My aim is to make a single pack canoe 11’ 8” long by 27” wide and 10” deep (3.55m long x 685mm wide and 255mm deep). I always work in metric dimensions but for this project I have used a mix of imperial and metric – mainly building to the imperial measures in the plan and double checking every dimension in metric (because fractions make my head ache).
I made all the jigs etc described in the plans long before starting the canoe build, and the most interesting one was the steambox. I was loath to buy a half sheet of ply just for this, but I had some old bits of polycarbonate roofing sheet lying around and I thought that would be OK. It has three layers and so should hold the heat OK – and I will be able to see what goes on inside! The pictures tell the story of the construction. I bought a second-hand wallpaper steamer off Gumtree to produce the steam, making a hole in the bottom polycarbonate sheet for the hose and making a small plate attached with screws to prevent the hose dropping out.
When I came to use the steambox for the first time it worked OK. Steam arrived in the box about 10 minutes after switching on the steam unit and the temperature inside the box rose to ~97°C (206°F) within just a couple of minutes. I could clearly see the steam in the box and just about see the bits of wood too, through the condensation. The only issue was that the top layer of polycarbonate bowed upwards between the screws as it expanded more than the wooden frame, but I remedied that by slotting the attachment holes in the polycarbonate and putting washers under the screw heads to stop them pulling through. The tape covering the joints to limit the steam/water leakage did melt in some places but generally did the job – I only needed one bucket under the hose connection at the back to save a puddle on the workshop floor. However, if I were building a steambox again I would just buy the ply…
Cutting Gunwale Stock
I have a bench saw but I couldn’t manoeuvre the 2.5m long 2” thick boards into it accurately, so I elected to use a circular saw with an edge guide to cut strips off the side of the board. It worked OK. You can see in the photo that I placed the strips I had already cut on the other side of the board to provide some support to the circular saw base as the original board got narrower. I would usually borrow a planer-thicknesser from my father but with the covid restrictions I was unable to travel to collect it, so I just cut the gunwale strips as close to the final dimension as I dared and got the belt sander out to remove the worst of the saw marks.
I couldn’t get 2” thick Western Red Cedar (WRC) boards longer than 2.5m so I cut six strips and scarfed them together to make the three gunwale boards. I cut the scarf at 20° and bonded them together using West System epoxy thickened with 404 high density filler – I had this epoxy left over from a previous job. When I cut the gunwale boards to length I ensured that the scarf joints were all offset so they didn’t sit together in the lamination.
Measuring the Rocker
I have always found string lines awkward so I splashed out and bought a laser level (£55 from Amazon) and it has been brilliant. It self- levels, and projects vertical and horizontal lines.
First I used the horizontal beam to level my work surfaces. I then blocked up the marked gunwale board as Brian suggests and took measurements from the top of the gunwale to the laser line at each end and each rib position. Zeroing all the measurements relative to the measurement at the centre location gave me the rocker profile for the rocker layout sheet. It was possible to measure to 0.5mm on the steel rule. The only tip I would give is not to use your shiny steel rule for this as the reflected laser light can make your eyes sting – my brushed finish steel rule was fine.
I cut strips for the stringers and keel from 1” thick WRC boards in the same way as the gunwales – using the circular saw. I then ran them through the bench saw to reach final size. Again it would have been ideal to run them through the thicknesser but some hand sanding had to suffice.
I used my new slick plane to round the corners. Brian, if you every find yourself wondering what video to make next I would suggest ‘how to set up a slick plane’! What a fiddle! But when it is set, it is great. I found it easier to pull the plane rather than push but maybe I’m just a weakling. I did get carried away and rounded the keel too – oops. I’m never going to get that flush joint with the stems now…
Laminating the Gunwales
I used West System epoxy thickened with 404 high density filler for this because I already had it and I was nervous about getting the boards together and clamped in the short working time of Gorilla glue (which I hadn’t used at that point). Again, I used the laser level to set up my work surfaces and check the curvature of the gunwale stack when it was assembled. I placed my clamps at 5” spacing and I could have done with a couple more towards the ends (that thing about never having too many clamps – it’s true). Following Brian’s clear instructions meant that the whole process went well and I ended up with a good lamination. I took the excess glue off with a rasp next day and ran the lamination through the bench saw to create two gunwales.
Morticing the Gunwales
I have a double edge guide for my plunge router and so I used that rather than making Brian’s jig. I got a spiral router bit and it does make a huge difference to chip removal. Despite watching Brian’s video just before routing I immediately fell into the trap Brian had described; cutting outside the 1” markings and wondering when the next mark would appear…TWICE! Actually, the second time I cut between a rib mark and that mark I had made to show where the screw was, meaning that I also hit the screw (cheap screw – 1, expensive carbide router bit -0, ☹). After that I put waste marks (a squiggle) between all the rib mortice marks to make it obvious where material was to be removed, and I erased the marks above the screw locations. No more errors after that. I remedied the false cuts with some WRC blocks and Gorilla glue.
Again, I got the laser level out to measure the sheer heights of the finished gunwales.
Making the Stems
As mentioned earlier, I didn’t have access to a planer thicknesser so to take the 1” WRC board down to 11/16” so I set the blade height on the table saw to 8mm and ran the board over it to make lots of grooves. This then gave me a depth guide when using the electric planer and sander to get down to a reasonably consistent depth over the entire board. I could only get 6” wide boards in 1” thick so I ended up cutting a section out of each stem, cutting a tongue and groove on the table saw and bonding the two together to create sufficient board width.
Building the Deck
Building the deck went pretty smoothly. As you can see from the photographs, I made extensive use of the laser for ensuring everything was straight and plumb. This was particularly useful when setting up the stems; – I marked a vertical (7°) datum line on both of the stems and used the vertical line of the laser to get them sitting upright while lashing them to the gunwales, and more importantly while attaching the keel.
Setting the rocker I found a little frustrating/worrying. For some reason I moved the trestles part way through the set up and got different rocker measurements without changing anything else. I was measuring the rocker by taking measurements from the horizontal laser line at the bow, stern and centre. I would average the measurements I got at the bow and the stern (because the deck wasn’t completely level) and then subtract the measurement at the centre to get the rocker. The table below shows how much the rocker changed depending on where the trestles were located relative to the rib locations. This made me conscious that I needed to be careful and consistent with my trestle locations for the rest of the build.
|Trestle Positions (rib locations)||Measured Rocker (mm)|
|5 ½ and 18 ½||20.0|
|7 and 17||25.0|
|8 and 16||26.5|
I also found it rather disconcerting how easy it was to move the stems sideways despite them being tightly lashed on. I tried using pieces of card to stabilise them. This wasn’t very effective and the card got in the way, although it did reassure me during all that ‘sighting along the keel’ when bending in the ribs. I think there must be a better way of achieving a bit more stability in the stems during the build – I’ll think of something better next time… or perhaps you will!
Cutting Rib Stock
I learned more about timber grain in half an hour watching Brian’s video than I had learned in my whole life; another excellent bit of educating! Without it I would have made such a hash of cutting the rib stock. As a result I was able to identify that some of my timber was rift sawn and some flat sawn and modify my cutting appropriately. All my cuts were on the bench saw because I didn’t have a planer thicknesser but I got them pretty close to the required dimensions. I bound them with tape and bagged them in polythene because I wasn’t going to bend them immediately. This was a good move as the offcuts took up all sorts of strange shapes and became much stiffer within just a couple of days.
I trial bent ribs 2, 7 and 12 (middle rib). After watching the videos many times it went pretty well. I bent these three ribs using the three basic methods Brian describes without any breakage, albeit with quite a bit more ‘adjustment’ needed to get them to be symmetrical than Brian’s seemingly effortless style (OK, he’s done a few thousand more than me). I didn’t have a belt that I wanted to consign to the workshop so I used a piece of seatbelt fabric cut down to 1” wide – I found this much less extensible than tie-down straps and about the same as my (good) leather belt.
This was my first time bending any wood and it was so cool – to be able to shape the wood into tight bends without it breaking and then for it to stay in that shape when it cooled was a revelation! It left me quite positive about bending the rest of the ribs the following day as I shut down the steambox for the night.
Bending in the Ribs
I got set up as Brian describes and went for it. I used an interval timer, setting it to go off every 3.5minutes, giving each rib 7 minutes in the steambox. This worked well for me – three minutes to bend the rib, fit it into the gunwales and adjust it was just about enough. The ribs were pretty symmetrical, I just struggled getting them to come up to touch the keel by the same amount. I broke Rib 1, proceeded along the boat, took a break and refilled the steamer after Rib 12 and went all the way to Rib 22 before breaking Rib 23 – all in just over an hour.
Then came the frustration. I re-cut Ribs 1 and 23, and then broke them again. I just couldn’t get them to that tight radius without them breaking. Using one of the broken ribs as a template I made a ½” ply form shown in the photo above. I cut more rib stock and bent a rib over the apex, pulling it into shape with the strap and getting my wife to position the clamps – that rib broke too.
Eventually I had some success (photos above) bending the rib stock around the form in a different way; I clamped the rib on the long edge and while pushing the rib stock and pulling the strap managed to form a rib with minimal break out at the apex. The key, I think, was to put as much compression along the rib stock while encouraging it round the apex. After a few minutes of cooling I released the clamps and the strap, put some gorilla glue in the broken out areas (with plastic film to stop any adhesion to form or strap) and again pulled the strap tight and secured with clamps for a couple of hours. This gave me a rib of the required shape which I could subsequently cut to length. I made a second rib for position 23 in the same way. Again, pushing the rib stock into compression while bending it round the apex of the form resulted in only a small amount of (repairable) breakout rather than catastrophic breakage I had seen every time I tried bending free-hand. I think I spent about 11 hours getting those two endmost ribs bent, all told.
The minor breakout on the other ribs was fixed with a little gorilla glue and local clamping.
On reflection I think that my rib stock, while OK, was just not good enough for those very tight bends at the end ribs. The weakness always appeared to be where the growth rings/lines were spaced a bit further apart – was that a good summer, short winter, wet year for the tree? More timber research needed into that I guess…