JeffH
03-10-2003, 10:57 PM
Ask and ye shall recieve....... Eventually! Some weeks ago I was asked to post how we had cut the wedge dovetails in our deck beams, so without further ado (there having been enough ado already), here's my best shot. If you want to ignore my stumbling attempts at technical description and cut to the chase, scroll to the end.
It was originally thought that this would be a fairly simple, straightforward operation. The "normal" way of attaching deck beams to carlins is to cut a sloping notch in the carlin, about an inch across, about an inch in at the top, sloping down the two inch thick carlin to flush at the bottom, and square on the sides. The coresponding male part is just a one inch wedge in the end of the deck beam, with about a half-inch shoulder on either side. The male and female parts fit snugly so all seams are tight and a fastening or two driven in from the inside (if you're Bud McIntosh, three galvanized nails). I'm talking in general terms here, not true demenions, since it's easier to visualize and easier for me to describe.
The Boss, being fond of dovetails, and since it was his time and money anyway, decided to do this one better and cut not a square notch but one that was dovetailed. This was originally thought to be a fairly simple and straighforward operation, consisting of making a sloping jig to fit on the carlin and using a dovetail bit instead of a square-cut in the router. The male part would be just cut as with a normal dovetail but finished with a cut on the sliding compound mitre saw so that it would have the appropriate slope. This was done with some scrap pieces of appropriate size, and it was found that no amount of turning things every which way would make the two parts come together in anything resembling the right way. Lots of head-scratching thus commenced, and all parties retired to the doodle pad to figure out why this was not working. Doodle #1 reproduced here:
http://www.imagestation.com/picture/sraid53/p0fd0dd6dede8bf6559787e747c157f91/fc8e8129.jpg
This is a drawing of what we were creating on the female end and what needed to be made on the male end to fit, as seen from the top, front and end-on. For the female part, the router was run up and down in a rectangular fashion with parallel sides. While this sounded good in theory, the shape of the router bit was making things difficult. A dovetail bit is, obviously, narrow at the top and widening towards the bottom. When the cut was started, the narrow part was near the surface, and as the router moved down, the bit basically rose out of the wood 'till the cut reached the bottom and the widest part of the bit was at the surface. This created a notch that was contstricted at the top and opening up in the bottom. Similarly, the male part was cut with parallel router movement and trimmed off with the mitre saw, and thus created a dovetail that was wide at the top and narrow on the bottom. This is the direct opposite of what was needed to make things fit, so it was back to the drawing board to figure out how to straighten it all up. Thus, we now have doodle #2:
http://www.imagestation.com/picture/sraid53/pdbbee7a3610b39c22ec92d8d7c0664f2/fc8e8122.jpg
What we hit on was what is, I suppose, the fundamental Principle of Dovetails. Take, for example, a simple, square, mortise-and-tenon dovetail. On the female side, the bit depth is set and the router face is placed square against the beam. The router is moved up and down vertically until the notch is the proper width and the sides are perfectly parallel. Then, the router is turned around to the end of the beam to be fitted and again moved up and down vertically on either side of the beam end, creating a tenon that will, hopefully, fit perfectly in the notch. The critical part is that to create a male and female part that are the same shape, the router must move on the same relative plane for both peices. Here, then, was our failing. For the notch, the router was set up to move at an angle to the face of the carlin, cutting a sloping mortise. The tenon was cut square to the beam end, which would be also square to the face of the carlin, and thus at a DIFFERENT relative plane. Aha! Now, what to do? Keeping the router at the same angle for cutting both ends creates some obvious difficulties. We could cut the mortise as we were, with the slope, but cutting the tenon at the same angle would mean we would be cutting away the shoulder of the joint and defeating ourselves. On the other hand, if the router is turned horizontal and the tenon cut vertically, the mortise would also have to be cut the same way, eliminating the slope. Commence, then, some serious head-banging.
Another day (and a few aspirin) later, The Boss hit on this very elegant solution:
http://www.imagestation.com/picture/sraid53/p87d8290fafb94587e091f8998f03309f/fc8e811f.jpg
The sides of the sloping carlin-cutting jig were angled in, creating a V-shaped notch that, when viewed from the front, had a parallel face. This, essentially, dovetails in two directions. Very, very, neat.
The tenon was cut as before, with the router set square to the beam end, moved up and down vertically and parellel on either side. For those of you still able to follow along, this does indeed violate The Dovetail Principle. It was found, however, the difference was inconsequential, and the fit is beautifully tight.
http://www.imagestation.com/picture/sraid53/pc1f1da92b7f4fea4f37c43af97a25f66/fc8e811e.jpg
Now for the good stuff (those of you ignoring the babble can start paying attention again). Here's the jig set up to cut the carlin:
http://www.imagestation.com/picture/sraid53/p5f4edc5e5e2aac6d1171211ce600fdc5/fc8e8126.jpg
You'll note there are a couple wedges on the top and side. The top one adjusts for the fact that the deckbeams are plumb on the sides, and the carlin is following the sheer. Since the jig is now tilted, the slope of the cut won't end up square and flush at the bottom (this is hard to describe... One of those things you have to try yourself and see), so the wedge on the side brings things back square. If one was really clever and doing a lot of these, I suppose it could be set up so that one of the sides would slide up and down to adjust the jig so that it was perfectly square vertically. This should, in theory, eliminate the need for the side wedge.
If I was diligent and thourough, I would also have a picture of the other jig. Sadly, I don't, so you'll have to put up with my description again. Fear not, though, it's fairly simple. Take a bevel guage and find the angle that the deck beam intersects with the carlin. There will actually be two; the carlin is probably curved, so the horizontal angle for each deck beam will be different. There is also camber in the deck, and since the carlin is curving, each deck beam will intersect it at a different point in the camber, resulting in a different vertical angle (boats are such a pain in the butt in this respect). Using a sliding mitre saw, cut the beam with these angles. Now, take a peice of plywood with perfectly square sides. Cut a rectangular hole in the middle of it so the deck beam end will fit in it with room to spare all the way around. Find a way to fasten it to the beam so it is flush and square with the newly cut beam end (either fasten the plywood to the bench and move the deck beam around 'till it's square on the face, or (as we did) clamp the deck beam to the bench and move the plywood around). With the dovetail bit in the router, set the router face flat on the plywood and cut either side of the beam until the tenon is at the proper width. You can use a router fence and run it on the sides of the plywood (our method), or fasten a couple strips of wood to the face of the plywood, parallel to the side, and at the proper distance apart so the router will cut to the proper width (this will probably take LOTS of trial and error). Take your deck beam back to the mitre saw and cut the tenon to the proper slope to make it wedge-shaped. Voila!
I'll now cringe slightly as I await Dave's blasting of my terminology. If there's anything that I need to clarify (like the whole thing, I hear you say), just shout and I'll do my best!
Jeff
[ 03-11-2003, 05:55 PM: Message edited by: JeffH ]
It was originally thought that this would be a fairly simple, straightforward operation. The "normal" way of attaching deck beams to carlins is to cut a sloping notch in the carlin, about an inch across, about an inch in at the top, sloping down the two inch thick carlin to flush at the bottom, and square on the sides. The coresponding male part is just a one inch wedge in the end of the deck beam, with about a half-inch shoulder on either side. The male and female parts fit snugly so all seams are tight and a fastening or two driven in from the inside (if you're Bud McIntosh, three galvanized nails). I'm talking in general terms here, not true demenions, since it's easier to visualize and easier for me to describe.
The Boss, being fond of dovetails, and since it was his time and money anyway, decided to do this one better and cut not a square notch but one that was dovetailed. This was originally thought to be a fairly simple and straighforward operation, consisting of making a sloping jig to fit on the carlin and using a dovetail bit instead of a square-cut in the router. The male part would be just cut as with a normal dovetail but finished with a cut on the sliding compound mitre saw so that it would have the appropriate slope. This was done with some scrap pieces of appropriate size, and it was found that no amount of turning things every which way would make the two parts come together in anything resembling the right way. Lots of head-scratching thus commenced, and all parties retired to the doodle pad to figure out why this was not working. Doodle #1 reproduced here:
http://www.imagestation.com/picture/sraid53/p0fd0dd6dede8bf6559787e747c157f91/fc8e8129.jpg
This is a drawing of what we were creating on the female end and what needed to be made on the male end to fit, as seen from the top, front and end-on. For the female part, the router was run up and down in a rectangular fashion with parallel sides. While this sounded good in theory, the shape of the router bit was making things difficult. A dovetail bit is, obviously, narrow at the top and widening towards the bottom. When the cut was started, the narrow part was near the surface, and as the router moved down, the bit basically rose out of the wood 'till the cut reached the bottom and the widest part of the bit was at the surface. This created a notch that was contstricted at the top and opening up in the bottom. Similarly, the male part was cut with parallel router movement and trimmed off with the mitre saw, and thus created a dovetail that was wide at the top and narrow on the bottom. This is the direct opposite of what was needed to make things fit, so it was back to the drawing board to figure out how to straighten it all up. Thus, we now have doodle #2:
http://www.imagestation.com/picture/sraid53/pdbbee7a3610b39c22ec92d8d7c0664f2/fc8e8122.jpg
What we hit on was what is, I suppose, the fundamental Principle of Dovetails. Take, for example, a simple, square, mortise-and-tenon dovetail. On the female side, the bit depth is set and the router face is placed square against the beam. The router is moved up and down vertically until the notch is the proper width and the sides are perfectly parallel. Then, the router is turned around to the end of the beam to be fitted and again moved up and down vertically on either side of the beam end, creating a tenon that will, hopefully, fit perfectly in the notch. The critical part is that to create a male and female part that are the same shape, the router must move on the same relative plane for both peices. Here, then, was our failing. For the notch, the router was set up to move at an angle to the face of the carlin, cutting a sloping mortise. The tenon was cut square to the beam end, which would be also square to the face of the carlin, and thus at a DIFFERENT relative plane. Aha! Now, what to do? Keeping the router at the same angle for cutting both ends creates some obvious difficulties. We could cut the mortise as we were, with the slope, but cutting the tenon at the same angle would mean we would be cutting away the shoulder of the joint and defeating ourselves. On the other hand, if the router is turned horizontal and the tenon cut vertically, the mortise would also have to be cut the same way, eliminating the slope. Commence, then, some serious head-banging.
Another day (and a few aspirin) later, The Boss hit on this very elegant solution:
http://www.imagestation.com/picture/sraid53/p87d8290fafb94587e091f8998f03309f/fc8e811f.jpg
The sides of the sloping carlin-cutting jig were angled in, creating a V-shaped notch that, when viewed from the front, had a parallel face. This, essentially, dovetails in two directions. Very, very, neat.
The tenon was cut as before, with the router set square to the beam end, moved up and down vertically and parellel on either side. For those of you still able to follow along, this does indeed violate The Dovetail Principle. It was found, however, the difference was inconsequential, and the fit is beautifully tight.
http://www.imagestation.com/picture/sraid53/pc1f1da92b7f4fea4f37c43af97a25f66/fc8e811e.jpg
Now for the good stuff (those of you ignoring the babble can start paying attention again). Here's the jig set up to cut the carlin:
http://www.imagestation.com/picture/sraid53/p5f4edc5e5e2aac6d1171211ce600fdc5/fc8e8126.jpg
You'll note there are a couple wedges on the top and side. The top one adjusts for the fact that the deckbeams are plumb on the sides, and the carlin is following the sheer. Since the jig is now tilted, the slope of the cut won't end up square and flush at the bottom (this is hard to describe... One of those things you have to try yourself and see), so the wedge on the side brings things back square. If one was really clever and doing a lot of these, I suppose it could be set up so that one of the sides would slide up and down to adjust the jig so that it was perfectly square vertically. This should, in theory, eliminate the need for the side wedge.
If I was diligent and thourough, I would also have a picture of the other jig. Sadly, I don't, so you'll have to put up with my description again. Fear not, though, it's fairly simple. Take a bevel guage and find the angle that the deck beam intersects with the carlin. There will actually be two; the carlin is probably curved, so the horizontal angle for each deck beam will be different. There is also camber in the deck, and since the carlin is curving, each deck beam will intersect it at a different point in the camber, resulting in a different vertical angle (boats are such a pain in the butt in this respect). Using a sliding mitre saw, cut the beam with these angles. Now, take a peice of plywood with perfectly square sides. Cut a rectangular hole in the middle of it so the deck beam end will fit in it with room to spare all the way around. Find a way to fasten it to the beam so it is flush and square with the newly cut beam end (either fasten the plywood to the bench and move the deck beam around 'till it's square on the face, or (as we did) clamp the deck beam to the bench and move the plywood around). With the dovetail bit in the router, set the router face flat on the plywood and cut either side of the beam until the tenon is at the proper width. You can use a router fence and run it on the sides of the plywood (our method), or fasten a couple strips of wood to the face of the plywood, parallel to the side, and at the proper distance apart so the router will cut to the proper width (this will probably take LOTS of trial and error). Take your deck beam back to the mitre saw and cut the tenon to the proper slope to make it wedge-shaped. Voila!
I'll now cringe slightly as I await Dave's blasting of my terminology. If there's anything that I need to clarify (like the whole thing, I hear you say), just shout and I'll do my best!
Jeff
[ 03-11-2003, 05:55 PM: Message edited by: JeffH ]