GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter X Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Plus Icon Minus Icon Picture icon Hamburger Icon Close Icon Sorted

Community and Q&A

Ice & water UNDER or OVER the drip edge?

user-1135248 | Posted in Energy Efficiency and Durability on

I did a little searching before asking this, thinking that the topic
*must* have come up in the past, but didn’t find anything coherent.
We’re having this bit of disagreement between a GC and a roofer as
to whether ice & water shield over decking should flash over or
under the drip-edge piece at the connection to the fascia. The
roofer’s claiming that putting the drip-edge metal over the I&W
isn’t a reverse flash because water in an ice-dam situation would
flow upward, and the GC [who I’m tending to agree with] points out
that a small roof leak [or, say, inevitable condensation runoff
under metal roofing panels] would hit the I&W and come down safely
over the drip-edge and be ejected instead of soaking into the
plywood underneath.

*Note*: this is a situation where the ice&water was trimmed to
stop just shy of the edge of the deck, leaving a little lip of
the deck and/or the shadow boards at the rakes exposed. This
deck is the over-roof structure atop a roof&walls polyiso retrofit,
where everyone seems to assume the taped foam underneath is yet
another water control layer. [Hint: it isn’t, because tape laid
down in a hasty fashion has numerous fishmouths.] The GC’s guys
must have wanted to leave a “clean slate” of nailable wood for the
roofer, but personally I thought not leaving a little one-inch flap
of I&W *over* the edge was crazy — especially when it took another
month for the roofer to get to the site, leaving only the full wrap
of I&W on the deck *as* my roof in the interim [with occasional water
leaking into the soffit space, I might add, which didn’t seem to
concern them too much]. The question arose when it finally came
time to finish off these details and git ‘er done.

It seems to me that you’d want some way to flash against *both*
potential scenarios, which we kinda wound up doing by putting the
drip-edge over the I&W but then rolling out Triflex slip-sheet
under the [metal] roofing that lapped over the drip-edge piece.
But Triflex or even generic roofing felt isn’t really supposed
to be a water barrier, as I understand it, esp. since it isn’t
self-adhesive at all and gets *stapled* down.

_H*

GBA Prime

Join the leading community of building science experts

Become a GBA Prime member and get instant access to the latest developments in green building, research, and reports from the field.

Replies

  1. GBA Editor
    Martin Holladay | | #1

    Hobbit,
    I answered this question on the Letters page of issue #226 of Fine Homebuilding magazine:

    "Grace, the manufacturer of Ice & Water Shield, advises that its peel-and-stick membrane should be installed before the drip edge. There are several reasons for this. If ice ever backs up from a clogged gutter, the ice can get under the drip edge and reach the plywood. If the Ice & Water Shield goes down first, the plywood is better protected. Also, if the galvanized drip edge ever rusts, it is easier to replace when reroofing if it is installed on top of the Ice & Water Shield. When it comes to synthetic roofing underlayments in warm climates, the same rule applies, but for different reasons. Water can wick uphill on top of the drip edge, so it's better to have the drip edge on top of the underlayment."

  2. Expert Member
    Michael Maines | | #2

    Recently I went to a training session put on by W.R.Grace. The presenter agreed with what Martin says above, but also said that they also recommend putting a strip of membrane over the dripedge, as shown here: http://www.na.graceconstruction.com/template.cfm?page=/underlayments/dd_iws102.html&did=8.

    The standard installation is here: http://www.na.graceconstruction.com/template.cfm?page=/underlayments/dd_iws101.html&did=8.

    And additional good information is here: http://www.na.graceconstruction.com/template.cfm?page=/underlayments/roofing_faqs.html&did=8#roofdrip.

  3. user-1135248 | | #3

    I like the detail with the two layers, that seems like the right general
    answer as water could conceivably come from two different
    directions. I'll note that the "alternate" detail sends at least one
    layer of Grace over the edge of the deck, protecting the plywood
    from water arriving from above. Unfortunately that wasn't the
    case with my reno, as the builders inexplicably cut the IWS back
    to slightly *above* the edge of the deck and subfascia, allowing
    water to reach the edge of the ply and get down behind the fascia
    metal trim itself. Any ideas on why they would have done that??

    _H*

  4. Expert Member
    Michael Maines | | #4

    Are you sure it was cut back, and not just installed a bit high? If it was cut, it could be that the bottom edge got gummed up on itself, or had puckers, or they overhung the roll and were sloppy about cutting it back. Installing I&WS properly is a real art that many carpenters and roofers have not learned.

  5. user-1135248 | | #5

    While this is departing a bit from the original question, it seems
    appropriate to clarify the situation a bit. Since this was a DER
    with foam over the exterior of original roof and walls sheathing,
    an over-roof nailing plane was built up over the foam with fairly
    large overhangs. That entire top plywood surface was Graced, but
    for the most part matching the lower edges as it was rolled out.
    In spots where it overhung the plywood, it was actually trimmed
    away to match the subfascia and rake ladder surfaces supporting
    it. I watched as they took a cordless skilsaw and did the same
    all along both eaves, too. Hopefully I can post valid links to
    pictures -- this one shows one such trim cut behing made:

    http://techno-fandom.org/~hobbit/hse/log/oGeWNzGx/ren/1249trimrakeply.jpg

    This one shows the IWS on a return detail. The IWS, plywood, and
    shadow board attached to the fascia basically all end at the same
    point:

    http://techno-fandom.org/~hobbit/hse/log/oGeWNzGx/ren/1286lkick-ret1.jpg

    As the white trim metal got placed onto that same detail, it created
    a situation where water could come down the I&W and drop right behind
    the trim metal -- look just above the guys's right hand here:

    http://techno-fandom.org/~hobbit/hse/log/oGeWNzGx/ren/1288lkick-cornfail.jpg

    Basically, all the edges of the over-roof deck sat like that for a
    *month* before the roofer finally arrived and started applying drip
    edge. Water would get down behind the metal, run onto the perforated
    soffit pieces *on top*, travel toward the house, and get into the
    "rainscreen" space under the siding that was eventually applied over
    top of all the vertical strapping. Since taped foil-face is actually
    *not* necessarily watertight or flashed to the exterior, this could
    [and I think did once or twice] bring water in toward the structure.

    Ironically, in the third picture the guy is trying to work that little
    downward bend in the soffit trim that I posted about a couple of
    months ago -- this is about the point where he gave up trying to make
    the 90-degree corner work and went back to a regular L-bend. Even that
    wouldn't have helped, as water getting behind the metal can jump
    onto the perforated plastic pieces before reaching that point.

    Does this strike anybody else as poor technique, aka a massive reverse
    flash even if they thought it was going to get handled later?? It would
    have been so simple to leave an inch or two of IWS hanging over this
    whole mess, even with some release paper still attached so the roofer
    could choose what to do with it later, but have the structure protected
    in the meantime. Even a half-inch of I&W hanging over the metal and
    then all that buried under the drip-edge would have been far better.

    _H*

Log in or create an account to post an answer.

Community

Recent Questions and Replies

  • |
  • |
  • |
  • |