Entry 03 / Timber / Workshop diary
The oak boards that moved, and what I allow for now
The first piece of furniture I made from air-dried oak cupped so badly that a straightedge rocked on it. The boards were not at fault, and in the end neither was the joinery.
Buying air-dried boards off a barn floor
The mill two villages over cuts local hardwood and stacks it in an open barn, which means the boards are air-dried rather than kiln-dried and arrive somewhere near the moisture content of the outdoors. That is fifteen or sixteen per cent in an East Anglian autumn, and rather more after a wet week. Kiln-dried stock from a timber merchant comes in at eight or nine. Neither number is wrong, but they are very different starting points, and I did not understand that when I carried my first three oak boards straight into the shop.
Air-dried timber has advantages that keep me going back to the barn. It works more kindly under a plane, the colour is better, and the boards are wide because nobody was in a hurry to convert them into something easy to sell. The cost is time. A board bought in October is not furniture until the following summer, and a board bought for a particular project is a board that will be used too early. I buy stock now with no project in mind, sticker it in the rack, and take the oldest boards first.
A damp shop and a dry house
The shop sits at around eighty per cent humidity in January and drops to perhaps fifty-five in July. The house has radiators and stays nearer forty all winter. A board that has spent two years reaching equilibrium in the shop is therefore still carrying several per cent more water than it will hold once the finished piece stands in the sitting room, and it gives that water up across the first heating season, shrinking as it goes. No amount of careful joinery was ever going to prevent that from happening.
A cheap pin moisture meter changed how I plan. It is not accurate in any laboratory sense, but the difference between two readings a month apart on the same board tells me whether it has stopped moving. Anything destined for the house now comes indoors as dressed stock for three or four weeks before final fitting, stacked with air all round it and well away from a radiator. The boards do most of their arguing during that period, on a shelf where it costs nothing, rather than inside a glued-up carcase.
- Shop in January: around eighty per cent relative humidity
- House in January: around forty, with the heating on
- Two readings a month apart say more than one reading ever will
- Three or four weeks indoors as dressed stock before final fitting
The tabletop that cupped in February
The tabletop was four oak boards, a hundred and ninety millimetres wide apiece, jointed and glued into a panel about eight hundred across. I glued it in a mild November, planed it flat over two evenings, and screwed it down to the frame through holes that fitted the screws exactly. By the second week of February both long edges had lifted by nearly three millimetres and a hairline split was running out from one of the fixings. Every part of that failure was a decision of mine rather than a defect in the wood.
The panel shrank across its width, as any panel will, and the screws would not let it. Something had to give, and what gave was the weakest thing available, which was the board around a screw hole. Alternating the growth rings of the boards, which I had also neglected, would have reduced the cupping but not the shrinkage. The top is still on the table with a filled split near one end. I kept it deliberately, because a repaired mistake standing in the kitchen is a better reminder than a note in a book.
Designing so the wood can move
The remedy is old, thoroughly documented, and I had read it several times without absorbing any of it. Wood moves across the grain and effectively not at all along it, and the job of a design is to let a panel do that while still holding it where it has to be held. Slotted holes or buttons in a groove for a tabletop. A panel floating in a frame with room at its edges. A drawer bottom in a groove rather than glued in. Breadboard ends pinned through oversized holes, if they are there at all.
The habit that genuinely helps is arithmetical rather than clever. Eight hundred millimetres of oak panel will move something like six or seven millimetres between a damp shop and a dry January, and that is a distance you can measure with a ruler and plan around. Once the number went on the drawing, instead of being left to the joints to absorb, my mistakes stopped being structural and became cosmetic. That tabletop still cups a little every winter. It simply no longer splits anything while it does it.