Edge-to-Edge Joinery – even more

Edge-to-Edge Joinery – even more:
Having chosen the wood and prepared the edges, Rob Porcaro glues up the boards into a panel.

At this point, the stakes have grown with considerable investments in expensive wood, the labour of surfacing it and skillfully planing the edges, all to coordinate with the other components of your project. You want to get this right. To do that calmly and confidently, make each step rational in design and predictable in execution. Here’s how. 

Set up for success 
Though they do not seem to garner the respect of the glamour tools, clamps are important and worthy of substantial investment, particularly since they are employed at the crunch time of glue-up. A good set of clamps, such as parallel-jaw models made by Bessey and Jet, is a key to trouble-free panel glue-ups.
Start by making wooden supports. These will run parallel to the clamp bars and the boards will rest upon them. Rip two fairly thick (25–35mm) boards to the same width and a few inches longer than the final panel width. Cover the top edges with plastic packing tape to resist glue. Below, I will discuss how to determine the width (height) of the supports.
A large flat surface, such as a dedicated assembly table, would be ideal for the glue-up. Accurately ripped supports would then transfer the flatness of the work surface to the panel. However, such a luxury is not available in many small shops. A more versatile approach is to think of the pair of support pieces as winding sticks, and to actually view them that way. Place the supports in the positions they will be used for the glue-up on whatever surface you have available, and simply shim one of them to make their top edges parallel. In this way you have created a flat plane that will be transferred to the panel. 

What about alternating the clamp bars above and below?
Ideally, the clamp heads would only exert force parallel to the face of the panel. In reality, some of the force is directed towards the clamp bar, which tends to make it bow towards the face of the panel, potentially distorting it. To help ensure a flat panel, we have two options – either minimise or balance the aberrant forces.
For the first option, put all the clamps below the panel and make the height of the supports so that the boards are very close to the clamp bars, and in this way, minimise the bowing forces.
The other approach, which I prefer, is to alternate the clamps above and below the panel. In this case, define the height of the supports to make the panel sit at about the same distance from the bars of both the upper and lower clamps. In this way, the bowing forces are equalised on both sides of the panel. This arrangement also makes the assembly more stable for moving to set it aside later.
With properly prepared individual boards (even if they have a bit of bow), well-planed joint surfaces, parallel-head clamps, a good board alignment scheme and properly set up supports, there is no need for awkward over-under clamps, cauls or special blocks on the faces of the clamp heads.

Glue matters 
Most woodworkers use PVA glue, based on its reliability, ease of use, wide availability and, frankly, habit. My favourite is Titebond III, which has a relatively long open time for a PVA. However, at glue-up, one still must heed the words of the late, great basketball coach, John Wooden: ‘Act quickly but don’t hurry.’ If the thin glue layer is allowed to skin over at all, the joint line will be too thick and, worse, the joint is destined to fail.
Furthermore, the US Forest Products Laboratory’s Wood Handbook points out that increased glue viscosity requires more clamp pressure. They note that viscosity increases by evaporation – during the ‘open time’ before the joint surfaces meet – and by absorption – during ‘closed time’ after the joint surfaces are together but before final clamp pressure is applied.
Liquid hide glue offers a convenient alternative with a longer open time. On their websites, Franklin lists 10 minutes open time for their product (versus 8–10 minutes for Titebond III), while Old Brown Glue lists
30 minutes. Set up edges side by side and work on them together. Put glue in the biscuit slots first where the bulk of it will delay evaporation while you work on the edges. Apply glue to both joint surfaces to ensure good wetting and equal penetration on both sides of the joint.
I run a bead of glue along the edge directly from the bottle, and then spread it with a hog-bristle brush. (Available from Tools for Working Wood.) I crop the 38mm bristles to 20mm to push the glue faster, and use it nearly upright. Flux brushes are too floppy, small and slow, and using my finger inevitably seems to transfer glue to someplace where I do not want it. A roller or notched plastic spreader is a good alternative.
Aim for a light bead of squeeze-out along the full length of the joint, but try to avoid a lot of dripping. Squeeze-out is simply assurance that you have applied enough glue; there is no need to overdo it. On the other hand, if you produce no squeeze-out, you cannot tell if the amount of glue in the joint is just perfect or too little. 

Assembly 
It pays to do a dry run of the assembly process. Unique sets of diagonal hash marks across the joints will avoid confusion. Work out placements for the supports and clamps. Will you have good access to crank the clamp handles? Make sure all the joints close neatly with just slight pressure.
Open the clamps to the correct length, allowing extra room for the width of the biscuits. Consider how you will move and support the boards for applying glue, including on the second edge of a board that already has glue on the other edge. Plan to minimise open time according to the type of glue you are using.
At game time, I move quickly. I close the joint by first gently tightening the centre clamp, and then work outwards. After adding the upper gang of clamps and gently tightening them, I check the panel with a straightedge. If all looks good, I torque down on all the clamps, check again for flatness and make any necessary adjustments to balance the pressure.
For panels with four or more boards, consider gluing up in separate sections to be combined later. The job takes longer but the work is more relaxed. The intermediate sections can be surfaced to reset any errors in flatness rather than let them accumulate.
I prefer to remove glue squeeze-out when it is still rubbery by lifting it away with a sliver of wood cut to a chisel edge. Then I remove most of the remaining glue with a wet rag that is more than damp but less than drippy. At least in theory, removing the squeeze-out early like this may cause the outside of the glue line to dry faster than the interior. In fact, I have sometimes observed that the glue line looks a bit wider in the hour or two after assembly, but this has never persisted in a well-made joint. Therefore, I do not consider this a practical concern. In any case, removing a lot of fully hardened squeeze-out is an undesirable chore and can chip the wood.

Afterwards 
When is it best to remove the clamps? Unless you need the clamps for something else there is really no hurry to remove them. After all, to avoid a sunken glue line and depressions over the biscuit slots, the panel surface should only be worked when the added moisture and consequent slight swelling at the glue line and biscuit slots have dissipated. This can be verified with a pinless moisture meter or a small rule.
The Titebond III instructions recommend ‘clamp for a minimum of 30 minutes (longer is better)’ and ‘do not stress the joints for 24 hours’, which is how long PVA glues generally take to fully cure. Liquid hide glues take at least that long. I wonder if the joint may be significantly stressed from the changing moisture content early on, and also by the applied forces of manoeuvring the panel in the shop.
For these reasons I put the assembly aside and wait until the next day, up to 24 hours, to remove the clamps. Small panels such as a drawer bottom can be safely removed from the clamps sooner.
With these methods, the work required to flatten the finished panel is likely well within the range of easy hand-planing for small to medium panels. Work diagonally with the jack plane, and finish with smooth planing, scraping or sanding, as you prefer. For large work, such as a dining table-top that needs substantial correction, consider the services of a commercial shop with a wide-belt sander.
Keep in mind that the panel only needs to be flat enough for its function. Small imperfections that yield to light hand pressure are fine, especially if they will be eliminated by the structure of the piece, such as a
table frame. On the other hand, do not allow, for example, a panel to distort the frame of a light cabinet door.
By the way, have some fun by bashing an offcut against the corner of a hard table edge. Even if the impact is at, or close to, the joint line, the wood will fail before a well-made joint gives way. Nice.

Use the force
Are small-shop clamping methods capable of generating enough pressure to make a good edge-to-edge joint? Yes. Let’s look at the details.
The very useful Wood Handbook (a free download at www.fpl.fs.fed.us – search ‘wood handbook’ on the site) chapter 10, page 16 (2010 edition) recommends pressures of 100 pounds per square inch (psi) for low-density wood and up to 247 psi for the highest density woods. The book also states: ‘Small areas of flat, well-planed surfaces can be bonded satisfactorily at lower pressures.’
Bessey claims 1500 lbs. of clamping force can be generated with their K Body REVO clamps, while Jet claims
(I suspect conservatively) 1000 pounds for their parallel clamps. For example, 1500 pounds of force will produce 167 psi over 12in of a ¾in-wide glue joint. Spacing the clamps 8in apart on the same joint requires 1000 pounds of force to produce the same 167psi.
There is no need to calculate in the shop. In general, to create adequate clamping pressure, tend towards more clamps and more torque with denser species or thicker boards. And, of course, make good joint surfaces.
The other influence on clamp distribution is each board acting as its own caul. Thus, the transmission of force to the glue line depends on the width and stiffness of the board, especially for the outer boards of the
panel. With narrow boards, if the clamps are spaced too far apart, the ‘squish’ effect may result in too little pressure in the areas of the joint farthest from the clamps.
I long ago adopted, with consistent success in the shop, Ian Kirby’s practical guideline: think of the force as spreading from the clamp head at 45° outwards to both sides. Therefore, a clamp can usually be considered to reliably spread its force over a length of joint equal to twice the width of an outer board.
So, beware of using too few clamps when gluing narrow boards. Likewise, you may be able to get away with fewer clamps when gluing wide boards, but only if the total force at the glue line (and consequent average pressure) is adequate, as discussed above. 

Putting it all together
Here’s what to remember:
• Select reliable wood, and create visual and structural harmony among the boards, especially along the joint lines.
• Use efficient methods, always finishing with a hand plane, to make good joint surfaces with just a very slight camber.
• Employ a method to reliably align the boards during assembly. I suggest biscuits and, for thin boards, the notched blocks described in the second of this series of articles.
• Use parallel-jaw clamps and board supports that are easily tuned to create a true bearing surface.
• Rehearse the glue-up and get it done quickly.
• These joints will look great and are going to last because you have made them righteously. Relax and enjoy!