Roubo leg joint questions (2)

[ian maybury]

Jake Ekins posted on a similar topic a while ago when he asked about how much shoulder area was required, but the subject never really got going. I've floated it before too. I'd appreciate some thoughts on the following, especially from those using a Roubo for some time - it's getting down to the wire on my starting the build.

The traditional Roubo joint letting the legs into the top combines a through tenon, and a dovetail keyed lap joint let into and flush with the edge face of the top slab - the leg in effect has double tenons. As here: http://www.popularwoodworking.com/techniques/joinery/a-j-roubos-sliding-dovetail-tenon-joint

A key feature of the Roubo design is that the top slab doubles as the top trestle, the leg joints must as a result resist strong racking forces in use. The heavy top means that even moving the bench across a floor could create quite heavy loadings.

Some of the more modern variants use a (sometimes quite short - maybe 50% of the top thickness) blind tenon and oak dowel pins to make the joint. There's lots been built this way, and it presumably works just fine in that problems don't seem to be reported. This joint has the advantage of being hidden and so visuals are not a factor, and it lends itself to disassembly for transport.

On the other hand a through tenon is clearly going to offer more resistance to racking - given that the full length of the tenon bears on the sides of the relatively longer mortice in the top. After this the dowels are only be required to stop the joint sliding apart. It creates the opportunity too for a nice visual feature, but only if it's neatly done - although movement in the top could cause the tenon to project or to retract at times. Wedging of the tenon looks nice, but presumably eliminates the possibility of disassembly for transportation.

One bit that's bothering me is that I'm struggling to see what the traditional outer dovetail keyed lap adds to the joint. It's open on one side, is surely not very strong, and risks splitting away the edge of the top should it become heavily loaded. My sense is that it's mostly for show, and I'm not that bothered about following tradition - but maybe I'm wrong.

The result is that I'm considering heading down the single large tenon road this guy followed in his very nice bench build: http://chesapeakewood.wordpress.com/category/roubo-workbench/ One major advantage of this approach is that both the mortice and tenon can easily be formed while gluing up the legs and the top - which sounds like a great idea. I'm even thinking of wedging the tenons to get the visual benefit, in that realistically I can't help thinking that with a 4in + top a fully assembled bench is only marginally less transportable than one with removable legs anyway.

Plan B on the other hand is possibly to go with as long a blind tenon as can reasonably be fitted in the top without leaving a weak patch over the joint. This should deliver quite a lot of the strength of the through tenon, but avoid any potential issues caused by movement of it relative to the bench top.

It's not a show stopper, in that left to it I'll make a call and get moving. Your thoughts would be appreciated though as this isn't a road I plan on traveling again...

Thank you

ian

[Stephen Cherry]

Ian- I'm wondering what your objective is with this. Are you building a bench as an end in itself, or do you just want a solid platform for hand tool use? It seems to me that, unless it's just something that you want (and I am not critisizing this in any way) a 4" top is not necesary.

As far as the racking forces go, given reasonable height stretchers, the top should not see any racking from moving a bench because the stretchers should be strong enough to absorb all of this. I had a bench with a maple top (maybe 2.5 inches) and softwood bottom. The bottom was two trestles made of 4x4 construction lumber, with 2 x 8 stretchers through mortised and wedged into the trestles. For the moderate hand tool work that I did on it, it was beyond rock solid. I've got to think that with a hardwood base of similar dimensions, the bottom would be extreamely stiff, regardless of the joint from top to bottom.

Here is a fun video:
http://www.pbs.org/woodwrightsshop/video/2700/2705.html

[Mike Zilis]

Ian,I can completely relate to your thought process. I've got all of my bench materials ready and while I'm not making a strictly Roubo bench, I would consider it a Roubo hybrid or at least Roubo like. The only design decision I haven't made yet is how to attach the top to the base. I'm not against having an upper stretcher and still mortising the legs partially into (or through) the top. An upper stretcher can be both functional for me as a shelf or tool cabinet support and practical as to mitigate racking. I checked out the Chesapeake blog last week and really liked how accurate and simple the through mortises could be done. This is likely how I'll attach the top to my base. Mike

[David Kumm]

I built my Roubo using KD SYP 2x10 with a 4" top. Since the boards were all milled to 1.25 I glued two shorter pieces to the legs and then trimmed the cheeks, and dadoed the two table boards and glued them together. The legs consisted of 4 boards so the tenon was 2.5 and approx 2.5" deep. I drawbored the tenon and probably glued it. Don't remember but know I will never knock the bench down. Need a forklift to lift it anyway although I can move it out from the wall. The joint is so solid and has absolutely no movement. Racking is not even close to an issue so whether you go through or not I don't believe there is a functional difference. Whatever makes you feel good when you use the bench. Dave

[ian maybury]

Ta Stephen. Roy Underhill is entertaining, and has a pretty light approach to his bench making. That said there's probably a lot more work goes in behind the scenes then seems to be the case in videos that are pretty heavy on pizzazz.

To paint a clearer picture of the context in which I'm asking about Roubo leg joints. I have the materials for a solid 4in+ top bench in beech along the lines of the Chris Schwarz Roubo. The use is mixed hand and machine tools, but with a requirement for lots of clamping flexibility (a decent grid of dog holes - more than the single row used for traditional hand tool work) for assembling bent wood laminated frames. Tradition and bling are not big factors for me, so I'm mostly interested in subtle visuals and straight up function and performance in a well and neatly made bench.

There's loads of proven designs about including Chris' and the Benchcrafted which use blind tenons - some are maybe only 60% of the thickness of the top. These do seem to work.

Against that I don't want to run into subtle loosening problems or the like down the road, and think that a wedged through tenon better resists racking, and like the look. So that's probably makes it front running option.

I know too that the top could be thinner, but I have a hankering after the rock solid feel that comes from a thick top.

I have the equipment and skills so that pretty much any of the designs are feasible, and maybe most of all I know there's a wide variety of ways the bench could be built, all of which will likely be just fine from a structural point of view. I guess though I'm into striking what for me is an optimised balance - I'd like to select a leg joint arrangement that appeals to my preferred mix of function/ease of build and understated style.

This means that much like Mike I'm not for example very interested in including the second dovetail keyed lap/tenon to the front of the more traditional versions unless it adds something, and am attracted to the Chesapeake single through tenon arrangement shown in the earlier link for its mix of function, ease of making and low key but attractive visuals.

Adding a second stretcher or cross rails is as he also says functionally an option, but would seem a pity on a minimalist design like a Roubo unless it's really needed. So barring that somebody can say that the second keyed dovetail lap/tenon in the traditional Roubo leg joint adds something the Chesapeake single through tenon sounds like the plan.

My one caution about any arrangement involving through tenons is that there's potentially .010in vertical seasonal movement in a thick top, which suggests that a tenon may potentially project above or drop below the surface by that much. Which if it really does happen in practice would be a pity as it's likely to be right in the line where small parts will be clamped between dogs for planing.

Depending on how big an issue this is likely to be in practice it might just suggest that blind tenons are actually the best choice.

So the question in simple terms Stephen is: given these requirements and preferences - which is the best leg joint in a Roubo bench? The single through tenon, a blind single tenon , or maybe even the traditional double through tenon with dovetail wedged lap to the front.

Guess I should have been able to say that in about ten words....

ian

[Adam Slutsky]

Ian, I am also making a Roubo like bench and want it to a "Knock Down" version. I think that a single stretcher mortise and tennoned will be enough to prevent it from racking provided the tenons are stout. You could use bench bolts to attach the stretcher to the legs or attached using through tenons and wedges. If in doubt, you might consider a second stretcher dovetailed and attached with a lag bolt as per the 21st century work bench (popular woodworking October 2008 #171). The top can then sit on the base and be attached using dowels or non through tenons. A 4" top is not going to move.

[ian maybury]

Thanks for the inputs guys, it sounds like blind dovetails and pins are plenty strong enough.

Pardon my caution, but the size and weight of these benches moves them into a different space to more typical work.

Plus there's a lot of differing approaches about - this guy in Canada built a very nice Roubo literally from the ground up. It looks traditional, but all of his stretcher to leg and leg to top joints are dovetails enclosed by leg and top laminations. http://forum.canadianwoodworking.com/showthread.php?40372-Roubo-Bench-Build/page6&highlight=roubo Even the top was built by gluing the laminations up one by one in situ around dovetails formed on the top of the legs. Very creative, and in many ways easier than the usual 'build a top and then a leg assembly, then assemble (with lots of kicking and screaming) the two.' Definitely not a knock down model if he ever needs to get it out and he finds it won't fit through the doors though.

What's interesting is that while the popularity of certain formats creates the impression that the thinking has settled on a few alternative leg joint layouts, the rationale for each isn't typically set out by those using them.

I'm sitting down to draw mine up in a week or two , and will make final calls then.

ian

[Jake Elkins]

The inner engineer in me has too questioned the mechanics of the blind tenon in a thick top.

In order for racking to occur, the tenon in the top must be capable of rotation. Therefore, any effort used that minimizes the ability for the tenon to rotate will mitigate some racking potential. There are therefore a few design considerations:

1. Clearance between the mortise and tenon. Obviously the tighter the fit, the less ability for the tenon to rotate. With everything else controlled for, rotation of the tenon in a near piston-fit mortise requires deformation of the top (and/or tenon). Thus, choosing a wood with a high modulus will reduce racking potential.
2. Friction between the walls of the tenon and the mortise. The greater the frictional force, the greater force (moment) required to cause racking. The obvious choice here is gluing the joint.
3. Since frictional force is proportional to surface area, a larger tenon will reduce racking more than a smaller. This includes making the tenon as long as possible without compromising the strength of the top.
4. Rotation resistance will lessen if vertical displacement of the tenon is allowed. Ways to prevent vertical displacement include: (1) pegging/wedging the joint; (2) Gluing the joint; and (3, indirectly) including a shoulder on all four sides of the leg tenon. This last point seems to be largely ignored in several of the the recent Roubo build guides (Schwarz included), but I think that having a four-shouldered tenon will considerably help.
5. When using stout stretchers (pegged, glued, well-fitted joints), you require that all four top joints act in unison. As with the chain, the bench will only be as strong as the weakest joint. Sufficient stretchers help in this regard. However, doubling or tripling the number of stretchers will not completely compensate for sloppy joints at the top.

Again, these are just some of my own conclusions after several thinking sessions and soliciting for advice here (thanks!). For my own bench, which is still a WIP, I have decided on the following:
1. 2.75" long tenons (4.25" top). 1 3/8 shoulders along the long grain, 3/4 shoulders cross grain. 5.5" x 5.5" legs
2. Drawbored and glued. I realize that gluing the joint prevents me from ever dissembling the bench. But, after drawboring the joint, isn't this really a mute point?
3. Single set of 5"x2.5" stretchers

I am confident that these precautions will suffice for a lifetime (and then some) of abuse. But time will tell.

[ian maybury]

As it happens Jake I'm an engineer - which probably explains some of my nitpickery.

Barring a problem getting the bench through a door, my sense is that disassembly for transport isn't going to be that important - it's not going to reduce the volume all that much, and the weight will be the same.

The game on a blind tenon may be to as you say get a good close fit, having made it as long as possible while leaving enough material left above the mortise to handle the hammering or whatever it may experience.

The problem with relying too much on the shoulder and the pin/dowel is that there would probably be a tendency when the leg is loaded in e.g. the direction of the long dimension of the table top for it to pivot about the far corner of the shoulder. Which not only will apply force to the dowel, it'll do it through a very high leverage ratio. (the distance from the corner of the shoulder to the dowel is going to be only a small fraction of that to the end of the leg where the load is probably applied.

If on the other hand the tenon and mortice engage over a decent length it'll be the side faces of this joint that will mostly bear the load, and if the fit is good and tight it'll minimise any tendency for the above to happen.

It's not very woodworking, but if the expectation was that the pin might be required to take a significant loading there's possibly a case for using a lagbolt or even a socket head screw. (which has the advantage of having a small head - it's easy enough to tap wood). It'd always be possible to glue a cap over the top of it. The offset in the holes used to draw the joint up tight would probably need to be reduced compared to oak dowel though......

ian