Medieval woodworking part 4: plank chest 2

Medieval woodworking part 4: plank chest 2

My nearest and dearest tell me that I have a lot of tools (they’re too polite to say “too many”). Each one, I say in my defence, has a purpose. No two are identical.

However there’s no doubt that there is some redundancy in the ranks. ‘Efficiency savings’ could be made. Some are more flexible than others. Chisels, for example can do a host of different tasks but require more skill to do some of those tasks well. Other tools perform only one task but they’re easier to use. Let’s take cutting a housing (dado if you’re American) as an example:

  1. Knife two parallel lines, chisel a wall to guide a saw, saw to depth, chisel out the waste (using a mallet), use a router plane to level the bottom. Six tools, one of which probably didn’t exist in medieval times.
  2. Clamp or nail a batten across the board. Use a dedicated dado plane to depth. Only two tools (if you don’t count the batten) but the clamps and dado plane are relatively recent inventions. The dado plane is about as dedicated a tool as you’ll find. It just cuts housings. At a specific width. Not adjustable.
  3. Clamp or nail a batten across the board. Spend half an hour setting up an electric router. Spend ten minutes finding and putting on protective clothing (mask, goggles, ear defenders, chainsaw trousers). Spend 30 seconds in abject, gibbering terror cutting the housing. Have a cup of weak tea while you recover.
  4. Get good with a chisel. ‘Knife’ the lines with the chisel, chop out the waste with the chisel. Level the bottom with the chisel.

IMG_3081I had two housings to cut. If this were not a masochistic act of reenactment and I were using a milder wood than oak I would have taken option 2. In the spirit of the thing I tried to only use tools I have evidence for having existed at the time. Option 4 it is.

Oddly I didn’t take a photograph of the result. It works but it’s not pretty.

Here’s the start of the second housing using option 1. I convinced myself that a plane, as they say, is just a chisel in a jig. Any idiot can find a video of Paul Sellers making a router plane with a block of wood, a chisel and a wedge on YouTube. What’s that you say? They didn’t have YouTube in the thirteenth century? Arrant pedantry.

With four rebates to cut I tried the chisel method again. This is easier than the housings. It’s a lot like cutting half blind dovetails but without the worry about nice crisp tails. I did even it up with a shoulder plane. And then used a moving fillister for the other three.

IMG_3101I marked the semi-circles for the feet with a pair of compasses. All period-correct so far. I’ve seen evidence of bow saws from the period but not a turning saw as such so I decided to see if I could form them with a saw, chisels and gouge. Kerfing went swiftly. Chiselling the waste out was quick as well. Fortunately I have a very large carving gouge at a fractionally quicker sweep than the chosen curve. Again – very straightforward.IMG_3116 I don’t think that a turning saw would have saved me a great deal of time and I’m much more confident with a gouge than a thin blade in 3/4″ of oak.

I did cheat a little bit and chamfered one curve with a spokeshave. The other I did with a chisel.

The only joinery left was to peg it all together. At this point I gained a lot of respect for the medieval joiner. The rebates on the ends of the front and back panels are there to help when assembling.

IMG_3122

Tight when made but worked a little looser by the time the chest was assembled.

Nevertheless unless the boards are dead flat and the rebates dead square putting this together without clamps would be next to impossible. I spent a lot of time with shoulder planes fettling every joint. A modern woodworker with good glues and a rack of clamps would pull a half millimetre gap closed and drive home a fastening, content in the knowledge that it wouldn’t show its seams before the customer got it home. The medieval craftsman had to get the joints perfect just to be able to drill the holes for the pegs. Some antique dealers use the word “primitive” to describe pre-industrial furniture with little ornamentation. One should earn the right to use that word about someone else’s work.

Sat astride my low workbench, the boards resting against pegs and being held by my knees I managed to drill a couple of holes. With a dowel at each end of the board holding it fast the others went in more easily. With a now-familiar sigh of relief I went back to my joiner’s bench and my clamps for the other three ends.

There’s surprisingly little joinery in a six board chest. With a jointer, thickness planer and a dado stack in a table saw (I wonder if the dado stack will be legalised in the U.K. Post-Bexit?) I imagine it would take less than an afternoon. For the pre-industrial craftsman with a frame saw, chisel, mallet and auger it’s a somewhat greater endeavour.

With the edges chamfered and the rougher grain scraped I put on some finish. The use of linseed oil was widespread in this period but its use on furniture is, as far as I know, undocumented. In the C12th Theophilus1 wrote in his treatise for artists that using linseed oil on wood is ‘tedious’. Boiling linseed oil to speed up polymerisation was known in the Middle Ages, as was adding lead to do the same thing, but Theophilus was apparently using it raw.

1. Theophilus (C12th) On divers arts Republished by Dover Art Instruction (2000)

Advertisements

Medieval Woodworking Part 3: plank chest 1

One down. At least three to go. The list of projects for this demonstration gets longer every time I open a book. With the workbench finished I’ve turned my attention to some ‘flat work’. Staked stools and demonstration tools  can wait. There are at least three chests I’d like to build for this: a Viking piece based on the Mastermyer chest, a clamp-front and this project.

I’m writing this half way through building  it and have I realised that I’m going to have to do it twice. Once to have a complete chest to show and then again to demonstrate the process. Fortunately it’s about as easy as furniture construction gets: a ‘six board’, ‘six plank’, ‘slab-end’ or just ‘boarded’ or ‘plank’ chest. Six bits of wood and some fastenings.

I’ve taken the opportunity to try a variety of techniques. There are several surviving chests that use pegs and at least as many that are nailed from around the period we’re interested in. Copper rivets1 were also in use and since I’ve got plenty kicking around from boatbuilding projects I’m hoping to squeeze some in. Two chests may not be enough.

Marhamchurch plank chest on white

A starting point: riven oak late medieval plank chest. Courtesy of Marhamchurch Antiques.

This chest will come from a single rough-sawn board. It is definitely a small chest;  height and depth are dictated by the available timber. The inspiration for this piece is from two different chests (above and below) that have been sold by Marhamchurch Antiques in Devon. They’re a bit later than the period we’re interested in but the only strong sign of this are the trefoil arches on the first chest that form the feet. Other than that either could have been built hundreds of years earlier.

Marhamchurch flat sawn chestSix board chests are built with simple construction rather than longevity in mind. The grain on the sides is perpendicular to the grain on the ends. The movement of the timber will eventually cause the fastenings to loosen and the boards to split. Despite this many survive because there are ways around this:

  • Version 2Keep it small. The narrower the long planks the less movement there will be and so the fastenings are more likely to be able to resist the seasonal changes and the  boards won’t split. I estimate that the front board from the riven oak chest from  Marhamchurch Antiques is about 18′ high and even in riven stock that creates a lot of movement causing a split near the bottom.
  • Build it from cleft (riven) or rift-sawn stock. Movement across the grain in the tangential plain (at a tangent to the growth rings of the tree) is twice that in the radial plain (a straight line going from the middle of the tree to the bark). Radially cleft, or riven, timber was common in Anglo-Scandinavian England2. This means that the pieces will inevitably be narrower than flat-sawn panels and the movement will be significantly reduced. You can see this in the first chest to the right but not in the second which appears to be built of flat sawn boards. In the second chest the sides may have shrunk away from the notches at the bottom leaving a gap. In the first they have not.
  • Marhamchurch flat sawn chest strapsUse metal straps. Many surviving large six board chests have substantial metal reinforcements. However a lot of these are ecclesiastical pieces or come from the estates of powerful and wealthy individuals. They are not the focus of this demonstration. The second chest from Marhamchurch compensates for its flat sawn grain by reinforcing the joints at the top. It’s hard to know whether this was done at the time of manufacture or later as the joints worked loose with the seasons. The position of the strap over the nail and the difference between the two nails suggests the latter.

The design of this chest is dictated by the timber. Though the Egyptians had glue thousands of years before this time period I can find no evidence that the Anglo-Saxond or Anglo-Scandinavians did. Even if they did it was animal glue and would have faired poorly in the damp. That means that wide panels made up from a number of boards were not used. The front of a chest was as long and as high as the planks as the widest board you could get from a tree.

Forklift and boardTime spent picking through boards at the timber yard is well rewarded. (And so is being nice to the man with the huge forklift. I learnt a lot from him that day.) I wanted to select wide stock with vertical grain. Access to large-diameter fresh cut oak is rare so riven stock is out of the question and that left me searching for the widest boards sawn from the middle of the log.

With a newly commissioned low bench sitting in the garden I started by trying to work the boards to their final dimensions in a manner that might be familiar to a medieval joiner. Sawing went well. It’s little more than a giant saw bench and I’m comfortable using it as such. Heavy planing was more of an issue. Bent forward, pushing a plane with just my arms was hard and uncomfortable work. When standing at the traditional joiner’s bench many more of my muscles help out. The much bigger and more resilient muscles of my legs do a lot of the work.

Frustrated and realising that I would wake up the next day aching and with much less work done than I had planned I went back to my familiar friend – the joiner’s bench. There is plenty of time to play with the low bench between now and the  demonstration.

Stickered oak.jpgI planed each piece to just over 3/4′ leaving a little spare to remove once the boards had moved a little more. A couple of weeks in stick and they were ready for joinery.

This gave me a bit of time to decide on fastenings. Chris Schwarz3 asserts that nails are the the secret to surviving chests’ longevity. They “allow you to get away with serious crimes of wood movement… a nail will allow the wood to expand and contract, bending back and forth through the yearly humidity cycles.” This also means that they will loosen over time. But it’s the best of a bad job.

Dowel plate

Not a medieval dowel plate. Its laser cut descendant.

Nails were made by blacksmiths, each one hand forged. So if a joiner wants to join a chest with nails s/he must buy (or trade for) them. But a dowel plate is a one-off purchase and then you’ve got pegs for life.

There aren’t enough extant medieval plank chests to be able to draw a conclusion about what was used more – pegs or nails – but in the interest of self-reliance I’ve used the trunnels for this chest.

Next time: joinery.

 

1. “Radially split offcuts of other species include… an alder fragment with 30-40 copper alloy rivets embedded in it…” Morris, C., Wood and Woodworking in Anglo-Scandinavian and Medieval York p. 2221-2223
2. “Most conversion of wood for artefacts was done by radial splitting…” Morris, C., Wood and Woodworking in Anglo-Scandinavian and Medieval York p. 2104
3.

Even professionals get it wrong

imageBefore I had the time and space (and skill and confidence?) to build such a thing my wife and I bought an oak sideboard. It purported to be made in England from domestic oak and I have no reason to doubt that. It’s got big, chunky, machine cut dovetails and the drawer fronts are applied. We still like it a lot. It’s a simple design with little decoration, nice proportions and is solidly constructed. The flat panels and simple, turned wooden pulls might be described as ‘Shaker’ by those who need to put a label on simple vernacular furniture.

At the same time we bought a ‘matching’ blanket box. It has an issue that I should have spotted even then. We were clearly carried away with the sideboard and didn’t pay enough attention. Those furniture salesmen eh? Wiley Devils. Here are the tops of both. Can you see the problem?

On the left is one panel of the top of the sideboard. The panel is fitted into an open backed frame and is rebated so that the surfaces of the panel and frame are flush. As the seasons change the panel expands and contracts and sometimes (as now at the end of a warm, damp summer) protrudes beyond the back of the frame. Because the grain is orientated so that it runs lengthways it will never expand along the length of the piece and so will never distort the frame. Good thinking sideboard-maker.

For whatever reason (lack of longer lengths of timber? ignorance? malice? who knows) the blanket box maker has orientated the boards so that they run front to back putting all of the expansion and contraction along the length of the box. Here is the result:

image

By the end of winter that reveal is even. This tells us two things: the box was built in winter when the timber was at its least moist. And it was built by someone who didn’t know about (or care about?) seasonal movement.

Hoadley’s calculation for wood movement tells us that the change in dimension across the grain is equal to the initial dimension of the timber multiplied by the total shrinkage percentage of that species of wood multiplied by the product of the change in moisture content divided by the fibre saturation. There are going to be some estimates in this!

ΔDi=DS(ΔMC/fsp)

ΔDi=(31.5)(0.105)((0.28-0.22)/0.28)

ΔDi=0.95

This would be bad enough if I was working in centimetres but I’m using imperial! Almost an inch of potential movement. I’ve been fairly conservative in my estimates and we still end up with horribly large number. Some cabinet makers use a rule of thumb that wood will move up to 4% during the seasons. That’s 1 1/4 inches.

Note to self: Use narrow panels. Work out where the extra wood is going to go.

img_0324

When I made this panel at the start of the summer I broke the rules in the same as the way as the maker of our blanket box. The timber is flat sawn and so has more movement than vertical grain planks and the long grain runs in the shortest dimension.

Aware that I was breaking these rules I cut the groove in the frame deep and let the panel in half way into the groove.

It moves around a lot if you make it but it’s never going to show a gap or prise its frame apart.

 

 

Fold-down bedside cabinet

Fold-down bedside cabinet

Fold down bedside cabinetMost of my work starts with something I’ve seen: a plan, a boat, a piece of furniture, a lump of wood with something hiding inside it. This project started with an empty space.

The space was eight inches deep, two feet wide and eight feet high. Somewhere in it I needed to put a cabinet that would hold a couple of books, a phone, a glass of water etc. All the things  someone wants by their bed. But eight inches makes for a very odd cabinet.

I had no starting point. I’ve never seen anything like the thing I needed to build. Somewhere in a corner of my brain this was lurking:

foldingsinksm I had seen a folding sink at a boat jumble. It was in horrible condition but an interesting relic of a bygone age. This drawing is from Shipmate who are planning to start producing them again. Things are looking up.

I love metamorphic furniture. Kenwood House is a healthy walk from my house and has a few pieces including library stairs and a folding desk you can stand on to get books from high shelves.

Ideas were colliding. A plan wasn’t exactly starting to take shape so much as coalesce.

The design wasn’t based on beautiful proportions or the rules of classical architecture. It was based on how tall a glass is, how thick a substantial novel is likely to get, what height is convenient when lying in bed and how thick flush-fitting hardware is.

I had to consider proportions, of course, but I tried to keep the piece as small as possible and as large as necessary. When all the requirements were laid out on paper it started to design itself. My biggest decisions were how thick to make the frames and how to lay out the grain on the shelf and panels.

Once the design was finished it presented some very diverting joinery: lapped and through dovetails in the drawer, mitred dovetails in the shelf, tongue and groove joints to hold the carcass together and mortise and tenons in the frames. The panels are rubbed joints in very thin stock. Lots of fun.

Overall I’m happy with it. There are half a dozen things I’d change if I were starting again but the client is happy and therefore, so am I.

There’s the rub…

imageI don’t use rubbed joints very often though that may now change. I’ve always been a bit suspicious of their strength. Or otherwise. Last night, when I was finishing a number of glued panels, I ran out of workspace. The carving bench, which I use as an assembly table when I’m not carving, was covered with clamps and panels and the workbench was similarly festooned. I had one small panel left but no space to lay it flat. The plan was to glue it over night and then saw it into two, creating two 1/4″ pieces from one 3/4″ glue-up. But I didn’t have space.

I sometimes use rubbed joints for this kind of work – grain matching and an invisible glue line are important. Strength isn’t. The panel will float in a frame and receive very little force.

However it had to be strong enough to deal with the sawing and planing involving in ripping it in two. With the edges jointed I applied the rapidly cooling glue and rubbed the joint; it stuck in moments. I left it in the vice for the night and tidied up.

This morning, bleary-headed, I carried some boards out to the workshop, forgetting the panel in the vice. Whilst maneuvering them I gave the panel such a clump that the bench moved. Fearing the worst I gave it a wiggle and realised it was still in tact. Wondering how much the glue joint would take I lifted I up by the top board and manage to raise the front legs of the bench from the floor. I’ve no idea how much force it will take but I’m confident that it’s strong enough for any normal application.

Suddlenly  my panel clamps are shuffling their feet and trying not to catch my eye for fear of an early retirement.