Ice Dams on a Cape Cod: Where the Thermal Boundary Actually Is
A 1,900 ft² 1½-story Cape Cod built in 1962 has 400 ft² of kneewall attic behind two second-floor bedrooms. The attic floor holds R-19 fiberglass batts, compressed to about 2 in. where the crew walked. The kneewall has R-11 stapled between the studs with the kraft facing turned toward the attic, and the sloped ceiling bays over the bedrooms have no insulation at all - bare rafters are visible from the attic side. The owner reports icicles every February along the north eave, and the second-floor bedrooms run 6–8°F colder than the first floor. A blower door reads 3,100 CFM50 (about 12 ACH50 at this volume). Infrared scanning at 25°F outdoor shows the kneewall top plate, the band joist at the attic floor, and the sloped bays all near outdoor temperature, while the attic-floor batt cavities read warm in the middle and cold at the edges. The pull-down stair has no insulated cover, and the bath fan discharges directly into the attic.
1. Explain the mechanism producing the icicles and the cold bedrooms, and why blowing R-38 over the attic floor alone would waste money and could worsen the moisture picture.
Read the worked answer
Where the heat is going. Ice dams form when heat reaching the roof deck melts snow above the eave and the meltwater refreezes at the cold overhang. Here the roof deck over the bedrooms is not separated from the house by an effective boundary: the sloped bays are wide open to attic air, the kneewall top plate is unsealed, and the band joist runs cold to the rim. Warm room air convects up the kneewall cavities and rafter bays and warms the sheathing. The IR scan confirms each of those paths is at or near outdoor temperature - they are conducting, not insulating.
Why the bedrooms are cold. It is the same paths run backwards. The kneewall behaves as an uninsulated exterior wall, and the kraft facing is on the cold side, so it is a condensing surface rather than a vapor retarder in the right place.
Why more floor insulation is the wrong move. The flat attic floor is not the surface feeding the ice dams. R-38 over it does not touch the sloped bays or the top plates, so the icicles persist and the money is spent. It also leaves the bypasses unsealed, and once the attic side becomes colder the stack effect across those leaks is stronger, driving more moist air into cold sheathing that the bath fan is already loading. Order is: air seal, insulate, then ventilate correctly.
2. Set the thermal boundary for this house and describe how you would air seal and insulate the kneewall and the sloped bays, including what happens to the kraft facing and the bath fan.
Read the worked answer
Boundary decision. Pull the boundary in to the kneewall plane and the underside of the sloped roof, so the bedrooms and the kneewall attic behind them are all inside. Two consequences follow: every surface on that plane gets a continuous air barrier plus insulation in contact with it, and the attic-floor batts stop being a thermal control layer - do not add to them and do not leave a hybrid where half the attic is inside and half outside.
Kneewall. Remove the kraft facing; it is on the cold side and will trap moisture. Fill the stud cavity with unfaced batts or dense-pack, then install a continuous rigid-foam air barrier on the attic side with taped seams and the perimeter, top plate, and bottom plate sealed with can foam. Add thickness if you need the R-value. The interior drywall with a vapor-retarder primer is the inside retarder.
Sloped bays. If the roof is vented, keep soffit-to-ridge flow with baffles, then use closed-cell foam at the sheathing plus unfaced batts to fill, or dense-pack behind a sealed air barrier. Never leave the bay open to the attic.
Seal before you insulate. Kneewall top plates, band joist, wire and plumbing penetrations, and the attic access - build a gasketed, weatherstripped, insulated, positively latched cover.
Bath fan. Duct it to the exterior through a short, insulated, backdraft-dampered termination. Fix the moisture source before the cavity is closed.
3. Which diagnostics confirm the work, and what results would tell you the kneewall plane is actually tight?
Read the worked answer
A blower door before and after is the headline number - expect roughly 3,100 to 2,400–2,600 CFM50 if the top plates and band are genuinely sealed. But CFM50 is a whole-house figure and does not prove that the kneewall plane is the surface you fixed.
The targeted test is a zonal pressure diagnostic: hold the house at -50 Pa, then measure the kneewall attic pressure relative to the house. A well-sealed kneewall plane lets the attic sit close to outdoor pressure, near the full 50 Pa differential; a leaky plane holds the attic near house pressure. Record both numbers and the house-to-outdoor reference, or the result is meaningless.
Supporting checks: a smoke pencil or IR pass over the top plate, hatch, and band joist while the blower door runs; a flow hood or smoke test at the bath fan termination to prove it discharges outdoors; a look at the soffit baffles to confirm the vent path survived the insulation. Document pre and post CFM50, the zonal readings, and IR images with the indoor/outdoor temperatures and the scan time, then rescan on a comparable cold morning rather than a warm afternoon.