How much glue is the right amount
PVA woodworking glue covers roughly 250 square feet per gallon in a single spread, which is a figure the manufacturers publish and nobody uses because nobody measures glue-ups in square feet. Converted into shop units, a gallon is 128 fluid ounces and 250 square feet is 36,000 square inches, so one fluid ounce spreads about 280 square inches. That works out to a film around six thousandths of an inch thick, which is what a properly clamped PVA joint ends up as.
An example. Five boards, 48 inches long, seven-eighths of an inch thick, glued into a panel. That is four glue lines, each 48 × 0.875 = 42 square inches, so 168 square inches in total. At 280 square inches per ounce that is 0.6 fluid ounces on the joint, and with an allowance for squeeze-out and what stays on the roller, call it 0.69 ounces. Under a fifth of a cup of glue for a whole tabletop panel. People routinely use four times that and mop the difference off the floor.
The number is small enough that being generous costs nothing, and there is a real reason to be generous: a starved joint is a failed joint, and glue is cheap relative to the panel. What the number is useful for is epoxy, where you have to mix a batch and the pot life is running, and for deciding whether the four-ounce bottle in the drawer will finish the job.
Clamp spacing and why it is usually what decides the count
Clamps do not apply pressure to the whole joint. Each one applies it to a cone that spreads out from the jaw at roughly 45 degrees, so on three-quarter inch stock a clamp influences a zone a few inches either side of itself. Space them too far apart and the joint between them sees far less pressure than the joint under them, and on a long panel you get a visible open line halfway between two clamps.
Six to ten inches is the working range for typical furniture stock, closer on thin material, further on thick. That spacing normally demands more clamps than the pressure calculation alone would, which is the usual and correct outcome. The pressure figure becomes the deciding one on short, thick joints, where two clamps might satisfy the spacing rule but cannot together deliver the force a dense hardwood wants.
Alternate clamps above and below the panel. Every clamp applied to one face puts a bending couple into the panel, and a row of them all on the underside will bow the panel up in the middle. Alternating cancels it. Cauls across the width do the same job more decisively and are worth the ten minutes on anything wide.
Clamping pressure by species
The pressure required rises with density, because the point is to bring two surfaces into intimate contact and a denser wood resists being conformed.
| Species group | Target pressure | Examples |
|---|---|---|
| Softwood | 100-150 psi | Pine, fir, spruce, cedar |
| Medium hardwood | 150-200 psi | Cherry, walnut, poplar, soft maple, mahogany |
| Dense hardwood | 200-250 psi | Hard maple, oak, ash, beech, birch |
| Very dense or oily | 250-300 psi | Hickory, jatoba, ipe, teak |
Turn that into clamp force by multiplying by the area of one glue line. The 48 inch by 0.875 inch joint above is 42 square inches, and at 175 psi it wants about 7,350 pounds of total force. Spread over seven clamps that is 1,050 pounds each, which a three-quarter inch pipe clamp or a parallel-jaw clamp will produce with firm hand pressure on the handle and an F-style clamp will not quite reach. That is the calculation that tells you whether your clamp collection is up to the panel.
Note that the force passes through all four glue lines in series, so the area that matters is one glue line, not all of them added up. Four joints in a row need the same total clamp force as one, which is the pleasant surprise in this arithmetic and the reason panels get glued up all at once rather than a board at a time.
Face laminations are the opposite surprise. Gluing four boards face to face over 48 by 6 inches means 288 square inches per glue line, and at 175 psi that is 50,400 pounds. No row of bar clamps will do it. This is why bent laminations go into a vacuum bag, thick blanks go into a press, and people who try it with clamps get a lamination that is tight at the edges and open through the middle.
Where the glue itself changes the rules
- PVA. The default. Wants a close-fitting joint, high pressure, a thin line. Open time is five to ten minutes on a warm day, less on a big panel, which is the real constraint on how many boards you can glue at once.
- Polyurethane. Foams as it cures, which fills the space but adds no strength. Needs moisture, needs full pressure held until cure, and any gap it foams into is a void rather than a joint.
- Epoxy. The exception on pressure. It genuinely fills gaps, and squeezing it to a thin line starves the joint rather than improving it. Light, even pressure. Mix in batches you can spread inside the pot life.
- Hide glue. Liquid hide has a long open time and is forgiving on complex assemblies; hot hide grabs fast and is unforgiving. Both are reversible with heat and moisture, which matters if the piece may need repair in fifty years.
Whatever you use, the confirmation is the squeeze-out. A fine, continuous bead along the whole length of the joint means the film was even and the pressure reached everywhere. A stretch with nothing coming out is a stretch that was starved or under-clamped, and it is far easier to take the panel apart while the glue is wet than to discover it after the finish is on.
Questions people ask
How long do I leave a panel in the clamps?
Long enough to handle, not long enough to work. PVA reaches enough strength to come out of the clamps in about thirty minutes to an hour on a well-fitted joint at room temperature, but it has not finished curing and the joint will creep if you stress it. Leave it clamped an hour, then let the panel sit unstressed for a full day before flattening or machining it. Cold slows everything down considerably: below about 50 degrees most PVAs will not form a proper film at all, and a joint glued in an unheated winter shop can look dry and still be weak.
Do I need to spread glue on both edges?
On a well-machined edge joint in ordinary hardwood, one edge is enough. The clamping pressure transfers the film to both surfaces, and spreading both roughly doubles the mess for nothing. Spread both when the surfaces are porous or absorbent enough to drink the film before the joint closes, on end grain, on species that are very open-pored, and always with epoxy, where wetting both surfaces is part of how it bonds. If in doubt, spreading both costs a few cents and never makes a joint weaker.
Can I have too many clamps?
You can have too much pressure, which is a different thing. Excess pressure on a PVA joint squeezes the glue line thinner than the film wants to be and can starve it, though this is much harder to do than woodworking folklore suggests: it takes real force on a well-fitted joint. What is easy to do is bow the panel by putting all the clamps on one face, or crush the edges of softwood with unpadded jaws. More clamps, each moderately tightened, is safer than fewer clamps cranked hard. The exception is epoxy, where high pressure genuinely does starve the joint.
What if the boards do not touch along the whole joint?
Fix the joint, not the glue-up. PVA is not a gap-filling adhesive; anywhere it bridges a gap it forms a thick, weak, brittle line that will crack. A joint that needs to be pulled shut with clamp force is a joint under permanent tension, and it will open eventually. Hold the two boards together and hold them to a light: if you can see through anywhere, joint the edges again. A sprung joint with a hair of hollow in the middle is deliberate and fine, because the clamps close it and it puts the ends under mild compression. A gap you can see light through is not.