Bed progress

Legs joined to panel, sanded to 150, 2 coats black ink, second coat of Waterlox Original Sealer/Finish wet. The recessed panels are a different color because they had no finish applied when the pic was taken. The gloss is largely the finish being wet.

...an idea of color. Consider: this is with poor lighting, poor camera, and poor photographer. Add to that the color differences imparted by your monitor. It really looks black, grain is more obscured than I thought it would be.

End-grain cutting board

Walnut, cherry, and maple end-grain cutting board. Mineral oil and wax finish. This picture is after about a year of use.

Post tops

Left: the "finial version". Right: the "chamfered post" version.

Click image for larger picture.

Torsion box

Core:
I = 1/12 x (W x T^3); W = rib width, T = rib height


Skin:
I = A x D^2; A = breadth x thickness of skin, D = centerline of skin to neutral axis of box.
Itotal = I top skin + I bottom skin + I each core element



Deflection:
Y = (.021 x W x L^3) / (MOE x I), where;
Y = deflection (inches)
W = Weight/ load (pounds)
L = span
MOE = modulus of elasticity (1.5 E6 is avg for wood)
I = Itotal from above.


Torsion box pictured: 61" span, 3" x 3/8" web, 3/8" skins. Deflection: .018" (~1/64") @ 1000lb.

Glue lam, forms, biscuits and top cap

Here's one set of forms used to make the "glue lam" panel top caps. A stack of thin strips is glued into a sandwich and clamped between these forms so they'll dry in the desired shape. In the upper left of the pic you can see the resulting glue-lam top cap laying on top of the panel.
These same forms have already been used to pattern-route the curved rails, and they'll be used again as a caul to attach the top caps to the panel frames. There's a separate set for the headboard because the radius of it's curvature is different than that of the footboard.

I just bought a biscuit jointer. Biscuits don't add strength to a long-grain glue-up like this, but it keeps things aligned while gluing and clamping. The only effort here is to get the slots centered in each of the pieces, since they're different widths.
The trial fit.




The glue-up. The biscuits automatically align the top cap, which makes the glue-up as stress-free as a glue-up can be. The form is used as a caul to distribute the clamping pressure evenly across the curved surface. Of little note: since the curves are true arcs (instead of parabolic arcs formed by bending something and tracing it), the position of the caul left-to-right doesn't matter much. There's a bit more hanging off the right side than the left, but I didn't even notice until uploading this pic. If this were a parabolic arc, that would mean that parts of the caul would be exerting no pressure.

The top cap for the headboard isn't a very good glue-up, but I'm going to try to salvage it with epoxy or cyanoacrylate. I don't know if I have enough stock left to make another, and it's not a fun process.

Side rails

A router template is made to mortise the end-grain of the rails to accept the hardware. The mortise is not centered on the height of the rail, so it has to go on the "right way" - it's easy to screw this up and wind up with mortises at different locations on each end of the rail.
Since I didn't have a top-bearing pattern bit at the time, I made it to work with a bushing and a 1/2" bit: the window will guide the bushing, and has to be oversized the proper amount to end up with the right sized mortise (5" x 5/8" x 1/8"deep).



A brass bushing guide is installed in the router sub-base. The bushing will follow the template "window", which will limit it's travel. The bit depth is set, and ...
...the mortise is routed, then this bogus picture is posed hours later...


After squaring off the mortise corners by hand, the hardware is fit into the mortise.



This hardware has two little pieces of meat on the backside where the hooks are crimped. It won't sit flat in the mortise, so some shallow reliefs are drilled with a Forstner bit.



The screws which hold the hardware to the end of the rail are going into end grain, which is not very good at holding onto the threads. Some 3/4"d flat-bottomed holes are made with a Forstner bit behind the screw holes. Short pieces of dowel will be glued into these holes - when the screws are driven in from the end, they'll get a good "bite" into the side grain of the dowels. This is supposed to be a much stronger attachment, but I can't see how the rail hardware screws could back out once the whole thing's assembled.
Another easy thing to screw up: you don't want these holes to go all the way thru, and you want them all on the back side of the rail, so they face the inside of the bed frame.



Having no 3/4" dowels, I turned from some scrap, but it's so dark in the lathe corner I couldn't get a good pic of that.

Also not pictured: the rails had to be cut to length with a circular saw and jig, because they're too long to cut on my table saw. It was a simple jig to make, but yet another step.
One of the rails had an end check that wasn't entirely cut away when they were cut to length, so I glued it with poly glue (gap-filling).
I saved myself a bunch of sanding by planing all the edges of the rails, which led to a sharpening diversion.
I started trying out some finishing schedules on scrap. Right now undiluted black ink soaked into the raw wood seems to be the best coverage, but the top coat (shellac is the plan) will HAVE to be sprayed, since the ink is apparently soluble in alcohol (which is also the solvent for shellac). Maybe a water-based lacquer is in order.

Panels

Yesterday the dados were cut in the frame rails and stiles. Since there's a center stile (which has already been doweled to the top and bottom rails), the dados needed to be "stopped". The round slot cutter leaves an arching cut at the end of it's travel, so the stopped dadoes need to be squared up by hand. The panel will conceal them, so appearance isn't important.



The panels are more than 24" wide, and all 4 together consume almost 8' of length. This means there's no room for "spares" in one ($95!!) sheet of plywood. Hardboard is cheap, so templates were made to tweak the fit - a good thing, too, since some planing and fitting was required.
Not having enough hardboard of sufficient width to make the templates, the table saw cover was sacrificed. Obviously it was well-used and ready to be replaced, so it was cut up to make the template for the large panel. Since both assemblies are symmetrical, only one of each size is needed - but fit was verified in both sides just to make sure before cutting up the spendy stuff.



The frames are finally filled with panels.
The top rails capping each assembly are glued laminations - some stock is re-sawed to produce several thin strips, which are then glued up in to a sandwhich, and clamped in a form to dry in the proper shape. I'll have to get some pics of them in the forms.

Frame dadoes

3-wing slot-cutter. The bearing makes the resulting cut 9/16" deep, 1/4" is needed. The slot-cutter is bigger than the hole in the table so it's inserted from above, then the collet is tightened under the table - a royal pain.

Without a properly-sized bearing, another crazy jig is required to limit the stock getting any closer to the cutter than 1/4". Since curves will be cut, a radius needs sit over the cutter, and clearance allowed so the stock can be manipulated without wedging against the fence.


The bit shank length won't reach the whole cut (indexing from one side), so it's set up to take the required 3/8" of width in two passes, one on each side. This automatically centers the dado, but doesn't forgive ANY variation in stock width or feed pressure. The stock has already been flipped, the cutter is finishing the second pass, completing the dado in one of the center stiles.

In mid-cut on the curved top rail, it's obvious that burying the slot cutter in the fence wouldn't work.

Headboard and footboard frames

Footboard frame (slightly exploded)

convex and concave curves together


Headboard frame

Final version

I found dimensions on Thomasville's website and tweaked this a little more. The footboard is now a little larger - it reaches closer to the floor. The bottom edge of the rails is closer to the bottom edge of the headboard. The bottom of the mattress sits 1" below the top edge of the rails. The headboard and footboard heights match the T-ville version. The post caps still aren't pictured, but th ey will add a little less than 3"to the height of the posts.

With all these changes, the top of the mattress is now about 26".

T-ville's version has the slats 10" off the floor, so a boxspring would be required. This version has the bottom edge of the rail 10" off the floor, and the platform 7 1/2" above that (the rails are 8 1/2" wide).

Note that the footboard intercepts the mattress a bit.

Routing a true radius

Do some math, resulting in locating a pin on each end of the template stock.

Build a crazy jig for the pins to ride against.

Set the template stock against the jig boards such that both pins bear against the jig boards. Feed it thru left to right, keeping pins in contact with jig boards until the pin on the other end approaches the router bit. Don't route the pins off!


Viola! A true radius routed in the template. The footboard arch is a different radius that this one. The set-up is almost exactly the same: one jig board has to be swung "open" a bit, but it will result in a radius of something like 18 feet. That would be a really, really, really big circle jig.

Sharon's bed

Eye height 60" - getting into bed
Eye height 60" - headboard and footboard height, curvature, etc.
Trying to emulate the photo from Thomasville's CKing version. The curves were pinned directly from the image in SU, but the HB still looks too "crested" - due to being "queen-ified," I guess.

Other notes: mattress assumed 9" thick, mattress top at 30" high.

Last update June - see "older posts"

Shop shots

turned mallet

Artsy little pot

Someone on Woodnet says, "if it don't hold soup, it's art."

Prototype hall table, poplar, no finish

Home Office cabs, cases, maple, HD ply, garnet shellac

The right side base cabinet with the printer tray pulled out.

A very sloppy shot of the three drawers in the left cabinet base.

The drawers in the right desk base cabinet.