The envelope is pretty good these days and adding more polyiso to
the inside of the basement wall has helped with overall heat loss
a little more, but I think there’s still something going on in the
cinderblock wall that’s affecting slab temperatures near the edge
and I wanted to float an idea for how I can better deal with that.

This is long, but one needs to understand the background and present
setup. Executive summary: should I insulate just the above-grade
exterior part of a foundation wall with Roxul panels?

Referring to this picture — there’s already a copy of “mywall.jpg”
somewhere on GBA from another post so I’m not going to upload
another one — the foundation wall appears to be about 80 inches
high in all [10 courses] from the footing, and I’ve got a fairly
generous part of it exposed above grade anywhere from 20″ to 26″
visible. The “more polyiso” was added on the inside with the top
tied to the sprayfoam and the bottom fairly air-sealed at the slab,
per recent building-science agreement that “basement walls are wet
rocks” and keeping their moisture strictly out is the right answer.
It’s all foil-face and taped up, so the wall and sill area can no
longer dry to the inside at all and the only viable path is to diffuse
toward the outside through the cinderblock. Which appears to work
just fine now. Thus I wasn’t really going to consider any additional
work on the exterior that might bury the sill in a moisture tomb.

Over the winter I found that the edges of the slab, particularly
in corners, stayed fairly cold and definitely below the nominal 55F
earth temperature in these parts. Center slab, which is uninsulated,
remained solidly around 55 the entire time. [I don’t really bother
trying to heat the basement, so its general air temp ran about 57F
most of the winter.] So I think what’s happening is that the upper
part of the wall gets chilled down, now even more since the interior
insulation has been improved, and that the “chill” can air-convect
down through all the open columns in the cinderblocks right down to
footing level. From there it’s a less-than-R-1 hop to the edge of
the slab, so the reverse of all that could easily be a significant
heat loss path. The polyiso also runs inward a foot or a little less
over the slab at the bottom and the vertical application sits on that,
as I knew I needed to mitigate some of the slab-edge effect up front
and I wanted foil-face bearing on the concrete and not the raw
polyiso itself. The coldest concrete is of course showing up right
under the edges of that foam “mini-footing”.

I tried temporarily laying more foam down on the slab into some of
the corners, and over another couple of days the coldness marched
farther inward to the edge of *that* — not quite as cool, but still
a noticeable gradient dropoff as it vanished under whatever foam
was present.

Given that the exterior exposed foundation still looks about
as relatively warm as it ever did on an IR shot even after the
re-insulation job, I’m surmising that there’s a continual battle
going on between general earth temperature and outdoor ambient
around the grade boundary and that the air columns inside the CMU
wall tilt the balance toward letting things become colder lower
down. [That can’t be good for my incoming water line, either…]
So my idea is that I want to insulate the exterior of the exposed
foundation wall but just a little way into the ground as opposed to
digging up the whole thing, which should help remove the effect of
the large above-ground concrete heat sink. Tucked right up against
the metal flashing against the sill, it might also somwehat cut
down on the “big aluminum heatsink” effect of that on the lowest
run of sheathing and help make the sill area itself warmer too.

The people at Roxul assure me that ComfortBoard IS would be a perfect
solution for this, even at the 2″ thickness I’m considering, and
totally moisture-open so I wouldn’t risk compromising the sill area.
A small bed of half-inch stone at the buried lower edges might help
drainage, although optional since the area next to the wall stays quite
dry these days because of all the overhangs. It wouldn’t need insect
protection but given reports here from other exterior Roxul jobs I
would probably need to protect it against birds, and am thinking simple
application of 1/4″ galvanized mesh up to just shy of the flashing.
The green coated stuff might even make an attractive color combination.
I guess a couple of IDP type push-in fasteners or equivalent [is there
an equivalent?] per panel would be the right way to hold it all on,
but how would one go about sealing between the panel edges — if
that’s even necessary?

The Roxul panels would only go a foot or less into the dirt, and
there would still be a few uncovered areas such as across a solid
concrete front staircase, the basement bulkhead pit, and a couple of
windows. So the central question here: is doing this exterior hack
likely to gain me anything or am I just kidding myself?

_H*