properties holding sizes equal) it seems that the difference in the native mechanical values for each sample would dwarf the difference you might see in using radial vs tangential. It seems to me that having a proof-loading device that allows you to test every brace would have more practical benefit.

True enough except that you cut the brace stock taller than wide so if there is a preferred grain direction you wouldn’t be able to test that meaningfully without some prior experimentation. So within a given tree or tree segment there might be a much stiffer brace regardless of grain orientation but would there be a way to anticipate the optimum way of cutting braces from that segment. Here-to-fore it has always been assumed that Optimum it would be quarter sawed And a fair amount of trouble and even waist has been involved in making sure the brace is as well quartered as possible. My question is whether that assumption is warranted

[Alan Carruth]

Haasome wrote:
"BTW- horizontal shear is highest at ends of “beam” at the neutral axis; and zero at mid span of “beam.” ( center-loaded beam) Is this the scenario you are discussing? "

Yes. I will say that I can't remember having seen this type of failure very often, but then, I've mostly dealt with guitars and they normally used braces that are quartered wrt to the gluing face.

Guitar braces are normally taller in the center, and are trimmed off at the ends, often quite drastically. You might see an Upper Transverse Brace that is 8mm wide and 15 mm tall in the center that gets scooped down to 6 mm tall or so at the ends over 30-50 mm. The ends of the braces are typically inlet into the liner: a doubler, often around 6 mm wide by about 12 mm tall, that is glued around the inside edge of the top and back to provide some surface to glue the plates to, aside from the 2 mm thickness of the rib itself. The UTB carries a fairly high load for a guitar brace; it runs across the upper bout at around the widest point, and takes up a lot of the load from neck rotation. It's been a long time since I've done much repair, but I do recall seeing one or two that split horizontally in from the ends, very near to the lowest point, which is about where you'd expect it if shear is the issue.

Lutes use flat cut bracing, often with nine or so tall, narrow braces across the top. They don't use liners, except for the 'end clasps' inside an outside of the lower end of the body below the bridge, that hold the ribs together. The cross braces are not inlet, but are left full height at the ends and very carefully fitted against the inside of the top rib. The angle of the rib relative to the soundboard provides support.

In terms of function, top bracing is more important structurally than acoustically. The vibration modes in an un-braced top are surprisingly similar to those of a properly braced one, so it would most likely sound fine if you made a top that way. Arch top guitars, which tend to be thicker, and get a lot of stiffness from vaulting, carry a couple of top braces who's main acoustic function seems to be to make up for the stiffness that's lost when you cut in the sound holes. A 'flat' top would need to be significantly thicker than 'normal' to hold up under the loads, particularly the static torque of the bridge and the neck load. Since the 'horsepower' in a plucked string is pretty limited we use bracing to reduce the weight of the top. A lot of different brace layouts have been tried over the years, and many of them can be made to work well. They all do 'color' the sound to some extent, of course, but the objective differences can be minimal even for radically different layouts. IMO there is no 'magic' brace layout that insures 'great sound': it's all too easy to make a bad guitar using a good design and wood simply by messing it up.

I'll note that the two brace layouts that are most commonly used these days are the Torres-style 'fan' and the 'X' brace pattern that was adopted by Martin, both dating to the early to mid 19th century in various forms. Both concentrate material in the area between the bridge and the sound hole, where torque and compression tend to buckle the top inward.

One of the ways that guitar bracing tends to fail is through loads on the outside of the plates. Simple 'crushing' is not common, but does happen when instruments get sat on or jumped on. Impacts on the outside can cause failures between the glue line and the braces, particularly when they are not properly inlet.

The X brace layout, common on steel string guitars, introduces a stress riser at the X crossing, which is typically made as a lap joint. Torque from the out of plane string load puts a download force on this joint, and the brace can peel at the bottom of the lap cut where there is a large stress riser. Capping over the open lap joint, where it faces inward on the inside of the top, takes up the tension and prevents peeling failures. This has traditionally been done with a linen patch formed over the top of the braces, but many makers now use a wood fillet.

[BrunoBlack]

John and Alan, thank you both for sharing your expert knowledge and experience. It’s greatly appreciated. Very interesting ideas John. I look forward to seeing how this proceeds. And Alan, I always learn something from your posts.