One direction, not two
The single fact that explains most of what a timber frame does in its first few years is that wood shrinks across the grain and hardly at all along it. Take a species with eight percent tangential shrinkage from green to ovendry and fifteen hundredths of a percent along the grain. That is a ratio of better than fifty to one. An eight inch depth of timber losing sixteen points of moisture moves 5/16 of an inch; a twelve foot post over the same sixteen points moves an eighth of an inch, and it is eighteen times longer.
The practical consequence is that a frame cut tight in the summer and dried through a winter opens at its joints in one direction only. Shoulders that bear across the grain open. Tenons, which run with the grain, stay the length they were cut. Nothing is going wrong when that happens; it is what the material does, and a frame designed by someone who knows the material has already decided where the movement is going to go.
Why the stack matters more than any single timber
A third of an inch on one timber is not much. The problem is that the movement adds up wherever a load path crosses several timbers, and different parts of the same frame cross different numbers of them. Follow the load down at one corner and it might pass through a plate, a girt and a sill — three cross-grain depths. Follow it at another and the post might land straight on a sill — one. Both settle; they do not settle by the same amount, and the difference is what turns into a floor that is out of level or a door that stops closing.
| Path | Cross-grain depth | Movement at 16 points |
|---|---|---|
| Post straight to sill | 8 in | 5/16 in |
| Plate, girt and sill | 24 in | 1 in |
| Difference between them | 16 in | 11/16 in |
Those figures use eight percent across the grain and a thirty percent saturation point, which is what the calculator defaults to; substitute your own species figures and the pattern holds while the numbers change. This is why the calculator asks for a comparison path. The absolute settlement of a frame is mostly invisible — everything goes down together and the building is a fraction of an inch shorter. The difference between two adjacent paths is what people see.
Above the saturation point, nothing happens
Green timber has water in two places: bound into the cell walls, and sitting free in the cell cavities. The free water leaves first, and its departure changes nothing dimensionally. Only once the cavities are empty and the cell walls start to give up their bound water does the wood begin to shrink. The moisture content where that transition happens is the fibre saturation point, around thirty percent for many species. So timber at forty-five percent moisture content losing fifteen points does not move at all, and then loses the next fifteen with all the movement in it. The calculator clips both figures at whatever saturation point you enter, which is why a very green starting number does not produce a very large answer.
What the coefficients mean, and where to get them
Published shrinkage values are total shrinkage from green to ovendry, expressed as a percentage, and given separately for the tangential direction (along the growth rings) and the radial one (across them). Tangential is usually about twice radial, which is why flatsawn material cups and quartersawn material does not. Dividing the total by the saturation point gives a per-point coefficient, and multiplying that by the dimension and by the points of change gives the movement. That is all the arithmetic here is.
These are species averages from wood reference tables, and individual timbers vary around them — sometimes considerably, since a large timber contains the pith, juvenile wood, and grain that runs in more than one direction over its length. The along-grain figure in particular is unreliable in juvenile and reaction wood, where it can be an order of magnitude larger than the published value. None of these numbers are supplied by this page, and none of the outputs are a prediction. They are the consequence of the inputs.
Questions people ask
Why does the post not get shorter as the frame settles?
Because its height runs with the grain, and wood barely shrinks in that direction — for many species around a fiftieth of what it does across the grain. The settlement in a frame comes almost entirely from the depths of horizontal timbers being crossed by the load path: sills, girts, plates, floor beams. The uprights stay very nearly the length they were cut.
What does a fibre saturation point of 30 percent actually mean here?
It is the moisture content below which shrinkage starts. Above it the water is sitting free in the cell cavities and can leave without changing the size of anything; below it the cell walls themselves start giving up water and the wood shrinks. The calculator clips both of your moisture figures at whatever value you enter, so a green timber at 45 percent is treated as starting from the saturation point.
Where should I get the shrinkage percentages?
From a wood properties reference for your species, which publishes green-to-ovendry shrinkage separately for the tangential and radial directions. Use the tangential figure if the dimension in question runs along the growth rings and radial if it runs across them; if the timber is boxed heart or the grain orientation is mixed, the tangential figure is the conservative choice. This page supplies none of these values.
Is settlement in a timber frame a problem?
It is a design consideration rather than a fault, and how it is dealt with is up to whoever designed the frame. Some designs take it in the joinery, some arrange the load paths so adjacent points cross the same depth of timber and settle together, some build dry so there is much less to take. What causes trouble is the movement arriving somewhere nobody planned for it — against a stair, a chimney, a stiff finish, or a run of pipe.
Why compare two load paths instead of just reporting the total?
Because uniform settlement is nearly invisible and differential settlement is what people notice. A frame that drops half an inch everywhere is half an inch shorter and nothing looks wrong. Two corners that drop by amounts differing by half an inch produce a sloping floor and a door that binds. The comparison field exists to put that difference in front of you.