Anonymous wrote:

Thanks, Scott. I'm gaining confidence in the puzzle joint, so I've converted the drawings to all-puzzle joints (full size) and also made all puzzle joint drawings at 1/4.25 scale, so that a model can be made from 2mm plywood, and the tab and slot joints should fit exactly (I hope)

yes thanks for the input Scott,

Always nice to hear from someone who has actually tried something.

The CLC stuff looks rather nice, and the B&B stuff also looks nice, although I remain to be convinced about the B&B origami assembly method. Can't see it saving much and it is definitely not a one man operation.

I wonder David if some experimentation might be in order. I still have half a sheet of 2.2 (!) mm ply and could get a lump of 6 or 9 mm. 

I think maybe a bit of trial and error of different offsets might shed some light.

Could also try some stress testing of you think that might be necessary.

What you think ? 

David Tyler wrote:

 ...pending further information and thoughts from anyone who has used puzzle joints. I've retained the all-scarphs drawing.

I've built a few of the CLC craft, and we are building their Teardrop Camper right now, believe it or not - my first kit ever, and therefore my first puzzle joint panels.  The panels in place do have some good bends and twists.  Nothing like the bow of my Core Sound 17, which was designed by the Graham Byrnes that David cites below of B&B Yacht Designs, but a fair amount good 3D shape develops.  Incidentally, the CS17 didn't use scarfs:  the panels are carefully butt jointed with glass tape on both sides, all the panels in one shot in big stack screwed right through to keep them from moving during setup.  They've never had a failure if done by spec.  Concerns about the strength of the puzzle joints may be like worries about whether nuclear fusion or fission is a faster way to boil water for tea.

I haven't done destructive testing, but boy the puzzle joint seems strong if glassed on both sides - even just one side is impressive. The fit is tremendously close, nearly idiot proof even.  It also forces any crack propagation along the join into a circuitous route that takes it across 90 degrees plus of turn, forcing its way through glass fibers sometimes running with the weave, others with the weft, and often 45 degrees to them both.  Not that we want any crack to start, of course, but propagation is quite frustrated.  

Like I said, I haven't destroyed any of them, but I am pretty sure that when twisted or bent it would break outside the joint.  If I were planning on using them in a design of my own, I'd keep them in flatter runs just like any other joint in order to keep the changing properties of the wood from ruining a fair turn. 

Visually... meh.  It looks a bit too mechanical for my tastes, but that's only an issue if finishing bright anyway.

Gary wrote: - unless you are the type who inserts their jigsaw puzzle pieces with a hammer. 

The CNC company takes my dxf files of the profiles, and then sets a toolpath at a certain offset from those profiles, according to the radius of the cutter to be used. How would it be to specify that that offset is just a little less than the radius of the cutter, so as to avoid too tight a fit and consequent brutal hammering together? Say 0.1mm or 0.15mm?

Anonymous wrote:

One school of thought says that puzzle joints are not as strong as scarphs, but are quite strong enough. Further, that they may not be the very best option when the panel is stressed in bending, as for a hull panel, but are fine on straight, unstressed panels such as the internal, longitudinal components.

That sounds like a good compromise to me. 

I think any type of join will create a hard point in a panel that will be bent, but the question is to what degree, and whether or not it is material. The answer to that will be determined on a case by case basis I think.

The main point is that speed and accuracy will be hugely improved this way, and the opportunity for error will be greatly lower - unless you are the type who inserts their jigsaw puzzle pieces with a hammer. 

Scarfs are not a huge deal if you set up a jig for your router. The large Makita is readily adapted using a couple of long rods and a frame. But I am of the school that says that machines should do everything they are capable of. I would rather be sailing than building.

One school of thought says that puzzle joints are not as strong as scarphs, but are quite strong enough. Further, that they may not be the very best option when the panel is stressed in bending, as for a hull panel, but are fine on straight, unstressed panels such as the internal, longitudinal components.

So, I've put puzzle joints onto the latter, but retained the scarphs for the moment, in this drawing, pending further information and thoughts from anyone who has used puzzle joints. I've retained the all-scarphs drawing.