Image Credit: Judy and Ed via Flickr
More Q&A Spotlight
GBA reader Erik is building a new house in southwestern Washington state, and he’s thought through most of the details with care. But as the time nears to install the standing-seam metal roof, Erik realizes he may have overlooked something important in the construction details.
As he envisions it, the unvented roof would be insulated with the “flash and batt” approach — a 3-inch layer of closed-cell spray foam applied to the underside of the roof sheathing followed by insulation batts to fill the remainder of the 2×12 rafter bays. Above the 5/8-inch roof deck, Erik sees synthetic underlayment and then the metal roof.
“Now, as we get close to putting the roof on, with quotes from roofers in hand and a few selected based on their experience with standing-seam, I realize that one thing I did not fully understand is ensuring the sheeting can dry in at least one direction,” Erik writes in a Q&A post.
Erik sees four options for venting the roof:
1. Adding a layer of rigid foam above the sheathing, which would require a second layer of plywood over the foam to provide a solid substrate for the roofing.
2. Adding a vent channel under the metal roofing and using a vapor permeable underlayment, which also would require a second layer of sheathing.
3. Use of a mesh layer such as Cedar Breather between the underlayment and the roofing.
4. Use of a vapor permeable underlayment or 30-pound felt over the sheathing, followed by the roofing.
There are drawbacks with some of these options, including higher costs and, in some cases, awkward aesthetics. But Erik would probably choose the second option as the least objectionable. Or should he be worrying about this at all?
That’s the topic for this Q&A Spotlight.
Just make sure…
Weekly Newsletter
Get building science and energy efficiency advice, plus special offers, in your inbox.
This article is only available to GBA Prime Members
Sign up for a free trial and get instant access to this article as well as GBA’s complete library of premium articles and construction details.
Start Free TrialAlready a member? Log in
46 Comments
small gaps
Backer rod is a nice addition to the "small gap" debate. Hopefully someone has put a moisture probe in sheathing to provide data (vs more speculation).
On my roof design I went with
On my roof design I went with SIGA Majcoat (34.0 US Perm Rating) on top of the OSB sheathing. Then a standing seam metal roof that is factory designed to sit on steel purlins which gives it a 2" gap between the metal roof and sheathing. My roof could only dry to the exterior since there is 6" of closed cell spray foam underneath the OSB sheathing.
Question for Peter L
What is the spacing between the metal purlins? And is your roof "walkable"?
Antonio
Antonio
That (purlin spacing) I would have to look up on my records as that was determined by the installer and roof company. From memory, I believe it was every 4 - 5 feet. The panels are 24 gauge steel. You can always increase the amount of purlins but that will increase the cost and labor. The roof isn't going anywhere since it is locked in place with each panel, which then locks into the brackets on the purlins. It's one cohesive locked unit.
The panels lock in place and once locked they are very, very difficult to unlock/remove and require a special tool. Which once installed, they don't need to be removed as they will last basically the lifetime of the home. Standing seam metal roofs are the longest lasting roofs they have.
I have walked on the roof and there is some give/flex when you walk between the purlin areas. It's best not to constantly walk on the roof which most people don't walk on roofs once they are installed. The roofers said that they have never observed any damage/issues from walking on the roof with purlins.
Peter L
What were the details at the ridge, drip edge and gable ends that allowed air-movement under the roofing?
Malcom
I don't have a way of posting that, maybe a picture later. There are small gaps/holes where the drip edge meets the fascia. Enough to allow some air movement but not enough to allow for birds or small critters to make their way inside.
A standing seam metal roof on steel purlins like this is not airtight. There is always gaps somewhere in the metal fascia and soffits. It's impossible to create a metal fascia and soffit that is 100% airtight so some air will always make its way into the area below the metal roof and decking via the fascia and soffit areas and drip edge areas.
Akos, unfortunately your
Akos, unfortunately your assembly does not meet building codes, or building science recommendations for safe assemblies, due to the risk of moisture accumulation at the sheathing leading to rot and/or mold. How long has your roof been in service?
Edit to add: now that you've added information to your post, I'll add that at with a framing factor you're looking at more like R-28 or R-29 on the interior. 2.5" of aged polyiso, best case, is about R-14. So that gives you about 33% on the exterior. In US climate zone 5, prescriptively you need 40% of the R-value on the exterior. In zone 6 we need 51% on the exterior. I would be nervous about your assembly but it's your building.
Peter L
I guess my question is similar to the one I asked Jon. Is adding an air-space under the roofing useful if there is no path for the moist air to move to the exterior? Horizontal purlins mean no direct path from eave to ridge. Unless there are vents at the gables the moisture is trapped, and even then with no difference in elevation, there is little incentive for the moisture to move to the perimeter. It seems to me that without an airspace there would be no where for the moisture to accumulate, which might be a better situation.
Malcolm
As you can see in the attached photo, there is clearly a gap between the horizontal steel purlins and the metal roof, as the metal roof is elevated at the purlins. So there is definitely a path of air. No moisture is trapped. There is a direct path from eave to ridge as the horizontal steel purlins elevate the metal roof off of the purlins as the steel clips hold the roof in place but allow for airflow.
In addition, the roof is a standing seam metal roof so the roof is not flat on the bottom portion.
If moisture gets trapped in the OSB decking during install or after install. Without the metal roof being elevated on purlins. The moisture in the OSB is trapped and cannot dry via the interior (due to closed cell design) and rotting would occur. See the Oregon article about the SIP roof that rotted because the contractor put a metal roof directly onto the OSB and there was moisture trapped in the OSB.
airflow
The title asks "Does This Roof Need to Breathe? But I think that code and everyone agrees that a proper unvented design doesn't strictly need to. The question is if there is any benefit to it.
> Cedar Breather, a 1/4-inch-thick mesh designed by its manufacturer, Benjamin Obdyke, for use under shingles or shakes.
But also explicitly sold for use under metal roofing - where they evidently think it provides enough air flow to provide a benefit. Note that they show the use of felt underlayment and a ridge vent in their standing seam example.
I would expect eve-to-ridge backer rod to provide similar airflow/drying to Cedar Breather. If it's cheaper and also helps with oil canning, that's great.
More claims for airflow under metal (.6"):
https://www.achfoam.com/ACH/media/ACH/docs/Roofing/Naillbase-Roofing-4003.pdf
Peter
I'm sure the fault is mine and I'm just not understanding what I'm seeing, but it appears to me to show a solid purlin attached tightly to the roofing panels above. There is a sizeable air-cavity, but what I'm unsure of is the utility of any cavity if it does not have a) a mechanism to promote air movement, or b) a defined path for that movement - that is vents at both ends of the cavities and a difference in elevation so the air has an incentive to move.
The difference between this and sealed rain-screens, which do promote drier sheathing, is that they work through diffusing the moisture through permeable cladding. They also don't experience the night-sky cooling that metal roofs do, which will cause the cavity air to condense on the underside of the roof panels.
Maybe you do get some drying through your set up, but it doesn't seem to have been designed in.
"If you've got a roof leak,
"If you've got a roof leak, you've got wet sheathing — but that's the case with any roof. In that case, you repair the roof leak or install new roofing (and you make sheathing repairs if necessary)."
In these "hot" roof types, how do you know when a roof is leaking and needs repair or replacement?
roof repair
> In these "hot" roof types, how do you know when a roof is leaking and needs repair or replacement?
Also, how does one properly repair a roof that has closed cell spray foam adhered to the underside of the sheathing? This alone might be sufficient motivation to be extra conservative on moisture issues (eg, breathable underlayment and some venting).
Keep it simple
Had a similar quandary when doing my metal seam roof in Toronto (code in the great white north is only R31 for cathedral ceiling). Ended up going with the following slightly above code stackup (inside to outside):
-un-vented 2x10 rafters with batt insulation (R30 insulation, R24 including framing factor)
-plywood sheathing
-Grace ice and water
-horizontal 2x3 strapping on edge with 2.5" poly iso in between (~R12)
-synthetic underlayment
-metal roofing
I used 2' wide insulation and put the strapping so the insulation would not have to be cut, extra strapping aft the eaves overhang where there was no insulation.
The insulation was fairly straight forward to put up, strapping and all for a 1200sqft roof took about 1 day.
The metal roofing was simply attached to the strapping, the roofers had no problems with the setup.
Important to do dew point calculation, this works here as the mean temperature in the wintertime is not too low (the assembly is borderline in the coldest days of February).
drainage/ventilation mat
Zinc mechanical seamed roofs have been using drainage/ventilation mats for over 20 years, Enkamat® 7010
drainage/ventilation mat
Zinc mechanical seamed roofs have been using drainage/ventilation mats for over 20 years, Enkamat® 7010
Steel purlin - Malcolm
Malcolm,
In this buildup, the steel purlin uses steel clips to hold and elevate the standing seam metal roof, with a high rib. The standing seam metal roof does not sit tight on the purlin, it is elevated with a steel clip which are spaced every 1-2 feet.
With the SIGA Majcoat membrane, it has a 34.0 US Perm rating so it is definitely breathable.
Installing a metal roof directly onto the OSB Sheathing insures that there is ZERO drying/breathability of that OSB to the exterior. It can ONLY dry to the interior but if you have a SIP, closed cell sprayfoam, etc. buildups, it cannot dry to the interior so you are left with only drying to the exterior.
This roof buildup has been done by others who followed GBA articles and discussions, in making sure that the exterior OSB sheathing can dry to the exterior. A breathable (34.0 perms) but waterproof membrane like SIGA Majcoat membrane and elevating the standing seam metal roof off of the decking to allow any water than may be in the OSB sheathing to dry.
These steel purlins are also used in steel sided buildings that have OSB wall sheathing. The steel is elevated so that the OSB wall sheathing can dry if wet. They have been installing these roof and wall assemblies for 30+ years with zero sheathing issues. There is 2" of spacing between the OSB sheathing and steel roof.
Even at the soffits, the steel cladding is not sitting directly onto the sheathing. This provides a way for the sheathing to dry if it gets wet. See below photo.
I hope you are not arguing just for arguments sake and stir trouble just to stir trouble. Sometimes this happens on the internet forums. I hope this is not the case.
Peter
Why would you even imply I'm here to stir up trouble? I'm trying to understand how to properly detail such a roof so it has a drying path. That's what all my questions have been aimed towards, and so far, for whatever reason, you have been singularly unable to describe how any moisture that accumulates under there roofing makes it way to the exterior.
I thought it was a useful discussion, but I'm certainly now not going to continue it with you.
how breathable?
With walls, there is typically an optimal amount of external side perms, perhaps 10x the interior side perms, where more doesn't do much for outward drying and more perms increases sheathing wetting from the exterior. So with < 1 perm of wetting through some under-the-sheathing closed cell spray foam, I wonder if there is an advantage to ~10 (vs much higher) perms on breathable underlayment.
Maybe that's a lot to ask given that I haven't seen any moisture data on any breathable underlayment in a vented vent-optional design.
I tried
Malcolm,
I attempted numerous times to explain on how this build-up works but you would not accept my explanation. Your lack of acceptance of my explanation does not constitute that I have, "singularly unable to describe how any moisture that accumulates under there roofing makes it way to the exterior."
I sent pictures, explained in detail in numerous posts, and yet you kept repeating the same thing over and over. What else am I to do?
You might not agree with my explanation but your disagreement is your opinion. Opinions vary as do building methodology.
Peter, it seems to me that
Peter, it seems to me that Malcolm is looking for a standard, code-required path from eave to ridge of at least 1" clear, which you do not have. You obviously have a sizable, continuous space below the roofing, but if I understand correctly, the primary (or only) way for moist air to escape that space is via the small gaps between drip edge and fascia on the rakes?
Vented
On my own cabin from the backside I put 2x2's along both sides of the truss. I then ripped 1/4" osb to install between the trusses for a 1 1/2" air gap and spray foamed to the 1/4" osb. Not that costly or time consuming. Works great for cathedral ceilings and shingled roofs also. If there is ever any need for deck replacement the foam is not stuck to the deck.
@Malcolm
I think you are missing this... "...The standing seam metal roof does not sit tight on the purlin, it is elevated with a steel clip which are spaced every 1-2 feet..."
The roof isn't attached directly to the purlins but via clips. These clips have height to them (we don't know how much) which provides a gap between the roof and the purlin. This gap allows some vertical (eave-to-ridge) movement of air between the purlins and the metal roof.
airflow vs diffusion
Keep in mind that we are talking about such tiny amounts of moisture that code allows one to ignore it - it's OK for it to accumulate in the sheathing where it eventually dries inward.
As is sometimes said about walls, air flow can move 100x more moisture than diffusion. Far more with low perms (like closed cell foam) and something that isn't designed to be airtight (eg, metal roofing).
John Clark
At the risk of prolonging what suddenly seems to have become a pointless discussion, there is no way 24 ga metal roof panels with purlins at 4 ft oc are supported on clips. The panels sit on the purlins and the clips hold them down. Any gaps are incidental, and there is no provision for the air to enter or exit at the eaves or peak.
Jon is entirely right. We aren't trying to ventilate a roof in these conditions, just perhaps remove what little moisture may accumulate in the sheathing. My initial concern was that providing an air-space may cause as much moisture to accumulate as it could remove - that as Michael put it "You have created a condensation machine". Air spaces aren't inherently drying, especially those under metal roofs experiencing wide temperature swings.
A picture is worth a thousand words...
See below clip. Bottom flat part of clip is screwed into steel purlin. The standing seam raised rib then clips into the top of the clip. It is locked in place in 2 places, top and bottom. There is a minimum of 1/2" of elevation from the steel purlin to the bottom of the lowest part of the standing seam roof. In other areas, there is over 1" of spacing between roof and decking.
So once again, the panels DO NOT sit completely on the purlins. That's how they are engineered. I can shine a light and look from the drip edge to the ridge of the roof since the decking is elevated off of the purlins.
Peter L
Ah - I've finally got it, and I understand your frustration with my posts. I couldn't understand what I was looking at having never seen anything like it. If you have a consistent 1/2" gap above the purlins, you essentially have one large air space with no real impediment to air moving freely between bays. Which (and I really hesitate to say this) brings me back to my first post. How do you detail such an air-space to facilitate drying?
Air-spaces in themselves are not inherently a drying mechanism. An easy test of this is to place two glasses, one completely full and one half empty, on a counter and place a saucer on top of each one. Over time neither will lose a noticeable amount of liquid. In the absence of other factors, the air-space makes no difference.
So how can roofs dry? Through diffusion, and this is out as metal is impermeable, or through air-movement. The two factors necessary for air movement are sufficient openings at the soffit and ridge, and a sufficient change in elevation between the two to cause the air to move (which is why flat or low slope roofs aren't ventilated).
Another long way round of asking the question: I wonder how best to detail roofs where the air-space above the sheathing so that these factors are present and they dry effectively?
metal roof underlayment with openings at drip edge and ridge
https://foursevenfive.com/vented-high-performance-standing-seam-metal-roof/
Nathaniel
Apart form the pitch of the roof, this still faces the same questions. A 0.3" mesh underlayment and 3/8" vents at the drip edge don't make for a vented-roof from either a building science or building code standpoint.
They also suggest that moisture can drain down the membrane and exit via the vented drip-edge. I don't see that as a desirable place for water to be dripping down.
Air gap between the metal and deck will sweat bad, rust the undersides and void the warranty. Think about a humid day , there's a thunderstorm then right back to 92 degrees... You've just built a condensation collector
But without any venting, it will still get condensation (metal roofs aren't completely airtight). And it will take longer to dry.
Vented metal roofs are very common, with some experienced builders (like Matt Risinger) deliberately paying more for venting when the design doesn't require it. I find no reports of underside moisture problems from venting or any warranty disclaimers. Do you have a link to data showing that vented metal roofing (with the rest of the roof design being the same) fail more often than un-vented?
I see numerous examples suggesting it isn't an issue. For example:
https://www.youtube.com/watch?v=ZkCZWdgPnYM
And that ventilation helps:
http://www.dorken.com/en/our-products/products/commercial/delta-trela.php
John R ,
I don't have an online source for this. I specialize in water infiltration remediation. I have spent my 25 year career thus far surrounded by and working alongside great contractors and engineers to diagnose and find permanent solutions for leaks. I speak from experience. Here in Pennsylvania, I can definitely put my name on this as a fact. I'm sure different climates and roof system configurations could make an air gap viable.
Colin rappa construction inc.
Is rusting under the panels a problem people see during tear of older metal roofs?
How do metal roofs fail first?
Metal panels come in all sorts of configurations, some leave a gap, some are flush. The roofers in my non-freezing climate seem to all quote "high temp self adhering membrane" with flush mount panels, which seems like a recipe for condensation that can never dry out. The vendors seem to specify "#30 felt", which would allow drying to the inside, compared to the self stick stuff.
To demonstrate take any two sheets of heavy plastic, put some water between, and see how long that takes to dry out.
Bryce,
My experience is only in my own climate zone here in the PNW. The failures I've seen on metal roofs here are primarily fourfold:
- Failure the exposed fasteners causing rot in the underlying substrate.
- Failure of the applied finish, which is usually cosmetic, but can cause some rust.
- Failure of the roof panel due to contact with incompatible materials that cause a galvanic reaction.
-Failure of the sealants and gaskets used on laps and penetrations.
I come back to my point about condensation. It is what occurs when moist air meets something below it's dew point. Where is the air between a sheet of metal and the impermeable underlayment it is secured to? To me it's like asking whether you will get condensation between the sheets if you put down two layers of plywood on a floor. Condensation pre-supposed an air-space.
Airflow (say from wind and the non-airtight roofing) can be sufficient to cause condensation, even in tiny gaps. I can also glance at #33 and see that metal roofing does have some air under it.
Surely if there is sufficient air-flow to move moist air though those tiny sporadic indentations, there is also enough to safely remove it.
Unlike all the work and testing that went on to determine how rain-screen walls function, adding battens or proprietary products for under roofing seems to have been done based on hunches.
And in the same vein, rain-screens on walls came about as a result of problems they were experiencing. Where is the evidence of problems on metal roofs installed directly on solid substrates?
> Surely if there is sufficient air-flow to move moist air though those tiny sporadic indentations, there is also enough to safely remove it
Surely not. Wetting (involving condensation) and drying rates are not symmetrical.
Agreed that there is a lack of data. And IMO, the metal roof itself isn't the concern - it's the wood under the underlayment (see caption on first picture above) where there is plenty of evidence of problems. Specifically, do some of these options result in dryer sheathing? Seems like a simple, basic building science question that hasn't been answered.
a) insulation, OSB, near zero perm underlayment, metal roof
b) insulation, OSB, permeable underlayment, metal roof
c) insulation, OSB, permeable underlayment, small vent space (eg using mesh), metal roof
Jon,
I agree - but now we are talking about something quite different than most of what went before. The idea would be to promote outward drying for the benefit of the sheathing below, not drainage planes, or ventilation for the condensation on the underside the roof.
I guess my first question would be why pick on metal? We don't worry about promoting outward drying on the much more common shingled roofs - and that seems to me sensible. As RDH points out, there may be some benefit to permeable underlayment, but in general it should be subsumed by the more important role of keeping bulk-water from penetrating the roof. https://www.rdh.com/wp-content/uploads/2017/08/RCI-2015-Problems-with-and-Solutions-or-Ventilated-Attics-GFINCH.pdf
My worry would be that variable-perm membranes are two way paths. They open up to allow moisture to move from wet to dry. If the cavity you create under the roof does become a condensation trap due to night-sky radiance, or some other mechanism, then what stops the underlayment from wetting the plywood underneath, rather than aiding it to dry? That would be a particular concern here in the PNW where we can get months when we do not get any solar drying.
I also wonder why it is necessary to dry to the top-side when ventilated roofs appear to work well in almost every climate? What is the problem we are trying to solve by having drying to both sides? In the absence of roof leaks, are vented sheathed roofs showing damage?
Common belief seems to be that moisture removal upwards through metal and shingle roofs is small and therefor irrelevant. Not much data, but what there is indicates that the latter half isn't always true. Drying/wetting ratio (not always absolute amount) is important.
> drying to both sides?
I assume you are referring to a case with an under-sheathing vent. Now that is something quite different than what is being discussed. Even so, the rdh.com link shows that upward drying still makes a difference (and suggests that permeable and fully adhered underlayment would be best).
"and suggests that permeable and fully adhered underlayment would be best"
I didn't read their conclusions to say that. The two underlayments they tested were roofing felt and an impermeable adhered membrane. The roofing felt did aid drying through the shingles, but they suggested the benefits were outweighed by the better resistance of the membrane to bulk-water intrusion.
Oriented strand board is the absolute worst choice for a roof substrate. It can get wet once and swell, delaminate and loose it's structure.
For vertical walls, maybe... But only if it is properly protected from water infiltration. A building should be 100% impermeable before the facade is installed, further where standing or whicking water may occur the flashing should include a rubber preferably or asphalt membrane to seal penetration, And a capillary break.
Back to the metal roof, breathable or sealed underlayment doesn't make much difference. The underside of the metal will have condensation. Adding an air gap will encourage moisture.
The main goal of the underlayment is to prevent scratches on the metal.
> Adding an air gap will encourage moisture.
There is lots of data showing that full size venting reduces sheathing moisture. Do you have any data that shows that any air gap has no effect or increases sheathing moisture? Or even that the metal itself is wetter?
Are you suggesting we no longer design walls to dry to the outside?
The main goal of underlayment is surely the same as any other WRB, not just to discourage scratches.
I am back to contemplating this again after having installed my roof (from top layer to bottom) in the PNW a year ago with 24 guage standing seam, 10mm rainscreen, Gaf Deck armour, OSB, Attic with air space and full insulation on the attic floor.
This morning I noticed a ton of frost on the roof. Sometimes here in the PNW we get so much condensation on the roof surface in the morning that the gutters will run! As soon as the sun goes down we get condensation almost immediately visible on the surface of the steel.
I am dying to know how much condensation is under the steel panels and I have even entertained cutting a small hole in the sheathing from the attic side and inserting an endoscope (camera) to examine the air space underneath.
I suspect there isn't much (hope I am right). I also suspect that with the sunny days that follow the frost (because it comes when there are clear skies) the dark coloured steel sheathing heats up enough to move the little moisture out from underneath with convection--it certainly is enough heat and sun to completely dry the top surface of the wet roof after a few hours!
I think there needs to be a large air surface for the steel to collect a ton of moisture (the way it does on the top of the steel cladding). Underneath the hips and ridges of the roof (I checked with the endoscope) there is some moisture, but it is a lot less than what is on the surface of the roof (less air flow under there). I suspect that underneath the steel cladding (where there is even less airflow still) that there is even less moisture--again the small amount will probably escape through the seams as the roof heats up in the sun, or into the space between the 10mm gap and the steel panels slowly moving out by diffusion and wind/air pressure to the seams and hips and ridges.
Anyway, I don't know if my roof was done right the way it is--maybe I should have gone with an impermeable underlay to keep the moisture from going towards the sheathing and attic, maybe I should not have included the air gap (10mm rainscreen at all)--but there did seem to be an advantage to having the ability to dry to both sides (attic and under the steel cladding by way of air gap).
ALSO--and this was important, and something no one has discussed. By NOT having the air gap under the metal, the dewpoint of the air would be driven further down into the roof assembly--like for example around the sheathing or underneath it. There did seem to me to be an advantage to creating that thermal break that the 10mm rainscreen creates between steel and the sheathing keeping the dewpoint higher up in the roof assembly, or maybe not in the assembly at all. If the steel can be kept at the same temperature as the ambient air, and there is no source of moisture under it but a dead cool air gap, then it theoretically should sweat very little. As we know, ALL roofs can condense, and they often do at the level of the sheathing in the attic--where there is more air space and more moisture in the air to condense on the cold surface of the sheathing. Anyway, this was my thinking--that the dewpoint in the roof assembly would be driven higher with the use of an air gap--the higher it is out of the assembly the better--and the ability to dry and pass moisture to both sides seemed also to be a net benefit in this situation (so I went with a breathable underlay) so moisture can move down towards the attic space and out the vents, as well as up towards the warm metal surface warmed by the sun and out the seams etc.
I guess time will tell if I start to experience mold etc.
You would think that someone would do a side-by-side test of sheathing moisture content with permeable and impermeable* underlayments under steel. And an additional test with a 1" air gap, no deliberate air gap and maybe even "sealant between steel and underlayment to eliminate any gap".
* - although nail/screw holes may mean that the sheathing sees some of that condensing moisture.
After reading this article and the corresponding comments, I'm less certain than when I started! In an almost identical situation to the Q&A, and cannot decide if I should do permeable underlayment + air gap or just go with an impermeable underlayment, spray foam under OSB, and metal roof right on top.
If there's been any additional thoughts on this, I'd really love to hear them.
Log in or become a member to post a comment.
Sign up Log in