Individually Styled and
Crafted
|
and needed a dining
table to complement
a great room with a high ceiling and lots of glass overlooking a
beautiful creek. For some months we worked together to design that table, meeting when we could around
their busy globe trotting schedules. We wanted an "organic"
design that would go well with the very natural views from the room,
and
glass in the top was a prerequisite. Together, we came up with a
very nice, and unusual design and I completed the table in February. It
is solid cherry with two glass inserts, and incorporates a "backbone"
of laminated cherry and red oak to provide the strength required. You
can see it in the photo to the right.


consistently for all the chairs I needed
to make templates to trace the shapes. This is where CAD once again
shines. I was able to easily draw templates of the legs from a front
and side view, print them out, and then use the paper templates to
transfer the shape to more durable 1/4" plywood. I can save the plywood
templates if I ever have to reproduce these chairs. The photo to the
left shows me deftly scissoring the paper templates, and the photo to
the right shows the finished plywood templates. Now I can lay the
templates on the candidate boards and see what works. Once the lumber
is selected and rough cut to size, I can let it sit for a while to
settle. Wood is always moving, especially after pieces are cut from
large planks, so a settling in period is necessary, especially for
larger dimensions.
large large at first so I could later
shift the template a bit to miss any defect I might find. Another part
of cutting these large bends is that it is easy to go through a lot of
lumber if not thinking efficiently. Were I to cut a single back leg in
way that would make working the piece the easiest, it would take almost
four board feet of lumber. At the price of thick cherry, that is a lot
of dough, like $30 per leg! So I try to shift the template and stack
the curve of one leg inside the curve of another to reduce the lumber
used. It still takes a lot of lumber to make curved parts, but thinking
efficiently can cut the cost to the client dramatically. The final
tricky part of cutting all these chair parts is
that the four chairs will be a set, so as I said earlier it is
advantageous to cut all the parts from the same tree. Cherry is not
cherry is not cherry. The color and grain varies with the locale in
which the tree grew, as well as the way the tree grew and the part of
the tree the lumber came from. Fortunately, I was able to get all the
parts cut from the same tree's lumber so that
all the chairs will match. To the left
is a shot of me bandsawing some of this thick cherry for a front leg.
Once all the parts were rough cut, I set them to settle for a while,
as mentioned earlier. It is important to let air flow freely through
the stack, so although the photo to the right looks a little strange,
you have to admit that air can flow around all the legs. The astute
observer may notice that there are a few extra parts in the stack. It
is prudent to make a couple extra of very unusual and difficult parts,
because invariably there will be a problem with one leg or a rail, and
it is a whole lot less expensive to cut the extra part now, as part of
a production run, than to re-set the tooling to cut a single part later
if it becomes necessary. My policy is to charge the client only for the
actual parts used in the piece(s); I save the extras, if there are any,
for a later project (like a replacement chair) or to use as
demonstration pieces to show how a particular joint or finish looks.
that reason, to prevent changes in the
laminations due
to creep. But I'm getting ahead of myself. The first task was to make
the form around which the layers would be bent and glued. I used medium
density fiberboard (MDF) to make the form, actually laminating layers
of the MDF to get the required thickness. The shape of the form, seen
in the photo at left, was exactly that of the crest inside curve. It
was a simple matter to make this curve, again using my CAD program to
print out a full scale drawing of the
crest that I could trace onto the MDF. I also had to construct some
curved gluing blocks that would apply even pressure on the laminations
when making the curve. In the photo you can see me clamping the layers
of the gluing blocks together to make the required thickness to match
the form. While
the glue for the form was drying, I made the layers of oak and cherry.
Using the bandsaw, I cut strips of cherry and oak boards, a little
oversize so I could later precision thickness them and ultimately trim
the glued up crest to size. For this run of four chairs, I cut a total
of 44 strips, plus a few extra. The photo at right shows how this was
done. After band sawing the strips, I ran them all through my
stationary drum sander, kind of a
thickness sander, to take out the saw marks and make each strip just a
hair under 3/32 inch thick. The photo at left shows the sander at work.
I used 100 grit sandpaper on the drum, which quickly removes material
and does not tend to burn the wood, yet provides a smooth surface with
a little "tooth" for gluing. If you look closely a the photo, you can
see a 4 inch hose coming off the top of the sander. That is a dust
collection hose which leads to a six inch steel duct which leads to a 2
horsepower cyclone dust separator. When sanding, or when doing any
cutting operation, dust collection is a must. Not only does it leave
the shop cleaner, but it protects my lungs from permanent damage. If
you ever visit my shop, be sure to ask to see the dust collection
system -- you will be impressed!
easier. (I wonder why we never completely
shifted to the metric system as talked about so much years ago. It
really is easier making measurement calculations in metric!) I
estimated that I would need 78 grams of mix (that just looked right in
the mixing dish), so I made that much and brushed it onto the fronts of
five strips. I then used a glue spreading roller, which is simply a
four inch wide foam rubber roller with a handle, to even out the glue
on the strips. I then stacked the strips and added the sixth strip on
top and placed them on the form. Using clamps, I gradually made the
bend, starting from the center and working out, and clamped the work
down hard to dry. Here is a photo of the clamp up. Tomorrow morning I
will be able to take off the clamps and see how I did. Then just repeat
seven more times! By the way, I had some glue left over -- next time I
will mix up only 52 grams. At least I over-guessed and did not have to
mix more glue in the middle of the glue up. That would really have done
a number on my stomach!
had to obtain reference faces on the
legs on which the mortises for the rails would be cut. As I discussed
earlier, the easy way to make these legs would have been to use a
gigantic block of wood for each leg and simple square the block up and
choose the reference faces from the squared blanks. They I would simply
cut the mortises and bandsaw the shape of the legs. But that would
waste a lot of wood, so I cut the legs first on the bandsaw, oversized
and approximating where the reference
faces would be, then I faced two adjacent sides where the seat rails
would intersect. I found it very quick to face the reference faces
using the oscillating belt sander, seen in the photo at right. This is
simply a big stationary belt sander with a precision table on which to
rest the workpiece. If the table is set 90 degrees to the sanding belt,
it is easy to make two faces 90 degrees apart. I also used this sander
to shape the convex sides of the legs closer to the finished
dimensions. The concave sides of the legs were shaped on the
oscillating spindle sander, seen in the photo at left.
these legs in such a way that mortises
can be cut, either by hand with chisels or by using machinery. This
gave me pause as I recounted my options. Fortunately, a new tool had
just been released by the manufacturer that could cut mortises in situ using a machine that looks
an awful lot like a biscuit joiner. Now, before you ask me what that
is, take a look at the photo at right and you will see the machine in
operation. It is trademarked the Domino, because it cuts a mortise in
which a floating tenon can be installed, and the floating tenons look
like dominoes! By the way, a floating tenon is a small block of wood
that is inserted into mortises cut in both the leg and the rail. In
conventional joinery, the tenon would actually be part of the rail. I
must admit that I was not initially sold on purchasing
the Domino, as I am used to traditional joinery, but I researched the
tool on the internet and all I found were rave reviews. Just to be
sure the joints would be strong, I made up a dummy joint using a couple
of scraps of wood and tested its strength. With a single
domino in an 'L' shaped assembly
mimicking one rail to leg joint I put my full weight on the end of the
'L' and the joint held me up just fine. I did this several times to
make sure. Now, this is a weight test far in excess of anything a chair
would see, but chair rail mortise and tenons usually fail by another
mechanism. With years of people pushing a chair out from a table
without relieving some weight off the chair, the back leg to side rail
tenon often loosens up. To combat this, the tenon is often pinned in
place with a small dowel, so this is what I will do here, as well, to
ensure the Domino joint never comes loose. Also, I can use multiple
dominoes, as in the photo at left, increasing strength proportionally.
The added benefit of this method of joining is that set-up is quick and
the mortises can be cut quickly, saving my client money. All in all, a
win-win situation.
work the back rail, which was a little
more complicated. It would be curved in two dimensions. The top edge
would be curved to match the dish in the seat cushion and the front
edge would be curved to match the dish of the back support slats, a
shape designed to fit the back of the occupant. Additionally, the back
rail would need mortises for the back slats to fit in, and both the
front and back rails would need through mortises
for a
center rail running fore and aft under the seat cushion. I would attach the seat cushion to that
center rail, and it would provide cushion support and stiffness to the rail/leg assembly. I used this center rail technique on chairs I built over 25 years ago and use daily,
and it works well. Getting back to the mortising, after drilling the
holes at the ends of each oval shaped mortise, I chiseled out the waste
between the holes. I did this for all
four slat mortises. I repeated this technique for the center rail mortises, using a 3/8" bit this
time. The
photo to right shows me drilling the center rail mortise holes and the
photo to the left shows me chiseling out the waste between the holes to make an oval shaped
mortise. By the way, the center rail got a double mortise and tenon joint which
is why you see two mortises in the photos. When I discuss cutting the
tenons, all this will make more sense, but for now just stay
with me.
After getting all the mortises cut, I
bandsawed the curves. The photo to the left shows me cutting the
front side curve (on the back rail). I pretty much breezed through the
explanations in these two paragraphs, but I assure you that there was
a lot of figuring and
double checking to make sure that
everything would be where it should once the chairs were assembled. I
think that is what makes chair building with lots of curved components so
difficult. There are very few
reference edges, that is, there are few square sides so it is not a
simple matter of just sawing the parts and putting them together. It is
more like
figuring the positions of the parts in thin air with nothing to refer
to. Chair making is for the adventurous! But, let's continue. With the
rails all shaped, I started work on the center
rails. As I said, these used a double tenon, and the photo to the right
shows me chiseling waste from between the tenons. These tenons were
tricky, because of that reference edge problem I keep talking about.
The
front and back rails were set perpendicular to the floor, while the
side
rails tilted back at one degree. So the center rail had tenons that
were horizontal to the floor, that is, cut perpendicular to the front
and back rails, but the center rail itself tilted at one degree. So, in
reference to the center rail, itself, the tenons were tilted one
degree.
Does that make sense? At least you did not have to figure out how to
cut them! For now, I only cut the tenons on the front of the center
rail. I waited until the chairs were dry assembled to measure the exact
length of that rail before cutting the back tenons.
square to the centerline of the legs so
that the crest would fit squarely. By assembling the back rail to the
back legs I was able to make this measurement and mark the legs for the cut, then using a hand saw and
edge belt sander I squared up the ends. Now there was the issue of
cutting the crest ends to the correct length. Remember, the crest is a
curved laminate, so figuring where to cut required refitting the crest
to the form I used to make the laminates in the first place. I had
marked the correct cut points on the form, so I could transfer the marks to
the crest then make the cuts on the radial arm saw. Now it was time to
cut the joints, themselves. These were very complicated joints. They
had to accommodate fore and aft stresses on the crest, side to side
stresses, tensile stresses, and they had to housed the upper "sub
crest" with mortises to hold the back slats. If you refer to the CAD
drawing above this will all make sense. Instead of showing you the step
by step process of making the joint, followed by a photo of the joint,
itself, let's start with a photo (at right) of the completed joint
being assembled so you know where this is going. You can see how
complicated this joint is, and remember, I had really no good reference
from which to cut the joint. There was some machine work in cutting the
crest, but only hand work in cutting the ends of the legs. On the
crest, I first cut the basic tenon shape as shown in the series of
three photos that follow. This was done on the table saw and radial arm
saw. First I cut the tenon cheeks, then the shoulders so I ended up
with one tenon on each end of the crest.


Next I cut the
tenons in two, with a space between that would form a mortise to house
the sub crest. I did this cut by hand, using a Japanese pull saw. The
photo at left shows me cutting that joint. This
was not a simple straight cut, but it had to go extra deep on the
inside of the crest, but not
show on the outside of the crest,
to form the sub crest mortise. After the cuts were made, the waste had
to be chiseled out to complete the joint. With the crest jointed, I took on cutting the
joints on the top ends of the back legs. First, I clamped up the legs
to the back rail
and temporarily clamped the crest
to the tops of the legs, then I laid out
where the cuts had to be by transferring the positions from the crest
joints. After extending these marks on all sides of the legs and adding
other marks I needed, I used the Japanese pull saw to make the cuts, then used a chisel to
remove the waste, shown in the second
and third photos to the left. The
leg joints were mirror images of the crest
joints, as you can see in the photo I showed you earlier. Here is what the joint looked like fully
assembled (photo to right). As you can see, the shape is still rough,
and I had to do spokeshave work to
get the shape to the final
configuration before final assembly and glue up, before the sub crest
was installed. So, using a few clamps I dry assembled the back legs,
back rail, crest assembly and shaped the crest joint areas to very near
the final size, only to be touched up after gluing. The photo to left
shows this process using the spokeshave.
With the parts pretty much
shaped, I could glue up all the mortise and tenon joints for the rails
and legs. The photo to left shows all four
chairs glued up on my gluing bench. After the glue had cured, I took a detour to strengthen the
rail to leg joints. If you noticed, this chair
design has
neither arms nor leg stretchers, making the back leg to side rail joint
extra critical. All the torsion is concentrated at this point. Although
many chairs use this design, including ones I have used as my everyday
dining chairs for 25 years, I wanted these chairs to be really strong,
since I would not be around to observe them after delivery. There is a
tendency for this joint to loosen up, so in addition to pinning all the
tenons with dowels, I developed a
glue block with long stainless steel screws to attach in the inside
corner between
the side rail and back
leg. I did this for the joints at
the front legs, as well. The blocks are glued to the side rails and
screwed into the legs. The photo to the
right shows these blocks in place. I
trimmed the blocks after the glue had cured, and then I took the
opportunity to
craft the seat bottoms to which the
foam and upholstery would be attached. These seat bottoms would be a
laminate of two pieces of 1/8" thick hardboard, the kind used
for pegboards in workshops. First, I made a cardboard template of the
place the seat bottoms would fit on the chair, shown in the photo to
the left. Then I rough
cut pieces of hardboard and
laminated them into a dished shape. I used a pair of forms shaped like
the curves of the front and back seat rails. I simply took two pieces
of 2x4 and bandsawed the curve of the seat in each piece so later I
could simply clamp the hardboard laminates between them to form the
dished seat bottom. The photo in the middle right shows how I glued up
these laminates. The finished seat bottom laminates were then cut
according to the cardboard template and were trimmed to fit nicely
inside the rabbets of the seat rails. The photo to the right shows one
of the seat bottoms in place.
shape the back of the person
sitting. With the subcrests dry fit, I could measure the length of the
back slats. These slats were to have a curve to approximate the back of
the sitter, but of course, everyone is a different size and fit, so I
made three different versions of one of the slats and tested them with
both myself and my wife. Shown to the right are the three versions, and
I chose the one with the least curvature so it would not push the
sitter too far forward. As it turns out, this was the one closest to
the curve I had put into my original shop drawing of the chair. I
needed a total of 16 of these slats, so I made 18 for good measure.
Then, after dry fit checking and some tweaking of the fit, I glued the
crest, subcrest, and back leg joint with the slats installed
(photo to left). This joint must be
strong, so I was careful to make the fit as perfect as I could and the
interlocking of parts and gluing as secure as possible. As it turned
out, the joints glued up very strong, but I did not like the feel of
the curved crest on my back (although my wife though the fit felt nice
- see what I mean about different shaped bodies fitting differently in
chairs?). So, I used my spokeshave to cut a little inside curvature on
the crest to better fit the back of the occupant. The photo to the
right shows me making that cut.
center rail. Two bolts would come up from
the bottom of the center rail and into the "tee" nuts embedded in the
seat bottom to hold it tightly to the center rail. Next I glued 2 inch
foam to the seat bottoms, using a fast setting spray adhesive, then
trimmed the foam to the exact shape of the
seat bottoms. Using a sharp blade, I then
chamfered the sharp corner of the foam so the upholstered seat would
have a rounded edge. The photo to the right shows this stage of work. I
wanted the upholstery to have a smooth loft, so I then applied two
layers of batting to the foam. The first layer was only around the
edges, padding the edges of the laminated seat bottom to reduce wear on
the upholstery fabric at this point. The second layer of batting
covered the entire seat and was taped to the bottom of the laminate.
The finished product is shown at left. If you look closely, you can see
the green foam and the first layer of batting around the edges. With
the batting in place, it was time to install the upholstery fabric.
This is where it got dicey. I did not want to screw up any of the
fabric my client had provided, as we ordered exactly the right amount,
so I borrowed some scrap upholstery fabric from my wife and practiced
with that until I knew how I would proceed. Then, with stapler in hand
I went about installing the finished fabric. Since the seat bottom had
notches cut out around the legs, this was not a particularly straight
forward process, but the seats turned out well. Now, as you
may know by now, my furniture is
finished all around, bottoms, backs and insides included, and my
upholstered chairs are no exception. So, I cut pieces of fabric large
enough to sew to the bottom of the seats so that when the chairs would
be turned over, a nice upholstered bottom would be seen. That gives me
closure, and I will not build a piece that cannot be turned over and
enjoyed, although on on some of my work turning it over would be quite
a feat! Again borrowing sewing supplies from Kathy, I sewed the bottom
fabric to the seat upholstery, which had the added benefit of allowing
me to tighten up the corners a bit for a more finished appearance. The
photo at right shows me sewing, and I didn't stick my finger once!