Roubo Build: Order of Operation

[Brian Holcombe]

The dovetail joints are really cool, that is one of my favorite parts of the Roubo bench.

It's very practical in use if you want the leg to remain flush with the front of the bench.

[ian maybury]

Hi Curt. Sorry, just saw yours now. I'm not an authority in that i'm just starting my own bench build - but i'd be cautious about using a simple lap joint in place of the dovetail where the outer face of the leg passes up through the top. My rationale is just that it's very easy for a strip of wood to bow a little, and if it bowed outwards (and if used glue wetting the inner face might be enough to create some tendency to do this) there would be nothing to mechanically hold it in place. Each joint is composed of three facets each having the grain at 90 deg to each other - so if glued it might or might not hold up long term. The dovetail if well fitted would presumably be pretty effective to mechanically hold it in. It also provides extra sidewards location of the top.

The Benchcrafted split Roubo is probably the design most commonly built these days using crosswise high level rails/stretchers. It's probably not be a bad place to look for inspiration.

Not 100% sure what you have in mind re. the splines, presume it's to let them into a slot in the top surface of the above cross rails, and to cut mating crosswise slot(s) in the the lower surfaces of the halves of the split top. The BC design instead relies on a single closely fitted blind tenon at the top of each leg (which lets the legs into the top) to locate the tops. Even if you use the dovetails it seems unlikely that you will want to leave these out anyway, so maybe you already have the required sidewards location and don't need the splines? The BC bench uses lag bolts up through the cross rails to hold the top down, which seem simple and solid - but splines if used might get in the way of these. Gluing solid wood splines into the bottom of the top wouldn't be good anyway given the likely seasonal movement caused by the differing grain directions.

There's a potential issue that needs consideration if using a BC-like arrangement, or a variant - where the outer edges of the tops are fixed by tight fitting tenons relative to the outer faces of the legs inside and out, and a high cross rail is also fitted. Seasonal movement will result in the top(s) changing width by a significant amount, so that unless provision is made to accommodate this damage will probably be done. The trad Roubo gets around the problem by using no upper cross rails (the legs tilt out of vertical and very slightly rack/splay the lower rail joints to accommmodate the movement), while the BC design accommodates it in the gap for the tool rack that fits between the halves of the tops. (they ask that the rack be fitted leaving whatever potentially fairly significant clearance is likely to be required - as determined by the dimensions/moisture content/likely environment/wood species etc from the tables)

The BC split Roubo could probably also with suitable dimensional adjustments be modified to use the traditional exposed dovetail, and a second inner tenon in each leg too - fixing the tops to the legs presents no problem since movement is accommodated in the centre gap as above.

The other option would be to use a one piece top, tightly fit the vise/near side leg tenons in the top, but leave enough slop in the back pair to accommodate the movement - although in this case the rear edge of the top will move in and out a little relative to the vertical face of the rear legs.

Another variation on the exposed dovetail leg joint is shown in the Roubo book - the drawing of the German Cabinet maker's Bench shows the socket receiving the exposed dovetail at the top of the leg stopped some distance under the top surface of the bench top. I'm not too sure how that's meant to work though, since as the top varies in thickness with seasonal movement it will (since the lower surface of the top is resting on the shoulder of the dovetail) presumably open and close the joint where the upper end of the dovetail buts against the receiving socket in the top. Alternatively the top could rest on the upper end of the dovetail/top line of the joint, and some clearance be allowed for movement between the shoulders and the bottom surface of the top. Maybe this was regarded as acceptable, or maybe i'm missing something...

My best guess (since so many of these benches have been built) is that while differing joint layouts have been used, all (presuming correct design) seem to work pretty well as people don't seem to be running into trouble...

[Brian Holcombe]

How about rising dovetails and a stretcher.

[ken hatch]

I've built a number of benches, some out of construction grade wood, some out of hardwood. As I've posted before I think the through dovetail/tenon while it fits the classic design as shown in Roubo's book introduces problems that are unnecessary without adding strength or preventing racking and I do not care how many hours a day you use the bench. I also prefer a split top for several reasons, seasonal movement, keeping the face flush with the front legs, and ease of build to name a few. I'm not an engineer and if I were one I doubt I would test the strength difference between the increased surface area of a through dovetail/tenon vs. a housed tenon draw bored or pegged with double 12mm or 19mm oak pegs. If there is a difference I expect the housed and pegged tenon would be stronger and less prone to racking. It's a mechanical joint vs.increased surface area...you pays your money and you take your chances, it is obvious which one I pick.

From an aesthetic view, the classic configuration can be attractive but whatever, it kinda depends on why you build the bench.

As always with everything wood YMMV.

[ian maybury]

I probably wasn't clear Brian (trying just to figure the pros and cons of some diffrent layouts), but as yourself i was suggesting the dovetails are one option with high rails/stretchers (blind tenons are another) - but that they need a split top, loose tenons at the back or some other arrangement to accommodate movement.

Based on a similar logic to Ken's and as earlier in the thread I'm planning cross rails/high stretchers in my own build (and will probably for multiple reasons split the top, but could yet go one piece), and agree that while the through dovetail looks great that it doesn't seem to add anything much structurally. (other than a continuous leg face through the vise clamping area - which may not be of much practical benefit) I could yet be tempted to incorporate them, but it'd be purely for show and so it's unlikely.

I have a sneaking feeling the dovetail joint might originally have been used on a stretcherless version of the Roubo where the dovetails and long tenon kept a very thick cantilevered leg vertical in an equally thick top, but didn't have to accommodate movement - it seems less likely that they relied on racking the joints in lower stretchers (maybe they floated at one end?) to accommodate movement. (or maybe they did - it seems to work OK anyway)

It seems like it doesn't matter what arrangement is thought of for the joint - there's always some potential (but minor) downside...

[Warren Mickley]

There are some weird and fanciful ideas in this thread. There are maybe a dozen benches illustrated in Andre Roubo's three volume work. There are some variations among these benches, but the top is a single plank of hardwood, bark side down, and the legs are tenoned into the top. There are no cross bearers or stretchers directly beneath the top.

From my experience there is no problem with an exposed tenon.

[Derek Cohen]

The legs of my base are 5" x 3" Jarrah. I had planned to have a 2" blind tenon into a 3 1/2" thick oak. As it was, I increased the height of the bench by adding to the shoulders of the tenons, and this reduced their length to 1".



The bench top dropped onto the base, and with a wiggle and a jiggle, the tenons dropped into the mortices. Rock solid.





I did pin the top, but it really was not necessary with the mass involved. The moral of the story is that much of the joinery we see is likely to be overkill. I have used this bench now for three years and there has not been a hint of anything but solid.

Regards from Perth

Derek

[Patrick Harper]

Derek, I've thought about leaving the tenons loose and seeing how things worked out. I might just do it. If I run into issues, I can always draw-bore the top to the legs.

Did you assemble your base before mortising the top, or the other way round?

[Derek Cohen]

Patrick, I built them alongside one another. I did have a bench to use. That one had been around for 18 years.

The bench was morticed to receive the legs, and the mortice and tenons for the front and rear stretchers were completed. Below is the dry fit.




The top was actually completed first. The parts of the base were assembled, and then the top was flipped over and placed on the base.

There is a record of the build here: http://www.inthewoodshop.com/ShopMadeTools/index.html

Regards from Perth

Derek

[ken hatch]

Derek,

I agree, with the mass of a 100mmX600mmX2500mm slab of hardwood setting on top of and joined by housed mortises and tenons to a two to three hundred lb. base, there is not much chance of movement with or with out pegs. I peg mine mostly just because I peg most M/T joints. I expect it is overkill.

ken

[ian maybury]

I guess in one sense the whole point of these benches is overkill!

All due respect Warren, but that was said - that through tenons (actually tenon and dovetail) are fine with no high stretcher.

You clearly have chosen a personal route (e.g. prefer to be guided solely by Roubo's text), and to diss some aproaches as 'weird and fanciful' - but the fact is that there are other variations on the theme in common and it seems successful use. What's more nobody on this thread is arguing 'this is the only way to do it', just that some designs may not be feasible, and that even the various successful ones have pros and cons. It's been about trying to understand what some of these pros and cons are.

Derek's blind tenon based approach is one that has been successful (it was broadly the approach taken by Chris Schwarz with his original Roubo bench too), as is the more traditional through tenon and dovetail variant CS has been building more recently. The split top high rail/stretcher approach used in the Benchcrafted design is another.

One point about the latter is that by tying the upper ends of the legs together it dictates a requirement for some arrangement (other than simply allowing the legs to splay) to accommodate the in an extreme case approaching 1/2in of wood movement in a full width 24in beech top depending on the conditions and starting state - such as the split top and gap between the halves that it uses. It's very purposely done that way, and in the clear knowledge that the gap is required, and must accommodate the movement. It delivers benefit heowever in terms of ease of build, disassemblability, transportability, and the option to incorporate a tool rack. Which may be of no interest to some, but that's fine - they are free to consider other designs.

The basic i've been trying to communicate is that even the trad Roubo arrangement has some pros and cons, as do all of the others (even presuming they are properly engineered) - but none of these pros and cons are show stoppers...

[Brian Holcombe]

I'm ok with my rising dovetails being labeled as fanciful, they are.

[Steve Voigt]

…You clearly have chosen a personal route (e.g. prefer to be guided solely by Roubo's text), and to diss some aproaches as 'weird and fanciful' - but the fact is that there are other variations on the theme in common and it seems successful use. What's more nobody on this thread is arguing 'this is the only way to do it', just that some designs may not be feasible, and that even the various successful ones have pros and cons. It's been about trying to understand what some of these pros and cons are.

…

The basic i've been trying to communicate is that even the trad Roubo arrangement has some pros and cons, as do all of the others (even presuming they are properly engineered)...

As Warren said, there is quite a bit of variation in the benches that Roubo illustrated, though none use stretchers. So, it is not a question of "only one way," but rather of definite limits on how many ways.

I agree that some of the ideas here are weird and fanciful, or maybe just plain silly. What I'm perplexed by is how you feel qualified to write about the "pros and cons" of the through-dovetail vs. other approaches, when by your own admission you haven't built one of these benches or even used one for any length of time. You can write 50 paragraphs of silver-tongued speculation, but it's really not worth anything unless backed up by personal experience.

My own experience is that I built a Roubo bench that is mostly traditional, but has a few "improvements" based on my (at the time) uninformed speculations about what would work better. I like my bench a lot, but the improvements are the only part I regret. If I build another bench in 15 or 20 years, I'll get rid of the improvements.

[Curt Putnam]

Steve, what is it that you no longer like (the improvements)?

[Curt Putnam]

I'm ok with my rising dovetails being labeled as fanciful, they are.

Brian, I'm not sure what you mean by rising dovetail????