One of my neighbors had a bad experience with a well-meaning contractor recently while attempting to get a bench built into a corner below an open staircase. Result was: no bench. I am not very good as a finish carpenter but I offered to do my best in exchange for some old redwood beams. Anyway, here is how it went:
The design mandate was to match the staircase in look and feel and make the overall dimensions of the "L" shaped bench legs 60" long by 36" wide, with a seat depth of 14 inches. 8/4 red oak, stained with Minwax English Chestnut and sealed with satin polyurethane was agreed to. The bench was to rest on 4 or 5 "T" shaped braces protruding from the wall.
 |
| finished bench top with square |
The brace idea was to build up a stack of 3 layers of 1x4 red oak and sandwich in a flared tenon cut out of the heavier 8/4 arm of the brace. This was to be glued together, forming the T shape. The "arm" would stick out of the wall horizontally and be screwed to the bench top. The vertical part of the brace would be set inside the wall and screwed to a stud with 3/8" lag screws.
 |
| sandwich/stack |
 |
| test assembly |
 |
| 5 arms for the braces. 8/4 red oak. |
The arms of the braces all needed to be drilled and counter-bored for the 5/16" lag screws with washers that would fix the arms to the top above. These lags were 2.5, 3.5, and 4.5 inches long due to the taper. Great care needs to be taken not to mix up the screws or they could go through the bench top upon final assembly. But that never happens.
 |
| ready for installation |
The top required me to glue up two pieces of stock to get the 14" width required. Take time laying up the pieces so grain matches tolerably. The top is joined with a splined miter of sorts. I resorted to using a router with a fence to hack out the mortises. Using a 1/2 x 2" long straight bit, and referencing from the top face on both pieces, I nibbled away to a depth of 2 inches and then squared up the corners with a chisel. The tenon/spline is more red oak planed to fit the mortise snugly, but not not so tight as to put distorting pressure on the sides of the mortises. Not a good location for a compression fit. Glue will do the holding work. Glue one side only, making sure there is no glue on the exposed tenon or miter face, then set it aside.
Next came the finish planing, fitting, and chamfering of faces and corners. Using a Japanese chamfer plane (kaku-men-ganna) makes precise chamfers simple. Once the top dry-assembled with a good fit on the miter and the top surfaces were co-planar, I did a dry fit and clamping to verify the fit and work out issues with clamps, blocks, movement, etc. With that done I did the final gluing, taking care to uniformly spread a thin layer of glue on all walls of the mortise. Gorilla glue was used in this case, but use your favorite strong glue (there is a lot of leverage on this joint so next time I would use a mitered half lap).
 |
| 2 pipe clamps hold the miter together during gluing. |
With the glue set, I needed to
transfer the actual wall contour at final bench height onto the back
edge of the bench so it would fit closely to the wall. The wall was far
from flat and the corner far from square. In the top picture you will
see a try square of sorts. This I positioned into the corner and then
traced the wall contour onto the square with a pen. Then I trimmed the
square to the traced line, and lastly I transferred the contour onto the
bench and then trimmed the bench to the contour. Painful but necessary
for a decent fit.
With all wood cutting complete I began the finishing process of stain followed by 4 coats of satin polyurethane with light sanding with 600 grit in between the last 2 coats. Sanding after the first coat could easily damage the stain and there is little point in doing it. There is a bit of filling that needs to occur with the porous oak, so plan on at least 3 coats. Following the last coat I rubbed down the surface with synthetic steel wool (0000) to make the sheen uniform and eliminate any small bubble craters, etc.
Installation:
The demolition was really just a surgical removal of as little sheetrock as possible to expose the studs we were to fasten the braces to. First we located sheetrock screws in the area affected with a small magnet and backed them out. We marked out the cuts with tape and used a new utility knife blade to make cuts.
Since I had transferred the wall contour onto the bench, I could not alter the contour of the sheetrock without introducing gaps galore. That sheetrock which would be covered by the back edge of the bench top had to remain intact. To accomplish this, we removed the sheetrock pieces and cut holes through them where the arms were and then put the sheet rock back in place carefully before installing the bench top.
Since all arms had to be coplanar and level before marking and screwing down the bench top, use a level after securing the first arm and reference off it. While the miter joint is strong, clamping down the top with lag screws on an uneven set of rigid arms could stress the miter and open it up. Also, sinking lags into the joint area itself could cause problems, so avoid it.
Stud placement was not ideal so in two cases we sandwiched in a short section of 2x4 to move the brace over to a better location. Lag screws, bolts, and glue hold the bench braces to the studs so movement is minimal. Things get tricky in the corner with brace overlap, so watch it.
I will post a final picture once the sheetrock cleanup and paint work is completed (by a professional).
