Laminated Roubo Bench
Over the years my off-the-shelf bench gradually had more and more bits screwed and bolted to it to try to make it stiffer and less prone to racking. In the end it also had boxes of tools and nails on the undershelf to weight it down but I still ended up dragging it back to the start line after any hard planing. The cost of a replacement was the issue – not the cost in itself but more that I’d rather spend the money on materials for future projects than on a bench. The solution came in the form of a skip full of joists after part of my house was demolished. They were brought into the workshop and hoarded for three months to dry out a bit and acclimatise. A slab-top laminated Roubo was the only choice to build since I’d seen it in Chris Schwarz’s workbench book. I could have created the frame to top joints by leaving the through mortises in the lamination but somewhere in the back of my mind I had the idea that this was a trial run for a really nice bench in hardwood later with a proper slab top if the right timber ever finds me. The best sides of the boards were easy to find. These were the ones without thousands of nail holes. I thought about patching the other defects, but then realised I’d rather be making furniture than patching knots on a bench. I did take the time, though, to align the board grain direction from right to left in the final benchtop for flattening in the future. I used dominos to keep the boards aligned and glued up three or four at a time. I laminated the slab up to the width of my planer thicknesser – 410mm – and was able to get this flat and of even thickness with one pass over the machinery and one visit to the chiropractor.
Creating the dog strip
The next stage was to create the dog strip and in a gung-ho moment went straight to the drill press and put a line of 20mm holes 75mm on centre for the whole length of the bench. Then I realised the legs would have blind dog holes in them – I’m sure there’s an expensive solution for that somewhere. I had a spare vice screw from the old bench and wanted to use it for a wagon vice on the new bench. I took the threaded nut and cross drilled and threaded it so I could attach it to a vice block. I then machined grooves in the front strip and the strip between the dog strip and the main slab to take the block. Shellac and wax on all the moving part surfaces before gluing stopped any squeeze out being a problem when gluing up to the main slab.
The wagon vice needed an end cap that would allow the main slab to expand and contract and also resist the movement of the vice. The end cap was dovetailed to the front strip and a large groove was cut to make the rest of the end cap into a giant breadboard end for the slab. The rest of the end cap was secured to the slab with 100mm long, 10mm coach screws through slotted holes.
The wagon vice sub assembly was glued up with the dovetail first and then the dog strip and backer strip. The vice traveller block had to go into the assembly at this point. Unlike some wagon vices this isn’t bolted in from underneath. The main slab was cross cut at one end to give a neat end for a tenon cut with a track saw. The track saw was then used guided across the end and the side to create the tenon. After fine tuning, the slab and L-shaped vice assembly of three strips was glued and the end cap screwed to the main slab. The end cap joint was left dry to allow expansion and contraction.
On the level
After the end cap and wagon vice assembly was fitted it had to be brought down to
the same level as the rest of the top and given an initial flattening before the leg joinery. I used a Sergeant No.3 set up as a scrub plane to waste most of this away followed by a jack and jointer planes. With an 8in radius on the scrub plane it was very quick.
Legs and stretchers
The old joists from the skip had provided the material for the slab top but I was still short of four legs and stretchers. Searching the skip came up with a lot of 2 x 6 and 2 x 8 offcuts and these were laminated into 5 x 5 legs and 5 x 3 stretchers. I made sure to cut these well oversize and make the depth of the shoulder between the two tenons dictate the height of the top. I wanted a tight joint and enough excess to be able to beat the joint apart from the topside without damage to the top. Roubo has a fairly steep angle on the sliding dovetails at the top of the legs. I was going to use the bandsaw to cut these and to accommodate the tilting table I made these 45º. I could then cut from either face of the leg and be sure to get a consistent result. It also meant the marking out was simpler as the same gauge setting could be used from each face. With the cheeks cut I cut the dovetail shoulders with a hand saw and the waste between the dovetail and tenon with a chisel, undercutting a little in the centre.
With the slab upside down on sawhorses the line of the legs was established with a knife line and square from the front edge. The legs were spaced to put a dog hole squarely in the middle of each. With the weight of the legs they didn’t need to be held in place to trace a knife line around each. The bevel of the knife was held against the leg to give a very shallow line. The legs were then removed and the lines used to set gauges. The gauges were then used to deepen the lines and transfer the lines to the other side of the slab. Squares were used to transfer the sides of the leg and the sliding dovetail. After the layout was complete I chamfered the tops of all the tenons and sliding dovetails to prevent blowing out the top on dry assembly. With the slab on edge the sides of the sliding dovetail were established with a saw. The knife line was deepened to a bit of a ‘knife wall’ popularised by Paul Sellers. One too-deep knife wall made the saw cut off line as the 45º cut meant the deeper the knife wall, the further from the line it’s possible for the saw to be.
Legs and stretchers
Once the sides of the sliding dovetail were established it was easy to whack out the waste between as the strength of the material was gone with the saw cuts, creeping back to the line to keep it neat and using the gauge line for the final paring cut. Cutting into laminated material means that normal wood behaviour of a split running back, possibly past the gauge line, doesn’t necessarily apply. The through mortises were tried first with a Forstner bit in a cordless drill. This worked fine but there was a lot of subsequent chisel work. For the last three mortises I used a Festool Domino with a 14mm bit. With a little marking out on centre lines this gave a mortise within a hair of the gauge line and hardly any paring of the sides of the mortise. The scale of the joints between the legs and top mean that sometimes unusual tools can be used. To clean up the corners of all the mortises I used a small Japanese saw with a front raker tooth. This efficiently got all the fuzz out of the corners.
The dry fit required the use of as big a hammer as I could find. The 2.5lb dead blow hammer was very useful here. For reversing the dry fit I found the deadblow from below onto the oversized tenon got the joint open enough to get in a crowbar. The crowbar gave a lot more control to the movement than just whaling away with a huge hammer – fun though that was. With all four legs dry fitted and in place I cut two 2 x 4 battens to help in marking the stretchers. The battens were clamped to the legs and the stretchers were clamped on top of the battens. The shoulder lines were then marked out directly. Any small error in the leg position is allowed for by cutting to fit in this way. Legs were then removed from the top – easier to say than to do – and mortised. I used a 40mm Forstner bit in the drill press with a fence to ensure all the mortises were identical. The corners were then chiselled square.
The stretchers were cut to length and shown to the mortises. The tenon size was marked directly from each mortise and the marks transferred around to the shoulders. The cheeks and shoulders were then cut out by hand and tuned with a router plane to an easy sliding fit. The legs were then drilled for two drawbore pegs each taking care to avoid hitting the peg from the adjacent mortise. The joints were then reassembled and the drill bit used to mark the centre of the holes on the tenon. The joint was disassembled and the mark transferred slightly closer to the tenon shoulder. How far the mark is transferred can vary and I make a test joint for any new size/timber combination and see what works. A full 1/4in offset was too much but just under worked fine.
I then made up 20 or so 10mm drawbore pegs from straight grained ash. Several spares are needed as not all will make it and sometimes enthusiasm will break some. Making pegs with a dowel plate means an even bigger hammer. This time a 4lb lump hammer was the best tool for the job. A larger hammer allows the same impact force to be applied with more control in the same way a sharper tool allows more control as there is less effort applied. The ends of the pegs were tapered on the disc sander and the holes in the tenons were chamfered slightly with a countersink. These all help the peg find its way through the tenon and out the other side. Drawbore pins can be used to test the fit and ease the path for the peg but I don’t have any and have never found them necessary. The benefit of the drawbore technique, apart from its huge strength, is that it allows joint to be a little loose and a frame such as this to move a little to allow the top to go on. With all the joints and the sliding dovetail/mortise and tenon for the leg/top joint glued up with fairly slow setting glue I put the top onto the frame. The best hammer for getting the top onto the joint is the top itself. Picking up each end and dropping it on the floor very quickly and effectively drives the top onto the legs. There are videos on YouTube of this process and although fun it shouldn’t be done in sandals. A light application of alcohol to the respective parts helps the process. Once the top is on and the clock is still ticking on the glue, the drawbore pegs go in. The pegs are glued and then driven into the holes. The peg starts going straight in, then dives in towards the stretcher, then back the other way. It’s important to keep hitting squarely and not snap the peg at this point. The best tool I found for this was the 4lb lump hammer with a good swing. Looking at people who swing hammers for a living – blacksmiths, those men with the huge mallets that install council paving slabs – they seem to have a technique of using the weight of the hammer to do the work with a swing that comes from the shoulder or elbow rather than just the wrist. When the drawbore pegs came through the other side the glue was left to set and then they were trimmed flush. The front face of the frame was then planed flush and a fairly heavy chamfer cut with a chisel, spokeshave and block plane on all the front arrises. This was to stop the sharp edges attacking me or my work in the future.
The end result
The face vice for the bench is my old record 53 bought for £10 many years ago and probably far older than me. I cut the majority of the mortise for the rear jaw with the domino and finished up with an auger bit and chisel for the final depth. An old bit of teak (Tectonia grandis) – I think – made a nice chop for the vice and this also hid the saw marks in the front jaw. Lined with leather for grip and finished with a thumbnail ovolo on each end and an 1/8in bead top and bottom. I bought four new dogs for the new bench from Veritas. The prairie dogs make use of the blind holes over the legs. The solid brass dogs I’ve put in the wagon vice and the dog strip. To aid pushing these up I’ve put a heavy chamfer on the underside of the holes to make it easier to get a finger in. I saw this on the Instagram site of Ed Sutton and thought it was a really good idea. With the bench all done I filled the worst of the defects on the top and gave the whole thing a quick coat of shellac to seal it without making it too slippery. First impressions are that it is rock solid and should last for years; at least until the perfect bench top slab finds me.