Leak in EPDM low slope roof
gatter
| Posted in General Questions on
I”m looking for a forensic building consultant to help diagnose a leak in an EPDM roof with no obvious holes or tears. Thanks
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Search and download construction details
Replies
Building science corp. would be great but maybe beyond your budget or maybe not if this is this is a 200-acre warehouse.
https://buildingscience.com/
Have you considered the possibility that the “leak” is from condensation.
If this is a smaller roof and you don’t have the budget for travel your location would be important to know.
Walta
HI: I'm in Columbia County, NY just west of the Berkshires in Massachusetts. I don't think it's condensation. We have closed cell spray foam on the interior of the exterior sheathing of both the walls and the roof, and then dense pack cellulose. The water is literally running down the inside of the wall onto the floor.
What is the temp and humidity of the air in the attic?
Do the walls ever get wet when it has not been raining?
Walta
gatter.
You need to post pics of the roof to show edge details and whether you have parapets or scuppers to deal with. If you have roof penetrations, those would need to be detailed along with a cross section to show how thick the foam is relative to the dense pack. Do you have can lights? Vents or duct work running overhead?
Also, the description of running down the inside of the wall is vague. Which wall? In relation to what part of the roof? Is it isolated to one stud bay in the wall? Etc. If you have an iPad or iPhone there seems to be a sort of infra-red option. Not really like a good FLIR , but what with cold weather in your area, any water puddling should show up. Worth a try.
Condensation should distribute widely if due to the rafters conducting cold inside the foam layer. If it gets into a gap-py dense pack region, however, a membrane containing the dense pack, might be able to gather the condensate and send it on to the downside of the ceiling slope. I think your region was down well below zero recently, so did the water start after that time?
A bit more detail would allow more of us to at least make some suggestions to check out before engaging consultants. Might be simpler than you think.
It seems to be common to use a thin layer of foam as part of a flash-and-batt system. That can work in specific situations, but in a low-slope roof in a cold climate, it's hardly any better than using no foam at all, because the interior surface of the foam can still be below the dewpoint temperature, resulting in condensation.
Do you know how thick your foam and cellulose layers are? Does your ceiling have recessed lights or other penetrations? Do you track your indoor relative humidity levels? Do you have snow or ice building up on the roof, especially where a low-slope roof meets an upper story wall? (These are questions a forensic investigator will ask as well.)
My guess is that the cause is a leak at flashing tape around the perimeter. A hole in EPDM can be very small but still leak.
HI All: I"m attaching dwgs showing the roof plan as well as the building section at that area and some pics of the roof and the scupper as well as pics of the interior bays where we had to remove the soaked dense pack. (This is painful!)
The area of the roof is approximately 29' x 62'. The roof drains to a scupper on the north side. . In the drawing of the section the EPS rigid insulation continues up to the top of the top plate . In the field the contractor stopped the EPS at the top of the concrete wall and then dense packed the whole cavity in that location. The ceiling has no recessed lights. The contractor put ice and weather shield over the roof deck to protect the roof while we waited for the roofer to do the EPDM. The roofer carried the EPDM up and over the parapet wall. He didn't carry it all the way over but he put a metal cap over the parapet top. We first noticed a problem in February of last year. At that point the roofer felt the problem was that he had not put the metal cap on the parapets yet. After he did that we monitored the interior and it seemed that had fixed the problem. Unfortunately it returned this February when we've already dense packed the wall and put up sheetrock. What we think is happening is that there is a small leak in the EPDM and it's finding its way to the ice and weather shield below and then running down to the low point in the roof and finding ways to get into the wall underneath the low point of the roof. It seems to be happening particularly when there is ice on the roof. So maybe that's like ponding water that stays there for a while. We're planning on putting a humidity sensor in the wall if we can ever figure out how to solve this other than reroofing the entire roof. Thank you for any suggestions! I appreciate it.
Is the Ice and Water Shield in direct contact contact with the EPDM or is there foam or another layer between them?
Have you done a blower door test? If so, did you or they look at the parapet area for air leaks?
Your insulation ratios appear to be reasonably safe from condensation risk, with about 45% of the total in the foam layers when in your climate zone 40% is the threshold.
I'll second Walta's suggestion to ask Building Science Corp if they would be willing to look at it, as it's not that far from there offices west of Boston.
I should add that the water is coming in from the top of the wall framing, so it's not anything related to the concrete wall on the north side.
The ice and water shield is on top of the roof sheathing. Then there's (2) layers of 1 1/2" polyiso and then the EPDM. We haven't done a blower door test yet. Do you think that would provide a clue? Thanks!
That's good news about the adjacency, since I+WS offgassing makes EPDM brittle, but it seems to only be an issue when they are in direct contact.
I've consulted on one project with a leaking parapet that resulted in fully stripping the exterior siding and sheathing, and I've heard of a lot more--we just don't see that many of them here in Maine. They are notoriously hard to air-seal which may be causing or exacerbating the problem. But it could also simply be a hole in the EPDM, as you suggest.
Thanks!
The weak spot on these roofs is around the scuppers. The first test would be to plug the scupper outlet and fill that area to see if it leaks in.
This is good info about finding leaks on flat roofs:
https://www.wbdg.org/resources/integrity-testing-roofing-and-waterproofing-membranes
The simplest DIY start is getting a decent IR camera and scanning the roof with it.
If the grace ANYWHERE is in contact with the EPDM it will cause it to soften and rot in no time. From the drawing if the parapets were installed over the grace and the grace is not run up them, if that is indeed the case you should be fine. If you have grace on the parapets directly under the EPDM you are in trouble.
Thanks!
Deleted
Step 1: get an infrared camera out there. There is no way a reputable EPDM sub does not own one...
Step 2: What are we looking at in photo 2? Is this the parapet before the metal cap was installed? Does this show the EPDM on the roof running vertically up the parapet, and then halfway across the top horizontal and stopping? And is that the ice & water shield that is the other layer of black material that continues across the top of the parapet beyond the edge of the EPDM? If so, that is an issue and may be the problem.
Step 3: Inspect those metal parapet caps, especially if the EPDM does not run up, over the top of, and 2" or so down the outside of the parapet. The photos you posted do not appear to show the metal caps.
Often, the metal caps are attached ON TOP (incorrect location), and it leaks right through the fastener. Or attached on the outside where there is no sheet membrane and leaks right through the fastener.
Sometimes the metal cap is attached with a fastener of dissimilar metals. Galvanic action eats away at the metal and creates a hole.
Sometimes the front and back vertical legs of the metal cap are not long enough, or the top horizontal pieces is too wide for the parapet width, and wind just drives water under there.
The drawing shows the metal cap canted inward towards the roof, but if there is no blocking ripped for the cant, chances are there is no cant on the metal cap. I've seen leaks directly through the fasteners if they applied on top, leaks at the butt joints between the metal when there is a combination of insufficient laps at the metal joints and insufficient slope of the cant.
If the fasteners go through two pieces of metal, and then through the EPDM or not, and into the blocking at the top of the parapet, heat and cold will cause expansion & contraction, and the lapping pieces of metal move in opposite directions. The fastener starts to work loose by making a larger hole. Sometimes there is also oil canning at the top of the metal, and when it rains water just collects water and pours in.
The correct way to do this is to cant the nailer so the metal cap follows the nailer cant. Then slot the sides of the cap, and attach fasteners with a thick compressible (EPDM or other UV resistant material) washer through the side of the cap, so the cap can expand & contract without requiring the fastener to move.
NRCA (National Roofing Contractors Association) has industry standard recommended details for all of this. That is the MINIMUM you should expect.
Thanks!