carbon fiber instead of truss rod

[Rick Turner]

Given that there are close to a million and a half Martins out there, I don't think luthiers will be out of the reset biz anytime soon!

[labraid]

You know, I keep laughing to myself, something in all this really is very funny. I mean, this whole CF rod placement thing being unresolved still. I mean, I'm not just an obsessive fool... The anecdotal evidence of aforementioned dude, whose lack of secured CF rod in the slot, placed directly on the "accepted" neutral axis under the fingerboard, and the neck that bowed like mad? No one else mentally juggling that? Something that flies in the face of everything we take for granted in our CF work, and... that's it?
My last ditch effort........

[labraid]

Ok, I have a plan. I'll present my hypothesis and the method I intend to use in testing this.... If it sounds to others like it would prove things one way or another, I'll go ahead with it. Give me a few days...... We're in the middle of a freakin' move, God I must be insane.

[Rick Turner]

I don't have a problem with his having had a problem. He used the stuff all wrong, so of course he had a problem... I don't see anything out of line with what Dave and I have been saying. Put CF in the neutral axis, and it won't do much except take up space. Put it at the very back of the neck, and it won't stretch. Make a fingerboard out of it (my patent assigned to Gibson from 1989...) and it is pretty darned good in compression. Add one or more CF webs to that (also in that patent), and you've got what I call a structural fingerboard, and yes, some of the CF is in the neutral axis in that structure which is what I do now. But it wouldn't be cost effective to do anything much different. With my guitars, the structural fingerboard is so stiff and stable that the neck could be practically anything. Also it puts stability and predictability right were you need it...supporting the frets. It's all about the fret tops when it comes to playability. That conical or cylindrical surface line with a gentle relief curve is what we're going for. Anything at all that stabilizes that is a good thing.

We're picking nits here with worrying too much about the neutral axis. There is one, it's there in any neck. I'd worry more about the fret tops and the line they take, and I believe I do know how to control that.

[labraid]

> I don't see anything out of line with what Dave and I have been saying.

This is where the neutral axis is worth picking nits on, IMO. Dr. Cohen recommends to all that they put the CF as deeply into the neck as possible, and not directly under the fingerboard. Why? Because under the fingerboard would be the neutral axis and therefore provide no net benefit to stability. My only problem with this recommendation is that I have seen the large amounts of CF that DC uses in his necks, and believe the man they are in fact very sturdy. The sad truth, though, is that DC could put his CF assembly almost anwhere, similar to what you say Rick, and it would have the very same effect -- he has essentially, and very wisely, hedged his bets. His recommendation to others, however, that if they were to use only one single rod alone, that it be buried very deeply in the neck... this recommendation, while not mal-intentioned, could be very harmful due to the fact that neither is DC using this configuration, but he admits, I believe, that he does not know where the neutral axis is with any detailed precision. Furthermore, no one person at this point is willing to gaurantee anyone else that the NA is the most stable region given the non-bending forces which distort the NA. His recommendation of deep-setting one's CF is based not only on traditional compression rod technique, but also on an unverified assumption.... Yes an assumption that deep down in the neck is the most instable region of the neck, an area which needs to be tamed more than any other in the profile. For one, though, a CF rod is not a compression rod, and two, the region of the neutral axis is not gauranteed any sort of stability outside of pure bending forces. Three, anecdotal evidence from an amateur builder's oversight (how many of us have had such a fortunate and telling error!) shows that a 1/8" wide x 3/8" tall cavity directly under the fretboard did not in fact contain any area of stability -- rather, and to the contrary, it was evidenced to contain an area significant in upholding the overall straightness of the neck.
Does my argument seem to be taking on some sort of precision at this point?

[Rick Turner]

The reason for putting CF as close to the back of the neck as possible has nothing to do with what part of the neck is the most stable, it's because CF resists stretching better than just about any material known to man, and when a neck bends forward, the back of the neck is stretching. That's why my original patent was for a hollow CF neck. Putting a high modulus material as close as possible to the surface that is under tension is the most efficient use of the stuff. I choose not to do that these days, but rather make a structural beam out of the fingerboard and CF that is quite stiff and strong enough for my purposes. I know I'm not using the materials as efficiently as I could, but there are other reasons that override pure engineering here...like the ability to carve a wooden neck.

[Dave Cohen]

Brian, no assumptions whatsoever are involved. As I said earlier, I have the math chops to do the calculation which you are talking about. I choose not to spend any time on the calculation because I don't think that it would be a good use of my time at the moment. Imo, it is not even that difficult a calculation. All one would have to do is make a scale drawing of the neck cross section(s), and use numerical integration to determine the shearing moments. Where the moment is zero, you are at the neutral axis.

"Stability" is not really the proper word for describing what happens when a neck bends. If it bends into a position and stays there (at least for a while), it is at equilibrium; it has the lowest possible potential energy (save being able to return to straight), and it is as stable as it's gonna get (at least for a while). When a neck bends under string tension, the material exactly on the neutral axis neither stretches nor compresses. Go a tiny bit south of the neutral axis, i.e., toward the back of the neck, and the material there has to stretch. That means that the fibers have to slide past each other in the lignin matrix. Go further toward the back of the neck, and the material has to stretch more. Go all the way the the back of the neck, and the fibers have to stretch the most. The terminology is that the shear is greater the farther away from the neutral axis you get. It also means that the restoring forces farthest away from the n.a. are the largest. Same thing happens above the neutral axis, i.e., toward the fingerboard, except that the material has to compress instead of stretching. Now, if you are gonna use some very high modulus material that is very resistant to stretching or compressing to reinforce the neck, you want to put it where the shear is greatest to get the greatest effect, and that would be all the way at the back of the neck. There, you are supplying the greatest restoring forces possible to counter the shear forces. Since we (a) don't want it to be visible, and (b) glue seam failure would risk catastrophic failure of the neck, we put it about 1/8" to 3/16" (ca 3-5 mm) up from the back of the neck.

Btw, no one is recommending that the reinforcement be at the n.a. To the contrary, it should be as far from the n.a. as practically possible. That is why I put my CF as far toward the back of the neck as possible, and Rick puts as much of his "structural fingerboard" as possible above the n.a. Once again, those are not assumptions at all. Another thing; I am using two 1/8"x3/8" CF composite rods in my necks, which is less than some recommend. I just configure them differently to get as much stiffening as possible.

Oh, and another thing: "non-bending" forces do not distort the n.a. The neutral axis is a construct that exists solely because of bending. All it is is the dividing line between where the shear results in compression and where it results in tension, i.e., where the shear is zero.

http://www.Cohenmando.com

[amowry]

While I don't have the ability to determine where the neutral axis is, I've always put my CF as far back in the neck as possible because it makes intuitive sense to me. We have the ebony fretboard at the front of the neck, which to me represents a structure that is fairly resistant to compression. If I'm going to put another reinforcing structure in the neck, it makes sense to me to put it as far back as possible, so I'm in a sense making the beam as thick as possible. I think the back of the neck is in tension, but even if the whole neck were in compression, it still makes sense to me to have the CF as far back as possible.

Say, just for the sake of argument, that the maple in the neck has no structural role at all, and is just there for filler. If you have a neck whose strength comes mostly from a fretboard and a CF bar, wouldn't you want those to be as far apart as possible? I know that's an oversimplification, but in my simple mind it seems quite clear, and holds true regardless of the location of the neutral axis.

[labraid]

Well, my method is going to be:
5 identical maple rods, each 4 parts maple in cross sectional height and 1 part ebony -- loosely but constantly mimicking a neck profile. A critical section of each profile will be removed, first from the top 1/5, the ebony, second, 1/5 below the ebony, and so on until piece number 5 will have a section removed from its back. Each will be strung up with identically tuned and equal length sections of musical wire on both front and back faces. A steel cable simulating regular string tension will then be applied, each piece in turn and identically so. Dimensional change will be measured by how much each wire changes pitch -- shrinkage indicated by a drop in pitch, elongation by a rise. Deflection will be measured by the space between test surface and steel straight-edge at direct center of the piece.
Consistent with what I have been advocating, it is my belief that the piece 4 (1 being where ebony from the face is removed, 5 being where back wood was removed) will show the least forward deflection of the 5, illustrating that compression forces in a neck outweight tensions caused by bending, and also that the implied location of the neutral axis in relation to the voids (where CF rod would usually go) is less a factor in neck stability than currently taken for granted. Lastly, we will be able to actually hear and measure which faces will show net elongation, and which will show net compression. Here it is my belief that top strings in general will lower in pitch more than bottom strings will rise. Again, I believe piece number 4 will come out with the lowest disparity between pitch changes.

At least I have a reason to go and buy a nice table saw now.

ps, I'll need at least three months.

[Rick Turner]

Andrew, you are absolutely correct.

I have CF probably in the neutral axis just because it's convenient. I dado 1/8" x 1/2" CF rods up into the fingerboard about 3/32"...just below where the fret slots are. When I'm really on a trip, I don't use normal wood for the fingerboard; I use "PakkaWood" or "Diamond Wood"...laminated birch veneers impregnated with phenolic resin. The compressive strength is pretty impressive on that stuff. The CF rods are then dadoed a bit more than 3/8" down into the neck, so I suspect they pass through the neutral axis, but what the heck, there they are. Then I have a two way truss rod between the CF bars. Belt and suspenders. No problems with the necks...even in the Antarctic climate.

[labraid]

Alright, here's a visual of the conundrum, does anyone else see a problem with the figure on the right?

[Rick Turner]

What I see is that you're trying to show us that the side of the beam closest to the load of the weight and the side farthest away are both in compression, and I'm not buying it. That's not how things bend. You're on this compression/compression kick, but it doesn't matter how you apply the load, if you're bending a beam, one side is in compression, one side is in tension. This is absolutely classic engineering.

Try using an archery bow for your example...

[labraid]

Well... I guess I am wrong on this one. What can I say?......

[Dfyngravity]

I have been following this thread for the beginning and it all seems very interesting. I have been thinking about a few things, but I am not sure if they have any effect or not or if the theory holds any water what so ever.

It seems to me that the closer you get the strings to run parallel with the neck, the less you have to worry about the neck bending. Also, the shorter the neck is the less you have to worry about it bending.

Taking that into consideration, a new mandolin such as a classical mandolin like a bowlback that uses lighter strings, has a shorter neck, and typically lower action with the strings running closer to parallel will be less likely to have neck that bows forward.

So by using this theory, a typical bluegrass mandolin that has heavier gauge strings, a longer neck and action that is further away from being parallel to the neck, the neck is more likely to want to bow forward.

Does any of this make sense or have any value? If so, I would think you would want to not worry so much whether or not to use a truss rod vs. CF and try to minimize the effects of the string tension on the neck. For instance, if having the strings running closer to parallel with help relief tension on the neck, then you could increase the neck angle slightly. This would allow to you keep the same bridge height but the strings would closer to parallel to the neck and the action would be lower. But I am not too sure what you would do with the length of the neck.

Anyways, these are just some things that popped in my head when I was reading though the thread. They are probably way off but I figured I would ask and see, since there seems to be a few here that have a lot more knowledge than I do.

[Dave Cohen]

Brian, no need for me to pile on now. I can see where you misled yourself, though. One thing is that you left out the little red vectors from your earlier diagram. The resultant force is the sum of the (red and blue) component vectors, which will point away from the top of the neck. That is, it will point up off the fingerboard. The other thing is that the origination of the vectors is at the top of the neck - actually above it, at the top of the nut and over the fingerboard - and on the basis of vector algebra, you can't claim that the force vector is doing the same thing at the back of the neck.... or anywhere but at the top of the nut, for that matter. Sorry that I didn't think to point that out earlier.

http;//www.Cohenmando.com

[Michael Lewis]

Well folks, this has been an entertaining, educational, and enjoyable session. Thanks Brian, Dave, Rick, Andrew, and others (you too) for the lively discourse.

[Joel Spaulding]

Brian , Dave and Rick;

As usual you have individually and collectively informed, educated and entertained throughout this thread.

Any time you three want to discuss this and similar issues until the wee hours of the AM, let me know and I'll cook and provide ample quantities of stout, pale ale or beverage of choice. Sorry I can't fly all of you to KY ( I guess that's all y'all. )

Brian - Cape Bretton is a pretty magical but somewhat remote locale - hope to return for another visit someday. Best of luck with your move.

[J.Sloan]

Thought I would bring this post back to life only to ask one simple question. For those with experience with CF necks, do you feel that it is crucial to use epoxy for gluing in carbon fiber rods, or would it be ok to use tightbond? I realize that epoxy is stronger, but once the rod & filler stick is glued in and covered, the CF doesn't bend anyway. I've always used epoxy in the few mandolin CF necks that I have made....but the stuff is so messy. I got into the habit off taping off the neck just in case of any spills or runoff. Maybe I could take a large hole syringe and slowly squeeze it in. But, if I can get the same results with wood glue....well...you know. I'm all ears and stand corrected if epoxy is supreme in this question. Up until now, it's all I have used anyway. Thanks!

[Dave Cohen]

I'm not sure that epoxy is much stronger for bonding wood to wood than are yellow AR glues such as Tightbond. On the other hand, the CF composite bars used for neck stiffening are usually CF/epoxy composite, or sometimes CF/polyester composite. The surface often needs to be made bondable by roughening the surface w/ sandpaper. Otherwise, no type of glue will stick to it for very long. Once the surface of the CF composite bar is made bondable, epoxy is the closest chemically to the composition of the stuff in the CF composite bars that holds the CF fibers together. I don't know if Tightbond will work in your application or not, but my first guess is that it will not work as well as epoxy. Tightbond is sold as a wood glue; it needs the somewhat porous surfaces of wood to be effective.

[fscotte]

I make my channel so tight that titebond is fine. I have to literally use a clamp to push my CF rod in after I put down a layer of glue. If your channel isn't so tight fitting then maybe a gap filling glue like epoxy would be better suited. It's likely one of those things that wouldn't rear its ugly head until 80 years later anyway.

[John Kelly]

On the three necks I have done with CF reinforcement I have used Titebond, making sure the channel for the CF has been routed to a tight fit (as fscotte says above) and the filler strip is clamped down really firmly while the glue is drying. I wrap the CF rod in masking tape too before inserting it into the slot. No problems so far but we are only talking about a couple of years or so, so I do not know long-term results. The steel rods I used before the CF have been in existence for around 12 years now on the first instruments I built and it's only on a bouzouki that I have used an adjustable rod - not on mandolin or octave.

[bpatrick]

Does anyone see a problem with using (2) !/8" x 1/2" carbon fiber rods glued together in the same channel instead of (1) 1/4" x 1/2" rod?

[Tavy]

Does anyone see a problem with using (2) !/8" x 1/2" carbon fiber rods glued together in the same channel instead of (1) 1/4" x 1/2" rod?

No, that's what I use. Just make sure that they're a tight mechanical fit in the channel in addition to being well glued in.

[J.Sloan]

I'm going with 2-- 1/4" carbon fiber rods for my current F4 build. I put them close enough together to not hinder the fingerboard profile once it's tapered down(1 1/8" nut) but I only did this because it's my first mahogany neck and it's a much softer wood than the red maple that I usually use for F5 and A5 mandolins. Hopefully, this will keep any future twisting from happening as well.

sloanmandolins.com