Prairie Sofa and Love Seat, Part 2

It's been a couple of months since Part 1, and there's been a lot of progress. So much, that as of this writing, I'm almost done, but I'll save that for Part 3. I left off in Part 1 with readying the design and acquiring the rough lumber. But since then, both of those things changed.

Frame and panel style sides

The design is now different. In the plans, and my original intent, I was going with the frame and panel style sides. That design depended on getting one important type of material for the panels: 1/2" quarter sawn white oak plywood. I needed 1/2" plywood, not 1/4" (which was readily available) or 3/4" (also readily available), because I wanted to have a 1/4" reveal on the outside, and flush panels on the inside. With the stile and rail stock being 3/4" thick, the panel material had to be 1/2". Sure, I could have gone with solid wood, but after going through that dance with the recent wardrobe build, I wasn't in the mood to do that again so soon. The advantages of going with frame and panel, with plywood for the panels, were many...

  • I could build and assemble everything, then apply finish (i.e., no pre-finishing needed).
  • I could glue the plywood into the frames, adding strength to the piece.
  • it would be faster to build and finish.

So, I looked around for 1/2" quarter sawn white oak plywood, G2S (good on 2 sides), but simply couldn't find any in the Chicagoland area. My local hardwood dealer gave me a quote for making a sheet up from two sheets to 1/4", with some MDF in the middle to pad it out (two 1/4" piece squeeze down to 3/8" when pressed to gather in their plywood machine), but that would cost $250/sheet. And I needed two sheets. That was pushing the pricey end of the range for project, considering a sheet of 3/4" QSWO ply G2S costs $134. I found some other sources in adjacent states, but with shipping the cost came to $250 per sheet, so that was a no go.

I decided that instead of using solid wood panels, I would try something different. There is a variation on this style of sofa that uses square spindles to fill the area between the upper and lower rails. In these pieces, I like the spindle design and the frame and panel design equally. Going this route meant a few things...

  • I would need to do pre-finishing of the spindles
  • I could use material I had on had, no plywood needed
  • It was going to take longer to assemble
  • I could keep the rail material thicker than 3/4",  (which was desirable, as I wanted to add to the overall strength, to compensate for losing the strength of the glued-in plywood panels)

"Spindle" style sides

Test build of the "spindles" sides, with a decorative cutout slat in the middle, which I later ditched.

To see if the spindle design was feasible, I did a test build. The results were good, so I decided to go for it. The disadvantages of all the additional work and having to pre-finish were more than offset by not having to buy plywood. The design had begun to grow on me. I now had a good idea of how much extra work the spindles were (not as much as I'd feared), and I was able to settle on a good mortise and tenon size. In the test build, I toyed with adding a decorative center slat with a pierced design in it, but ended up ditching that idea, as I felt it was piling on too much decoration, and would take away from the other details of the pieces.

The next change involved the wood itself. As I discussed in Part 1, I was very pleased to be working with the urban forested lumber sourced from the University of Chicago. But much to my dismay, as I started milling up the lumber, it was evident that many pieces were lacking in the classic quarter sawn white oak figure. Some of them were fine, but too many were angled just far enough away from being true quarter sawn, veering into rift sawn territory, that they weren't showing the medullary rays, a.k.a. "ray flecks" I wanted to see. That meant I had to buy some more deeply figured material from the hardwood dealer. While far from a total loss, it was disappointing I wasn't able to get everything from this one source. But, that's just the nature of working with wood: sometimes it's hard to know just what you're working with until you start working with it. 

Almost quarter sawn white oak, with a "rift" appearance - not enough figure!

Quarter sawn white oak showing great figure with ray flecks, that's what I want!

With a new batch of lumber, I was ready to proceed with the build. The first steps were to laminate the legs. One hallmark of this sort of piece is that legs are usually put together in such a way that the quarter sawn face appears on all four sides of the leg. There are several ways to do it, one being four mitered sides with a square core, and another being a simple rectangular glue up with with 1/8" veneers on each side. I opted for the latter, as it is much easier to do. In addition, it allowed me to use some of the less visually optimal lumber for the center pieces.

Four mitered pieces attached around a core, left, and pieces with two 1/8" veneers, right

The glue-ups went well, and I was able to get some great looking faces for the legs. For each set of four legs, I was able to the glue-ups all together in one big glue-up sandwich, which was a ice time saver, and made efficient use of my clamps.

Below, showing some of the "ugly duckling" wood used for the centers of the legs, and veneered faces, and the four-at-once glue-up...

top view of a rear leg, showing the two tenons meeting on the inside.

With the leg blanks glued together, I next laid out the joinery for the mortises in each leg. In the original "panel style" version of these pieces, the side stiles were simply glued into dados in the side of each leg. With approximately 22" linear of long grain to long grain glue surface, that's a very strong joint. But by moving to the spindle design, that glue surface isn't present, as there are no side stiles. Instead there are just two rails going into the legs. That meant I had to lose the stub tenon/dado for the joint, and instead use full depth (or as deep as I could make them) mortise and tenon joints. In those parts of the legs where tenons were coming in from two directions, I had to work with the tenons running into each other. But luckily, there was enough depth to where the joints were going to be very strong with a lot of "meat" in them. With that problem solved, I proceeded to mill all the rails, cut the mortises and tenons, and fit the basic structure together. I took care at this stage to make sure all the joints were seated nicely, with no gaps at the shoulders, and that the fit was snug but not too tight. I didn't want to fuss with the fit later while seating the spindles. For the legs, I had picked the best faces for the showy parts, and the not-as-good faces for the areas that would be against the wall, and marked them with a triangle, also marking an (S) for sofa, and an (L) for love seat. It seems obvious, but clearly marking out parts as you go, especially when building two almost identical pieces at the same time, is a lifesaver. I ended up referring to these marks constantly throughout the whole build. I also made sure to mark each of the rails as they were fit, making sure there wouldn't be any mix ups later.

When the joinery was complete, I did a dry fit of each piece. This is always an exciting moment, when all the flat two dimensional pieces finally come together and form these three-dimentional sofa and love seat looking things! At this stage the legs still needed shallow 1/2" x 1/2" dados cut on the sides to accept the corbels, and this was the perfect time to mark the legs as there was no chance of confusion while they were standing in the dry fit. I also marked the rails for the 1/4" x 1/2" dados that will accept the spindle mortise strips. Again, doing it now while the pieces were in place greatly reduced the chance of marking them incorrectly while they were disassembled.

The next stage involved setting up for the spindles. From my test build at the beginning, I'd decided to go with 5/8" square spindles, set 1" apart on center, with 1/4" square x 1/4" deep mortises. Of the several combination's I'd tried, this seemed to look the best. At first, I thought I was going to need much larger tenons, and my first test had them at about 1/2" square. But that meant the dado to house the crenellated strip that would make the tenons would cause the remaining sides, or walls, of the rails to be less than 1/4" each, and that seemed too flimsy for my taste. 

The 1/4" square tenons were surprisingly strong, and more so when all lined up together in each side or back assembly. I had no worries this was going to be the right approach.

Now it was time to make all these things. Between the sofa and love seat, there were 206  12 1/2" long spindles, needing 412 mortise and tenon joints, with 824 faces. Each part needed to be jointed, planed, cut to length, have tenons cut, and each face hand planed. Later, they were sanded, had the grain raised, sanded again, had the dust removed, dyed, then shellacked. I spent a LOT of time processing those spindles.

I made the crenellated strips, cut the dados in the rails, and glued that together, taking care not to fill up the little mortises with glue. I only had to dig glue out of a few of the 206 mortises. I did a full dry fit of each of the two sides and the back, for each piece. This process made sure I discovered any malformed mortises or tenons, or any part that wasn't fitting, well before I would be doing final assembly. For the final assembly, the tenons would be going into the mortises dry, with no glue. Gluing them would have been tedious, and would have caused lots of squeeze out that would have required meticulous clean-up. The glue wasn't going to be adding anything in the way of strength on such tiny glue surfaces. The dry fit showed the spindles to be nicely snug in their mortises, and my fear that they would end up loose was assuaged.

This was as far as I could go without doing some pre-finishing. I made test boards to get the right balance of the dye and shellac, trying out a few different dye concentrations and types of shellac. I was going for a similar color/appearance to the mission pedestal table I did a couple of years ago. The final formula was full strength Transtint dark mission brown water based dye and several coats of dewaxed garnet shellac.

I started the pre-finishing process, working the two sides of each piece together, then later doing the backs. That split the work into four roughly equal sized labors (love seat sides, love seat back, sofa sides, sofa back). Managing the spindles individually was troublesome, so I constructed a couple of racks that would support each spindle by its tenons, letting the faces be free to air dry during the various finished steps. This also made moving them around the shop easier.

Spindle racks in action

At first, I used little dot stickers to make sure I didn't get any shellac into the tiny mortises, but after testing it out, I realized this step wasn't needed, and stopped doing it for the rest of the process. 

That's it for Part 2. Stay tuned for Part 3 coming soon.