Metal roof and condensation
Okay. This is driving me crazy and I can’t seem to find an answer no matter how hard I look. Is there anyone with a background in building science that could help?
Back story. I recently purchased a standing seam metal roof in 24G galvalume panels 1.5 inch seams, to go on a residential house with a “normal” attic. House has drywall on the ceiling and kraft paper beyond that on the attic side, on top of that is 16-24 inches of fibre glass batt insulation (2 layers) and then a normal attic air space before the joists and sheeting. For the roof, (after much research and talking to several contractors) the best idea seemed a recommendation to go with Gaf deck armour on top of existing 1/2 inch OSB seathing then apply the 10mm thick rain screen, then the metal panels–this is what we did. The roofer recommended the rainscreen to provide a channel to have the roof dry in case of condensation. The roof is vented in the attic with ridge vents and at the soffits with soffit vents. The roof was screwed down every 12 inches with clips through the mesh and underlay, and no sealant was used on the screws that were used to screw down the sheathing.
My question, I live in the pacific northwest where it is very humid and cold and rainy, and I woke up several mornings this Autumn to find the TOP surface of the roof literally drenched with condensation, or in one case frost. However, does this mean that the bottom side of the steel panels is also “drenched” with condensation because of the air cavity or the nature of metal? I have no way of seeing into this cavity between the steel panels and the 1/4 inch mesh underlay, but I am dying to know if there is any significant condensation in there? Or does this roof behave like a normal asphalt shingle roof–with condensation on the surface facing the atmosphere only? Does warm/moist air have to migrate from my living space to the underside of the metal panels to cause condensation? Or is this a phenomenon that will happen regardless of how well my living space is sealed when the dewpoint is reached outside and the attic air is hitting the underside of the seathing and working up towards the metal?
PS. I have read the thread on here where it was suggested that a metal roof not be laid on cedar breather (which is essentially what I used) because of oil canning concerns. This info appears to me to definitely be FALSE. If anything, laying the metal panels on the cedar breather naturally crowns the panel and there is very little oil canning at all. It provides a sort of “cushion” for the panels so that any imperfections in the roof sheathing are not noticeable at all in the panel. This was also told to me by the roofer who said that the cedar breather (10mm rainscreen) helps a lot with oil canning. Even with 18 inch wide panels and two very small strengthening ribs running the length of the panel, there is very little (almost none) oil canning on my roof. Just to add to this, several different people worked on my roof, and there was one particular person who laid the panels quite tight in one area and these panels are very slightly more crowned than the other ones–they literally bow a bit when I walk on them and step off–it’s a nice effect. Laying the panels quite tight with a fastner every 12 inches, and using a mesh underlay with a heavier guage of steel (24G) seems to be a good way to go. I also think the screws need to be driven in at a slight angle at the top of the panel (by the hips and ridges) to help with crowning the panel lengthwise. In areas where the screws were driven flat/hard at the hips to secure the panels, some oil canning exists.
In any event, was my roof done right? Will there be a lot of condensation under the metal panels??? It seemed to me that allowing the metal to breathe underneath was a good idea–hence the rainscreen and breathable underlay. Won’t this allow some movement of air and drying toward the attic side and between the mesh and the Gaf underlay? Some have said the mesh won’t allow significant movement of air, but doesn’t air move from low to high pressure? So when the wind picks up or air conditions change, or the metal heats up from the sun, wouldn’t this suck the moisture out from under the panels through ANY available gap no matter how small?
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
Tony,
When skies are clear, and you see moisture on your roofing early in the morning, that's normal. It's called dew. If the weather is cold enough, the phenomenon is called frost. It also happens on your car if you park outdoors. It also happens on your lawn (if you have one).
For more information on this perfectly natural phenomenon, see "Night Sky Radiation."
The amount of condensation on the underside of your roofing is probably less than the amount of condensation on the top of your roofing, for a variety of reasons. (The main reason is that the upper side of the roofing is exposed to more outdoor air and therefore more moisture.)
A small amount of condensation on the underside of roofing is normal. In almost all cases, this moisture evaporates before it causes any problems.
Tony,
You roof will be fine. It's not unusual for there to be enough dew on metal roofs in the PNW that they cause the gutter to flow. But if those roofs are built on purlins and you can see the underside, the amount of condensation there is very small.
I'm one of the ones that argued against a mesh underlayment for metal roofs in other threads. I don't think they effectively drain moisture and the gap created may increase the condensation on the underside of the panels. However the huge difference in the roofs we were discussing and yours, was that they had no mechanism to dry the sheathing, while you have an attic, allowing any moisture to quickly dry to the outside.
Martin, thanks. My reasoning was that there would be less condensation underneath the metal panels too--because of less exposure to cool moist air. However, by this reasoning isn't it then better to lay the panels directly on the underlay to limit the exposure to air and therefore limit the condensation even more (?) I guess the reasoning of my roofing contractor was that "IF" there "IS" condensation under the metal panels (or it does develop) the 10mm rainscreen would provide a gap large enough for the condensation to evaporate and dry out--whereas laying the panels on an impermeable underlay would inhibit this process. I imagine that if the panels were to be laid directly on the underlay (and it is breathable) then any moisture that develops could dry towards the attic and sheathing side as well (provided the attic is well ventilated).
By this logic it makes sense to use at least a breathable underlay (instead of an impermeable underlay) beneath the metal panels so that the moisture can dry towards the attic and does not condense between the underlay and the sheathing (causing rot of the sheathing), or between the metal and underlay (where it would have less chance of drying).
Ultimately, by using a breathable underlay, AND the rainscreen, my thinking (and the thinking of the contractor) was that you provide two possibilities for "drying of moisture"--any condensation or moisture coming from the house will work it's way through to the cold metal panel surface and then downwards into the rain screen where it has a space to dry--but it can also dry towards the attic side through the permeable underlay. So I think it makes sense to use both a breathable underlay and the rainscreen. The rainscreen also adds a "thermal break", even if it is a small one. The air gap serves as a bit of insulation. This would likely have an impact on the possibility of dewpoint being reached further down in the roof construction (like under the sheathing, or in-between the underlay). Even though I have no solid proof, the house does appear to be warmer to my wife and myself. We have yet to go through a summer to see if it is cooler in the summer--however, because of the low thermal mass of the metal panels and the air gap underneath, I most certainly think the house and attic will also be cooler in the summer with the rainscreen.
At any rate, I am thinking, and hoping, that a breathable underlay and also a rainscreen is the correct stuff to put under a metal roof.
One other thing. From what I can gather from a couple of studies, plywood has a much better capacity to store and release moisture when compared to OSB. So for anyone doing a new roof plywood would seem to be the way to go to further help the battle against moisture.
Malcolm, thanks for the reply. Wouldn't the rainscreen help even in cases where the metal roof is being placed on an unvented roof assembly. With a permeable underlay the sheathing could at least dry to the outside if it can't dry to the inside. It would seem to me that sandwiching wood sheathing between insulation and a vapour impermeable underlay is just asking for trouble since at some point in the roof assembly dewpoint will be reached eventually and if there is moisture present it will condense or rot over years of time.
From a logical standpoint it would seem to make sense to allow drying towards the outside and the inside where possible on metal roof assemblies--to allow the roof to breathe as much as possible no?
It also seems to me to make sense to provide as much insulation as close to the roof surface as possible (as well as in the attic). I initially wanted to lay the metal roof on foil bubble insulation, but opted for the rainscreen because it seemed more popular in this application--as it allows for moisture drying as well as a thermal break. I had considered using both the bubble wrap and the rainscreen, but was not sure what configuration would work best--wondering if the moisture would get trapped between the bubble layer and the roof, or the bubble layer and the sheathing. Plus there is the problem of trying to screw the roof down over top of two layers of material--which does not make it practical.
I imagine that if you really wanted to do it, the bubble wrap would go above the breathable underlay and below the rainscreen. I never considered purlins since that seemed to me to be prohibitively expensive and time consuming.
There seems to be a lot of confusion on how much air can move through a 1/4 inch screen--but I would think it is a lot--with convection, and sun heating of panels etc. My roof is seamed at the hips with small gaps where the panels meet and then the hip caps were run down the length of the roof and screwed to the top of the seams--so there would be some air escaping at the hips. If I were to do it over again I would use a product that vents at the drip edge so that it could push air up the entire length of the mesh from the bottom. My roof is hemmed tightly at the bottom so the air circulation there is not much--however, it will breathe through the hips and at the ridge where it opens up under the ridge caps. I imagine that air escapes through any available channel when the high and low pressure changes. Indeed, I can recall several times when I would hang a soaking wet shower towel to dry inside and the wind and air pressure change of the weather was so drastic that it would be completely dry in the morning inside the house--almost as if by a miracle.
Tony,
Theoretically yes. Venting above the sheathing can help it dry as long as two things are present: A path for the moist air to escape, and a mechanism to move that air.
To make them effective codes typically require from 1" to 3" of unimpeded space on vented roofs for the air to move through, and a certain volume of openings at the eaves and peak. I don't think a roof with 10mm mesh relying on gaps in the flashing to exhaust the air provides a path that would allow much moisture to be exhausted.
Ventilated roofs also rely on the stack effect to move air. That's why low-sloped ventilated roofs are so problematic. With a small gap filled with mesh, I don't know what mechanism would move the air through the cavity. There may be some drying of the sheathing into the cavity, and some small amount may manage to exit, but my fear is that the presence of the air below the roofing may allow moisture to condense at or above the rate it can dry.
Your example of a towel perhaps illustrates my point. The towel dries because it is wetter than the surrounding air, and is exposed to a volume of air sufficiently large to absorb the moisture. That same towel left in a sealed bin overnight would be almost as wet as when you hung it up.
Malcolm. Thanks for your points. More to consider and lot's of very valid stuff. I guess this is also what I am wondering about. How much space does moisture need to escape and dry effectively? If it is in a gaseous state, wouldn't it be very little?
Your points about the towel I mentioned are interesting. The towel I talked about gets hung in the same location every night, often soaked, and most days it is wet when I wake, or a little less damp. However, on certain days, when it get's cold and windy outside it dries completely. So there must be some mechanism in the air that is speeding up this process on certain days--perhaps high/low pressure systems(?) that literally work to "suck" the moisture out the same way air is let into a vaccum when you release the seal. I don't now enough about science to know how this mechanism works, but it seems to me that it is present (?) When I watch the dew on the grass in the morning it is stagnant and calm--but the moment the sun comes out and it is hit by the slightest amount of heat, it starts to dissipate into the air as if by magic. Now if there is less air it would dissipate more slowly, but the tendency of warm air to rise through cold or stagnant air is always present, and would be present even in very small airspaces no (?)
I can see how a small space under a metal roof (1/4 inch) would not be that well vented and a part of me thinks that it would just stay wet like you say. But when you add certain weather phenomenons like high pressure and sun warming up the metal creating convection currents would it not cause a lot of drying and release of moisture as the air makes its way from low to high pressure through any available crack or crevice (?) Some of my roof underlay was soaking wet when the metal panels were installed and my contractor seemed to think it would dry quite readily when the heat of the sun warmed the panels (?)
Tony,
In the absence of any research showing what's happening with the mesh under the roofing, or evidence of problems, I think we are safe calling it neutral - and not worrying about your roof.
One thing I should add is you may over time see some discolouration on the underside of your roof sheathing in the attic. That is endemic to vented roofs in the PNW, and doesn't appear to cause any problems. We just live in a very damp place.
Malcolm,
I agree, of course, that sheathing mold in the attics of the Pacific Northwest is normal and doesn't cause any problems.
I have to salute you as a PR genius, Malcolm -- you managed to avoid the use of the word "mold." I like your substitute phrase: "some discoloration."
Just practicing. Rumour has it Sarah Huckabee Saunders' job may open up.
If you observe what happens to a vehicle that is parked outside during night sky radiation dew point nights. The top of the vehicle is covered in moisture but the underside of the hood does not experience moisture. Even though the engine compartment area is not sealed, the metal hood underside doesn't see any moisture.
I have a raised metal roof on steel purlins (about 2.5" from decking to roof). We see dew a lot during winter and when I looked underneath the metal roof (have a removable panel), I have yet to ever see moisture underneath and on the membrane.
Peter. Thanks for this. I considered the example of the car hood, but I also thought that many of them are insulated and perhaps that’s why they don’t have condensation under them? Also the steel is very thick—like 20G isn’t it?
And thanks for your example of your roof. It helps to know what’s going on by someone who can actually look under their roof and assess it.
There isn’t much thermal insulation in the bodies of most cars. Usually the insulation is mainly for sound damping, and consists of strategically placed pieces to minimize materials costs.
20ga steel is pretty thin. Even thick steel is going to conduct heat very well. Not as well as copper or aluminum, but far better (many orders of magnitude better) than wood.
Bill
Malcolm. Thanks for the further reply. You brought up an interesting point about mold. When I recently purchased the metal roof I had the top portion of the roof resheeted because of mold (we live in a three level split so the roof has lots of hips and 2 levels to it, but only the top portion had evidence of mold). The mold was minor but I wanted to get it resheeted because the roof was already costing a bundle and I didn’t want the mold to get worse. Anyway, in the summer when they tore off the sheeting it was still in very good condition, and the mold had dried out completely and was hardly visible. I probably wasted $4000 resheeting(?). The osb was very solid when they tore it up as it was splintering everywhere. I ended up putting in a ton of soffit venting on the top portion of the house because I felt I needed to remedy the problem (added 6 more holes to the soffit) and now the attic breathes really well, but maybe I didn’t need to add any venting? Maybe I overdid it? And maybe it is worse now that the top attic has more vents at the soffit than the bottom level which is also a much larger level of the house? There are now 15 soffit vents on the top portion of the attic and only 10 on the bottom level. My reasoning at the time was that I needed to improve the airflow to keep it from getting moldy—and I figured if there was more vents at the soffit on the top level it would create a “sucking” effect on the bottom roof and probably vent the whole thing better(?)
Do you think I should plug some of the added soffit vents? It’s very difficult for me to assess how much venting I have because the ridge and soffit vents are covered with grates. When they resheeted I made sure to unblock many of the baffles that had been blocked when I purchased extra insulation years ago. They also put a large ridge vent on the upper roof deck (larger than the others on the lower portion of the roof). Circulation is definitely better up on the top level, and on frosty days there doesn’t seem to be any evidence of condensation (yet). But perhaps I overdid the venting?
I don’t want to have too few soffit vents because I understand it is better to have more soffit venting than less...I do feel like I needed to add some soffit venting as the ridge vents are quite large now (I only had 4 box vents on the whole roof before!!!).
Do you think I should plug some of the soffit vents I added and go back to the 9 vents I had on the top portion of the roof? The vents are holes in the plywood with grates on them about 6 inches wide by 12 inches long. Or do you think I should add more venting to the bottom portion of the roof? It currently had 7 but I added 3 over the years so now it has 10. I was troubleshooting leaks with the old roof and had to cut holes in the soffit so I just decided to install vents there.
Anyway, confused about all this crap. I thought more venting was better, but maybe it’s not. And now I have a very large ridge vent on the top roof (about 8 inches wide by 8 feet long) so I probably need the added soffit vents to keep the ridge vent from becoming a suction point and pulling in air no? The other 2 ridge vents are a lot smaller (3 inches wide by 6 feet long).
Tony,
Is that much ventilation necessary? Probably not, but it does no real harm. I don't think you can over-ventilate. Too much air-flow might cause a slight energy penalty by washing over your insulation, but not a MOLD problem. The proportion does seem to matter. Having more at the soffit and less at the ridge, as you say, helps maintain a positive pressure in the attic.
The high humidity on our PNW outdoor air causes minor discolouration on the underside of roofs even if they are completely open to the outside - like lean-tos, wood-sheds and soffits - but as Martin and your experience confirmed, in the absence of some other problems it doesn't get bad enough to do any real damage.
Malcolm. Is there ever a danger of too much moist air moving into the attic through the vents and slowly saturating the fiberglass insulation, or increasing the risk of condensation on the sheathing? I thought more venting can sometimes bring too much moist air in from the outside?
I am glad to hear about the mold issue so I won't be as alarmed when/if it shows up again. For me seeing is believing. When the sheathing was ripped off and exposed to the sunlight the dark grey streaks that looked so ominous in the attic during winter had mostly disappeared and the wood was splintering and dry/fresh.
With excessive ventilation the moisture levels in the attic should end up the same as the moisture levels outside.
That's the case in my shed, which is unheated and has large continuous vents at soffit and peak. The underside of the exposed roof sheathing has a few dark spots, Some things things stored there, like sandpaper, feel moist but dry out on sunny days. Common nails take decades to acquire a thin veneer of rust. Batts of insulation are not affected.
I don't know much about warm humid climates such as those in parts of the southern US, but I understand introducing moist outside air can be a problem there. Here it's essentially a cosmetic one.
One more thing that might help understand what's happening. This is a quote from Dana Dorsett in another thread:
"In the PNW and elsewhere, cold damp winter air is still usually much drier than the indoor air in absolute terms. Pulling that air indoors does not create a humidity problem."
Thanks Malcolm. I guess this means that venting the attic more rather than less, and leaving a window cracked open in the house is not any worse than overheating/over-humidifying your house through use of heat and cooking/showering?
Does this apply to summer weather as well?
I am in the heart of the PNW--with all this rain that is the main reason I wanted a metal roof--to get rid of moss and standing moisture, and mud and sand (from the asphalt shingles) in the gutters. It is really lovely to see the roof dry in a few hours after harsh weather. The gutters are a lot cleaner too and I won't have to pick moss off it like I did my old roof.
Tony,
When outdoor air is hot and humid, you have to be careful about inviting this outdoor air inside.
If you want to air condition your house on the day after you left the windows open all night, the air conditioner will work harder to remove all of the moisture introduced overnight. (This concern applies when dew points are high during nighttime hours.)
Similarly, venting a cool crawlspace with hot, humid summer air is counterproductive.
Finally, introducing warm, humid outdoor air into a vent channel under metal roofing can lead to condensation when there is snow sitting on the metal roofing. This phenomenon occasionally happens during spring weather, and there isn't much you can do about it. The roofing underlayment usually catches the drips, and the condensation eventually evaporates.
Martin. This is interesting. So you are saying in the summer it is better to close the windows on a humid/warm rainy day--and worse to have them open in the summer during a rainy day than during the winter rainy days (which are cooler)? We don't have AC, so typically leave the windows open all over the house a lot in the summer. Perhaps this is not good.
I also open the crawlspace vents in the summer and close them in the winter--perhaps another bad practice (?) We have a finished crawlspace (3 feet high) with foam insulation on the walls and plastic then a layer 0f carpet on the cement.
Tony,
If you don't have AC, leave the windows open in the summer -- especially at night. If you wake up and the house is cool, with a hot day ahead according to the forecast, close the windows in the morning and open them up in the evening.
Opening your crawlspace vents in the summer is a bad idea in North Carolina -- but in the Pacific Northwest, the idea is not quite as bad as it is in North Carolina. More information here: "Building an Unvented Crawl Space."