Frost at the Ridge, Ice at the Eave, Poly on the Warm Side
A 1.5-story Cape Cod with a 2008 cathedral-ceiling addition: 2x10 rafters at 16 in. on center, R-30 fiberglass batts, a 1-inch vent channel from soffit to ridge, ridge and soffit vents, 6-mil poly on the warm side, and painted drywall. It is January, outdoor 12°F, indoor 68°F at 38% RH because the owner runs the furnace-mounted humidifier at 40%. The owner reports frost on the underside of the roof sheathing in bands near the ridge and around four non-IC recessed lights, dark staining on the sheathing, drip stains at two ceiling penetrations, and a small ice dam at the eave over the addition only. Infrared shows the same areas as cold spots. The rest of the house has a plain vented attic with no complaints.
1. Two moisture pathways are in play. Which one dominates here, and what in the description tells you so?
Read the worked answer
The pathways are diffusion (vapor driven through the assembly by the indoor–outdoor vapor pressure difference) and air leakage (indoor air carried into the cavity by a pressure difference). Air leakage moves one to two orders of magnitude more water vapor per unit of hole, and it shows a fingerprint this house has:
- Frost is concentrated in bands near the ridge and around the recessed lights, not spread evenly across the whole deck. Diffusion is uniform across the field of the assembly; it cannot produce a bullseye around a fixture.
- The poly is intact and on the warm side. Six-mil poly is a strong vapor retarder. If diffusion were the driver, the poly would be suppressing it - yet there is still frost, so the water is arriving by a route the poly does not cover: the holes.
- Non-IC recessed lights in an insulated ceiling are unsealed penetrations, and the ceiling plane is also pierced by top plates, wire and plumbing penetrations, and the ridge-area drywall joints. Warm 38% RH air hits the cold sheathing, drops below its dew point, and freezes.
- Ice damming at the addition's eave only points to a warm roof deck - from air bypassing the insulation and, plausibly, from compressed or displaced batts at the eave blocking the vent channel. Note that a small ice dam is a separate symptom of the same air leak plus an eave problem, not evidence of a different mechanism.
Diffusion is present but secondary; the humidifier setting raises the stakes by raising the indoor dew point, but removing the humidifier would only mask it.
2. The homeowner's contractor says to pull the poly out. Should it come out? What would that fix and what would it not fix?
Read the worked answer
No, and removing it is a net negative in this assembly.
- What removing the poly would do: slightly increase the drying capacity of the cavity toward the interior in summer, and slightly increase diffusion into the cavity in winter. That is the opposite of what this building needs in January at 38% indoor RH.
- What it would not do: it would not stop a single one of the air leaks. Frost around the lights and at the ridge would persist essentially unchanged, because that moisture is being carried by air, not vapor pressure. You would have spent money, created a mess, and left the mechanism intact.
- Worse: the assembly is a vented cathedral ceiling with fiberglass. Without the interior retarder, the cold sheathing becomes a much easier condensing surface whenever the cavity is cool, and you have made the assembly more reliant on the vent channel doing its job perfectly.
The correct sequence: air seal the ceiling plane first - the four recessed lights get replaced with airtight, insulation-contact-rated fixtures or airtight boxes built and sealed from below, then top plates, wire penetrations, and the ridge-side joints get sealed. Only after the air barrier is continuous do you evaluate whether the poly is appropriate. Fixing the air barrier is the intervention with the return; the vapor retarder is a control for a pathway that is not the dominant one.
Also worth stating plainly: the poly is not defective. It was installed on the correct side and it is doing its job. The house has holes in a different layer.
3. Lay out the repair sequence for the roof and the ceiling, and say what you do about the humidifier.
Read the worked answer
1. Verify the vent path before you add anything. Confirm soffit vents are open and not blocked by batts at the eave; a vent channel that is choked at the bottom does nothing at the ridge. Confirm the ridge vent is actually cut open.
2. Air seal from the interior. This is the highest-value step. Airtight-rated lights or sealed boxes, top plates, penetrations, and any drywall-to-rafter gaps. Do this before adding insulation so the new insulation is not bypassed.
3. Repair the insulation. R-30 in a 2x10 cavity is roughly full depth; the failure is more likely compression at the eave and voids at the ridge than an R-value shortfall. Correct the installation, maintain the 1-inch vent channel, and do not stuff batts against the deck.
4. Address the eave. The ice dam is fed by the warm deck and by meltwater refreezing at the cold overhang. Air sealing plus venting addresses the first; the eave detail (vent chutes, and in a severe case a membrane underlayment) addresses the second. Do not just add heat cable.
5. Then the humidifier. 38% at 12°F outdoor is aggressive and is buying the owner nothing but condensation potential. Set it so that window condensation does not form at design conditions - typically in the low 30s or below at these temperatures - and tell the owner the setting must track outdoor temperature, not stay fixed at 40%. This does not fix anything, but it stops the building from being tested at an artificially severe indoor condition.