The kitchen window panel

HARDWARE RUN, Home Depot, $89.86, and the panel's remaining parts are mostly in hand: keeper steel (48 x 3/4 x 1/8 flat bar), five lines of #10 screws including the FLAT heads the magnets need, a black 3/8 aluminum tile edge profile that reads as a trim-language candidate, matte black hinges, and the microcement applicators, which means finish work is close. One conflict caught on arrival: the keeper bar is 3/4 wide and the groove is 3/4 wide. The spec said keeper NARROWER than the groove, 1/2 in a 3/4, for locate tolerance; a full-width bar jams. Either the groove goes to 1 wide (one more router pass) or the keeper gets narrower stock. The 1/8 thickness is also double the 1/16 the stack assumed, so the groove depth check runs again before routing: 3/16 magnet + 1/8 keeper = 5/16, not 1/4.

THE REAL PANEL IS IN. Cutout made against the template, panel standing in its final home between counter and uppers, window cranked open through it, evening light doing the rest. One day, notebook to fitted: template cut and proven, dimensions into the model, blank cut, cutout cut, dry-fit. What it still waits for: the cured rail frame with its groove, keepers on the back, magnets in the cart, seal on the bottom edge, and the microcement that turns it from plywood into a wall.

GLUE-UP. Titebond III on, the second layer going down staggered over the first, and the whole thing clamped with everything on site: bar clamps, F-clamps, and a kettlebell earning its keep as a flatness weight. The frame is being glued around the cassette itself, which locks the fit for good, the same template discipline the panel got. Overnight under the clamps; the groove waits for a cured, flat frame.

POPLAR RIPPED and the rails DRY-FIT, and the dry fit is being done the right way: around the actual blind cassette, pulled from the window and laid on the bench, not around the notebook numbers. The strips box the cassette with the roller and screen in place, which sizes the opening off the one part that cannot negotiate. The corner laps read clearly in the photos. Next: cut the layer pairs to final length, glue up flat, diagonals before the clamps come on.

RAIL FRAME dimensioned in the notebook and the poplar rip is next: frame outer 49 x 24 1/2, opening 45 x 20 1/2, rails 2 x 1.5. The numbers self-check, since outer minus opening is 4 both ways, two 2 in rails. The frame takes 23 ft 2 in of 2 in strip, well inside the 80 ft the 1x5 poplar yields. Glue-up plan drawn in section 5: two 3/4 layers, long members staggered so every corner locks as a half-lap, squared by diagonals, clamped flat, groove routed after. One relationship worth writing down: the panel cutout (41 3/4 x 15) is smaller than this opening (45 x 20 1/2), so the panel laps past the rail's inner edge and becomes the reveal, and the magnet annulus is the 2 in rail face outboard of it.

CUT DAY. An outdoor cut station under a canopy: sawhorses, sheet stock staged against the shed, layout squared off the notebook onto the real sheet, and the REAL PANEL BLANK cut to size. The window cutout is the remaining operation: radius the corners first, then connect them, show face managed for tearout. The test template from earlier today is the reference the blank gets finished against, not the notebook numbers directly.

A TEMPLATE FRAME is up around the kitchen window, and it is the reference the poplar gets cut to. It is construction stock, fitted in place around the window with the pane opening through it, which proves the frame geometry: outer size, the reveal to the window, and that the awning pane and its latches clear the frame when open. The poplar rails copy this frame, then take the blind pocket and the 3/4 magnet groove the template does not have. Same discipline as the panel: numbers to template, template to fit, corrected template to final stock. Before the poplar is cut, mark onto this template anything the fit taught, and check the groove and the blind pocket do not collide in the same 1.5 of rail depth.

The kitchen panel is CUT and DRY-FIT, and the window dimensions are finally measured rather than assumed. Off the cut sheet: panel 56 x 27 7/8, cutout 41 3/4 x 15, with 5 7/8 above the cutout and 7 below, which sums to 27 7/8 exactly. The sketch does not give the left/right margin split; the drawing assumes centered, 7 1/8 a side, until measured. Every drawing on this page now uses the real numbers. Two consequences. The panel is lighter than designed for: 6.5 sq ft net and about 11 lb against the assumed 10 sq ft and 17 lb, so the magnet arithmetic now says 4 would hold it and the 8 drawn are 4x margin; fit fewer if it does not buzz. And the panel is 27 7/8 high, not 36: it spans counter to upper cabinet, not counter to ceiling, so the U-bracket detail stays but the shadow-gap question moves to the cabinet underside.

Home Depot run, $27.12: Titebond III 16 oz for laminating the rails, a roll of magnet tape, and an assortment of zinc round-head wood screws. Two notes for whoever builds this. The magnet tape is not the panel fixing. Flexible ferrite tape runs about 1 to 2 lb of pull per linear foot on steel, so a 30 in roll is a few pounds against the 34 lb this panel wants, and it is a soft compliant mount where the panel needs a hard one. The cup magnets are still to be ordered. Round-head screws will not work at the magnets. A cup magnet has a 90 degree countersink and needs a flat-head screw to sit flush; a round head stands proud and breaks the 3/16 plus 1/16 equals 1/4 stack that lets the panel bed flat. These screws are fine for the L brackets.

Rail stock bought at Mr Plywood: 4 lengths of 1x5 poplar, 10 ft each, 40 linear feet for $114. A 1x5 is 4.5" actual, which yields two 2" strips per board, so 80 feet of strip and about 40 feet of finished 2 x 1.5 rail once laminated. That covers the kitchen window and both lounge windows with margin. Poplar over a construction 2x3 for three reasons: it is dried to furniture moisture rather than 19%, it is clear so no knot lands in the magnet groove, and it holds the bracket screws better.