"Strong" is a very ambiguous concept. I think the idiot who hypothesizes that you can adequately compensate for a poor fit by filling gaps with epoxy because epoxy is stronger than wood is pretty much talking out his ass. In every published adhesive test one of the immutable conclusions is ALWAYS the looser the fit the weaker the joint - period - regardless of the class of the adhesive.
Here is how I would address the above premises:
1.You don't need a tight fitting A-joint with a good axial force. This is true... kind of. The connection between a shaft and a butt is plenty strong enough, as is the A-joint connection in a jump/break cue - and those joints don't have any adhesive AT ALL. However, the indisputable fact that a sloppy joint is weaker than a close-fitting one should be all the motive we need to achieve proper tenon/socket fits. Only a hack who embraces the concept "good enough" would say otherwise (IMO)
2. Joint screws do not need threads . What for ? The epoxy will hold it fine. This is incorrect. The bond between epoxy and a non-porous surface (metal) is generally weak, and the weakest aspect of that already-weak bond is it's ability to resist a shearing force. Both the action of screwing/unscrewing a cue AND the action of drawing that connection tight - producing great axial force - result in a shearing force being applied to the interface and can easily break the epoxy bond, allowing a joint pin to be pulled right out of the butt.
(there is a well known cuemaker friend of mine who used to believe premise #2, and simply epoxied a smooth-barreled pin into his butts. In the early 90's, after many of his cues had found their way to Japan, the word got out that his pins were coming loose. The cause was the differential expansion factors between the wooden butt and the metal pin, and this was no doubt exacerbated by the humidity difference between his location and Japan. He changed this practice immediately)
3. Thin butt plates don't need to have threaded stem going up the cue. What for? Epoxy is stronger than wood. Again, a surface mounted butt plate will be held to the cue ONLY by the epoxy bond itself. The silly mantra "Epoxy is stronger than wood" fails to account for the fact that even if this is true, we aren't concerned about the "strength" of the epoxy OR the "strength" of the wood. What we care about is the strength of the bond between the two, and THAT is far weaker than the mechanical connection provided by the stem construction. Without the mechanical connection the butt plate is no more secure than a tip glued on with epoxy, and tips fly off all the time - even the epoxied ones.
4. Half joint ? Don't need to engineer it to have some axial force holding it. What for ? Epoxy is stronger than wood. Under repeated impact the joint collar is constantly being compressed and then springing back, within the bounds of it natural modulus of elasticity. Those values will be different for the joint collar material, the wood of the butt, and the cured epoxy - and because they each have different moduli of elasticity there will be a minute shearing action going on every time the cue hits a ball. Over time the epoxy may continue to be "stronger than wood",,, it just won't be bonded to the wood anymore. A mechanical bond (i.e., screw threads) does two things; it reduces the level of shearing action by breaking up the smooth interface between the collar and the wood, and it prevents the joint collar from sliding off even in the case of total adhesive failure.
5. Short ferrules don't need to be press-fitted. In fact, give it some .010" clearance for epoxy. After all, epoxy is stronger than wood. Addressed in point #4 above.
TW
