The rockers are not cut from a solid piece but for strength
are made from thin strips that are laminated together against a form. Each rocker has 9 layers of cherry and one
layer of katalox.
The rockers represent the last major pieces that have to be
built. So to start I pulled out the last
piece of cherry reserved for the rockers and checked it for flat and square
which, good news, it only had a minor bit of cup. A few passes through the thickness sander
took care of that.
The katalox however was in no way straight or flat. It had a nice combination of cup, a gentle
bend along the long edge with a dog-leg near the end along with a few waves
thrown in plus just a tad amount of twist.
In that one board there was just about every kind of contortion you could
imagine. Fortunately I only need 2
strips 1¾” wide, just over an eighth of an inch thick and about 54” long plus a
couple more 8” long. To clean up this
mess I started with a long straight edge to give me my rough long dimension of
the 54”. With a straight pencil line to
start I used the bandsaw to make my first cut.
Next, I used the power hand plane clean up the bandsawn edge to give me
a true straight edge. That left me with
some cup, the twist and waves. The
thickness sander took care of the cup and waves. The twist is not great and when I laminate it
up because the finished piece is thin and flexible it will get flattened. However, because of the twist I had to cut
the strips with the bandsaw rather than the table saw. To give me a little buffer I cut them a
little thick and will finish up with thickness sander.
There are a couple of ways to cut the multiple strips of
cherry needed. One is to set the rip
fence at the desired width and cut away.
This works good most of the time but as the pieces needed get thinner
your fingers get closer and closer to the blade. I have a push block that will allow me to cut
pieces as thin as ¼” with no problem.
However, these pieces are just .17” wide which rules out using the push
block I have. I could make a special
push block but am a little leery of having a piece this thin between the blade
and the rip fence with such a big fairly heavy piece on what is normally the
waste side. The cherry blank is 9” wide,
almost 2” thick, 55” long and weighs about 25 pounds.
The other way takes a little more time but works well. Its advantage is that the thin layer is on
the outside or what would normally be the waste side of the blade. This means that it cannot get trapped between
the blade and the fence. The
disadvantage is the after every cut you have to reset the fence and it can be
difficult to get pieces of a consistent thickness. The solution to getting a consistent thickness is a simple
jig that sets in the miter slot. A screw
on the end is run in or out to set the piece thickness.
To use, the rip fence is adjusted so that the
main board just touches the screw head then the jig is removed and the cut is
made. The process is repeated over and over until you get the
needed number of strips. In my case that
is 18 plus a couple of extras or 20.
The pieces are rough cut at .17” but the final dimension is
just over 1/8”. The extra is 45 thousandths of an inch is so I can
run the cut pieces through the thickness sander to remove the inevitable saw marks and
make sure the surfaces are parallel.
I keep the layers in the same order as they are cut. This assures that when the strips are
laminated back together in the rocker form the grain will be consistent across
the finished piece. In order to be able
to put them back in their original order I number and mark the end of each
board and group in sets. Red for one and
green for the other. These are the
stacks with the katalox inserted where it goes.
When I glue up the 10 pieces in the form it is controlled
chaos. Because of the limited working
time I have with the glue. I only have
about 10 minutes from the time I start spreading the glue to when I need to
have the 21 clamps tightened down. This
means everything needs to be laid out and ready to go before I start applying
the glue.
The glue goes on first, I have
about 6 square feet to cover and it needs to be applied with no missing or thin
spots. Next is to stack the layers up,
quickly wipe excess glue off the joints and set the stack in the form. Starting in the middle I work toward one end
keeping the layers flush and aligned then do the other half. Once everything is tightened down I wipe the
glue squeeze-out off and let it sit overnight.
Then next morning I will pull off all the clamps, inspect the form for
damage and go through the process for the other rocker. Using one form for both rockers assures that
they will be identical.
Once out of the form I mostly flatten one face on the rocker
with the power hand plane checking with a square to make sure the adjacent
faces are square with each other. I will
use this face as a reference for adding a secondary lamination.
Under each of the four chair legs where they are attached to
the rocker is a secondary stack of 7 laminations made up of 3 cherry layers, 1
katalox and then 3 more cherry. This will provide the material for the transition from the
leg to the rocker. To locate I set the
chair on the rocker roughly where it goes and mark a line about three inches
either side of the leg.
The laminations are then cut, glued in place on each of the
rockers and allowed to cure overnight. I
probably could machine them after about four hours in the clamps but since they
are bent and under some stress I wanted to give them overnight to cure.
Once this secondary set of laminations has cured I have two
choices in milling them to the correct width, either the planer or the
thickness sander. I chose the thickness
sander because there are lots of grain directions and curves going on here and
I am just a little paranoid about the planer ripping a big chip out of the
rocker or leaving me a big snipe on the end.
Neither is a concern with the thickness sander – it just takes little
longer to get there.
After getting the rockers to the right width I set the chair
on them and checked to see if the legs sat tight on the rockers. No real surprise here, they didn’t fit
tight. To get a tight joint I took a
compass set just slightly wider than the gap and used that to scribe a line
around the leg.
Grinding to that line with the disk sander will not provide
a final tight fit but gives me a working setup for continuing. The joint will need to be fine-tuned once the
dowels are fitted in place that tie the chair to the rockers.
I mentioned in the last post that I would be doing the 99%
final sanding and I have been working on that, at least from the seat up. It was not quite as bad as I thought it was
going to be. Although I swear I have
completely sanded away any and all of my fingerprints. I had taken everything to 80 grit so the next
step was 120. This took me about 3 hours
to go over all the chair’s surfaces. The
next step is 150 followed by 220 where I quit until all the woodworking is
done. Just before finishing I will hand
sand with 320.
One problem that showed up early on when I cut the seat to
shape was a knot on the seat’s left side.
As the project progressed final shaping did not remove the defect
although on closer inspection it is not a knot but more like a bark inclusion.
As it will still be there when I am done I
have to do something with it. Right now
I have cleaned out the loose material and have come up with three options. First, completely remove the defect and fit
in a cherry patch. Second, fill the hole
with black epoxy and third fill the hole with a mix of turquoise granules set
in a clear epoxy.
Next up – Installing the Rockers
Getting the thin laminate strips to exact and consistent size was a challenge for me, Dave. Your method of cutting the thin strips on the OUTSIDE of the tablesaw blade was very helpful and a whole lot safer than trying to cut the strips between blade and fence. Not having a drum sander, it was still a challenge to get the strips glue ready. I resorted to a simple DIY planer sled designed for thin stock.
ReplyDeleteI built my chair out of cherry, and in the end, I wished I had paid more attention to color and grain matching of the rocker laminates to better approximate a solid rocker encasing the maple accent strips.