Individually Styled and
Crafted
Fine Hardwood Furniture
by
Andrew Pitts ~ FurnitureMaker
Work in ProgressPersimmon Shelving Unit with Drawers
Article Completed 21 December 2009
Individually Styled and
Crafted
|
had done some research and discovered that persimmon is
from the same family as ebony and shared some characteristics, so it
could be fabulous wood. I also learned that persimmon is a difficult
wood to dry and that I could expect some degradation, such as cracking,
during drying. We decided to give it a shot, and if during milling it
looked like the tree would not yield sound lumber we could call it
quits at that
point. So, with trailer in tow, I went to the site to recover the logs.
Fortunately, my client owns a tractor, so 
with his tractor and my trailer winch we got the logs onto
the trailer. The photo at right shows just how large the two log
sections were. A few days later I set up to mill the logs. I
wanted to get them into the kiln right away for a couple of reasons.
First, I wanted to arrest any further decay by drying the lumber below
the point where decay thrives. But, I wanted the drying to be slow, and
a solar kiln in the middle of winter dries really, really slow! During
the milling, my clients visited to watch the process, which can be
pretty exciting, especially when your log is opened for the first time.
In the photo at left you can see just how nice the wood looked on the
mill.
Immediately
after
milling I charged the solar kiln with the lumber - often
I will air dry the lumber
for
a while, 
but as I mentioned, I wanted to kiln it right away.
The photo to the right shows the charge in the kiln, along with some
red oak I milled at the same time. The oak in the kiln afforded the advantage of
providing some heat sink so that the kiln would not heat up too much
during the day. The more lumber in the kiln, the more there is to heat
up during the few hours my kiln receives direct sunlight so the maximum
kiln temperature will be lower before the sun goes behind the trees. It
took until early June for the wood to dry to the 7-8% moisture content
necessary for furniture kept indoors around here. When I
removed the wood it looked good, for the most part. I would still have
to work around a lot of cracks, but I assessed that we would be able to
build a piece of
furniture with it! The photo at left shows the best pieces of all the
lumber, and you can see where the tie to ebony comes in.
Turnpike, but I made it. After they loaded the big box onto
my trailer using a big fork lift, I asked for directions home. I think
the directions must have been pretty good, if I had actually taken them, but a wrong turn
early in the homebound trip sent me through some of the nicest parts of
west Richmond, but
unfortunately not the best way to go with a 17 ft trailer and a heavy
load. 
Lesson learned - just stay put here on the Northern Neck
whenever possible, and when I do have to travel afar, don't forget to
bring a map! When I got home, I could barely pick up the big box with
my tractor (take a look at the box on the tractor at right), but I
managed to get it into the shop and onto the new cement slab I had
poured this spring. It took several days to uncrate, move, electrically
wire, and hook up dust collection for the machine, but it looks pretty
good in the shop, now and I am ready to make wide boards flat.
about 12.5 pounds per square inch. So, the top would see a
total force of 12.5 psi x 1900 sq in = 23,750 pounds, or about 12 tons.
The bottom of the
form would see the same force, so the forces would equalize. However,
the form itself would have to
withstand
these
12 tons of
force without
crushing. Hmmm ...
12 tons ... that's about six
Volkswagens! The sides would see great force, as well, more like 2-1/2 tons.
Better make the form strong! The photo at right shows the form upside
down, and you can make out the bottom and 
the ribs. The photo beneath
shows the form flipped right side up. The way I made all the ribs the
same was to make one exactly right then use it as a template to form
with a straight bit in my router, as seen in the photo to left. I
added some cross ribs between
the outer most ribs and the next ribs in, and then onto the ribs I
screwed two layers of 1/4 inch plywood as seen at left. I put this form
into the bag for a dry run, and when the vacuum got to about
half-way,
the form started creaking. At three quarters vacuum it creaked enough
that I dropped the vacuum, removed the form from the bag, and put a lot
more screws into it. Then I tried again and made full vacuum, leaving
the form under vacuum for several hours. No crushing! Now I am ready to
actually glue up some laminations.
decide which plank would go where, and in the photo at right you can see my
jointer room clobbered with all the planks. It made walking around
tough! I decided
upon
a plank that I could resaw into thin strips, but since that
particular plank had damage in the core I ripped it into two smaller planks, jointed a face
flat, and then resawed it into strips about 3/16" thick on the bandsaw, as seen at left. My objective was
to get 1/8" thick strips after sanding them to thickness on the
horizontal drum sander, as seen at right. I used 100 grit paper, since
that is a good roughness for gluing. Fortunately, I ended up with
enough wood to laminate both faces of the large curved side
piece. Next, I decided upon the veneer layout. This process was easier
said than done, and it took me a while to get it right. Here's the
problem; I was veneering both sides of the curved core, and I wanted
the inside grain pattern to match, as much as possible, the outside
pattern so it would appear as though you were looking at a single
curved plank of wood. Now, with the shop cut veneer I was using here my
best hope was to at least get the match close, and knowing that nobody
could actually see both sides of the plank at one time I took advantage
of that fact. Additionally, I was planning to book match the
strips as much as possible, and at the same time work around the
defects. I tried and retried different
combinations until I had the grain patterns optimized, then
carefully planed the edges of the veneer and laid them up on a large
table, temporarily taping them in place relative to one another using
blue masking tape, the kind that pulls off easily. When I was convinced
that I had both side's veneers correct, I used veneer tape to assemble
the pieces, taping on the unglued side (the show face) of the veneer,
as seen at right. Veneer tape is very thin paper tape that has a water 
activated adhesive, like on the flap of an envelope. As the
tape dries on the wood, it shrinks a little and helps pull the joint
tight so when the veneer is glued to the substrate it will appear to
have no glue line, or at least that is the objective. The photo at left
shows the concave face veneer taped up into a panel. I've got to add
that with an asymmetric panel like this, I really have to keep my
situational awareness about the orientation of the veneers. If I got
confused, I would tape up the pieces as a mirror image of how they
should be, and if not caught before gluing, the intended show face
would actually become the glue face. But this day I was not confused,
and the layout looked good. Next came the making of the core for the
panel. The chosen wood was tulip poplar because it is plentiful,
economical, tight grained, and easy to work with. The reason I used
solid wood for the core was twofold. First, I wanted to ensure that the
panel would shrink and swell seasonally similarly to the shelves, so
keeping grain orientations the same was paramount. I had thought of
using 1/4" plywood for the core layers, which certainly would have been
less expensive than making the core layers from planks, but a plywood
core would be
exceptionally stable and resist the movement of the shelves
mortised into it, making for danger of cracking of the shelves during
the winter when the heat is on and the air inside the house gets really dry. Secondly,
although the edges of the panel would be ebonized (stained black), I
still wanted to retain a grain texture on the edge and plywood would have to be filled and
smoothed, lest it look like plywood. So, although the poplar grain is
tighter than the persimmon grain, it was still a good compromise to use
poplar. If I had an unlimited supply of persimmon, I would have simply
use that for the core, but my supply was pretty tight for the job. I
first had to select
the poplar planks, face one side on that nice new big
jointer, and then resaw the planks into pieces a bit shy of 3/8" thick,
as seen at right. I went for 3/8" because the bandsaw leaves marks that
have to be planed or sanded away to achieve a final thickness of 1/4".
Since 1/4" is thick enough to use the planer, and the planer is far
superior to the drum sander for thicknessing I was able to use my 15"
planer to do the job, as seen at left. For those not familiar with the
planer, it has a rotating drum with 15" long knives (sharpened like a
very wide hand plane blade) that cuts the top of
the plank as feed rollers move the plank along a machined
table. This little Makita 2040 planer, no longer made, is sweet and
does a nice job. With the laminates planed to thickness, I took a light
cut on the drum sander to roughen the surface for glue and then started
the layout to get layers in the same shape as the veneers. I had to
stager the joints so that the seven layers of poplar would glue up to a
strong panel, and I wanted the joints fairly tight as well. The photo
to right shows all seven layers on my saw table. Now came the hairy
part - the glue-up. The plan was to use Unibond 800 resin glue to make
a total of seven poplar and two persimmon layers into a big sandwich.
Of the seven poplar layers, only four were full height. The last three
layers were cut progressively shorter so that after gluing them I could
smooth what would be the inside, convex face of the panel before
attaching the inside veneer. That would give the panel it's thicker at
the bottom than at the top look. I was unsure of how the gluing would
go, and because I have to work fast and get the work clamped before the
glue sets I decided to
keep it simple and glue only the bottom veneer and the
first layer of poplar core. As it turns out, that was a good move
because I had some difficulty keeping all the parts aligned and got the
assembly finally clamped just in time. These kind of glue ups are a
real sweat exercise! The photo at right shows the work on the form and
in the vacuum bag with the opening of the bag clamped shut. You can
also see the vacuum pump off to the right. It is a venturi pump, using
compressed air to draw a vacuum and I can achieve about 25 inches of
vacuum, or about 12.5 psi pressure on the work. I learned some lessons
with that first glue up, and in the second glue up I also attached two
additional layers of poplar, although this time there were actually two
glue layers to contend with.
lot of contrast between the work, which was in the shade,
and the bright deck behind. After power planing, which is just a rough
shaping, I brought the piece inside and clamped it to my bench. Then I
used a smooth plane and then a jack plane to really smooth it and make
it truly flat across the board and with a smooth continuous
curve, as seen at right. A smooth plane is what you normally think of
as a hand plane, and a jack plane is just a longer version. I like to
use low angle planes, where the blade sits at a lower angle than in a
conventional plane. If you are interested in why, shoot me an e-mail and we
can talk. While I planed I noticed a curious thing. The side seemed to
be bowed up slightly across the plank at the lower (thicker) end. Now,
I knew that the form was true, and I knew that the laminates were all
the same thickness, so why the bow? Hmmmm.... here's my theory. I am
vacuum clamping, and at this point I had gone through four iterations, adding layers as I
went. That means the bottom two layers saw the vacuum four times,
usually over night, so they were under vacuum for something like 48
hours. The next two layers for 36 hours, etc. In a vacuum, the partial
pressures of gases in the atmosphere (a vacuum) are zero, since there
are no gases in a vacuum, while the gases inside the wood, and that
means potential water vapor from the moisture in the wood, have
what we would call a normal partial pressure exhibited before the wood
went into the vacuum bag. So, there is a natural diffusion of the water
vapor out of the
wood (we call that drying) and into the vacuum in an attempt to
equalize the partial pressures inside and outside of the wood. When the
pressures 
inside and outside equalize, there is no more diffusion.
But, my vacuum pump keeps pulling any gases inside the bag out, so the
pressures of the gases inside and outside the wood can never equalize
and drying goes on and on. I think the layers that were under vacuum
for the longest dried the most, and that means they shrinked more than
the next couple of layers, etc. This naturally put stresses on the
piece between pairs of layers, such that the concave side, or the first
layers glued, tried to shrink more than the next layers, etc.
Additionally, the bottom layers can dry out more easily because the
glue between layers acts as a barrier to water diffusion, so possibly
the bottom layers continued to dry with each clamping, while the inner
layers could not dry as fast. Of course, my theory could be completely
bogus! The bottom line, though, was that the plank was not flat from
side to side, so I planed it almost flat. If you take a look at the
photo at left, you may notice the horizontal "lines", four of them,
that show the junction of the short layers to the layers below. You can
see that the lines are kind of U-shaped, especially the bottom most
line. That is what happens when I plane more at the center than at the
edges in an attempt to flatten the plank from side to side. I hope that
after all the clamping is complete, the plank will settle down and
partially return to a shape where the face is truly flat from side to
side. for that reason, I left a little of the bow in the plank. As a
last resort, the final layer is a 1/8" persimmon veneer, and with such
a thick veneer I can do a little touch up with a plane to make sure the
surfaces that will accept the shelves are nice and flat.
total width of the bottom approached 24 inches, and none of
the persimmon was near wide enough for that. Additionally, since the
persimmon dried so wildly, I found that I had to make the bottom from
four narrow slabs in order to maximize its thickness. The original plan
called for a bottom 1-1/2 inches thick, but I had milled all the lumber
to 5/4 (1-1/4+ inches). I realized that I would either modify the plan
for a 1 inch thick bottom, or make up a filler piece to glue to the
underside of the front of the bottom to
give the illusion of thickness. I'm still thinking about
that. In the mean time, I sawed the lumber to length using a portable
circular saw, as shown at left. This is a really messy operation, so I
do it downstairs alongside the big jointer. I have to open the garage
door and cut outside, or if it is raining turn on a fan to blow the
dust outside. But, at this point the planks are kind of wild as they
twist and bend, so I cannot saw them on a stationary machine. I have to
wait until the pieces are cut more to final size before I mill them on
the jointer and planer, or rip them on the cabinet saw. But, when they
are manageable I joint and plane them better to size. At right, I am
planing a slab of the bottom on a 20 inch planer. The bottom, although
the largest of the shelf pieces, is also the least visible, so I did
not worry about grain orientation and simply made up the wide slab from
the four most straight pieces I could. In the end, I glued up the four
pieces and the bottom is nice and true.
the top from a single plank, but here the grain mattered. I selected the
clearest piece of lumber I could, but it dried with a severe cup, and had I flattened it on the
jointer and planer I would have lost a lot of thickness. The solution
was to rip the plank lengthwise, joint and plane the two pieces flat,
then edge the pieces so they could be reglued to
make the plank look almost like it did before ripping, but
flat! The photo at right shows me ripping the plank on the bandsaw,
which made a thin kerf so little of the grain was lost. Then the pieces
were flattened on the jointer, similar to the way I used the jointer at
left. Then I planed the pieces like I did with the bottom in
the photo above, and when they were nice and flat I 
edged them on the jointer as seen at right. By running the
two pieces along the fence on their opposite faces I was able to cancel
out any error from 90 degrees in the fence angle and end up with pieces
that fit together to make a flat panel. The photo at left shows the
pieces sitting together, before gluing, and you would be hard pressed
to tell that you are actually looking at two pieces of wood. After
glueup it was near impossible to tell. It is important
to make sure that the two pieces lay along exactly the same
plane when glued, and for this insurance I decided to use small dominos
as tenons about every foot along the length. These tenons don't add any
strength to the joint, but they do an outstanding job of registering
the two pieces so the end glueup is almost flat and requires little
additional planing. The photo at right shows me using the Domino tenon
machine to cut the mortises in the edge of the top pieces. By the way,
I used this same technique to register the four pieces of the bottom
before gluing. I made the shelves without having to rip the halves
apart, except for the tops of the drawers, the pieces that form the two
"countertops", which were cupped so badly that they required this
procedure.
edges correctly. The back edges were the only ones that
were to be straight and true. On the long side I had to use the
circular saw to true the edge, and to mark the cut line I used a thin
strip of wood and clamped it to the piece to mark along with a pencil, seen
at right. Again, sawing was a dusty operation, and I was also afraid of
small, sharp pieces of wood and glue flying into my face, so I doned a
full face mask to make the cut, as seen at right. On the smaller sides
I used a cabinet saw to true up the back edge. With the back edges
true, I cut the top and bottom edges to the 
correct angles so when the
sides were set vertically the top and bottom edges would be horizontal.
To figure out the correct bevel angle for the saw (the
dusty circular saw), I propped up one end while the side was on the
bench, concave face
down. I had already used my CAD program to figure out how far to prop up the pieces, so I simply measured the
correct bevel angle for the saw while the sides were propped up.
Cutting the curved front edge required sketching the curve (again,
using cues from the CAD program to mark the surface of the wood) and
then using a saber saw to cut the curve, as seen at left. I finished up
by using a spokeshave to fair in the curve nicely, as seen at right.
the section of the tree just below the main crotch, where
the grain was a bit wild but the planks were anything but flat. In
order to do the bookmatching from the 5/4 lumber (nominally 1-1/4" thick), considering
the warpage in the planks, I knew that if I could end up with a
finished panel thickness of 3/8" I would be doing OK. The panels had
some cracks in them, so after thicknessing and trimming them 
almost exactly to size, I made up some butterflies to seam
up some of the larger cracks. These butterflies, which were hour glass
shaped pieces of walnut, would hold the cracks together, as seen in the
photo at right. Aside from the structural attributes, butterflies add
some interesting accents to an otherwise plain surface. You may have
noticed the lower part of the piece is wider than the top, so I needed
a 7th panel down there. For this I made up a single book
matched panel from some figured persimmon, seen at left
(I'm sanding it with my Festool Rotex sander, a really nice tool). With the panels roughly made, I worked
on the frame mortises that the panels would fit into. I decided to use a 5/16" end mill bit to make a
5/16" slot, 1/2" deep for each panel. Then I would shave away 1/32"
from each 
side of the panels at the last 1/2" of the ends to make 5/16" tenon ends. The
photo at right shows a frame piece on the router table and if you look
closely you can see the 5/16" end mill bit in the router (click on the
photo to enlarge it). Also, in the photo you can see some of the
mortises that will hold the frame together. Those were made with the
Domino machine. The photo at left shows the same router table being used with a
rabbeting bit to shave the 1/32" from the last 1/2" of the panel to make the tenon end. Notice the auxiliary fence I
made to keep the bit away from my fingers and provide a fence for
guiding the work. The cut could be made using the
bearing on the bit to guide the work, but the aux fence is
much
safer. With all the machining done, I needed to make
the back fit nicely into the sides, so I needed to trace the actual
curvature of the sides onto the back frame and then cut the pieces on
the bandsaw before gluing up the back. So, I dry assembled the entire
back and set the side pieces onto the back to trace the curves, as seen
at right. At this point I was really only interested in the frame
fitup, but the panels were just laid on top of the frame to check the
look of the grain. 
using a card scraper to remove some saw marks from a piece of the frame. A card
scraper is a thin card shaped piece of spring steel, typically
something like 3" x 5" or so (but it can be shaped to a curve, as well)
that is sharpened and used in a unique way. Instead of sharpening to a
chisel like edge that is pushed into the wood to slice it, the scraper
is first ground with a perfect 90 degree edge, then using a burnisher (a bar
of steel), the corner of the edge is "pushed" over to form a small
burr, or hook of steel. Then the card is held a few degrees from vertical,
flexed with the fingers and thumbs, and is pushed along the wood. The
burr cuts a minuscule shaving, which curls away from the wood. The
scraper can be used in any grain direction - it will not tear out
unruly grain, and that is the strength of the tool. Rather than chance
tearing grain with
a hand plane, the scraper makes quick work of small defects and results
in a surface that requires only very fine sanding, like P400 grit or
finer. By the way, as kind of a tool junkie I own a nice cabinet
scraper and a scraper plane, which are nice holders for scraper blades
to make surfaces really flat. But, maybe because I learned with a 
simple card scraper in the first place I get better results
with the simpler tool and find myself always reaching for it. I have
developed a good feel for when the card scraper is cutting right, and
know how to tune it up for the best cuts. The only real downfall to the
card scraper is that it gets very hot in use, and my thumbs, in close
proximity to the cutting edge are the victims of that heat. Ouch! After
planing, scraping, and sanding, I checked the surface for scratches by
applying some mineral spirits (paint thinner) to the wood, as seen
at left. Any scratches will show up while the surface is wet. In addition,
the mineral spirits will raise
the grain to the extent that any spirit based finish will
(which is not much), and since it was decided to use an
oil/varnish/beeswax finish, a quick sanding before finishing is all that will be
required. Finally,
with all the parts machined and sanded, I glued up the back. First I
glued the bottom frames, let them set up well, and then glued up the
top frame to the bottom. The photo at right shows the final glue up,
with the assembly on the shop floor and my longest bar clamps to bring the mortise and tenon
joints home.
shelf and side would be correct. The first step was to do a
very accurate layout and ensure all the angles would be correct. The
photo at right shows my setup. I took four saw horses and set two old doors on them
(my house's former
front doors),
shimming the doors until I had a planer surface. Then, I set the newly glued up back onto
the doors, and then set the sides on edge along the back in the same
configuration they would be in for the finished piece. If I had cut the
back edges of the sides true and flat, it would be a simple matter of
laying out lines where all the shelves would fit so I would have an
accurate picture of where the mortises had to be cut. Well, at this
point I was able to discover some inaccuracies in the trueness of the
back edges of the sides and I corrected them with a hand plane. Now,
with the sides vertical and the back horizontal and the orientation
correct, I marked on the back the position of each shelf and then used
a framing square to draw vertical lines on the sides showing the line
of the mortises. Finally, I used a long straight edge and connected the
tops of the lines I had just drawn so I could draw lines across the
edges to the outside of the sides, these lines being coaxial with the
mortises. I compared these lines to those calculated with my CAD
program to make sure they were close to what I had predicted. Once I
was satisfied that the layout for the mortises was correct, I
contemplated how to cut the mortises, themselves. Now, there are lots
of ways to cut mortises, but in all the 'how to' books the mortises are
being cut in nice square and relatively small stock, not in immense and
curved laminations. Was I crazy to design a piece like this, or what? I
had hoped that I would have a revelation before this point on how to
cut the mortises, but now I was against the wall. I could not simply
use a chisel and guide block to chop the mortises, because what is
square, here? None of my conventional methods would work, either. Well,
I won't bore you with all my thinking on this, but if you watched Dr.
Shepherd on Grey's Anatomy spend 26 hours in the operating room staring
at a tumor trying to figure out how to operate on it, you get the
picture of what my workshop looked like when I was figuring out how to
cut these mortises. Fortunately for my client, I stopped the billing
clock during these long deliberations - it would be unfair to charge
them for my insanity in designing such a complex piece of furniture!
What I finally came up with would be accurate and doable. I would use a
drill press with Forstner bits to drill out holes, the diameters of
which would be the same as the dimensions of the mortises. Forstner
bits are special drill bits that cut a hole straight and true, even if
the surface the bit is entering is not square to the direction of the
bit travel. Then, with a corner chisel (a chisel cut with a 90
degree angle between faces, like a corner) I would "simply"
square up the holes I had cut. Easy! (not!). To get the angle of entry of
the bit correct, I would have to shim the work until the lines I had
drawn were coaxial with the travel of the bit in the drill press. Here,
I 
could compare the amount of shimming necessary against the
CAD calculation to double check the setup. In the photo at right, I am
drilling the mortise holes in the shortest side piece, and you can see
the shimming
arrangement. The far end of the work is sitting on a block
of wood, raising it until the bit enters at the right angle. The photo
at left shows the same thing with the longest side, which due to it's
length had to be shimmed on a table made from my saw offcut roller
stands with planks stacked across them. Amazingly, this all worked! Now
all I had to do was get out the corner chisel and go to work. So, after
marking exactly where the chisel cuts would go (it was critical to have
the mortises square so the tenons would fit exactly), I began chopping
mortises as seen at right. As I said, for most of the mortises I used a
corner chisel to make sure the corners were square. Then I would come
back with a bench chisel to clean up the mortises. It was critical that
the holes be square, especially on the outside face where they would be
visible. If I did not keep the holes nice and even through the wood,
that was not a big problem, but that outside face was a bigee.
mortises cut, the tenons were marked to fit exactly into
the mortises. I used a marking knife to make exact lines where the
tenons would be so that they would fit right. Especially critical
were the "open" tenons where the short lower side intersected the short
right side shelf that formed the top of the upper right drawer pocket.
If this mortise and tenon joint was sloppy, it would look it in the
finished piece. So, I carefully cut the joint and dry fit the pieces,
with a considerable amount of tweeking with chisels and files to get
the fit exact. At right you can see the dry fit joint. After gluing the
joint, I will trim off the excess tenon material and make the faces
flush. This particular joint did not get wedges, as all the through
tenons received, so there was no room for error. However, with the
through tenons
I would be installing two vertical wedges per tenon, so if
the tenon was a hair wide or narrow, the wedging action would correct
that. At left I am using a Bosch saber saw to cut the tenons. I had
marked the tenons with a knife, and I am 'splitting' the knife line
with the edge of the saw blade. That is actually easier than it sounds
and is critical to a well fitting joint. Instead of trying to cut a
pencil line, which is a thick line, I get one side of the saw kerf to
split the knife line. In this case I am using a Lenox C450S blade, 10
teeth per inch in my Bosch saw. This blade has extremely little set in
the teeth, if any, which means that the teeth are not alternately bent
to one side or the other as in most saws. This is great if a very
smooth and exact cut is desired, although it does not cut as fast as a
blade with less teeth per inch and a lot of set. I was quite surprised
how exact I could make the cuts with this setup. I did have to check
that the blade was exactly perpendicular to the shoe of the saber saw,
though. For a smaller job I would use a hand saw, a Japanese pull saw
to cut the tenons, or use a setup on the table saw, but these shelves
were immense so I had to bring the saw to the work and use power! As
well as cutting the tenons, I also cut two kerfs in each tenon 1/4"
from each of the sides for the wedges.
pieces fit together seamlessly. First, though, I had to cut
a small rabbet on the top and bottom of each shelf tenon so that the
mortises would not be at all visible on the inside of the cabinet. The
photo at right shows me using a chisel to pare away material on the top
of a tenon. I did the same on the bottom of the tenon so that the tenon
height was a little less than the shelf thickness. I had to do this for
all the tenons, and there were lots of them. For a smaller
cabinet this operation could easily be done on the table saw, but for a
six foot long shelf, holding it vertically while guiding it on edge
through the saw, or using a dado blade to do the work horizontally
seemed just a
little dangerous. The price of big work! After paring the tenons to
size I dry fit each joint and made little cuts here and there to get
the fit smooth and as tight as I could. Then I assembled the whole
carcase dry, that is, without glue, to check the fit and make marks for
the shelf supports. I know this sounds easy here, but this was days of
work to get done correctly. The photo at left shows the dry fitup with
the shelves propped into the correct relative positions. Note that the
shelves
were all cantilevered, that is, each was only mortised into
one side. The other side, at the center of the carcase hung from the
top shelf via a
support structure. This support structure was a series of short
vertical boards, each cut to fit between adjacent shelves. One
end of the support would be screwed into a shelf and the other end
doweled. That way, I could install the shelves starting at
the uppermost shelf and work down. The photo at right shows the first
shelf installed with the support for the second shelf clamped in place
after gluing the dowels. By sequentially installing shelf, then
support, I build the shelf structure from the top down. Of course, all
the tenons had to be wedged as I installed each shelf. In the
photo at left I am installing a wedge (walnut looked best)
into one of the kerfs I had cut in a tenon. After the glue on the wedges dried, I used a
flush cut saw to trim the mortise flush with the sides, as seen at
right. A flush cut saw has it's teeth set on only one side, the side
away from the surface, so the surface is protected from becoming
scratched. One of the difficult parts (OK, it was all difficult) was
drawing the sides to the bottom two shelfs tightly. These shelfs were
full width, six feet long, and with curved sides there was no way to
clamp the opposite side together to make the joints tight. My solution
was to drill holes for screws to pull the sides onto the shelves, then
plug the screw holes with plugs cut from scraps of persimmon. This
method worked very well, and I left the screws in place to give even
more strength to the joints (as if they needed to
be stronger). I could hardly see the plugs after trimming
them and sanding. Speaking of sanding, I guess I should mention that I
sanded each part
before assembly and I spent what seemed like days with my Festool Rotex
sander bringing the surfaces to P320 smoothness. The Rotex is a random
orbital sander with an additional setting for action more like a
conventional disc sander. With the shelves all glued in place, I
screwed the back on and plugged the screw holes. The screws went in
through the top shelf and up through the bottom shelf, with a pair of
screws directly through the back into the upper drawer opening tops.
The photo at right shows the final glued up carcase. Next, the drawers.
with formula e to set the compound miter saw to the correct
angles. Later, I will post my calculations with a link from here
(mainly so I can remember how I did all this next time!). The photo at
right shows me finally making a cut with the compound miter saw, which
cuts a bevel and a miter at the same time. It was very easy to get
confused, so I checked and rechecked before each cut. Once the cutting
was done, I laid out the cut lines for the dovetails, which were hand
cut. I know of no jig that can cut these dovetails with a router, since
they are all cut on beveled pieces, each with a different 
bevel angle. Hand cutting was not only necessary, but faster
than trying to make some kind of adjustable jig. The
photo at left shows me making marks for the dovetails,
using a marking knife. I also used my 1:8 dovetail marking guide to
make the angles of
each tail. I marked and cut the tails first, then marked them onto the
pins. There is always lively debate whether to cut tails or pins first,
but in this case the tails seemed like the easier to cut first and use
to mark the pins. At right I am sawing tails with
a Japanese pull saw. This saw cuts a very clean and
accurate line and is the only saw I use for dovetails. At left you can see
me chopping away the waste between tails (in the side pieces) using a
chisel and a guide block. I cleaned up the sides of the tails with a
chisel to make sure they were nice and flat, then laid the tail boards
(sides of the boards) along the ends of the pin pieces (front and back
of the drawer) and used my marking knife to mark the cut lines for the
pins. Again, I used my pull saw to cut the pins, then used a router to
hog away most of the waste between pins, as seen at right. I used a
chisel to cut the base of the pin piece to the correct angle and to
smooth the sides and fit the dovetails for a smooth fit.
drawers would smell sweet, and I left the cedar unfinished.
For the same reason, I pre-finished the insides of the drawer sides and
fronts (after disassembling them) with shellac polish. Shellac polish
smells nice, and complements the
smell of the cedar. After the finishing, I glued up each of
the drawers. I don't have a photo of the glue-up, but with all the
angles involved
there was quite simply a mass of clamps required! With the drawers
glued, I spent time fitting them to the unusually shaped drawer
pockets. First, I trimmed each drawer to height, allowing a small
amount of space for growth when the weather becomes more humid. At left
I am making height cuts on the two square faces of the drawer. After
that I used a hand plane to match the height on the two angled sides.
Then I used a sander with P36 grit (really rough) sandpaper to rough
shape the angled sides to the curve they would match, as seen at right.
Following the rough shaping I used progressively finer sandpapers until
the sides were smooth. It took a long time to fit all the drawers, but
once done I finished the sides with shellac (for the same reasons as
the insides of the drawers) and finished the fronts with the
oil/varnish/beeswax mixes.
to attach. The pulls are attached to the fronts with
3/8" dowels. Then to the pulls themselves. I have the routine for
making handles down pretty well, so it took little time to cut the
blanks to size with all the correct angles and drill holes for the
dowels. I then used the bandsaw to cut away the major waste, as seen at
right. The handles, starting to take shape, were then shaped using the
oscillating spindle sander, as seen at left. Way back I used a
whittling knife to shape handles, but after a lot of cut thumbs I
realized that the oscillating spindle sander could do the job more
safely, less painfully, and a lot faster. I
bet if the old timers had oscillating spindle sanders, they
would have used them, too! The sander enabled me to make the six
pulls nearly identical, but I could only get them moderately smooth this
way. I had to do a lot of hand sanding to bring the
surfaces to the smoothness I require, so I sat down in the Adirondack
chair on the deck and went to sanding, as seen at right. I like to sand
this way because the light is so bright I can see the defects and sand
them out. I sanded to P600 grit before finishing with shellac polish.
The photo at left shows a finished pull attached to a drawer front. All
that was left to do was apply some wax to the drawer sides and slide
them into the pockets. I installed a strategically placed stop (a
neoprene faucet washer screwed in place) to register the drawer faces
to the front of the carcase and prevent them pulling out all the way,
as well. The stop is screwed into the underside of the top of the
drawer pocket after the drawer is inserted half way. I used this method
because the drawers are relatively short as drawers go and they are
heavy. We don't want them falling out and causing sore toes (or worse)!
Below is a photo of the finished piece. Click on this link
to see more photos (details and the back of the piece).