A neck joint test--wood is strong!

Decided to test a bolt-on neck variation today and was surprised enough to share.

I have niggling and possibly irrational concerns about possible short grain failure on the heel of a metal-cross-doweled bolt on neck. As the heel becomes smaller and more svelte, the cross dowel moves closer to the inner face of the heel, and I imagine tightening the connector bolt one day and hearing a CRACK as it pulls through.

Even if that doesn’t happen, imagine the horror of having the joint let go during a bar fight.

In a modest defense of my rationality, having made and studied furniture for decades I’ve seen plenty of short grain failure, for instance in the tails of 14 degree dovetails that were common in factory furniture for several decades.

Returning to petty worries, there is the scenario of someone trying to install a heel strap button and drilling into the cross dowel, pushing harder when it resists, and at minimum making a poor installation if not outright *$&% up the heel.

I wondered if one could have the benefits of a bolt on neck by substituting wood for the metal cross dowel (or the alternative bronze screw-in fitting, which I'm not partial to).

No doubt this has been done, but I haven’t seen a writeup, so after a conversation with a good friend who builds over on the big island I designed a suite of tests:

- instead of a metal cross-dowel, 3 sizes of wooden dowel (1/4, 5/16, 3/8)
- dowel holes 3 distances from the end grain face (3/16, 5/16, 7/16 between nearest points)
- 3 species of dowel wood (birch, mahogany, and something really tough).

Dowels would be glued. Instead of a connector bolt, I used a hacksaw-shortened no. 8 Power Head screw, a very strong coarse-threaded screw with a connector bolt-like wide head designed for hanging cabinets.

Being hasty though, I grabbed a neck scrap, drilled 1/4” holes at a few eye-balled distances, pressed a cheap birch dowel into the one closest to the end (the weakest, and without glue), cross drilled it from the end grain (uke body) face, and drove in a shortened screw, passing loosely first through a cross-ways stick that I could grab to pull on. Weighing it before assembly, the joint components were lighter than a metal cross-dowel and connector bolt by a shade under a half ounce.

Putting the assembly in a vise, I tugged on it in the extraction direction, expecting this flimsy version it to pull through easily. Nope. Then a good steady pull, increasing to I could muster in “upright row” position until I was light-headed from strain…and nothing.

I measured the distance of the dowel to the end grain face—7/32” to the nearest edge of the dowel, or 11/32" to the centerline of the measly 1/4” dowel. Hmphh.

Taking it outside, I rigged a climbing sling to hang it from a beam--adding a little twisting force as well because of the arrangement--then body-weighted it (not quite 200 lbs). Not a peep; nor when I jerked on it forcefully with slack to provide impact force.

Taking it back inside, I rigged it in the vise so I could beat on the assembly with a 12 oz hammer (in the extraction direction). Nada.

Finally, I tightened the screw with all the force I could muster with both hands. I couldn’t strip it, the square drive ultimately levered out of the screw. It wouldn’t give up.

This wasn’t in the least scientific, but I was well impressed. It seems as if it would make a very stout yet easily removeable neck joint, and it could probably be made stronger yet (subject to testing) with a 5/16” dowel glued in, and perhaps hardened after the first threading with super glue or thin epoxy, though frankly these seem like overkill now. (Not gluing in the screw of course).

What do you think? While few if any here may share my concerns about cross-dowels, short grain failure, heel cracks, and so forth, I’d be very interested to see if you see disadvantages or problems down the road that haven’t occured to me, or if you know more of the history of this joint in lutherie. Letting my devil’s advocate loose hasn’t produced much:

The dowel in heel arrangement is well within the guidelines for glued (or unglued) cross grain construction (i.e. cross grain wood movement problems).

Should it ever strip--which I was unable to do with both hands on a straight line ratchet driver—it could be repaired without removing the fingerboard in more than one way—shrinking the hole with glue, or drilling it out and replacing with a cross grain plug for instance. This latter would be an alternate way to do the entire joint, though it would be fully dependent on the glue joint and have no mechanical interlock between the wood parts. If the fingerboard was off, it could obviously be repaired to original condition.

A glued in dowel would also reinforce the heel against cross-grain failure—I feel like I’ve seen this in some Collings neck cross sections for instance. That of course is an extremely traumatic impact, but hardly unheard of in repair circles.

While an errant strap button screw could still contact metal, it would have to be errant in a very exact way, and is certainly less likely.

By all means point out potential weaknesses, and thanks as always for your insight.

Now that is a detailed post. Love it. Thanks. A couple of thoughts:

1) We are talking ukuleles here and not furniture.
2) Maybe the fear of cross grain failure is not really that big an issue?
3) All of your methods of neck attachment will work just fine. All are good.
4) Gotta ask yourself: Are we possibly over thinking things here?
5) Maybe just concentrate on getting the thing on flush, plumb and perfectly inline.

Totally agree with Sequoia, you are over thinking this. A short grain failure on a uke neck heel is very rare, even on cheap ukes with odd ball woods.

Reasonable comments, all, sequoia (and now BlackBearUkes). Though I am interested in the questions I asked, the original motivation to post was amazement--how strong this seemingly minimal connection was!

Someone should have told me we weren't talking furniture though;-).

Silliness aside, it's woodworking, and an deep understanding helps greatly in making a sweet and long-lived instrument (or drawer). How wood moves, how stuff fails and can be repaired, and so forth all are dimensions of caring about our medium, tradition, and those who will use and care for our creations in the future.

Along with over-thinking--a kind understatement, by the way--I agree whole-heartedly with plenty of attention to flush, plumb, and inline...



As you say, the concern about short-grain failure may have been misplaced--as I speculated near the start of the post. However, now I've got pretty good confidence it's misplaced. Maybe the old hands here have already been through this, but it was an eye-opener for me.

I may still test a metal cross dowel though. If I can hang my car from it, I'll post a picture.

As someone who has made thousands of glue dowel cabinet face frames, cabinet doors, hand rails etc. I believe that glue dowel failure is far, far more likely that short grain failure. That is if you use a mortice and tenon joint and don't over tighten. Glue dowels always fail on the crossgrain if they do fail. The problem with the way you are doing it is each time you run a screw in and out you damage the supporting wood. You are making wooden threads with a screw. Wooden threads are going to fail long before the steel cross dowel threads. When you glue in a dowel cross grain of dissimilar woods. Say mahogany and iron wood I doubt if you are gaining anything over using steel unglued, as crossgrain glue failure or at least weakening is very likely over the decades.

Not a dowel joint

I should have posted pics to be clearer, here they are:



1st photo is the setup for body-weighting the screw to see if it would pull or break out. Yes, my body weight is overkill, but by then I was curious how strong this thing was.

2nd photo is after I'd driven it in super hard trying to intentionally strip it. Here you can clearly see the puny 1/4" cross dowel poking out, how close it is to the joint face, and the wide-head screw clamping down on the handle I attached in order to test the joint.

Michaels response came while I was composing this, still digesting it.

And thanks to all for your responses!

I think your experiment has merit. I sometimes fret as well when I use a barrel bolt in Spanish cedar or brittle mahogany. But since I let the back cover the heel and glue down the fretboard on most of my ukes, I don't think the neck can move enough to put abnormal stress on the wood in the heel. A piece of furniture with sides connected at right angles can get a greater leverage if you bump into it.

Thank you for taking the time to experiment and and sharing the results with us. As an unscientific amateur builder I find it comforting to know how strong a screw tapped into wood will hold. I may try this on my next build.

Bob

I can't understand why you picked on the neck joint ..there are much more delicate areas of a uke to test..bridge, sound board & backs. Etc:

Timbuck said:

I can't understand why you picked on the neck joint ..there are much more delicate areas of a uke to test..bridge, sound board & backs. Etc:

Click to expand...

Timbuck, I just happened to be talking neck joints with a friend who has built them several different ways, and got curious about how to get most of the metal out of the heel, still have the advantages of a bolt-on, and perhaps make a sleek heel stronger. I hadn't seen it discussed on the forum, and thought an experiment would be fun.

Had I looked here http://www.acousticguitarforum.com/forums/showthread.php?t=377225 or on Frank Ford's site--as I did this evening--I would have found that a coarse woodscrew thread into a wooden cross-dowel has been in common use for quite a while by high quality guitar builders at all scales of production. A no. 10 sized hanger screw/bolt (roughly 3/16", one size bigger than what I used) is a time-tested choice on highly stressed instruments.

When I removed the 1/4" dowel from my test block this morning, even though there was very little dowel left where the screw passed through, there was no visible deformation anywhere, in spite of loading it way beyond anything a uke would experience, and having run the screw in an out any number of times.

As Michael points out, metal threads are stronger than wood, but the wood threads seem to work because the very coarse pitch (wide spacing) of the threads leaves enough wood between successive turns to resist pull through. The threads are also a much deeper than those designed for metal.

My test obviously doesn't examine long term deformation under load, but the feedback from other instruments suggests that it will continue to work.

Based on the track record from these other instruments, I plan to use this joint for a while to see how things go. I will switch from the Power Head screw of my test to a no. 8 or no. 10 hanger bolt--it will only need to be driven in once, thereby allaying worries that multiple ins-and-outs will damage the thread path.

I'll also move up to a 3/8" or 1/2" dowel for more thread engagement. Birch dowels and mahogany necks have worked well in many thousands of Collings guitars, so I won't worry about making dowels out of some rock hard wood.

No doubt I'll sleep an extra millisecond or two each night having soothed (mostly) my concerns about errant strap button screws, and the cross grain reinforcement that the dowel will provide to the heel;-).

Above all, this is NOT an attempt to convince anyone else to use this joint, nor am I dissing any other neck joint choice. In fact the test suggested that short grain breakage was much less likely than I thought, though I'm still curious to try it with Spanish cedar and Port Orford cedar.

I am a very inexperienced uke-maker and in no position to give advice! I'm just working through this particular learning curve in light of several decades of other types of finicky woodwork, and I am very grateful for your responses, and for the treasure trove of info many have shared on this site.

This is such an interesting post, but I have to agree with Ken: Why go for the neck to body connection? Is this really the vulnerable place? I once posted here that I thought the compound dovetail joint was over engineered on an ukulele and got severely flammed so I won't go there (ducks, covers and runs). Which got me to thinking: Where does the ukulele fail? I've built a few ukes and I'm no luthier/repair person, but it seems to me the neck to body joint is low in the failure department. Maybe a pro repair person will chime in. Going off thread a bit, here is what I think are the places an uke is most likely to fail:

1) Fret/nut/saddle set-up issues: buzzing, high frets, low frets, protruding frets, worn frets, low nuts, high nuts, etc. etc.
2) Bridge detachment. Shouldn't happen, but it does.
3) Top cracks, back cracks, side cracks, joints opening, braces detaching, things generally falling off, etc.
4) Peghead detachment: Snapped off pegheads. Very bad. Happens a lot. This joint is way more vulnerable than the neck to body connection.
5) Finish issues: Peeling, cracking, delaminating, clouding, too thin, too thick, etc.
6) Neck issues: bad set, twisting, racking, bowing, etc. etc.
7) Tuner issues: Too tight, too loose, too heavy, too light, etc.
8) Sound board issues: Lifting, Bellying, warping etc.

and low down on the list

Neck coming off: Detachment, catastrophic failure, etc.

I just think that torquing forces and pounds per square inch pressure on an uke don't really justify over doing this joint. Steel string guitar: Another kettle of fish.

Apologize for the long post.

The way i see this is that it is no different than a regular bolt on neck, other than, the fact it's a screw into wood dowel. I don't think the OP was suggesting this is a new method all together, or stronger, better neck attachment, just that coarse thread screws into a dowel hold like a mutha f**ker. I've never liked the idea of a steel/brass/aluminum threaded post-a-majig in the neck with a bolt. And I've really never liked the brass insert piece either. This just seems so uke-like for the "bolt-on" concept. I am going to give this a try for sure. Thanks for posting it!!

I have no faith in those screwed into end grain brass inserts. However if you are doing a mortise they too would work fine until the glue dries.

A question I have for the crowd. If you use a mortise and bolt on how much glue do you use before you send your instrument into the world. I have been using a small amount of the LMI white for this joint and red cap for most of the rest of the instrument thinking that I could make loose the neck joint with a little steam without the rest of the instrument coming apart. But I have never actually tried and I probably am dreaming and would likely have to take the fingerboard off first. I use hide on the fingerboards so it would probably come loose anyway.

Thanks for this enlightening thread.

Some people I think just have to learn that a thread is about what the thread is about and if it isn't about something else then it isn't about that.

Back to the thread in hand. The Gougeon Brothers of WEST System Epoxy fame put a lot of engineering work into thread retention in a marine environment and for maximum strength they recommend putting in a screw, taking it out, drilling out the hole much (but not all) of the way, filling the hole with Epoxy and reinserting the screw, using that little bit in the end you didn't drill out to hold everything in place while the epoxy sets. This is the recommended method for fitting mission critical components in a boat and is also I am told the way most ski bindings are fixed.

This method is probably anathema to most "you have to be able to take it apart" purists, other wise known as the people who finish up having to fix things, not if but when they break. On the other hand it is damn strong and is less likely to break and would satisfy both criteria if used to fix hang bolts or the screw in insert thread popular with some.