Can lighting roof moisture problem
This winter I bought a 2 floor condo on the top floor of a 3 story brick building built in 2012 in Chicago, with hot humid summers and cold winters. During the hottest days, when it reaches 90+ degrees, I have found water around the recessed can lights on the top floor, which is directly up against the flat roof. I have also noticed a water stain on the drywall about 6 inches away from the ceiling, which I assume must be a wooden rafter.
The 2nd floor is actually a dryvit enclosed 600 sq foot “penthouse” on the 4th floor with its own flat roof. This roof has 2 vents on each end and the cans on each end do not show the same moisture problem. There is a return HVAC duct on one end of the penthouse, but the can lighting nearby is least affected by the problem. The roof is a single ply rolled product (I think modified bitumen but I am not sure—it is gray in color)
Also of note is the AC system is single stage single zone, so this 2nd floor room never gets below 78 degrees on these hot days, which makes it surprising that anything could condensate at all.
Here is what I have tried:
1) Had 2 professional roofers inspect the roof for leaks. None were found.
2) Pull the IC-AT rated can lights out from below. Remove the damp fiberglass insulation found inside, and replace with new insulation. Caulk seal between can and drywall, add LED retrofit kit which has less leaks than the previous LED bulbs.
3) Placed standing dehumidifier in the room, brought internal room humidity down to 45% (from 55%)
Any ideas on what the next step would be? I am considering the following:
1) Add additional roof vent in the middle. My thought for this is that if the ends of the roof seem to be OK, maybe this will help. However, I logically think this would just introduce more humid air to the roof space and make the problem worse.
2) Paint the roof with an elastomeric white roof coating such as https://www.homedepot.com/p/Henry-4-75-Gal-687-Enviro-White-Roof-Coating-HE687406/202091034. If this problem only happens on the hottest days, maybe the solution is to lower the temperature in the roof. This could have another benefit of cooling the room (which is of course always hot due to the single zone single stage AC over 2 floors..but that is a different issue entirely) I’m also not sure how much of a difference this would make anyway (if the claims are real).
3) Pull out all the can lights entirely and replace with fixtures, re-insulate (access is only from below). There are also overhead speakers built into the ceiling that I assume would need to be removed as well in this option.
4) Remove the vents, lay down rigid foam insulation on top, and put a new membrane on top. I think this is the nuclear option – just give up on the current roof and try again with an unvented one. I imagine this would be the most expensive route as well
Any other ideas that I am not thinking of? One thing I am surprised about is why the previous owners who lived here for 5 years claim they have never seen this before. While it is possible they are lying, I don’t see any signs of drywall repair on the ceiling.
Really appreciate the help, I am at a loss for what to do here.
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
Can you tell us your name?
Here is a good article that should be relevant to your situation: https://www.greenbuildingadvisor.com/blogs/dept/musings/insulating-low-slope-residential-roofs
I suspect the big issue is having can lights and speakers in a low-slope ventilated room. So you need to remove them and make the drywall layer as air tight as possible. You also may need to take other action to improve the performance of your low-slope roof, but I think posters here will want to know more about how it is constructed.
Thanks Steve - My name is Bryan Leavitt, a homeowner trying to figure out what to do about this problem. I've talked to a couple of general contractors and neither seems to know what to do here.
I agree with you that the can lights + speakers are a bad idea -- If I had designed the home they would certainly not be there. But given that they are already there (and that the can lights and speakers on each end of the roof are not leaking at all), is there any solution that would keep them?
Bryan,
Like you, I am a homeowner. Let me ask you this. Will your coop or HOA allow you to re-roof the "penthouse?" You could then seal the roof "attic" and install R-48 insulation above the sheathing (along with a decking layer and a new membrane.
As for the existing ceiling and roof, I haven't a clue but might guess that the roof vents are allowing enough air movement between the conditioned space and outdoors to dry whatever condensation is developing in those areas.
Yes, if that really was the only / best solution to this I could certainly convince the HOA to allow it, but nobody would be happy about it. Especially given the roof & building is only 5 years old.
If those roof vents are allowing enough air movement, maybe then the solution is to add more? That was my proposed solution #1. I am concerned that it will actually just introduce more humid air and make it worse, however.
By the way, the final option is to simply ignore the problem, but I have no idea how bad the problem is behind the drywall, or if doing nothing will cause the fix to be even more expensive down the road.
Bryan,
Just to clarify, I don't think adding more vents to a poorly sealed roof and ceiling assembly is a solution. If you do nothing, you will probably end up with a rotten roof. But lets see what the construction experts suggest.
Note that it's possible for air to enter from outside, deposit moisture and then return to the outside - no amount of interior side air sealing will prevent this.
As Steve says, more details about the complete ceiling/roof design/layers would be helpful. This is a flat roof vented attic with fiberglass insulation?
Any AC supply ducts?
Bryan,
Here are the facts:
1. Every house needs an airtight ceiling on the top floor. Your recessed can lights are preventing the ceiling from acting an air air barrier. This is a fundamental flaw.
2. Installing white membrane roofing to make the roofing colder will almost certainly make the situation worse. Because of night sky radiation, white roofing is associated with moisture problems -- more so than black roofing. For more information on this phenomenon, see "Insulating Low-Slope Residential Roofs."
3. Solving moisture problems in a low-slope roof with "improved" ventilation is a fool's errand. Most so-called "vented" low-slope rules lack enough of an air space above the insulation to be effective, and adding more vents makes the situation worse.
4. The best solution is to install an adequately thick layer of rigid foam above the roof sheathing, followed by new roofing, and to seal the vents.
Thanks for the reply Martin.
What do you think of the option of doing nothing at all? How could I tell if the problem is serious enough to matter vs just ignore? (As you can tell I'm trying to avoid spending serious money on replacing a 5 year old roof)
In reply to #2 -- I noted that the study in that article is from Arizona, where I understand winters can be below freezing at night but then 70 degrees during the day. Does the same recommendation hold true in Chicago (hot humid summer, cold dry winter) ?
In reply to #3 -- What of the can lights by the current ventilation that seem fine? Do you think the installers simply did a better job creating a large enough air space there?
Bryan
"Because of night sky radiation, white roofing is associated with moisture problems -- more so than black roofing"
It's more the lack of daytime solar gain than the nighttime radiation that causes moisture problems with white roofs. Black roofs radiate in the deep infra-red at about the same as white roofs, but white roofs reflect a lot more of the DAYTIME solar spectrum than black roofs.
A white roof reduces the peak roof temperatures in Chicago in much the same way as it does in Arizona or Florida, and will not improve a an air conditioning condensation problem- it may make it (slightly) worse.
In Chicage/climate zone 5A installing at least 40% of the total R of impermeable insulation (such as closed cell spray polyurethane foam) on the underside of the roof deck is one way of getting there without insulating above the roof deck.
Bryan,
As I just noted on another Q&A thread, condensation occurs when warm, humid air contacts a cold surface.
Whenever there is summertime condensation, the first mystery to be solved is: What is the cold surface? Candidates include air conditioning ducts and roof materials cooled by night sky radiation.
I'm not sure what's going on in your roof assembly. But if you have already noticed damp fiberglass insulation, you seem to have an ongoing problem.
Since you have had the roof inspected and the roof is not leaking, your problem is caused by condensation.
I also note that it is "currently" occurring only during the summer when you are conditioning the living space.
I have a question about where exactly you see any moisture. Is there moisture on the inside of the can? Was the wet insulation in the ceiling space outside the can?
I am going to answer your question based on the answers to those questions being no and yes -- your moisture was in the ceiling, not inside the light.
First some condensation theory. It is not about hot meeting cold as many would explain it. Condensation occurs when warm air is cooled to below the dew point temperature. Air can hold moisture. The percent of moisture air can hold depends on its temperature. The cooler the air the less moisture it can hold. When air has a certain percentage amount of moisture in it and the air cools below the dew point temperature (the temperature of air that can hold that percentage of moisture) the moisture condenses - or forms dew or water drops. The condensation does not 'really' occur on the cold surface, it occurs 'in' the cooler air. It is just most often seen by us when that moisture forms a cloud, is enough that it falls from the sky, or collects on a surface such as the inside of a window pane in the winter or a drinking glass or your light can in the summer.
So your problem is - there is moisture in the humid air (from the humid Chicago summer) that is inside the ceiling space and that moist air is being cooled below the dew point. The air in the cool air-conditioned space is keeping the light can cool and it, in turn, is cooling the air near it.
* Note: the opposite can occur in the winter when the air in the ceiling is cold causing the can to be cold if there is too much moisture in the room. The moisture will then be seen inside the can.
- Caulking the can does nothing since the moist air is still near the cool can and is still being cooled below the dew point.
- Replacing the fiberglass insulation will not stop the problem since the new fiberglass will still allow moist air to get near the cool can.
- Sealing the vents will not help since the moisture will penetrate the construction and still be in the ceiling space.
- Dehumidifying the room will not help since the problem is the moisture in the ceiling.
- Changing to an LED, a lower wattage (cooler) light bulb, does not remove the moisture, and can actually make the problem worst since the can could be colder.
What you need to do is keep the moist air away from the cold can in the summer. Or raise the temperature of the air in the room.
(Or in winter keep the cold outside air away from the can to keep the can warm. This is done in winter by having the can on the living space side of the insulation and not having too moist of an interior.)
Increasing ventilation in an attic or ceiling space 'can' help because it keeps the air moving so the air is near the cold can less time and doesn't cool off as much. BUT this is problematic with no attic, a low slope which will not provide the airflow needed, and an insulated ceiling/roof space further impeeding air flow.
You could put an insulating 'box' around the can that is impervious to air and/or moisture. The fiberglass is not. This is why Martin mentions the importance of an airtight ceiling. BUT this is also problematic since recessed lights are susceptible to heat buildup and 1) they are designed in such a way as to limit the heat, 2) there are building practices and regulations about certain combustible materials being too near a can, and 3) why it is so important to never put a higher wattage bulb in a light, especially a recessed light, than specified. Not to be an alarmist, but it could be a fire hazard.
Because of all this, recessed lights in a flat roof, especially a low slope one, are problematic where there are cold winters with possible moist air inside and even more so where there are hot humid summers needing air conditioning. It is not an issue where they are in a ceiling between floors (surrounded by conditioned space) and is not as big an issue where there is an attic above providing more air flow and insulation options.
I hope this all helps you evaluate your options. But here is my recommendation.
Remove the recessed lights. Install a regular ceiling electric box with a good insulation and a moisture barrier above it. Plaster up the ceiling. Then install a nice looking surface mount fixture that will be completely inside the air-conditioned and heated space -- and that will not be part of the division between inside and outside. If you like the idea of recessed lights because they point the light you can use a museum-style directional light or track lighting.
Thanks Bob for your thorough reply.
I attached a photo of the can lights on the very first day I noticed the problem. The insulation was on the ceiling space behind the can (the cans are rated IC-AT, so the insulation was firmly up against the walls of the can)
I also attached a photo of the can light after I replaced with an LED airtight retrofit, caulked around the edges and added as much insulation as reasonable above and around the can. At this point I have not noticed the huge "flooding" of water from the first photo, only minor amounts of pooling around the edges of the retrofit, none felt through the drywall.
The first thing I did when moving into this home was replace all the incandescent bulbs with LEDs. Maybe that is why the previous owners never saw this problem: their bulbs were so hot that it never allowed for condensation. Maybe a simple solution is to go back to the less efficient bulbs ?
I posed the question above but would like your take -- at what point is this a "serious issue that must be addressed" vs something that can wait? If I do nothing, what is the worst that can happen here, assuming the amount of visible water is hardly noticeable, and not through the drywall?
I would hazard a guess that the worst that could happen is you will get mold and mildew in the ceiling and if it gets wet enough pieces of plaster or sheetrock could sag or fall in the area of the light. It would require redoing the ceiling in that area or for mold opening a larger section to do mold abatement and then rebuild. Taking a chance it is possible to get some types of mold (some types of black mold especially) that are a health hazard, and for some allergic people, any mold could be an issue.
An even worst scenario would be enough dampness that rot might set into some of the lumber but I suspect you would see other problems first.
I am not a carpenter, mechanical, mold, licensed engineer, or whatever person, I am just a former electrical person and now physics teacher so this is, of course, all imho.
> The air in the cool air-conditioned space is keeping the light can cool and it, in turn, is cooling the air near it.
Bryan: You say the interior is 78F. As far as I know, Chicago outdoor dew points are pretty much never much above that, so you shouldn't have condensation. You could put a wireless sensor above the can to see if something is causing a higher then expected dew point.
Conceivably, for the summer, you could do what is often done to keep crawlspaces dry - block other outside vents and install a small exhaust fan sufficient to pull a little bit of interior air into the above ceiling space. This will dry the space and prevent condensation. Unlike a typical attic fan, the cfm is not enough to cause an energy problem.
You can cause a similar reversal of summer stack effect airflow by pressurizing the interior space.
Jon,
Yes, during the hot days when it is 90+ outside, the interior of this space (even with the AC blasting) cannot get below 78. This issue is likely due to poor HVAC design. The 1st floor is cold while the top floor cannot keep up.
I imagine the attic space can get seriously hot when the sun is beating on it all day -- could this create a local dew point high enough for the condensation? If not maybe this problem truly isn't condensation at all as you are suggesting.
The house currently has an exhaust-only ventilation system, (there are 3 bathroom exhaust fans, and a kitchen hood exhaust). Could this be the problem? Would adding a supply side ventilation intake and thus applying positive internal pressure to the entire home help?
Martin, and others, have suggested that ventilation to solve this issue is a fool's errand because it simply introduces more humid air into the space.
Additional heat doesn't effect dew point - unless it causes some water source to evaporate. Moist air from outside and dry air from inside are very different - the latter lowers the dew point.
Exhaust only ventilation increases infiltration - yes, this makes summer moisture problems (and temp stratification) worse. Give positive pressure (verified with measurements) a try.
So it sounds like you think it is not condensation? What else could it be? It only shows up on very hot days.
Is there an easy way to test a positive pressure system before installing something permanent ? Are there any other side effects to this? Maybe int he winter months?
I think it's condensation, but it's not clear to me how the dew point gets as high as it must be to produce significant condensation against a 78+F degree surface. Condensation from night-time cooling being re-evaporated (causing localized increased moisture)? This may be consistent with no problems near vents.
A fan in a window (on the lower floor if floors are open to each other) can produce positive pressure.
In Winter, a slight negative building pressure is beneficial.
Bryan,
As Jon wrote, the mechanism for the condensation is a little mysterious. It may be a combination of night sky radiation (which cools the roofing and the roof sheathing) and building depressurization (which pulls humid outdoor air into the roof assembly through cracks) -- but this is a hypothesis. And I don't have a lot of confidence in this hypothesis.
The water is only seen around the afternoon to late afternoon, when I assume the sun has had a chance to really heat the roof up.
Jon I think is on to something...the stack effect has to be very strong during these times -- 72 degrees on the lower floor and over 100 in the attic air space. Exhaust-only ventilation system compounds to this effect. The "exhaust" vents must be acting as intake vents instead.
If I added a central fan integrated supply ventilation (outdoor duct to hvac return, motorized damper + aircycler control) could this be calibrated to provide enough pressure to counter the effect just enough (and then dial it back in the winter to allow for more negative pressure) ? This article suggests maybe not: https://www.greenbuildingadvisor.com/blogs/dept/musings/designing-good-ventilation-system "The relatively small airflow of most supply-only ventilation systems (75 cfm to 150 cfm) will have little effect on this situation other than to shift the neutral pressure plane down slightly, in all but the very tightest of homes."
On the other hand, maybe all I need is a little bit in order to get back to even -- It might be just barely sitting on the "edge" of the exhausts acting as intakes vs out.
I did some more investigation. Pulled out a recessed can, pushed aside the fiberglass. See the photos. It looks like there is about 12 inches of space between the drywall and the OSB roof sheathing. 9 inches of which is fiberglass, and a couple inches of air. There looks to be I joists running up against the OSB, suggesting that the vents are further ineffective (even if 2 inches of air space were enough). I also notice some black areas which may be mold growth, but I have no idea, it could just be dirt.
I am thinking of following this plan:
1) Remove both vents
2) Add 3 inches of poly iso, then 3 inches of XPS on top to keep it warm during the winter (not sure if this is an appropriate amount or not. I am in climate zone 5, chicago)
3) Cover plywood
4) Cover water & ice shield
5) Cover modified bitumen
How does this sound? Do I need to be concerned that the insulation is not in direct contact with the roof from below? How do I make sure that this can sufficiently dry inward, given that this is currently a "vented" roof design?
Just so I understand this better, is the main mechanism for reducing condensation in this case simply the closing of the vents? Because the attic air space should be cooler in the summer after applying this rigid foam insulation -- won't that be counter productive and actually cause additional condensation?
Thanks
I'm not sure why XPS + polyiso rather than just polyiso? XPS performance drops to R4.2/inch over time as it's high global warming potential HFC blowing agents (more than 1000x CO2) dissipate out. Even well aged 2lb roofing polyiso won't be sufficiently derated for temperature in this stackup to be a condensation problem at 6".
At 6" polyiso would meet code-min performance on it's own, even without the fluff in the rafter bays. But unless you're pulling the fluff out, it's better to fill them with mid-density fluff (for air retardency) and put sufficient foam up top for dew point control (40% of the total R, minimum for Chicago's climate). At 12.0" if you top off the rafter bays with cellulose (drilling from the roof deck side) you'll end up at about R45 in the rafter bays, and would need R30 up top. You could get there with 5" of foil-faced polyiso, or 6" of fiber faced 2lb roofing iso (which is widely available for cheap on the reclaimed foam market.)
By keeping the roof deck warmer in winter the roof deck doesn't accumulate a significant moisture load over the winter. In summer the cooler roof deck isn't a problem, because it's average temp will still be warmer than the dew point of your indoor air, and you will have blocked off access to the much higher dew point outdoor air. I suspect you're dealing with the consequences of inadequate venting and air leaks from the interior, which put an inordinate amount of moisture into the roof deck in winter when it was much colder than the dew point of your indoor air, and don't have enough ventilation flow now to remove that moisture in a timely (seasonal) manner.
Also, the air & water tightness of the structural roof deck is far more important than whatever you put above the foam. Put the Ice & Water Shield on the structural roof deck, and use a more appropriate underlayment for the modified bitumen you're putting on the secondary deck above the foam.
Interesting, so you think the moisture is built up from the winter humidity (I keep the home at 45% humidity with a whole home humidifier in the winter), and took this long to release into the drywall? I never noticed anything during the winter, and I did not think it could take 4-5 months for the water to "release". I guess that might explain why the flood of water was on the first 90+ degree day, and has since gotten less and less of a problem, now just a small amount.
I was thinking XPS on top of polyiso to keep it warm based on your comment following this article:
https://www.greenbuildingadvisor.com/articles/dept/musings/how-install-rigid-foam-top-roof-sheathing
Do you think I shouldn't worry about this and just use polyiso alone?
Is adding additional dense pack cellulose necessary? Why does this article recommend against it? https://buildingscience.com/documents/insights/bsi-043-dont-be-dense
What if I did nothing and left the fiberglass batts as shown in the photos above, while completing the 6 inches polyiso above? What is the risk?
RE: ice & water shield on structural roof deck: I was thinking of just doing this directly on top of the existing modified bitumen roof, without removal. In this case, will the modified bitumen act well enough as an ice and water shield to not need an additional layer ? If so, maybe I do not need an ice & water shield at all?
Bryan,
We've just learned that you "keep the home at 45% humidity with a whole-home humidifier in the winter."
Here's some advice: if you have a wet-ceiling problem that seems to involve condensation, the first step is to disable the humidifier and hit it with a sledgehammer. Don't even worry about the whole winter-versus-summer issue. Just do it. It will feel good, and it will help save your building from damage. No home needs a humidifier. (Some patients with specific medical problems may need a humidifier in the bedroom, but if you don't have a specific diagnosis that require a humidifier, destroy it.)
If you go ahead with your plan to install 6 inches of rigid foam above your roof sheathing, all your problems will stop -- as long as your roofing contractor knows how to seal all the ventilation openings, and as long as the air sealing work is verified by a blower door test.
(By the way, you seem to have misinterpreted Joe Lstiburek's article on dense-packed cellulose. The problematic roof assemblies he talks about in that article were unvented roof assemblies without any exterior rigid foam. If a roof assembly has an adequately thick layer of exterior rigid foam, the roof sheathing stays above the dew point, and there won't be any problems -- even if the rafter cavities are filled with dense-packed cellulose.)
Gotcha, I misunderstood the article on dense pack.
As far as 45% in the winter -- it gets very uncomfortably dry here in Chicago in the winter..it can get into the teens for RH. I think low RH is just as much of a problem as high RH -- we live in a home for comfort after all. That said, I think this winter I will lower it to 40%, or even 35%.
I also did not notice any problems during the winter itself, so I thought it was fine at the time. I am surprised that it took this long to show up (because of course the humidifier is not running in the spring / summer)
Ok, so 6 inches rigid foam above. If I left the fiberglass in place with the 2-3 inches of air space and don't add additional cellulose as Dana recommends (out of cost concern) is there a risk of this not working, or do you think the foam is good enough?
An wintertime indoor humidity of even 40% RH @ 70F is considered high in building-science terms, and 45% near the high end of human health range of 30-50% recommended by health professionals. In summer 45% is fine, but in winter hold the line at NO MORE than 35%.
The dew point of 70F/45% RH air is 48F, so any time the roof deck is below 48F the entrained air in the cavity will be at a lower dew point defined by the roof deck temperature (the moisture in the cavity air ends up in the roof deck as adsorb), and there will be a vapor pressure difference drawing moisture from the room air into the cavity via diffusion.
In Chicago the average daily outdoor temp is below 48F for about 160 days in a row, from the beginning of November into the first weeks of April. (The average temp of the roof deck is lower still during much of that period due to nighttime radiation into the sky).
https://weatherspark.com/m/14091/11/Average-Weather-in-November-in-Chicago-Illinois-United-States#Sections-Temperature
https://weatherspark.com/m/14091/3/Average-Weather-in-March-in-Chicago-Illinois-United-States#Sections-Temperature
The dew point of 70F/35% RH air is 41F, and the outdoor temp in Chicago is below that line for "only" 120 days, from about Thanksgiving into mid-March. Not only is that a shorter moisture-absorbing season, it's at a lower drive, since the difference in roof deck temp and indoor dew point is less.
The foam without completely filling the cavities is "good enough" from a moisture point of view, but filling the cavities with cellulose by drilling from above is dirt compared to the cost of the foam, and "worth it". Some would insist on dense packing the cellulose (which adds to the expense), but in this stackup there wouldn't be much additional benefit to gain from dense packing, since the range of seasonal temperatures at the roof deck will be small enough to limit the moisture cycling that causes open-blown cellulose or wall cavity cellulose to settle.