Sweating Ducts in a Vented Attic: The $18,750 Machine and the $8,400 Enclosure Fix

The call came from the heating and air contractor, not from the homeowner. He had already done the obvious repair, twice over, and it had not worked. His ducts were still sweating. He wanted to know whether foam on the metal would end it.

The house is a large old one in Algiers Point, across the river from the French Quarter. The complaint was narrow and entirely reasonable: my ducts in the attic are wet, water is dripping, and now the metal is staining. Not optimize my attic. Not monitor my dew point. Stop the sweating.

That distinction is the whole article. A homeowner's stated concern is a specification. Everything a contractor proposes should be traceable back to it, and anything that is not should be labeled as an extra, priced separately, and declined without penalty.

The prior sequence, in order:

1. The rigid metal ducts in the vented attic sweated, and the original fiberglass duct wrap got wet and stayed wet.
2. A heating and air contractor removed the wet wrap and installed new foil-faced fiberglass duct wrap.
3. He then mastic-sealed the seams, laps, and joints of the new jacket as tightly as he could, to keep humid attic air from reaching the metal again.
4. The ducts sweated again. The new wrap got wet. The metal began showing rust staining.

He deserves credit on two counts, and the article is weaker if it skips them.

First, the strategy is the one the insulation industry publishes. The North American Insulation Manufacturers Association states that condensation on a sheet-metal duct surface “is controlled by using a sealed vapor retarder of sufficient perm rating,” and that insulation with sufficient R-value “along with a low permeance vapor retarder will minimize, and in most cases prevent, condensation problems.”¹ That is what the published guidance calls for.

Second, he chose the right material. Lawrence Berkeley National Laboratory ran duct sealants through accelerated thermal aging designed to swing duct surfaces from attic-hot to condensing-cold. Cloth-backed tapes showed “visible signs of failure within about 3 days.” Mastic did not: the report records that “the metal backed tapes, the aerosol and the mastic show no visible of measurable signs of degradation” over the test period.² He used mastic, not tape. That is the durable choice.

So this was not sloppy work, and this article is not a critique of him. It is a demonstration that the strategy was aimed at the wrong layer. Both NAIMA statements above are conditional on sufficient R-value plus the vapor retarder — not the vapor retarder alone. And there is a second problem, quieter and more expensive, that only shows up when you look at the perm numbers. It is taken up in the water already in there.

The mechanism has one sentence and no controversy in it. Building Science Corporation: “When an attic surface temperature is lower than the attic air dew point, condensation will occur.”³ The Department of Energy's Building America program states the duct case directly: “When the surface temperature of the duct falls below the dew point of the surrounding air, there is a potential for condensation on the duct.”²⁶

Now put New Orleans numbers on both sides of that inequality.

The air. The ASHRAE climatic design table for New Orleans International gives a 1% design dew point of 78.0 °F and a 0.4% design dew point of 78.7 °F.²⁷ Inside a vented attic it is worse than outdoors during the day: Building America's monitoring of hot-humid attics found the attic dew point dropping to around 60 °F at night and rising “rapidly into the upper 80s during the day,” fluctuating across a 50 °F to 90 °F band.²⁶ BSC's own vented-attic research measured Florida attic dew points reaching 85 °F.³

The surface. Air leaving a residential cooling coil is 50 to 55 °F.³ Sheet metal is a conductor. With too little insulation on the outside of it, the outer surface of that insulation sits far closer to 55 °F than to the 85-plus °F dew point around it.

A surface below dew point in air above dew point condenses water. Every cooling cycle. Building Science Corporation is blunt about where the responsibility lies: “The attic venting is responsible for the sweating ducts.”²⁸ And it is common, not exotic — “In hot humid climates condensation on ductwork and air handlers located in vented attics is common.”⁵

There are therefore exactly two levers. Lower the dew point of the air touching the duct, or raise the temperature of the surface the air touches. That is the entire decision this house faced, and the two proposals split cleanly along it: one bought the air, the other bought the surface.

A code footnote most people miss

Louisiana requires R-8 duct insulation for supply ducts outside conditioned space.²⁹ That is an energy minimum, and it is not a condensation-control value. NAIMA reproduces the ASHRAE 90.1 duct insulation table with its own footnote attached: “The required minimum thicknesses do not consider water vapor transmission and possible surface condensation.”¹⁰ Their duct manual says it again in plainer terms — “R-values as required by ASHRAE 90.1-1989 may not be sufficient for condensation control”¹ — and states the remedy in one line: “Adding insulation is the best way to control surface condensation.”¹

A code-compliant duct in a New Orleans attic can sweat all summer and still be code-compliant. Anyone who tells a homeowner the duct is “up to code, so that is not the problem” is quoting the wrong table.

Louisiana enforces the 2021 International Residential Code, effective January 1, 2023,³⁰ and the entire state is Climate Zone 2A. Three code paths matter here, and all three couple sealing an attic to insulating and air-sealing the roof deck.

Sealing the vents is not a standalone move

Louisiana's amended energy provision permits blocking attic ventilation in exactly one context. The exception reads: “When the thermal covering at the roof line creates an unvented attic… There shall be no outside attic ventilation and all openings must be blocked with rigid material and are sealed.”²⁹

The permission to seal is downstream of a thermal covering at the roof line. Take away the thermal covering and the sentence does not apply to your attic.

IRC R806.5, the unvented attic assembly section, sets the same condition from the other direction. Item 1 requires that “the unvented attic space is completely within the building thermal envelope,” and the insulation options require air-impermeable insulation “in direct contact with the underside of the structural roof sheathing.”⁷ An attic with a bare roof deck and insulation only on the ceiling is not within the thermal envelope, and has no insulation in contact with the sheathing.

Where dehumidification does appear in code, and what it is standing in for

Mechanical dehumidification is not absent from the codes. It appears in two places, and in both it is conditional on an enclosure that has already been built.

In the unvented attic provisions, dehumidification appears as an alternative to supply air in an assembly that already has air-permeable insulation installed below the roof sheathing.¹¹ It is a conditioning option for a completed unvented roof assembly, not a substitute for building one.

The crawlspace section is the closest true analogy, and it is instructive. IRC R408.3 allows continuous mechanical dehumidification of an unvented crawlspace at 70 pints per day per 1,000 square feet — but only inside a sealed enclosure with a Class I vapor retarder on the ground and joints lapped and sealed.³¹ Code lets you dehumidify a crawlspace. It does not let you dehumidify a vented crawlspace and call it done. The vapor retarder and the sealed perimeter come first, and the machine conditions what is left.

That is the pattern across every code path: the enclosure is built first, and the equipment conditions the enclosure. The mechanical proposal here inverted it — buy the equipment, seal the vents, and let the equipment substitute for an enclosure that was never built.

The scale of the openings you would be fighting

IRC R806.2 requires a minimum net free ventilating area of 1/150 of the vented space, reduced to 1/300 where at least 40% and not more than 50% of the vent area is in the upper portion of the attic.⁴

Worked for a 1,500 square foot attic:

• At 1/150: 1,500 ÷ 150 = 10.0 ft², or 1,440 in² of net free opening.
• At 1/300: 1,500 ÷ 300 = 5.0 ft², or 720 in².

A pair of ducted dehumidifiers moves on the order of 200 to 400 CFM combined. Pushing 400 CFM through 5 ft² requires a face velocity of about 400 ÷ 60 ÷ 5 = 1.3 feet per second — a barely perceptible drift. The code-mandated openings are enormous relative to the airflow the equipment can process. You are not dehumidifying a room with a door ajar; you are dehumidifying a room with a wall missing.

And one Louisiana-specific prohibition

Worth knowing because it rules out the other common variant of this fix: Louisiana's amendment states that “In Climate zone 2A Supply and Return ductwork shall not be buried in insulation.”²⁹ Burying the ducts in deep loose-fill — a documented Building America technique elsewhere — is off the table here. The foam has to carry the thermal job on its own.

The search was deliberately for the affirmative case: a study, a field report, a manufacturer's engineering bulletin, anything establishing that mechanically dehumidifying an attic whose roof deck has neither insulation nor an air barrier is a defensible strategy. Building Science Corporation, NREL, Oak Ridge, Lawrence Berkeley, Florida Solar Energy Center, DOE Building America, ACCA, the Florida Building Commission, and the dehumidifier manufacturers themselves.

There was none to find. No code path, no peer-reviewed study, no national-laboratory field study, and no manufacturer guidance supports it. An absence is weaker evidence than a finding, and it should be labeled as an absence — but for a technique carrying a five-figure price tag and a permanent service obligation, the absence of any published basis is itself the answer.

Here is what the literature does say.

A vented attic exchanges its entire air volume several times an hour

Florida Solar Energy Center's review of tracer-gas measurements in real attics reports 1.7 to 2.3 air changes per hour in August, 3 to 4 ACH under moderate wind, 2.9 ACH in a long-term test, and 0.9 to 1.8 ACH in two others — with modeled daily rates ranging “from about 1 ACH at night to 4 ACH during windy afternoons.”¹⁵ Building Science Corporation puts a code-built attic higher still: “The air change in a perfectly built and vented attic (code 1:300 ratio) results in an average air change rate of 3 to 6 ach.”¹⁷

Lawrence Berkeley adds the part that matters most for anyone hoping to fix this by blocking the visible vents: “ventilation rates in conventional vented attics are roughly an order of magnitude greater than those measured inside homes,” and unintentional venting alone runs on the order of 1 ft² of leakage per 500 ft² of roof area even without designed vents.³² Oak Ridge is equally direct: “even if the whole-house air leakage is fairly small, the attic may still experience quite high air exchange with outdoor air,” and “it is critical that in hot and humid climates the attic is air sealed to the outside carefully.”³³

Blocking the soffit and ridge vents in an old framed roof does not produce an air barrier. It produces a slightly less vented attic — and, per Berkeley, one whose remaining leakage is still an order of magnitude above a house.

The one test of almost exactly this configuration

The Florida Building Commission commissioned a side-by-side attic vent performance comparison that included a test cell with the vents sealed, the roof deck left uninsulated, and R-30 insulation on the ceiling — the configuration the mechanical proposal would have created, minus the dehumidifiers. No moisture failures occurred during the test period, which is the honest headline. But the report records that “Sealing attic vents resulted in roof deck surface RH to increase from 68% RH to 82% RH,” that deck moisture content rose from 11.5% to 13% with peaks at 17.5%, that cool-weather deck relative humidity averaged over 80%, and that “Currently the attic vent sealed test configurations are not permitted by the Florida Building Code.”³⁴

No dehumidifiers were used in that test, so it cannot settle whether equipment would have held the deck drier. What it does establish is that sealing the vents without insulating the deck moves the roof sheathing in the wrong direction, and that the test configuration is not code-approved where it was performed.

The roofing side has a view too. GAF's technical bulletin describes an acceptable unvented roof as one where “the thermal, moisture and air control layers are located at the roof deck,” and while it does not void warranties simply for an unvented deck, it excludes damage caused by the unvented condition.³⁵ On a historic house with a roof the owner intends to keep, that exclusion is not a footnote.

The manufacturers say seal it first

This is the part that should end the conversation, because it comes from the companies that would be selling the equipment.

The one major model marketed for attics carries this instruction in its own manual: “In order to efficiently control humidity levels, the area in which the dehumidifier is to be operated must be free of water intrusion or excessive fresh (outside) air infiltration. Before installing… water intrusion and air infiltration problems should be addressed.”¹³

Competing product literature names the application as sealed attics specifically — “Dries a specific area that has a moisture issue (basements, crawl spaces, sealed attics, etc.).”³⁶ And a manufacturer FAQ states the general principle for trying to dehumidify one part of a connected volume: “The humidity from the rest of the house will overwhelm the efforts to dehumidify a single room.”⁶ Substitute “the outdoors” for “the rest of the house” and you have this attic.

And the equipment cannot survive the hours when it is needed most

Every published operating range for this class of equipment tops out below a Gulf Coast attic's afternoon temperature.

New Orleans' 0.4% design dry bulb is 93.6 °F.²⁷ Add the FSEC attic rise and the attic is 116 to 129 °F on the afternoons that matter. One brand's units are outside their rated range; the other brand's controls lock the compressor out entirely above 104 °F.¹⁴

Set aside the code and the temperature lockout. Assume the equipment runs at full nameplate output, in a vented attic, and ask whether the numbers close. This is arithmetic, and every input is stated.

The moisture the vents deliver, per day

Moisture removal required equals the mass of air exchanged times the difference in humidity ratio between what comes in and what you want to hold:

required removal (lb/h) = attic volume × ACH × air density × (W_outdoor − W_target)

Inputs, each labeled:

• Volume. 1,500 ft² attic footprint at 4 ft average clear height = 6,000 ft³. Geometry assumption.
• Density. 0.075 lb dry air/ft³, the standard-air convention. Hot attic air is lighter, so this overstates the mass flow and therefore overstates the required removal — the assumption is conservative against the argument being made here.
• Incoming moisture. At the New Orleans 1% design dew point of 78.0 °F,²⁷ saturation pressure interpolates to 0.4788 psi from published steam-table values,³⁷ giving W = 0.622 × 0.4788 ÷ (14.696 − 0.4788) = 0.02095 lb/lb.
• Targets. A 60 °F dew point gives W = 0.01104 lb/lb. A 55 °F dew point — the level actually needed to stop condensing on a 55 °F duct surface — gives W = 0.00933 lb/lb.³⁷
• Conversion. One pint of water = 1.043 lb.
• What is excluded. Steady state, ventilation moisture only. Roof-deck sorption and desorption, diffusion and air leakage up from the house, and re-evaporation of condensate are all ignored. Every one of them adds load.

At 1 ACH — a still night — 196 pints of nameplate capacity covers the 103 to 120 pint requirement with margin. At 2 ACH the requirement of 205 to 241 pints per day equals or exceeds the entire nameplate output of both units. At 4 ACH it is two to two and a half times their combined nameplate.

And nameplate is generous. These ratings are taken at 80 °F and 60% RH. Manufacturers' own second rating points show what happens as conditions move: one 155-pint model falls to 104 pints at 73 °F, a 70-pint model to 47 pints, a 100-pint model to 74.¹⁶ The rating condition is not the attic.

The honest bound: the equipment can win, but only if the airflow term collapses to something like 0.1 to 0.3 ACH. That is not a vented attic anymore. It is an air-sealed one — which is the thing that was not built, and the reason all of this is circular.

The equipment also heats the attic it is drying

A dehumidifier is a heater with a drain. The physics is not in dispute and the manufacturers publish it: air leaves the unit “warmer and drier than the air entering,” with a published discharge temperature rise of 10 to 30 °F.³⁸ Allison Bailes puts it in one line: a dehumidifier “is a device that converts latent load into sensible load.”³⁹

Quantified for one 98-pint unit drawing 670 W,¹⁶ with both intake and discharge in the attic:

• Electrical input: 670 W × 3.4121 Btu/h per W⁴⁰ = 2,286 Btu/h
• Water condensed: 98 pints/day ÷ 24 h = 4.083 pints/h × 1.043 lb/pint = 4.26 lb/h
• Latent heat released: 4.26 lb/h × 1,052 Btu/lb⁴¹ = 4,480 Btu/h
• Total per unit ≈ 6,770 Btu/h; two units ≈ 13,500 Btu/h, or about 1.13 tons of heat added to the attic at 12,000 Btu/h per ton.⁴¹

If the units are throttled or locked out, the latent term shrinks and the floor is the electrical input alone: 4,572 Btu/h for two, about 0.38 tons. So the range of heat dumped into the attic is roughly 4,600 to 13,500 Btu/h.

That heat is not free. Raising attic air temperature raises the temperature difference across the duct walls and across the insulated ceiling, and both drive load into the house. Building America measured duct thermal losses and gains in unconditioned spaces at 10% to 45% of heating and cooling load, worst “at peak demand conditions when the attic is hottest.”²⁶ ACEEE's summary is blunter: ducts in attics “add ½ to 1 ton to the cooling load, and increase peak demand by about 1 kW.”⁴² The same manufacturer's manual warns of the reciprocal effect on the cooling equipment: “THE DEHUMIDIFIER WILL HEAT THE AC COOLING COILS WHICH DIMINISHES THE AMOUNT OF WATER THE AC SYSTEM WILL REMOVE WHEN OPERATING.”¹⁸

The equipment fights the symptom while making one of the symptom's drivers worse.

One number to keep honest

It is tempting to say this is two extra tons of compressor running around the clock. It is not, and overstating it would be the same sin as the proposal. Two ducted residential dehumidifiers are a 500 to 950 watt each class of appliance — 1.34 kW connected for the pair, against roughly 2.39 kVA for a 2-ton condensing unit. That is about 0.56× a 2-ton condenser, or roughly one ton of connected load, not two.

The real problem is not the peak draw. It is the duty cycle. A properly sized air conditioner reaches setpoint and cycles off. Two dehumidifiers in a vented attic have no setpoint they can reach, so they run — and 1.34 kW that never stops uses more energy in a year than a 2-ton compressor that cycles.

What that costs to run

Arithmetic, at published watts and a published tariff:

• Two units at 670 W = 1.34 kW.
• Continuous: 1.34 kW × 24 h × 365 d = 11,738 kWh/yr. At 50% runtime, 5,869 kWh/yr.
• Entergy New Orleans' own published residential bill computation totals $133.43 for 1,000 kWh before state or city taxes, fuel adjustment and all riders included — 13.343¢/kWh all in.²²
• 11,738 kWh × $0.13343 = $1,566/yr continuous. At 50% runtime, $783/yr.

For scale, EIA puts the Louisiana residential average at 12.66¢/kWh, up from 11.44¢/kWh a year earlier — a 10.7% increase in twelve months.⁴³ The figures below carry no escalation, so read them as a floor.

Everything to this point argues that the mechanical proposal could not have held the attic dew down. This section argues something stronger: even if it had worked perfectly, it could not have dried the ducts.

Look at the assembly the previous contractor left behind, layer by layer, in perms.

At 0.02 perm the facing is not a vapor retarder in the loose sense. It is five times tighter than the 0.1-perm Class I threshold.⁴⁴ And sheet metal is itself effectively a vapor barrier — the Spray Polyurethane Foam Alliance notes that “metals and glass… restrict diffusion so much they can be considered true vapor barriers.”⁴⁶

So the wet fiberglass sits between two vapor-impermeable faces. Mastic-sealing the last seams did not add a barrier to a drying assembly. It closed the final opening in a cavity that already had barriers on both sides.

That condition has a name and a warning attached in the federal guidance. DOE: interior vapor retarders must be avoided in hot-humid climates, “otherwise a double vapor-retarder condition would result with the potential to trap moisture within the cavity,” and “an exterior enclosure surface which is not vapor permeable will not allow drying to the exterior.”⁴⁷ Building Science Corporation states the failure mode in one sentence: “Vapor barriers were originally intended to prevent assemblies from getting wet. However, they often prevent assemblies from drying.”⁴⁸

How little moisture moves through 0.02 perm

A facing manufacturer's technical bulletin does the arithmetic in the other direction, for vapor moving in: a facing at 1.0 perm passes about 50 gallons of water per week into the insulation, one at 0.09 perm about 5 gallons, and one at 0.02 perm about 1 gallon per week.⁴⁵

Diffusion is symmetric — the same membrane throttles vapor trying to leave to the same order of magnitude. But the driving force outward is far weaker. That inward figure assumes the enormous vapor-pressure difference between a hot humid attic and a cold duct. A dehumidifier can only lower attic dew point modestly, which shrinks the outward drive well below the inward one it is compared against.

Which is the answer to the specific question this house raised. Two machines lowering the dew point of the air outside a mastic-sealed 0.02-perm jacket cannot pull liquid water out from under it. Not slowly. Not eventually. The rate is set by the membrane, and the membrane was chosen to stop exactly that transfer.

Wet insulation does not sit still — it makes the condensation worse

Water in fiberglass displaces the air the R-value depends on. The National Insulation Association: wet fiberglass “can temporarily lose a substantial measure of its thermal resistance… Water is more conductive than air,” and the added weight compresses the batt, which reduces thickness and reduces R-value again.⁹

That closes a loop. Less R-value means the outer surface runs colder relative to the metal, which keeps the assembly further below dew point, which condenses more water, which lowers the R-value further. It does not stabilize on its own, and no amount of attic dehumidification interrupts it, because the loop is happening inside the jacket.

The rust staining

Corrosion under insulation is a formally standardized failure mechanism, not a maintenance opinion. The peer-reviewed review literature defines it as external corrosion of metal beneath insulated equipment caused by water ingress, states that it “is initiated by the retention of moisture/water within the thermal insulation and usually accelerates when the insulated equipment is operated under cyclic temperature and wet–dry cycles,” names condensation explicitly as a source, and reports that “in the case of carbon steel, corrosion rates under the insulation can be up to 20 times higher than those in naturally aerated atmospheric conditions.”²¹

The same literature is clear that the jacket plays both roles. It is “the first external line of defence” against water getting in — and once water is in, the gaps at the insulation-to-metal interface “retain moisture, making the underlying metal vulnerable to corrosion,” forming “an ideal reservoir for water/moisture retention.”²¹ Practitioner guidance adds that water ingress is the number one cause, that “once trapped, that moisture initiates corrosion,” that thermal cycling accelerates it “especially in coastal or humid climates,” and that stains on a jacket “should be investigated as soon as they're discovered.”⁴⁹

For HVAC specifically: “rust in ducts is a problem indicator, showing quite clearly that the duct system has been wet.”⁵⁰

Being precise about what that proves: rust proves the metal has been wet. No located source establishes that visible staining alone proves sustained or repeated wetting rather than one long event. What it is consistent with — given that corrosion under insulation is initiated by water retention and accelerated by wet-dry cycling, and that a cooling duct in a hot-humid attic cycles every time the system runs — is exactly the condition the rest of the evidence describes.

Which is why the wrap had to come off

Two industry rules settle it, and neither belongs to the contractor performing this work.

The National Insulation Association gives the drying window: insulation wet with clean water “likely can be left in place and allowed to dry,” but “this drying process should take no longer than 24 hours… If the insulation does contain contaminants or takes longer than 24 hours to dry, it should be replaced with new insulation.”⁹ Fiberglass inside a mastic-sealed 0.02-perm jacket, on a duct that recondenses water every cooling cycle, has not dried in 24 hours. It has been wet long enough to rust the metal.

NAIMA's duct standard is shorter: “Duct systems in service that are found to be wet should be replaced.”¹⁰

And the reverse move — wrapping more insulation over the existing jacket — is specifically warned against: “Adding another layer of insulation can create a condition where the original vapor retarder is now cold enough to be a condensing surface. This can lead to wet insulation,” with the only acceptable case being ducts carrying heated air only.⁵¹

The scope, as written:

• Remove and dispose of the fiberglass duct insulation from the existing rigid metal duct.
• Dry the metal first — heat on for two hours before work, to bring the metal above dew point so foam is not sprayed onto a wet surface.
• Apply 1.5 to 2 inches of closed-cell polyurethane foam at R-6.5 per inch, seamless and continuous, to all rigid metal ductwork from the supply plenum out to the pressure box collars, with depth verified every two feet.
• Continue the same closed-cell foam over all 13 supply-air boots.
• All sheet metal covered. Debris removed.

Four reasons that is the correct control layer, each from published guidance rather than from opinion.

1. The thickness matches the published condensation-control threshold

DOE Building America's spray foam guide gives the number: over R-6 duct, “an additional R-9 (1-1/2″ of ccSPF) for a total of R-15 will prevent the dew-point from being reached in nearly all typical conditions,” and closed-cell at that thickness “is also an effective vapor retarder inhibiting moist attic air from penetrating into the insulation where the dew-point might be met.” The same guide rates open-cell foam on ducts “Not acceptable” and closed-cell “Preferred.”⁴⁷ At R-6.5 per inch, 1.5 to 2 inches is R-10 to R-13 of added resistance on bare metal.

PNNL's Building America Solution Center specifies the same technique and the same coverage: “Apply at least 1.5 inches of ccSPF to all duct surfaces, including trunks, branches, and register boots.”⁵² Building Science Corporation reports it works over R-6 and R-8 ducts in hot-humid climates “at a little more than 1-inch thick. But low-density foam does not work” — and adds, “don't forget about the boots and penetrations.”²⁸ Thirteen boots, all foamed. That was not thoroughness for its own sake; the boots are where the metal is closest to the room air and where the published guidance says failures show up.

2. It is more vapor-open than what it replaced, on purpose

This is the part that surprises people, so here are the numbers.

The foam is roughly 30 to 40 times more open to drying than the facing it replaced. That is not a shortfall in the specification; it is the point. The assembly keeps a drying direction. Building Science Corporation states the design requirement as an absolute: “We want the assembly to dry inwards from the control layers — and to dry outwards from the control layers. Always. Everywhere.”⁴⁸ BSC's working criterion for duct encapsulation is that the insulation be “less than 1 perm or enclosed in a layer or surface that is less than 1 perm”²⁸ — and 0.63 to 0.83 perm satisfies it while staying well clear of 0.02.

3. There is no seam to seal and no cavity to fill

The previous strategy asked a hand-assembled membrane to stay continuous across hundreds of seams, laps, hanger points, and takeoffs in a hot attic. One breach undoes it: “Even the best vapor retarder can be rendered inadequate from leaky seams.”⁴⁵ Detailing requirements include saddles under every hanging strap, maintained thickness at every seam and joint, insulation over the tape rings, and repair of every puncture in the outer jacket.⁵⁴ Even the lamination process degrades the number: the manufacturer's own footnote states 0.02 perm is measured “for facing material prior to lamination. After lamination, permeance values may be higher.”⁸

Sprayed foam has no seams to fail. SPFA: “because spray-applied polyurethane foam is seamless and closed cell, air leakage is less of a concern than diffusion.”⁴⁶ The same document makes the contrast explicit for seamed materials: “While steel sheeting may be considered a vapor retarder (virtually total), a steel roof deck usually has so many seams and holes that it may be considered a vapor transmitter.”⁴⁶ Substitute “hand-sealed foil jacket” for “steel roof deck” and that sentence describes what was on these ducts. Closed-cell foam is air-impermeable at 3/4 inch by ASTM E283 testing,⁵⁵ and DOE notes it is used as a structural adhesive in modular housing plants — it bonds rather than wraps.⁴⁷

Stated as an inference, not a citation: the corrosion literature attributes water retention to gaps at the pipe-to-insulation interface that act as reservoirs.²¹ An adhered, gap-filling foam should eliminate that reservoir space. No retrieved source states that in those words, so treat it as a reasoned expectation rather than a published finding.

4. Nothing about the building had to change

The attic stays vented. No vents blocked, no roof deck touched, no code exception invoked, no historic roof assembly modified, no shingle warranty question raised. The intervention is confined to the duct, which is where the problem was.

First cost is the least interesting comparison, so start with the honest one and then extend it.

First cost

• Mechanical: $18,750 — plus an electrician for two attic circuits, plus stairway sheetrock, plus the new pressure boxes and new duct insulation its own notes required separately.
• Enclosure: $8,400, paid in full, with nothing owed afterward and no separately noted work.

On the first-cost line alone that is roughly 2¼× ($18,750 ÷ $8,400 = 2.23), before the add-ons and before the first year of monitoring. It is also worth noting what $18,750 buys per unit: $9,375 each installed, against a published national installed range of $1,700 to $3,500 for a 90 to 100 pint whole-house dehumidifier¹² — 2.7 times the top of that range and 5.5 times the bottom of it.

Ten and twenty years

Inputs: $418/yr monitoring as quoted; electricity at the arithmetic above; filters at $10 to $30 each replaced at least annually;¹² replacement on a 5 to 10 year service life, with market-rate replacement at $1,700 to $3,500 per unit installed.¹² Nominal dollars, no escalation, no discounting.

A middle case for the ten-year figure — continuous runtime on electricity, market-rate replacement at the midpoint: $18,750 + $4,180 + $15,660 + $400 + $5,200 = $44,190 over ten years, about $4,420 a year.

What else the recurring column buys

The proposal added eighteen or more new components above a finished ceiling: two compressors, two fan motors, two control boards, two condensate pumps, two sets of float and safety switches, two overflow pans, two drain lines, two Wi-Fi monitors, and two dedicated circuits. Every one of them is a failure mode in a space the International Code Council itself describes as one where failures are not readily observable.

The ICC's commentary on its condensate pump provision is worth quoting in full, because it was written about exactly this installation location:

“Condensate pumps are often located in attics and crawl spaces and above ceilings where they are not readily observable. If they fail, the condensate overflow can cause structural damage to the building, especially where the overflow will not be noticed immediately.” And: “These float controls are commonly not connected, and in other cases, the pump might not be equipped with an overflow switch.”

ICC commentary on 2015 IPMC/IMC §307.6, which requires attic and crawlspace condensate pumps to shut down the equipment they serve on failure.¹⁹

The IRC requires secondary drains or auxiliary pans wherever overflow would damage building components.⁵⁶ The proposal included pans, pumps, and switches, which is correct practice, and the manufacturers require them.¹⁸ The point is not that the design was careless — it is that all of that protective apparatus exists because the design puts a water-producing appliance above a historic plaster ceiling, and none of it was necessary before the appliance was proposed.

Two more facts belong in the recurring column. Public-adjuster guidance notes that when a condensate pump fails, “water accumulates in the pan and overflows onto whatever surface is below,” that the most common denial basis for such a claim is neglect, and that “the HVAC unit, drain pan, condensate pump, or drain line that failed is almost never covered under a standard homeowners policy.”²⁰ That is one respect in which a paid maintenance agreement has genuine defensible value — the documented service history is the defense against a neglect denial. It is also, precisely, the mechanism by which the equipment makes the agreement necessary.

And on service life, the fair statement is that nobody knows. The commonly cited industry figure is 5 to 10 years for a whole-house dehumidifier.¹² The ASHRAE service-life table that everyone quotes has no line item for dehumidifiers at all, and ASHRAE's own technical FAQ cautions that the table “is based on a limited survey conducted in 1978” summarizing “only 68 survey participants and is, therefore, of questionable statistical validity.”⁵⁷ No located source gives a service life for a dehumidifier run above its rated ambient temperature. Shorter would be the reasonable expectation, but that is an inference and it is labeled as one.