With some trepidation, I’ve started on the first big test of this building method – making the inner stem. I suppose making things on a CNC mill is a big time-saver if you have a big-iron mill with automatic tool changing, and you’re making repetitive parts out of consistent stock. This project has none of those aspects.
I decided to start with the inner stem, or apron – its not a part that requires a high level of finish, and most mild screw-ups can be hidden between the outer stem and the planking. It does require, however, a two-sided multi-step milling setup, and the two sides must be precisely symmetrical (if nothing else). Starting to see why picking up a chisel and just whacking away at it is faster?
I’ve given up on reclaimed wood for the backbone of the boat – I briefly entertained laminating and scarfing pieces out of 3/4″ white oak plank, the best I’ve been able to find around me in Brooklyn. The alternative has been hunting through the Structural #2 and Better at the big orange box – I’ve managed to find a few 2x10s and 2x8s with enough clear, tight-grain material to make the stem and keelson – we’ll see what happens after that.
The key with planning a milling setup, I’ve found, is to crowd out operator and material error. The machine is very accurate in relationship to itself, but the moment I start relying on my own measurements the inaccuracies start to multiply. The answer is to always have the machine making the jigs as well, and checking the machine position against the digital model along the way.
I started by cutting the three pieces of the inner stem out. Let me be clear – the puzzle joints of these pieces are nowhere near what an accomplished carpenter could make on a bandsaw. They are, however, just about accurate enough for a nice juicy epoxy joint (between 1/16″ and 1/32″). I also cut a caul that would serve as the final shaping jig. This is a technique that I found in a book about making archtop guitars – basically, a female jig that holds the male part exactly while it is being worked on. The caul also ensured that the final glued assembly would match the intended shape and dimensions – it also happened to be the perfect amount of clamping pressure for the epoxy joint. The addition of packing tape underneath the piece was enough to create a nice snug fit for the gluing operation.
Setup for gluing in the routed ‘caul’
Not tight enough for furniture, but plenty of room for epoxy
After the glue dried overnight, I planed off the excess glue, and reset the stem blank for the next shaping pass. The important purpose of the caul (despite its usefulness as a gluing setup), is that it accurately locates the piece on the mill work surface for further operations. After a roughing and a surfacing pass, I had one side cut to the appropriate rolling bevel. To save some time, I chose not to surface the piece completely smooth – I figured a bit of “tooth” would be good for the epoxy.
The roughing pass
The second side was a bit more challenging – I had left two flat areas to the piece would sit level when flipped over (with the first cut face down), but I also needed to cut a mirror-image caul. I discovered when I clamped the piece in the new caul that it required a bit of leveling to make sure the uncut face was completely level. In the process of clamping it down with pits of MDF, I inadvertently discovered what I will call “MDF cam clamps” – basically, a small block of MDF is set against one side and screwed down to the work surface with a single drywall screw. When the piece has been thus clamped from all sides, I discovered I could tap the MDF slightly to the side to tighten everything up – the MDF is soft enough that it will deform before marking the stem.
The rough final finish is visible here. With 45 minutes (more passes), this would be quite smooth.
MDF Cam Clamp
The second side went off without a hitch. Now on to the keelson!