Wood Drying and Shrinkage
by Steve Reznek

How many times have you lost a great piece of wood because it split as it dried? Or how many times have you made a segmented bowl and had the joint move? This last is bad enough when you can only feel the lack of alignment. It is disastrous when the joint actually comes apart.

We live in New England. It can get pretty hot and muggy in the summer; and it gets cold in the winter – or at least it used to. When we heat up the air in the house the relative humidity drops to zilch. This of course is bad for the turnings. If their diameter is small, not much happens. However if they are large, two things can happen and neither is all that great. They may distort and worse luck, they check. (I wonder why wood turners call it checking when what it is doing is splitting!!)

Of course we all know the reason. But this article will first try to put a few numbers on what is going on and then suggest some things that you can do that might help. Of course if you haven't had problems or already have a preferred cure, you can skip the rest.

Let's start by talking about cells, drying and shrinkage. (I don't mean in the Seinfeld sense.) Think of a log from the trunk. The cells in wood are generally very elongated cylinders. For most cells, the long axis is up and down. However, a minority of the cell point from the center out. These are in the "medullary rays". These rays can give rise to some of the more interesting patterns in turned objects. I am not sure if they are the cause of the tiger stripes in maple, but I am sure they are the light "cross grain" streaks you often see in oak. When the cells dry they shrink, of course. But for all practical purposes, they don't shrink in the long direction. The cylinders get skinnier, not shorter.

Call the direction along the trunk "longitudinal". Call

the direction from the center out, i.e. across the grain, "radial"; and the direction around the trunk, i.e. within one of the grains around the tree, "circumferential". Some typical numbers for shrinkage are 0.1% in the longitudinal direction, 4% in the radial direction, and 8% in the circumferential. Think of a board. The length of the board is longitudinal, or at least mostly so. If the board has an interesting pattern, the width is circumferential, and the thickness is radial. If the board shows a lot of grain, then the width is radial and the thickness circumferential.

Now a little bit about moisture in wood. There are two types of water in wood. The free water is in the capillaries and the bound water is inside the cells. The free water is what sprays all over you when you turn really green wood. Although there is a heck of a lot of it, this stuff is not the bad actor: trees don't shrink in the winter, when the amount of free water is greatly reduced. The bad actor is the bound water, which is in the cells. There are no cells pointing circumferentially, i.e. around the tree. When the cells lose their water they shrink and therefore the 8%. The medullary rays have cells pointing radially and they reinforce the wood and reduce shrinkage in that direction, hence the 4%.

One final note: The moisture (bound water) content of dry wood of course depends on the relative humidity. It also depends on temperature, but only slightly, and the type of wood. At 20% relative humidity the moisture content is in the single figures, say 5%. At 90% relative humidity it is about 25%. The "standard" content for dry wood is 12%, and shrinkage is measured either between 100% relative humidity and 12%, or between 100% and 0%. Listed below are some shrinkage values for different woods. The values are for drying from 100% to 12% moisture.

OK, these numbers are typical for the tree type and vary with a lot of things, like how fast the tree grew, etc. etc. More importantly wood moisture in a dry winter home may not be 12% and in a humid summer home it is not 100%. But they are good guides to the relative importance of shrinkage.

Let's first look at the usual non-segmented, non-endgrain turning. In one horizontal direction it doesn't want to shrink at all, and in the other it wants to shrink somewhere between 3.5% and 7.0%. Ah ha, that is why they go oval! In the vertical direction, they want to shrink between 2.5% and 4.5%.

Now of course what happens depends not only on how much it wants to shrink, but how strong the wood is. Strong woods, with small shrinkage, made into small turnings do little. Strong woods with high shrinkage go oval. And weak woods, with some shrinkage, check. "Strength" does not mean how hard it is to bend the wood, i.e. stiffness. That goes into the shrinkage value. It means how much bending it will take before it breaks. Somewhat contradictorily, strong woods bend and weak woods don't and therefore snap.

How do you minimize the problems? There are a couple of things you can do. The most obvious is to reduce the height of the turning so that you are as far away from the pith as you can be. This minimizes the difference between radial and circumferential. Remember this difference is the same size and the difference between longitudinal and radial. Also the wood is far and away much stronger in the longitudinal direction, so it can take the "oval" shrinkage.

Remember thin walls can bend more easily than thick ones. So again somewhat contradictorily, thin bowls are strong, i.e. don't check, while thick bowls do.

The other common thing that is done is to dry very slowly. This lets the cells of the wood "creep" and distributes the stress of shrinkage; but it does not eliminate it. Finally there are a number of tricks that are reported to work, but how they work is usually not explained. The first is to boil the boil the wood. We all have heard of steam bending, well boiling does the same thing. I think what it does is to give the dry cells some ability to slide past each other and relieve the stress of shrinkage. The other thing that is in the literature is to soak the wood in dish detergent before you dry it. I guess it does the same thing, i.e. lubricates the cells and allows them to slip past each other. (Creepy, no?)

OK, finally segmented bowls. There are two no no's. The worst is to have the circumferential direction glued to the longitudinal. This is a "cross grain" joint. Almost as bad is side grain to end grain, i.e. radial to either circumferential or longitudinal.

If you make the standard multi layered "ring" designs, you can have more or less longitudinal to longitudinal joints and the larger the number of segments, the more closely aligned the grains will be. This is why many people do it – no expansion except through the (thin) wall. There are three problems of course. What do you do with the bottom? Making eight, twelve or sixteen pieces come together at a single point is really tough. Second, the "brick work" structure strikes some people as not too attractive. And third, you cannot make the bottom out of a single piece of really cool wood.

An approach that I use to get around some of these problems is to make all my segments align in the same direction, so I have radial to radial, or longitudinal to longitudinal or circum. to circum. joints. In a simple case I make the bottom piece out of a single (square) piece of wood. I cut strips (cross grain and with grain) and glue up the top so it looks like a tic-tac-toe board. Of course the center square is large. A large number of designs are possible and some quite neat patterns can be made. Of course I still have the problem that the different woods shrink a little differently. But you can't have everything.

Some guiding rules for segmented turnings:

1) Keep your diameters small

2) Use really good glues (super glue is terrible)

3) Choose your woods carefully for small shrinkage and the ability to bend

4) Make the walls thin

5) Don't do the no no's and most importantly

6) Only sell your bowls to people who promise to keep the humidity up in their houses in the winter!

There are many good books on wood drying and shrinkage. One quick one is the introductory chapter of of "Turning Green Wood" by Michael O'Donnell.

+++
Posted March 2007