Buried Ducts Risk Condensation in Humid Climates

15 Comments Read/write comments

Condensation on the insulation jacket of a duct running through a crawl space

Water vapor from the air condenses on air conditioning ducts in humid climates. It's as normal as poorly insulated bonus rooms making occupants uncomfortable or cigarettes causing lung cancer. Condensation on ducts is most common in crawl spaces and basements, where the air is more likely to have a higher dew point. But it also happens in conditioned space and occasionally even in attics. That's why you have to be careful about burying ducts in attic insulation in humid climates, especially in coastal areas. Since the 2018 International Residential Code (IRC) and International Energy Conservation Code (IECC) specify how to get away with burying ducts in humid climates, let's take a closer look.

The problem with buried ducts

The potential for duct condensation is a humid climate, cooling season problem. My friends out west regularly bury their ducts in air permeable insulation (e.g., fiberglass or cellulose) and don't worry about condensation. The air in the ducts may be as low as 55° F, but their summer dew points are usually below 50° F. There's just not enough water vapor to condense.

In humid climates, that's not the case. The photo above shows condensation on a duct in a crawl space in Hilton Head, South Carolina. Outdoor dew points can hit 80° F in the summer there and average in the mid 70s. With 55° F air moving through supply ducts, you've got to have well-installed insulation on a duct with significant R-value to keep the surface temperature above the dew point.

When that supply duct is in a vented crawl space, it's really hard to keep that surface temperature above the dew point because the temperature down there is already close to the dew point. When you have a supply duct above the ceiling insulation in a vented attic, however, the surrounding air temperature is higher than ambient in summer. Thus, the surface temperature of that duct most likely will be above dew point. That means no condensation.

But if you bury that supply duct in fiberglass or cellulose insulation, which allow attic air to reach the duct's surface, all bets are off. As you pile more and more insulation on top of the duct, the surface temperature of the insulation jacket (a vapor barrier) drops.

Since attic air permeates into the ceiling insulation, the dew point within is presumably the same as in the attic space above the insulation. Further, if the attic is vented, the dew point of that attic air should be about the same as the outdoor air. Right? The result is a duct surface temperature that's below the dew point of the air in the ceiling insulation, and that means condensation.

Water spot on a ceiling from condensation on the duct above

As it turns out, ducts buried in ceiling insulation in humid climates do sometimes have condensation problems. If you live in a humid climate, especially locations near the Gulf of Mexico or Atlantic Ocean, you may now have the reason for those mysterious water spots on your ceiling. Maybe. They could also be from roof leaks.

What the research says about buried ducts

Well, that sure sounds reasonable, but is it confirmed by what actually happens and by research? A Building America paper from 2016 titled, Compact Buried Ducts in a Hot-Humid Climate House (pdf) has an answer for us. (Notice I used "an answer" there, not "the answer.") Home Innovation Research Labs did a study on a buried duct system in a home in Lady's Island, South Carolina (a bit north of Hilton Head). One of the main questions they wanted to answer was:

What is the minimum level of duct insulation to prevent condensation at the outer jacket of buried ducts in hot-humid and mixed-humid climates?

So they set up this model home with a buried duct system and installed a lot of sensors. The ducts had the standard R-8 insulation you find on attic ducts but then were buried in R-30 ceiling insulation. Here's what they found:

  • No condensation on the ducts
  • Colder air delivered to the house
  • Simulated simple payback of 3.1 years & simple return on investment of 32%

Another really interesting thing they found was that the assumption I mentioned in the last section doesn't seem to be warranted. Remember when I said that the dew point within the attic insulation is presumed to be the same as the dew point in the attic above the insulation? Well, that's not what they found. The graph below shows two things: (1) The dew goes through daily cycles (similar to what I found in a spray foam attic in Atlanta), and (2) the dew point peaks decrease as you descend into the ceiling insulation and get closer to the drywall. This is good to know.

Dew point temperatures above, below, and with attic insulation

Now, this is all good stuff, but it's just a start. For one thing, the report is based on only one month's data (August 2015). For another, they set this house up specifically as a test house and thus made sure to get as close as they could to their design conditions. (They did miss on duct leakage, though. Their goal was 1 cfm25/100 square feet of conditioned floor area and they hit 4 cfm25/100 square feet.)

It may well turn out that R-8 duct insulation is good enough for general use, but I'm not convinced yet. Neither, as it turns out, is the International Code Council. In my next article, I'll explain the new buried duct requirements that make their debut in the 2018 IRC and IECC. Here's one little teaser, though: The codes decided R-8 was good enough for most of the country but some climate zones require more.

Read part 2:

Buried Ducts Allowed in 2018 Building Code


Related Articles

A Line in the Sand — The Dew Point Duct Duel

7 Tips & Tricks for Blower Door and Duct Leakage Testing

Humidity in a Spray Foam Attic

Home Performance Flaw #212: Ducts Placed High in the Attic



Both of my parents were smokers. Both died of lung cancer.


NOTE: Comments are moderated. Your comment will not appear below until approved.


Feb 14 2017 - 7:31am

This problem is not limited to the ducts alone. I worked with another builder several years ago that built a home in Wilmington NC that was directly on a creek near the Intracoastal Waterway. He was using a vented crawlspace with the air handler in the crawlspace along with all of the ducting. In August is was like a tropical rain forest under there. Not only were the handler and all of ducts dripping water as shown in your photo, but the entire underside of floor above the insulation was dripping like you had sprayed it with a garden hose. With the AC on, the floor above was very cool and way below dew point. Mold had started already. We had to remove all of the insulation and seal the crawl space. Then we brought in a mold remediation company to treat the underside of the floor and the joists. The owner also had us install several dehumidifiers to ensure there would be no further problems with humidity. It was extremely fortunate that we caught it when we did before the owner had moved in. It is still amazing to me how fast we are moving away from vented attics and crawl spaces in humid climates.

Feb 14 2017 - 9:25am

I've seen things like that, too, Thomas. Vented crawl spaces in coastal climates are difficult beasts to control.

Feb 14 2017 - 7:48am

I'm glad you bring this up, Allison. Plain old duct insulation still makes me nervous. Even Cleveland has relatively high dew points, we are usually in the mid-60s in summer with plenty of spikes into the low-70s. I know the initial research on deeply buried ducts looks ok, but I share your trepidation.

We did one project in 2015 where we installed a new duct system, part of it was in the attic. We coated it with 1-2" of closed cell, then buried it under cellulose. No issues so far. http://energysmartohio.com/case_studies/1900-net-zero-ready/

Feb 14 2017 - 9:35am

I didn't bring up the topic of buried, encapsulated ducts, Nate, but I think that's a safer way to do this. Building America has done work on that as well.

Feb 14 2017 - 8:59am

Couldn't the air inside the ducts also play a part in the condensation issues. Running systems at lower air temps can also, I believe, cause condensation issues on the surface of the duct insulation. From poor design, set-up of new systems or an attempt to squeeze more humidity out of the conditioned space wouldn't the lower temps contribute to condensation where it might not otherwise happen?

Feb 14 2017 - 9:37am

Bob, yes, if the supply air temperatures are lower than they should be, that will increase the likelihood of condensation on the ducts. Typically, supply air for an air conditioner should be between 55° F and 60° F to cool the house. If the house has a restrictive duct system with low air flow (Many do!), that causes the air temperature to be lower. Same with a dirty filter. Fix the air flow problem and get the temperature in the right range and condensation problems sometimes disappear.

Feb 14 2017 - 2:19pm

For many energy efficient homes with a perfectly sized AC or HP, I still recommend supplemental (stand-alone) dehumidification. I am getting better at predicting which homes will have little AC run time and thus get into some high humidity situations. But I get a little push back from those who do not agree with the expense of a dehumidifier and feel a variable speed AC or HP will take care of the humidity since they have special dehumidification modes. Slowing the fan, coordinating compressor speed, monitoring output temps to keep the coil near but never freezing and when you read the brochure... wow! Smart dehumidification. I don't like it. In a home that is already too moist, you now have very cold air temps, the supply registers are rusting and thanks to this blog, you now know what is going on under the covers. (Insulation covers, that is.)

Feb 16 2017 - 3:20pm

A properly designed system with variable capacity equipment can improve humidity control, but I agree this stuff is oversold. You can still have a humidity load when there's no sensible cooling load. The best way to address that is to seal the envelope and avoid over-ventilating. In most cases, supplemental dehumidification is an expensive band-aid.

Feb 14 2017 - 9:06am

I have measured humidity within a vented attic, and been surprised to find it reported significantly lower dew point than outdoors. Maybe 5 degrees F. Looking for a theory why.

Roof pitch is steep, leading to greater volume inside the attic. There is radiant barrier and 1:125 passive ventilation, in keeping with some traditional wisdom. It seems to work very well in terms of energy efficiency. Could the wood, especially warm/hot roof decking, be absorbing enough moisture to account for this? Have not ruled out measurement error, but don't think so.

Thank you.

Feb 14 2017 - 9:40am

Mark, do your data show the attic dew point staying below the outdoor point all through the day? Does it cycle as in the graph shown in this article? What kind of insulation do you have?

Yes, a lot of materials in the attic are hygroscopic, which means they can adsorb water vapor. Wood definitely is. Cellulose insulation is, too. What the data above and the data from the spray foam attic article I wrote last year both show is a daily cycling of dew point indicating adsorption and release of water vapor. I imagine your attic does that, too, but it would be nice to have confirmation if you have the data.

Feb 14 2017 - 10:55am

Yes the data shows dew point lower than ambient all through the day. The insulation is fiberglass on the floor and not much of it -- was earlier on my agenda to improve that, but other improvements have been sufficiently successful that may not get done.

Suspect the daily solar cycle may result in some kind of engine which moves humidity.

If I were ever to build a new house, would do several things differently. Ductwork inside would be good, of course. Sealing of the envelope, including the ceiling-attic interface. The ventilated attic design has many decades of experience and out of conservatism I would try to optimize that -- so much research is yet to be finished with sealed and conditioned attic in a hot-humid climate. Thank you.

Feb 14 2017 - 10:57pm

I would add supply duct leakage into your attic dew point trending as a large factor, provided your supply ducts are in the attic. That and a less than airtight plane between interior ceiling and attic floor. Average Texas stick built home has plenty of thermal bypasses into attics that ar concurrent with notable air leakage between house interior and attic spaces.

Feb 14 2017 - 1:13pm

Allison, We have been intentionally insulating ducts since we changed to Fiberglass way back around 1979. Duct Coverage up to 36" as measured from attic floor is routine for us as we have zero problems (reported) in the (ALOT) Arkansas, Louisiana, Oklahoma, Texas region.

Feb 14 2017 - 2:42pm

1) The 1 CFM25/100. Was it post construction including an air handler or duct only with the supply and return added? AHRI allows an air handler case to leak 60 cfm25.

2) I have had improperly sloped B-Vent cause condensation issues and interesting IR images of the ceiling.

Feb 16 2017 - 3:36pm

Whenever there's risk of condensation on supply ducts, the end-to-end integrity of the vapor barrier (typically a foil wrapper on the insulation) is just as important as the R-value.

When burying ducts in a humid climate, condensation risk is highest if the duct is in contact with the ceiling drywall. I believe R-8 would be sufficient in humid climates if supply lines are supported just above the joist level. But this is difficult to achieve with run-outs parallel to the joists.

As an aside, supply ducts inside the thermal boundary (especially encapsulated attics/crawls) are still at high risk for condensation in humid climates. It all depends on supply temperature and (peak) ambient humidity. These factors can be mitigated by design, but the safest approach is to minimally insulate interior supply ducts.

Interestingly, neither the IMC or IRC require insulation on interior ducts, but a senior code official with N.Carolina Dept of Insurance once told me, if duct sweats and causes damage, the HVAC contractor owns it.

Add new comment