John Bullar - Furniture Maker

Wood Movement © John Bullar 2004
This article was previously published as part of a series on techniques by John Bullar in New Woodworking magazine. Thi Wood moves in mysterious ways. The last thing any of us wants to find out is that our latest and greatest woodworking project has self-destructive tendencies, but wood movement could cause this.
	Wood consists of layers of fibres, formed in rings just under the bark each year. These fibres naturally swell with water, but they shrivel up if they dry out, reducing considerably in width, as well as becoming slightly shorter. This causes the wooden boards themselves to become narrower, thinner and shorter (in that order of scale). All wood contains water. Live wood in trees or freshly felled timber is brimming with water but this has usually dried out by the time you buy it. In the long term, the amount of water that stays in dead wood reaches a natural balance with the amount of water in the air around it. This means wood is continuously changing its dimensions! A second set of fibres at right angles to the main vertical ones, grow in thin bands between the outside and the centre of the trunk. This second set produces the patterning known as Medullary rays, most noticeable in oak and visible to a variable extent in many other hardwoods. The cracks that form in felled logs, left to dry, usually follow the line of one of the rays.
The annual growth rings of tree trunks appear as a series of arcs, visible on the end grain of sawn boards. If wood has been sawn by the routine through and through method, the arcs are quite long, often stretching from one side of a board to the other. As the wood dries out the annual growth rings try to straighten themselves. In a log, this causes cracks to form but in sawn boards, it causes cupping. Seen end on, the board curves in the opposite direction to the rings. The real problems arise when we come along and join pieces together. Usually the grain direction of the joined pieces will be different and so they do not move by the same amount. Once a piece of wood decides that it wants to shrink, no amount of brute force will persuade it not to. It is a bit like when an irresistible force meets an immovable object, something is bound to snap - either one of the pieces of wood or the carefully crafted joint - and quite likely the temper of the craftsman too!
	In the atmosphere, water vapour is mixed up with Nitrogen, Oxygen and all the other gases we breathe. When it is warm, air can hold a lot more water than when it is cold. The amount of water in air is measured in ‘Relative Humidity’, which is the percentage of water relative to the most the air can possibly carry without it dropping out as dew. When cold air is heated, such as when it enters a warm building in winter, it becomes capable of holding more water vapour - if it cannot find any the relative humidity falls so the air becomes drier indoors. When dry air meets the surface of a piece of wood, it takes water from it to restore the natural balance. With the outside of the wood now drier than the inside, the fibres shrink and there is tension created that can cause cracks and possibly a change in shape. Expensive veneers are stored in sealed bags to prevent moisture entering or leaving.
Air-drying timber is the traditional method and is still widely used by small-scale operations and specialists. The trunk is sawn into flat boards, which are then stacked up and separated by ‘stickers’ to let the air flow between them. Usually this is done outdoors, but under cover so as to keep off rain and direct sunlight. The rule of thumb is one year per inch of thickness, which should take the moisture content down to about sixteen percent, which is fine for outdoor woodwork. However, it is too high for use in a central heated home so after roughly cutting air-dried timber to size it is usually stored in a warm dry indoor location for some more months. Kiln drying is a much quicker method, often reducing the time to a couple of weeks, and it is more controllable. Most commercial wood converters who can afford the investment use only kiln drying. The process is one of heating the wood up in a controlled atmosphere so moisture trapped in it is released as water vapour, which is blown off the wood surfaces by fans.
	It would be a mistake to think that using kiln dried timber solves all the woodworker’s problems. Because the drying is so rapid, unless it is controlled correctly it can cause serious defects. One common fault from poor kiln drying is where large splits are produced inside the timber, sometimes not visible on the outside at all. This is known as case hardening. The outside of the wood becomes rigid while the inside is still soft and swollen. Then, as the inside dries and shrinks it is pulled apart by the outside. One simple way of guarding against movement is to choose the right wood. When oak has been quarter-sawn or cut at right angles to the growth rings the ray patterns are large and clear. As well as making the wood beautiful, this is an indication that it will be stable and not prone to cupping on wide panels. In past centuries, oak cut this way had the special name of ‘wainscot’ and featured extensively in high quality buildings and furniture.
Woodwork in a central-heated house needs a moisture content of around ten percent. Measure the moisture content in a number of places to be sure it is evenly dry. Simple meters measure water content of wood by electrical resistance with a pair of pins that you stick into the surface of the wood while the meter passes a small electric current through them. These only measure the water content near the surface of the wood which may be very different from the inside. They also leave pinholes in the wood, looking like two woodworms have been for dinner. Some of the better modern moisture meters are pinless. They work by sending an electromagnetic field deep into the wood. These measure the moisture content throughout without leaving marks.
	To maintain the moisture content of wood around the ten percent mark, the air needs to have a relative humidity of about fifty percent. Measuring the water content in the air is straightforward. There are some very fancy electronic meters to do this accurately but a simple mechanical meter or hygrometer will do the job more than adequately. These are available from craft suppliers for a few pounds. One way to improve the dryness of your workshop is to install a de-humidifier. This works like an air conditioner and heater combined. A fan sucks air in from the room where it meets a cold set of pipes and forms condensation, which drips out into a collecting bucket. The air then passes over a hot set of pipes, gets re-heated, and comes out again drier. I installed a drainpipe off my de-humidifier into a water barrel outside the workshop so I do not need to keep emptying it.
In an attempt to reduce end slitting, wood-yards commonly paint the timber ends with oil paint before placing in the kiln. This reduces the loss of moisture out of the exposed end-grain, making its shrinkage rate similar to the rest of the board. Small pieces of timber for turning or carving are often sealed with wax over its ends, or sometimes over the whole piece, to prevent moisture entering or leaving the wood during storage. This is extremely effective, while the wax is present, but there is a risk that the wood will experience a very sudden change in moisture content when the wax is removed in the workshop. If you plan to work on expensive wood that has been protected in this way, make sure the workshop is suitably dry before you start.
	Cupping of boards that are not cut on a line passing through the centre of the trunk, is almost inevitable. Having accepted this, you can minimise the effect on wide panels such as tabletops by building up the width from narrow pieces joined edge to edge. If the ring directions alternate (so the end-grain pattern looks like a corrugated roof) the cupping effects will cancel out. Frame and panel is a long established way of building timber in large stable areas. The idea is the long-grain of the framework holds a consistent shape while the panel, trapped in a groove inside the frame, is free to expand and contract. The technique has commonly been used for building internal walls, as well as furniture and doors.
Dovetail joinery - the highest standard of woodworking construction techniques - locks end grain to end grain so any changes in width are shared by both halves of the joint. Dovetails eliminate sideways tension and prevent both pieces from cupping. John Bullar featured on the cover of New Woodworking magazine, demonstrating handmade dovetail techniques.	Cover Photo by Ben Daniels

Making a Cotswold Arts and Crafts Cabinet in Elm © John Bullar 2002
This project article by John Bullar was previously published in the book 'Carcass Furniture', by Guild of Master Craftsmen Publications Ltd Travelling home from the Axminster Furniture & Cabinetmaking show a couple of years ago, I stopped off at Cheltenham to stay with old friends. We took a trip into town and spent some happy hours in Cheltenham Art Gallery and Museum. This has a gallery devoted to the Cotswold school of Arts and Crafts furniture - inspiring pieces of work from mid 19th to mid 20th Century with a common theme of pleasure in unpretentious handcraft. I came away wanting to design and build some furniture that would express a few of the ideas I had seen. Straight away I got sketching while the memory was fresh - I would start by making a sideboard cabinet and follow it with a co-ordinated round table, both pieces for a large kitchen. One fundamental of the Cotswold style was to use wide boards to give uninterrupted figuring on surfaces. I very much wanted the sideboard top to be made from a single board without any joints, but I would have to find the right wood.
	Some of the finest furniture in Cheltenham gallery, even quite large pieces, was in English walnut, which is certainly beautiful and good to work, but large pieces are in very short supply. I have only ever been able to obtain enough good quality material to use it for small delicate work and haven’t come across English walnut boards wide enough for this kind of piece. Instead, I decided to use Elm, a beautifully wild organic-figured wood with bags of character. It has warm brown colours with occasional streaks of green and a profusion of cat-paws knots. The downside of Elm's wild nature is that it can move a lot while drying, twisting and buckling around the knots and nearly always splitting around the central pith of the tree as it weaves up through the trunk. Mature elm trees have been in decline in Britain now for more than half a century - the ravages of the fungus ceratocystis ulmi (better known to its enemies as Dutch elm disease) has brought down most specimens with a trunk more than a foot diameter. On the waney edges of my wide boards, just under the bark were telltale patterns of the galleries where the elm-bark beetles had hatched their fungus-carrying brood.
I had selected the boards very carefully when I bought them. They had already been air-dried - I had them kilned, then collected them within a day of coming out of the kiln. After several months storage in my workshop most of the boards remained completely flat while the ends had cupped by a few millimetres. I picked through the boards for the cabinet top and sides to give an appearance of continuous figuring ‘wrapped’ around the top corner joints. I wanted the finished carcass to be a full inch thick. With so many dovetails joints on the carcass corners, they cannot be forced together if they are at all tight, but on the other hand the slightest gaps would be very conspicuous. For normal dovetails on drawer sides and the like I usually knife-mark the top of the pins from the underside of the tails. Sadly, there is no way that I could cut the required accuracy through inch-thick dovetails using this technique. Instead I had to work out the following method so the visible surfaces would be marked directly off one another:
	Having prepared the faces, edges and ends of the four boards and chosen the front edges mark out the dovetail pins on the ends of the carcass sides with a 1:8 guide in the usual manner. With 29 pins and 30 tails into the 26-inch depth, the pitch of the pins works out at just less than one inch. In keeping with the Arts and Crafts style, the alternate pins are cut to 2/3 depth to enhance the visual interest of the joints. Mark these out with a knife-edged marker gauge on the end grain and after cutting all the pins full size pare back alternate pins to the line. For marking the tails the carcass side is clamped to the left-hand end of the bench with pins protruding above the bench surface. The corresponding carcass top or base is lightly scribed twice with the marker gauge for two depths of pins, laid along the bench and butted up to the side piece then clamped up tight against the pins using the tail vice. Next use a square to mark from all four corners of each pin to the scribed lines where the tail will be cut. Cut the tails in the conventional way, checking by with a calliper that they will be a good fit on the pins.
The doors are frame and panel construction. The stiles and rails are jointed using open slotted mortises and full size tennons. The vertical and horizontal cross members meet in a halving joint and have small tennons to engage in the same slots as the panels. All the door parts are cut from adjacent boards to keep them looking ‘book-matched’, the rails and stiles are straight grained and the panels from burred ends of boards. This cabinet uses internal pivot pins to avoid the look or feel of door hinges protruding from the front of a cabinet. These doors are hung on pivot pins made from large brass screws that pivot in washers that are trapped in holes in the carcass. Variations on this method have served me well for a number of years. When I first made these simple hinges I thought I could claim it as a new idea – until I saw a similar method on a seventeenth century chest (still bearing up well).d of the rail. The pilot hole can now be enlarged:
	The drawers are made in the conventional hand-dovetailed manner with elm fronts, quarter-sawn oak sides and cedar of Lebanon bases. As with the rest of the cabinet front, the drawer fronts are book-matched by choosing a pair of consecutive boards. The sides are joined to the front with stopped dovetails and to the rear with through dovetails. The cedar bases are slotted into the sides and rear. The handles are made from brown oak in the style of some traditional Cotswold Arts and Crafts pieces with upper and lower tactile concave surfaces for a finger and thumb grip. I turned the brown oak first to produce a matching pair of ‘wheels’ rather like heavy-duty castors about three inches diameter. From each wheel I bandsawed out a central strip and discarded it, leaving two sector shaped handles. Gluing-up up any carcass can be a scary business, but with 120 dovetails and four double tennon joints all coming together at once, good planning and slow setting glue are essential.
I just used two sash cramps and a trusty pair of old canvass band clamps, at the same time dashing around with a large square and pulling the carcass up square by small movements of the clamp heads. The workshop was quite cool that day and the operation probably took about half an hour. I don’t like to apply too much pressure with cramps, but with so many joints soaking up moisture from the glue it certainly took more force than usual. I must admit that if I were trying that operation again I would ask someone to assist me. After the glue has set, the moment of truth comes when you expose the clean pattern of the joints with a sharp plane. After cutting all the joints the rails and stiles are slotted to take the panels and cross members. I hand planed the chamfers on all the panels using a No. 78 rebate plane with a spur cutter which I keep sharp to give a clean profile. The door is then assembled in the normal manner.
	The carcass of this cabinet sits upon a pair of fairly massive pedestals, undercut to form front and rear feet, and with stepped fronts. These are cut from three by six-inch elm, so the front end-grain patterns are bookmatched. The cut-aways are produced on the bandsaw, together with a mortising chisel to remove the angled undercut of the feet. I used a finely tuned block plane where to chamfers where possible and, where the block plane would not reach, I turned to a long handled Japanese paring chisel to complete the finely cut chamfers. To make a simple back, vertical strips of cedar of Lebanon were sprung into slots in the top and base. All the cabinet faces were sanded with fine paper only; to preserve the hand planed surfaces and sharply defined mitred edges. I gave all the outside faces four applications of Danish oil, brushed off then wiped off again, diluted with turpentine in ratios 1:1, 1:2, 1:3 then 1:4, sanding finely between each. I brought the finish to a sheen with beeswax and turpentine polish, I also recommended it on the label for future maintenance to help develop a patina.

Making a Scottish Arts and Crafts style Chair © John Bullar 2003
This article by John Bullar was previously published in Traditional Woodworking magazine. The finished chair featured on the front cover. A friend asked me to make a special chair as a fiftieth birthday present for her brother, a tall Scotsman. This oak and leather seat with its high back was the result. It is the latest in a series of individual chairs I have made recently, using similar construction techniques but varying in detail and style. It combines rectangular and flowing shapes in a style based on the Scottish Arts and Crafts movement, a century ago.
	The frame sides have a flowing curve from top to bottom, progressively reducing in width (like a tree trunk) from the base. The arms sweep around and close in slightly towards the front as if to hug the sitter loosely. The back is 50 inches high - taller than most chairs. In the past, I have made similar chairs from ash and yew, which can take a very fine shape because they are such supple woods. Oak is coarser though, with less cross-grain strength, so it needs more of a heavy-duty construction. The frame parts were were marked out from templates or rods then cut from solid timber to a curved shape using a bandsaw. Cracks or sections of sapwood would get in the way of the cutting plan. Knots may add interest or they may introduce weakness, depending on how big they are and whether they are dense and solid or full of internal cracks. Short grain has to be accepted to some extent, but not if the frame members are thin or carry a lot of load.
After roughly shaping the components on the bandsaw, I faired the curves with a range of sharp-edged hand tools. A compass plane is best for long sweeping curves, but when the curves change radius or turn from concave to convex you need to continuously adjust the radius of the sole plate using the control on top of the plane. The compass plane will not fit into tight radius concave curves so I use spoke shaves for these. Old-fashioned wooden spokeshaves have a good feel if the narrow blade is sharp and finely set. They also have the advantage that you can re-shape the sole with a block plane to make them tightly convex, this makes them suitable for fairing very tightly into concave curves.
	The second method of making curves for this seat is laminating thin slices of timber together with glue in a shaped mould. I built a pair of moulds out of MDF sheets bolted together. The shape of the moulds was marked using the back edge of the chair frame as a template. This was cut on the bandsaw to make the rear half of the mould. The front half of the mould needed a strip, equal to the thickness of the finished slats, cutting from its front edge so it had the right radius of curvature to match the other side of the laminations. After cutting and bolting the mould pieces together I planed them with a compass plane to prevent them denting the oak. The back components of the frame, including the laminations, came together at the first glue-up. The front rail then joined the two front legs and when the front and back sub-assemblies had set, the side rails then joined them. The arms came later. Many furniture makers recommend cutting mortise and tenon joints for curved parts while the wood is still straight and square. Cutting the joints before the parts are shaped means you can use machines such as mortises, pillar drills and router tables to align them. In addition, if you use hand chisels for cutting the joints or cleaning them up, you can clamp the wood down flat on the bench and chisel at a true vertical.
The main joints between the sides of the seat and the legs carry the most weight. If anyone rocks back on a chair the torsion stress on the rear joints is enormous - frequently the cause of terminal failure for older chairs. The backward sweep at the base of these rear legs makes this stunt impossible I hope, but just in case some heavy individual tries it, I made these joints to maximum strength. This means the thickness of the tenon should be equal to the total thickness each side of the mortise. This way each half of the joint is equally strong. The tenon is double lengthwise so that the mortise socket has a central web to discourage it from opening up if it receives a twisting force. Horns protruding from the top of the frame are not cut off until the glue has set on the rails to eliminate the risk of the top joint splitting.
	The top matrix is a series of short blocks of oak trapped in place between three back rails, with stub tenons to help locate them. The top of the back is dual concave - it curves both from top to bottom and from side to side. The individual short blocks in the matrix are straight but the curve is taken up by a slight angle between the top and bottom of the rails. There is a big advantage in the Arts & Crafts tradition of cutting chamfers on all the edges (or arises) - it stops them from splintering. The chamfers also catch the light and emphasise the outside edge of the shape to give it clarity of definition. This is something that many pieces loose by sanding over the edge. Where the ends of chamfers approach a joint it is traditional to stop the chamfer with a chisel carved end / chamfer stop. This always strikes me as a poor option because it leaves a sharp corner that can easily splinter. A better option is to continue the chamfer right up to the joint. (Of course, it is no use cutting the chamfer before you cut the joint because that will leave the joint looking very gappy.) I cut the end of the chamfer with a very sharp chisel, watching out for grain that falls towards the joint - this would try to carry the chisel with it creating a monster splinter. The way to avoid this is to cut extremely finely with a freshly honed blade.
The exposed joints on top of the arms are, of course, decorative but they have a good function too. People will inevitably lift a heavy chair like this by its arms and I have found in the past that stopped joints mortised into a thin piece of wood can work loose. These through-joints use the full wood depth and wedges on the upper side splay the tenon out to fit tightly in the mortise. For the mortise in the arm, I cut the top visible by the conventional chisel chopping technique. If the chisel is sharp enough and the cut is light enough it makes a crisp un-torn edge, without any depression of the land around it. The resulting chair was well received by its new owner. It captures the style of Charles Rennie MacKintosh, the revolutionary Scottish designer of 100 years ago, understood by few in his day but now immensely fashionable.

Making a Twisted Dovetail Box © John Bullar 1998
This article by John Bullar was first published by Furniture & Cabinetmaking magazine in 1999 This artist's pencil box in American black walnut (Juglans Nigra) and English sycamore (Acer Platenoides) was made with joints based on the twisted dovetail, displaying tail-type wedges both on the sides and ends. A drawer only gets pulled one way but the force on the joints of a box pull both sideways and lengthways - normal dovetails wedge the sides together in one direction and rely on glue to hold the box together in the other direction.
	Dovetails only do half the job of holding the box together, and so I looked for a way to make a box with joints wedged both ways. The answer came when I was playing around with variations on the twisted dovetail - the Japanese nejiri arigata joint. Marking out starts by dividing up the width of the end between the total number of joints - but there the similarity with dovetails ends. For fine handcut dovetails you would probably try to make the pins very narrow and the tails much broader. But with this twisted dovetail joint, the pin on one face also forms a tail on the other face so they are all the same width. The cutting lines are made by marking the position of each cut, and then the angle of each cut. Unlike a dovetail, it is not easy to cut one piece first and then mark the mating piece off it.
To maintain accuracy with the joints, mark both corresponding cuts on the side piece and the end piece together as two complementary pairs. To mark the positions of the cuts on each arris butt the two pieces tightly together end to end. Use one press of the knife tip to mark both end and side pieces simultaneously at each measured distance, so there are no possible differences. Next, line up the end and side pieces side by side ready for marking the angles of the cuts. Working first on the end piece, align the edge of a sliding bevel gauge between the position mark on the arris and the centre point of the opposite edge, then marked against it. The sliding bevel is locked in position and turned over to mark the same angle across the endgrain, so the lines meet at the arris in a chevron pattern. With the sliding bevel still locked, the same angle is then marked from the corresponding position on the arris, and across the endgrain, of the side piece. This process is repeated for every cut so that the lines focus on one point, making the flared joint pattern - you could use a more distant point to lessen the short grain if you wanted.
	It's a good idea to number the cutaway socket pieces in order to lessen the confusion! Saw the joints with the piece clamped fairly low in the end-vice using a. fine-toothed dovetail saw with a very small set on the teeth. This cuts a kerf just 0.5mm wide. Because the set is so small it cuts a very straight clean edge, but it is unforgiving. The first couple of strokes are critical to the line it will follow - after that there is no way it can be pulled back into line if it starts to wander. It is possible to get it to run just down the inside edge of a knife line so that no paring is necessary to make a well fitted joint - at least that is the theory. After sawing down the sides of all the joints, cut the sockets across a couple of millimetres above the depth line with a fretsaw. Level up the bottoms of the sockets with downward chisel strokes, starting with a fairly coarse cut and finishing with a 0.5min slice up to the line. By clamping the piece under a stout guide block of right-angled hardwood you can guarantee a straight chisel cut. This guides the chisel in two ways - firstly, it accurately sets the position of the chisel cuts up to the line, and secondly, by holding the chisel back flat against the block, you know for certain it is slicing though at right angles.
Now the joints should fit together. But how? Of course they won't slide longways or sideways - that's the whole idea of this kind of joint! However, they can be persuaded to fit diagonally at 45º by moving both parts of each joint inwards together. It is probably best not to try fully engaging any of the joints before the final glue-up because if they are a snug fit there is the risk of damage while pulling them apart. If just the tips of the corners are engaged, it is quite easy to sight along the lines of the joint to check that it is going to engage smoothly. Glue the joints with PVA using an artist's brush. The four corners are engaged together, then the box is placed between a sash cramp and a couple of G cramps, not forgetting first to fit the bottom panel in its groove! Closing the joints is a matter of tightening the three cramp screws a fraction of a turn at a time to keep all the joints moving inwards together. Of course it is necessary to do this sensitively because any tightness would force the joints apart and is likely to split the wood. With all four joints closed up, remove all the cramps so as not to distort the wood whilst the glue sets. When the glue is dry, clean up the joints, using a low-angle block plane. A micro-chamfer, planed along the top and bottom arris helps to make them more durable - it stops the long fibres forming splinters and removes the temptation to sand over the sharp edges and make the shape look soggy - chamfers catch the light cleanly and help to visually define the edges.