Trying to find this leak for months.


Roof is approx 12'x 20'. Had a roll type roof made of rubber. Did not leak.
I decided to put metal above it anchored to 1 x 4 lats at 24" o.c.


I have a 20" roll flashing at top section that goes under house shingles 2 to 3 inches.
I sealed all the shingle seams and screws. All the metal seams and srews are sealed also.


It will leak/drip for days after it rains. Like the water slowly finds it's way to that location to drain.
And the location it drips (between the exterior hardie fascia and 2x6 sub-facia) I'm guessing it would be
coming form the top.


But I'm at a lost now... especially the way I have it all sealed.
Any ideas appreciated.


http://i819.photobucket.com/albums/zz111/lacogada/roofleak_zps4f2ef0c6.jpg (http://s819.photobucket.com/user/lacogada/media/roofleak_zps4f2ef0c6.jpg.html)


http://i819.photobucket.com/albums/zz111/lacogada/DSCF0032_zps768d8dbd.jpg (http://s819.photobucket.com/user/lacogada/media/DSCF0032_zps768d8dbd.jpg.html)


http://i819.photobucket.com/albums/zz111/lacogada/DSCF0029_zpsfdd7dd10.jpg (http://s819.photobucket.com/user/lacogada/media/DSCF0029_zpsfdd7dd10.jpg.html)


http://i819.photobucket.com/albums/zz111/lacogada/DSCF0031_zpsc21bff25.jpg (http://s819.photobucket.com/user/lacogada/media/DSCF0031_zpsc21bff25.jpg.html)

The first thing is that the flashing should have been 2'' under the 2nd coarse and completely under the first. Tar may look good but still be able to allow water in.

Did you set the cedar strips in silicone? If not water, can be backing up under them.

Chuck

Believe I follow what you're saying about shingle 1st & 2nd coarse.

I would say the flashing was 2" under 2nd coarse... top of flashing was not visible before putting tar.

In fact I had only put tar on horizontal seem where flashing and shingle meet, at the beginning. Later I added additional tar to vertical seams.

... and yes, cedar strips are sealed also.

It's a head scratcher.

Thanks for the suggestions.

Hi dannac,

I don't claim much roofing expertise but I have done a lot of study of surface tension when I developed my own gutter helmet to keep leaves out of my gutters (too cheap to buy the patented systems).
When I look at your top drawing, all looks fine and sealed etc.
Then I see your photo of the flashing and the large voids in between where the flashing lands on the high ridges of the metal roofing. I think that you have a quantity of water that travels from the top of the flashing at the low end- and via surface tension travels back under the flashing edge. You also have a significant amount of rain windblown back under that flashing edge, and it is a very difficult area to dry out, after heavy rains.
In a perfect situation, the flashing would have the same shape as the metal roofing at the point where it overlaps onto the metal roofing, and those large gaps between the ridges of the metal roofing wouldn't exist. There is a self-adhering rubberized material used in valleys and around skylights to provide a good seal in those applications. Could you use some of that type material back under those gaps. I definitely don't like those gaps and I think there is more water going in there via surface tension off the low end of the flashing and windblown rain, than you might realize. Good luck to you.

smearcase... thanks for all your thoughts and suggestions.

The voids you mentioned are actually filled with custom cut pieces of cedar.
I cut the angle for the ridges and made some pretty descent fits.
No more than 1/8" gaps... but all was sealed with tar are caulking.

The metal company did have some custom fit pieces of rubber, but they were not very rigid. Screwing the flashing edge through that rubber would have not been solid. That's why I went with the cedar, to have good support for my screws and not bend down the edge of the flashing.

In the last week n ahalf ... we've had approx 6" of rain, but no rain in last 2 days with 92 deg temps and its still dripping.

Thanks again, at this point I'm listening to all idea's.

The water either enters at the low end of the metal roof (I can't tell how those ridges look at the bottom edge) possibly by surface tension, especially at the metal roof ridges/joints at low end OR the water enters or is blown in at the high end of the metal roofing. Since it leaks for days, it seems more likely to me that the water enters at the high end and travels on the old rubber roofing (fairly flat) and possibly penetrates the roofing at the nail or screw holes that secure the 1x4's.
As regards surface tension, consider this short video: Gutters, Gutter Helmet, and Surface Tension - YouTube (http://www.youtube.com/watch?v=V0BaL4JWDbE)

Pic to show bottom edge. Metal extends pass fascia approx 2"

http://i819.photobucket.com/albums/zz111/lacogada/botmetal_zps95021c4e.jpg (http://s819.photobucket.com/user/lacogada/media/botmetal_zps95021c4e.jpg.html)

Never heard of surface tension, thanks advice and link.

This is the drip edge detail for a metal roof. (Drawings 22 and 23) :http://www.unioncorrugating.com/literature/ResidentialMetalRoofing_Trim.pdf
I can't tell for sure what you did at the raised joint and did you go back far enough onto the roof as suggested in the drawing?
All the voids between the roofing material and the flashing at joints and non-joint areas are opportunities for roll back or being blown back or rain in just the right direction to get behind the flashing. Note on the details (22 and 23) that there is no difference in drip edge shape at the joints and that it travels up the roof a significant distance.

Thanks for that pdf... it will come in handy in the future as I plan to metal the whole roof, and do away with that flat roof.

Surface tension... I can see how water may back up at that 2" overhang, but think it would stop at flashing the way I have the lower edge screwed down.

It would have to back-up 5 1/2" to get on the inner side of my flashing or 1 x 4.
And the water has to be on the inner side of the 1 x 4 (on the rubber roof) to leak where it does, between the ext fascia and 2 x 6 sub-fascia.

I believe you are correct that the problem is at the top somewhere.
3 days of no rain and still dripping.

Monday on my work computer I will show a good detail of where the metal meets the shingles, maybe you will spot something.

Thanks for your efforts, appreciate it.

That's a long time dripping. Sounds like a pool forming somewhere like flowing on the rubber and slowed by the 1x4's, as it works its way down.
Wouldn't it stand to reason that if the rubber roof didn't leak, the only way to get a drip is through a recent hole made through the rubber? Or is that faulty logic?

http://i819.photobucket.com/albums/zz111/lacogada/roofleak_zps587180ba.jpg (http://s819.photobucket.com/user/lacogada/media/roofleak_zps587180ba.jpg.html)
Not faulty logic... but I do not understand why it makes it all the way down and only leaks ( it seems ) after its made it pass the rubber roof.

Roof is 20' x 12' with 1x4 and screws at 24" o.c.
If coming from top it would come across plenty screw holes and places to leak.
No signs of leaking on ceiling under roof (3/8 ply)

There is 1 main leak... with the heavy rain we just had, it may drip in 3 or 4 locations within a 8' span.

Thanks smearcase