New workbench top glue up advice

[Rich Purdum]

I'm not sure where to post this so since the bench is intended to hand tool use I'll take a chance here.

I'm building a new bench more or less following Garrett Hack's Workbench of a Lifetime lamination approach for the bench top. I was able to find some very nice and affordable European beech 5/4 boards ranging from 8 to 14 inches wide. Rather than rip them into a uniform narrower size I chose (for better or worse) to leave them wide and arranged the three layers so the boards edges overlap. Everything has been ripped so the overall width of each layer is within a 64th of 23-1/4". The finish length of the "slab" will be 91" without the tenons for the breadboard ends.

Hack's approach to glue-up was to face glue the first three boards with the shorter middle board forming a "groove". The next three boards are again face glued with the middle board forming a "tongue" which fits into the previous three board's groove.

I had planned to approximate this approach even though I only have three boards per layer. However, in my dry fits on each layer, the panels are coming up very nice and flat so I'm thinking why not just do three flat panel glue ups and then face glue them one at a time to make my three layer top. I would "smother" each glue up with cauls and with lots of additional clamps along the edges. My biggest concern is clamp pressure in the middle of the top. I'm going to use Titebond Extend which should give me ample time to get things lined up for clamping.

What are the risks with the three flat panel approach? I've read other posts here where folks here have followed his approach to the lamination and it would be good to hear their thoughts on this too. If I need to provide more info just ask. This is my last bench and I want to get it as right as I can.

[Gene Davis]

You should do as Hack did, and avoid the big wide glue-ups. Without dead flatness and a serious press, and I mean a really serious press you cannot imagine nor afford, you cannot get the zero-glueline you think you can.

I'd have taken your pile of beech and ripped all to 3-1/4", with the offcuts edgeglued back up to get to that width, and faceglued them all up like a butcher block. Unless those beech boards were all quartersawn, in which case I would proceed as you did and be a Hacker. But presuming your beech boards were plainsawn, my method would give your top the quartersawn edge, an appearance I prefer.

After my butcherblock glue up, I would flatten the whole affair with a router sled running on rails, top and bottom, netting out at 3". But that's me.

Hey, you sought advice. Take mine for what it is worth.

[Sean Hughto]

The reason to rip them all to three inches and face glue them on edge is to effectively make a quarter sawn top. I've never done what you describe, but the cupping, twisting etc of all those flat sawn pieces on eac other doesn't sound optimal. But maybe in practice it will work out fine?

[Reinis Kanders]

Router sleds just are no fun and they are dusty. Laminating bench top is a good opportunity to use jack planes and jointers if one is serious about hand tools. I am slowly laminating my third benchtop, this time mostly with wooden planes and it has been a good learning experience.

[Jim Matthews]

If the edges you've already exposed are nearly flat, just get on with it.

It's a bench for working wood, not an Optical table for interferometry.
No sense complicating this.

[ian maybury]

I'm still looking at a big stack of beech and no sawdust, but am headed down the face glued on edge route.

Gluing up three large/bench top sized horizontal laminations would as Gene be risky without industrial scale equipment - they would each need to be very flat (the middle one on both sides), and need huge clamping pressure over the entire area on a dead flat surface to reliably deliver a decent joint line.

The alternative of configuring the boards in a tongue and groove format a la GH must be do-able, but would again need boards of precisely equal thicknesses with dead straight edges to go together tightly. If three are glued to form a groove and allowed to dry first, then any glue run out into the groove would need to be 100% cleaned away too (including fillets in the corners) or it would likely stop a joint closing, or jam it during assembly. Any bend in the middle one would cause problems later. Getting the next middle board (the tongue) fully bottomed in the groove would need straight mating edges, maybe slightly relieved corners, care in the handling of the glue so that it has somewhere to go and good clamping up too.

The by now almost traditional 'boards on edge' deal has a lot going for it. They only need to be flat (more convenient of course if they are accurately dimensioned and parallel all over too - but not a show stopper), gluing and clamping is fairly straightforward (done one or two at a time), and if the boards are of a carefully chosen and accurately planed to a set thickness (they don't have to be) the layout gives lots of flexibility in controlling the overall width of the top, building in slots for carriage vices, fitting a pre-made dog hole strip, incorporating a split top, building in the mortices to accept the legs etc. Another possible advantage is that it becomes feasible to glue up boards in groups of say three or four (depending on the required weight/width), and to if preferred run each group over/through (dare i mention it?) a planer thicknesser.

Getting off to a nice straight start when gluing the first (and relatively flexible) few boards togther must be important in any method - i'm on the lookout for a length of nice straight steel box section i can borrow to clamp them against...

[Brian Holcombe]

With large laminations, just do one at a time.

Make sure that you have the grain direction oriented in the same way, so that they will be easy to plane once they're glued up.

Use a glue spreader, do not rely on the glue to self-spread.

Jointing the boards for a slight spring effect helps. It eliminated joints that gap near the ends. I usually determine the amount of spring by how much I can press out myself without clamps. I dont use a lot of spring, generally about .010" or so (off the cuff estimate).

Laminating one at a time allows you to re-visit how parallel the faces are after each glue up, which will allow you to catch mis-alignment and ultimately have much less work once you are near completion.

[Patrick Harper]

I would do as others have mentioned and rip the boards to a consistent wide and glue them face to face. I did this for my Roubo and it worked extremely well. I skip planed them and then glued them up into 4 sections. I then milled up the 4 sections with hand planes and made a bench top. It was tedious with only handtools, but the final top needed only about 20 minutes worth of flattening.

Roubo Workbench Build (http://bloodsweatsawdust.com/2015/01/04/2015-roubo-build/)

[Rich Purdum]

First thanks to all for their inputs and comments. Since my boards are basically ready to glue, I'm going to keep heading down the "face up" versus "edge up" road. My boards started out with what the lumber guy called "hit and miss" quality. The faces were flat and parallel but there were wanes in some places. The edges were all over the map. I trued one edge with hand planes. Got to use my No 40 (a great little tool) and my No 8 and find out what a mediocre sharpening job I had been doing. (European beech is reported to be harder than eastern maple.) Once I had one true edge, I used my (gasp!) table saw to true the opposite edge and then my (gasp again!) thickness planer to bring them all to identical thickness. I have done trial glue ups putting all three layers together and all seems fine. The only layer that needs to be dead flat is the top one but the middle and bottom tested flat with a strait edge (as did the top).

Despite all this I decided after sleeping on it last night to take the more measured approach of face gluing the bottom and middle boards of the front edge and then working from bottom to top and front to back two boards at a time. I have 4 deep throat pipe clamps which together with cauls should give some pressure on the middle of the widest boards. This will take longer but it will be easier to control and make corrections if things start to drift out of line.

This bench normally be against a wall so I'm doing the attachment the bread board ends and attachment to the base to allow expansion to take place from front to back. The Garrett Hack article did mention wood movement but not exactly where the expansion would be allowed to go. Only place I can see is the small tool till at the rear of his table but I didn't notice any explanation of how this was supposed to work. I'm not incorporating a till since there is room to mount one on the wall above the bench.

Wish me luck

[Jim Matthews]

I would not use a breadboard.

I would build with a space in the middle, to hold a batten.
Bolt down the front of each "plank" but let the back float
on the stretchers - either with a screw in an elongated hole
or on a simple dowel.

Make a recess for the dowel that's a little oversize.

Most planing forces are front to back, so if the front is fixed, the top won't move much.


http://www.rpwoodwork.com/blog/2015/02/28/washers-for-wood-expansion/

[ian maybury]

It ought to be do able Rich. It'll need plenty of clamps as the joints resulting from the addition of each pair of boards will probably require clamping both vertically and horizontally. Sounds like you may get more widthwise wood movement, but no problem if it's properly accommodated. Maybe try gluing opposing faces (cup to cup, or peak to peak) so that any tendency to curl is (hopefully) counteracted by the opposing board - but it'll either way be resting on and attached to high level stretchers?

Think your plan to glue two boards at a time is a good one (presumably staggered as laying bricks in a wall) - that way you never form a closed groove, or end up trying to force another board into one to seat it. So the third (more likely fourth) board is fitted into a corner instead. You can as a result see that both glue lines are closing properly, and by clamping both vertically and horizontally you can also (presuming nothing like cured adhesive run out from the previous step is left to stop them seating) make sure it happens. Once you move on to edge jointing the second pair of boards to the first two you could clamp battens across the bottom, and include it in your vertical clamping to ensure the lower surface (and hence the assembly) stays flat. The top layer could be left until the bottom two layers are completed, and then be added one board at a time....

PS always intrigued at the stuff great stuff that pops up from the Lee valley site that i've not noticed Jim...

[Gene Davis]

The issue I see with a stacked up member, groove on one side, tongue on other, is that it is difficult to get each edge precisely in place, full length of subassembly, so as to ensure hair-thickness glue lines from end to end, all the way through.

When you mill a board on a shaper with T&G cutters, the geometry is all handled for you by the cutters. When you stack them up offset and face glue them, all bets are off on the edges.

Hate to say it, but Hack is a hack for doing it that way. There will be gaps.

But hey, it is just a workbench. It'll work just fine.

Problem is we are getting preached at by all these dudes building ruobos, with their massive piles of stock, super-precision in flatness by any means possible, preferably using either $2,500 worth of green-trimmed power tools, or almost that in a set of $400 hand planes. Then there is the star wars vise hardware.

My lifetime bench I built almost 20 years ago, the Tom's Torsion Box workbench-in-a-weekend, uses 2x4s and plywood, has a wood handscrew for an end vise, and a plain old 4x10 Jorgenson quick release face vise. For the 2x4 waffle grid inside all parts, stretcher, legs, and top, I used engineered LSL stud stock, almost twice as dense as framing lumber, hard, super straight. It is a heavy monster, and dead flat on top.

[Rich Purdum]

Well, I'm down to my last board to be glued which will be the "top" of the slab (or sandwich). The top is two wide boards (12 and 11 inches) each of which is flat sawn. Conventional wisdom would say to alternate rings up and down but since these two boards will be flattened (and maintained flat) as a unit I'm thinking it would be easier to hand plane if the cathedrals could be attacked in the same direction which means both up (or down). Not sure if there is enough of a description above to allow comment so if there are questions please post.

FWIW, the process has been slow in part because I'm using Titebond Extend and have given it the full 24 hours clamp time as suggested by the maker. With three boards per layer (except the top) and adding one board per session that's 8 days elapsed. The results are looking promising so far. The breadboard ends are actually aprons which will be bolted to the slab. The front apron has the dog strip to tie in with the tail vise while the rear apron is there for clamping just in case I need to pull the bench out from the wall for some reason.

For each "session" I have used a kitchen timer to do a dry fit first so I can be sure of making the 20 minute working time for Extend. Single handed, it's been a push at times. I've used both a roller and a spreader. Each has it's merits but I think for this project a wider spreader would have been best. Next time...