The problem: a through-the-wall package air conditioner that loses all
kinds of heat energy in the winter, both through metal conduction and
air leakage around a badly-done installation.
Solution: build an insulated box to simply slap over the entire thing in winter and seal against the surrounding interior wall. A quick set of on-site measurements gave the necessary LxWxD needed to clear it all around. After a little design noodling, a single 2x8 piece of XPS board would suffice. Went with 1.5" foamboard instead of the original idea of 1" since that's what HD had at the time, a happy accident that in retrospect made it much better.
This XPS was "furring ready", i.e. molded with a flange so that for doing
wall insulation, typical 1x3 wood strapping can sit between adjoining pieces
and nicely bridge the gap but remain flush with the outer surface.
For doing basement retrofits and such.
The cross-section is *not* shiplap, it's sort of a "hat" with the flanges
along each long edge pointing the same way.
Turns out this lent itself easily to much stronger glued box joints.
The end pieces each needed one additional edge rabbet to accomodate the
inner surface of the big face piece and have it all nest together.
Key to accurate foam cutting: thin and *very* sharp knives, a straightedge to guide cuts, and successive passes going a little deeper on each one. There's a school that favors serrated knives, but that makes cut edges more rough and I needed smooth faces.
|The four sides were glued together and aligned down on a flat table to simulate the target wall, lashed up with rope to hold it all together tightly, and left to cure thoroughly. Each corner received two beads of adhesive since two mating surfaces were available. The face piece temporarily dropped in to provide corner-to-corner alignment but wasn't glued in at this point; the side joints needed to be solid first.|
|Then, time for the face [or top]. Two beads along the edges to help hold against a slight inherent bow in the foam and make sure those joints were sealed; only one needed along the long top and bottom lines as stress would be straight inward against the wall anyway.|
The top was left weighted down to cure undisturbed for a couple of days.
In addition, a few bits of stiff wire were inserted into interference-fit holes drilled across joints to help "dowel" the side corners together -- especially where they would contact the wall.
Since the match to the wall would be close but couldn't possibly make
a perfect air-seal on its own, a gasket of squishy-foam went on all
around. I wound up using my
spinning staff as a roller to
press down the gasket sticky-back.
All the outer angles got a liner of Tyvek tape, as yet another redundant air-seal and to protect the foam edges a little better.
Conceptual fastening would be through a quick-n-dirty clamping bar
arrangement to help spread out load.
A string ties to the A/C grille inside and gets
led out through a small hole in the box face, through a close-fitting [and
additionally gasketed!] hole in the spreader, and gets fastened around the
two nearby screws with a stretchy element to absorb any dimensional changes
while everything's in place.
Probably higher R-value and better sealed than the wall it's going against, in fact.