Condensation on duct insulation: ceiling void dew point, insulation surface temperature, and choosing jacket and thickness

The situation

Say that during the rainy season, stains spread across the ceiling tiles of a top-floor office, and when you open an access panel the insulation on the supply duct is wet on the outside and dripping. Inside the duct is cold air straight off the cooling coil. The ceiling void is hot and humid from the roof and from outdoor air leaking in. Two things need checking: what is the dew point of the air in the void, and how warm is the outer face of the insulation with the thickness and jacket that are installed? If the surface is at or below the dew point it condenses. Even slightly above it, with no margin, the spots where the insulation is thinner, such as joints and hangers, get wet first. The conditions below are an example for illustration. A real check uses temperature and humidity measured in that ceiling void and the insulation product's own data.

Given

The outer face of the insulation picks up heat two ways: convection from the surrounding air (h_cv) and radiation from the roof slab and other surfaces around the duct (h_r). The radiative part is roughly proportional to the jacket's emissivity, so a shiny aluminium jacket takes in very little heat by radiation. That difference drives the whole calculation below.

Reading order on the chart

1. Plot the ceiling void point PA (30 °C, 70 %).
2. Move horizontally left from PA (constant humidity ratio) to the saturation curve: the dry-bulb temperature at that point, DP, is the dew point. Air touching the insulation surface cools without changing its humidity ratio.
3. Draw a vertical line at the supply air temperature (14 °C). That is the surface temperature with no insulation, far to the left of DP.
4. Draw a vertical line at the insulation surface temperature from the heat balance. Left of DP means condensation, within 1 K to the right means at risk, and further right is safe.

Step-by-step calculation

1) Dew point in the ceiling void. At 30 °C and 70 %, the humidity ratio is W = 18.80 g/kg, the vapour pressure 2.972 kPa and the dew point t_dp = 23.93 °C. The test is whether the surface stays at or above 24.93 °C (dew point + 1 K).

2) Bare duct. Sheet steel has almost no thermal resistance, so the surface is at the supply air temperature, 14.0 °C. That is 9.9 K below the dew point, and water runs off the whole duct.

3) 25 mm glass wool with new aluminium foil (ε 0.05). Both the conductivity and the surface coefficients depend on the surface temperature T_s, so the balance is solved iteratively. At the converged values the mean temperature is T_m = (14 + 24.23)/2 = 19.11 °C, k = 0.0354 W/m·K and the insulation resistance R_ins = t/k = 0.025 / 0.0354 = 0.707 m²K/W. At the surface, natural convection gives h_cv = 2.20 and radiation h_r = 0.31, for a total h_se = 2.51 W/m²K (h_cv = 1.42·(ΔT/H)^0.25, the laminar formula for a vertical plate H = 1 m high, switching to the turbulent 1.31·ΔT^1/3 above ΔT = 10 K) and a surface resistance R_se = 1/h_se = 0.399 m²K/W.

The heat reaching the surface equals the heat passing through the insulation: (T_a − T_s)/R_se = (T_s − T_f)/R_ins. Solving, T_s = (T_a·R_ins + T_f·R_se)/(R_ins + R_se) = (30 × 0.707 + 14 × 0.399)/(0.707 + 0.399) = 24.23 °C, with a heat gain q = 14.5 W/m². The surface is 0.30 K above the dew point: no droplets, but short of the 1 K margin. Verdict: at risk.

4) Same 25 mm, painted or canvas jacket (ε 0.9). Radiation rises to h_r = 5.61 W/m²K, so h_se = 7.41 and R_se = 0.135 m²K/W. The surface is 27.42 °C, 3.49 K above the dew point: safe. The trade-off is a higher heat gain, 19.1 W/m², than with foil. If the aluminium oxidises or collects dust and its emissivity becomes 0.13, the surface is 24.80 °C (margin 0.88 K).

5) Minimum thickness. Bisecting for the thinnest insulation that keeps surface − dew point ≥ 1 K gives 31.3 mm with new foil, 26.0 mm with oxidised aluminium and 9.9 mm with a painted jacket. Rounded up to standard thicknesses: 40 mm, 30 mm and 10 mm.

What the result means in practice

By thickness (ceiling void 30 °C / 70 %, dew point 23.93 °C, heat gain with the foil jacket):

A foil jacket helps cut heat gain but works against you on surface condensation. It receives so little radiation that the surface gets pulled toward the cold inside temperature. On condensation alone, this example needs 10 mm under a painted jacket but 40 mm under new foil. If the jacket can't be changed, add thickness. If the duct is already installed, look first for a way to bring down the humidity in the void.

The minimum thickness for the condensation criterion changes with the humidity of the void like this (30 °C, 1 K margin):

Going from 50 % to 80 % RH takes the minimum under new foil from 5.7 mm to 76.9 mm. While the dew point sits well below the void air temperature, a few millimetres are enough, but as the air approaches saturation and the dew point closes in on the air temperature, thickness cannot keep up. If the ceiling void is a return air plenum and holds room air (26 °C / 50 %, dew point 14.8 °C), the same duct with 25 mm and foil has a surface at 21.52 °C, a 6.7 K margin, and needs only 1.9 mm. That is why, when a condensation complaint comes in, you log the void's temperature and humidity for a few days before stripping insulation, and look for where outdoor air gets in (roof and wall penetrations, open access panels) and whether exhaust fans are pulling the void negative.

Condensation the calculation does not see: joints, hangers, vapour barrier

The calculation above is steady state on a flat surface with evenly applied insulation. What gets wet first in practice is where the insulation thins or stops: hangers, straps and supports that crush it to half its thickness, duct flanges and stiffeners, damper bodies and access doors, and joints where the tape has let go. Where the thickness is halved, expect the result from two or three rows higher in the table above.

Fibrous insulation such as glass wool lets air through, so it needs a continuous vapour retarder on the outside. A foil jacket doubles as that retarder. If the joint tape fails, humid air gets into the insulation and condenses on the cold inner face. That water wets the insulation, raises its conductivity and makes the surface condensation worse. Even when the surface check passes, a continuous vapour barrier is half of the fix.

The calculation treats the duct side as a vertical flat surface 1 m high. A shorter side has a higher natural convection coefficient and a warmer surface, and on a duct a few tens of centimetres high the difference can be enough to change the verdict, so the 1 m assumption is the conservative one. The surrounding surfaces that exchange radiation with the jacket (slab, ceiling tiles) are also taken to be at the air temperature. Where the roof slab runs hotter than the air, as it does on a top floor in summer, the surface ends up a little warmer than calculated. The top of the duct, on the other hand, holds a layer of cooled air with weaker convection, and a face close to the slab sees different radiation. The 1 K margin is there to absorb these differences and measurement error.

Common mistakes

• Checking against room conditions. The duct runs through the ceiling void, not the room. In this example 25 mm with foil looks generous against the 26 °C / 50 % room (margin 6.7 K), but in the 30 °C / 70 % void the margin is 0.30 K.
• Assuming a reflective jacket helps with condensation too. Aluminium cuts heat gain but makes the surface colder. At the same 25 mm, a painted jacket sits at 27.4 °C and new foil at 24.2 °C.
• Sizing right at the dew point. A surface a little above the dew point on paper drops below it where the insulation is crushed, at joints and on the top face. Keep a margin and round up to a standard thickness.
• Forgetting a change in supply temperature. Moving to low-temperature supply air, or lowering the coil leaving temperature, drags the surface down with the same insulation. Recheck the duct insulation whenever the supply temperature changes.
• Looking at thickness and ignoring the vapour barrier. Fibrous insulation gets wet from the inside when the barrier is broken. Even if the surface check passes, inspect the taped joints and the sealing at penetrations.

Try this calculation on the site

Insulation surface condensation: 25 mm with foil jacket25 mm with painted jacketOpen the condensation check on the chart

The dew point tool links fill in the 30 °C / 70 % void and set the insulation surface condensation section to a flat surface (duct, equipment), 14 °C supply, 25 mm glass wool and the jacket. They show the surface temperature, margin and verdict, the minimum thickness, a table by thickness and the full derivation. Change the void RH or the jacket to see where the verdict flips. Switching the shape to a pipe gives the same check for chilled water pipe insulation. The chart link draws the void air point with the foil jacket surface as a condensation check. The cooling coil calculation that sets the supply temperature is in summer cooling coil selection and reheat, condensation on walls is in winter exterior wall condensation, and the plotting principles are in how to read a psychrometric chart.

References

• ASHRAE Handbook Fundamentals 2017 (SI), chapter 1 Psychrometrics: humidity ratio, saturation vapour pressure (Hyland-Wexler) and dew point. The psychrometric numbers here were computed with this site's engine, which implements those equations at 101.325 kPa.
• ASHRAE Handbook Fundamentals 2017 (SI), chapter 23 Insulation for Mechanical Systems: preventing surface condensation on cold insulation, vapour retarders, and how jacket emittance affects surface temperature.
• ISO 12241: calculation rules for heat transfer through insulation on equipment and pipework. The site's insulation calculation follows this form (insulation resistance plus a surface coefficient split into radiation and natural convection). For flat surfaces the natural convection coefficient is not the ISO 12241 formula but the simplified air correlation from Holman's Heat Transfer. The equations and coefficients are shown in the derivation in the dew point tool.
• The glass wool conductivity (k₁₀ 0.034 W/m·K) and the jacket emissivities (0.05 · 0.13 · 0.9) are typical design reference values from the site's calculator. Use the product's test certificate and the jacket manufacturer's data in a real design. The 30 °C / 70 % void and 14 °C supply are example values.