Testing Spray-Foam Jobs
The article "Insulating Unvented Attics with Spray Foam"
(3/07) was very good, as far as it went. We have been using
spray foam in unvented roof assemblies for nearly 10 years,
overall with good results.
One thing to keep in mind, however, is that spray foam is not
foolproof. It's possible to install it badly — in fact,
in my experience, careless installations are surprisingly
common. Because the material is so good at air-sealing where it
is applied, the spots you miss can have an exaggerated impact
on the performance of the house. Even if you manage to seal 95
percent of the potential leakage points with the spray foam
— outstanding performance, by the standards of insulation
installations in general — you can cause real problems as
warm moist indoor air or freezing cold outdoor air gets
funneled into the 5 percent of the remaining leakage
The solution should be easy: Have your insulation contractor
test the house with a blower door just as his crew is finishing
up. They'll find the remaining leakage points when it's very
easy to fix them. The problem is that, at least in my region,
there's not a single insulation contractor who owns a blower
door. To me, this is analogous to a carpenter who doesn't own a
level or a tape measure.
So our company bought our own blower door. When we started
testing our projects, I was truly alarmed at the mistakes we
discovered that our insulation subcontractors had been making,
despite their best intentions.
At $2,500 to $3,000, a blower door is not a trivial
expenditure. Possibly you could team up with a few other
companies or a local builders' association chapter to share
ownership to reduce the cost. If you're truly serious about the
performance of your projects, though, and about quality
construction in general, it's a question not of if but
when you will make sure you have ready access to
Ridge Beams and Collar Ties
I wanted to make a few comments regarding the article
"Retrofitting a Structural Ridge" (2/07). This was a typical
retrofit of a structural ridge directly below the existing one,
and the major design factors were considered. Unfortunately,
one of the most important design aspects — the connection
— was neglected.
The author was correct in pointing out the importance of tying
the rafter to the ceiling joist at the top-plate location.
However, his observation that a structural ridge was needed
because collar ties were being eliminated was incorrect. Collar
ties have absolutely no relation to the strength of a ridge.
The IRC clearly states that as soon as the ceiling joist
— or rafter tie — leaves the top plate, a ridge
beam is required (R802.3.1).
In a standard roof system, the critical connection is between
the rafter and the ceiling joist. In that situation, all of the
roof loading is being transferred at the connection by shear
forces on the nailing. With a structural ridge, however, a
large portion of the roof loading is being "held" at the ridge
to prevent the walls from bowing outward. Surely the new ridge
will support the old ridge and the imposed loading.
The problem is the connection point of the rafters to the old
ridge. This connection is now under both shear and tension
stresses, whereas the existing through-nailing is ineffective.
In new construction this is dealt with by using a connector
such as a Simpson LSSU; in remodeling there are other
connectors that could have been used.
I was surprised to see that this point was not mentioned in
the article. Will this roof fail? Probably not. Is this a
correct ridge design? Absolutely not.
Thomas H. Wojick
Structure Tech LLC
Author Kipton Tewksbury responds: Mr. Wojick's comment
that collar ties have no relation to the strength of the ridge
seems to contradict the IRC (R802.3.1, 2006 edition), which
says, "Where ceiling joists are not connected to the rafters at
the top wall plate, joists connected higher in the attic shall
be installed as rafter ties, or rafter ties shall be installed
to provide a continuous tie. ... Where ceiling joists or rafter
ties are not provided, the ridge formed by these rafters shall
be supported by a wall or girder designed in accordance with
accepted engineering practice."
We could not clearly determine how the ceiling joists were
attached to the rafters because of the attic floor. According
to our engineer, verifying and strengthening the rafter-joist
connection was one of the things we could have done instead of
installing the ridge beam. But because of poor access and our
wish not to disturb the ceiling below, we opted for the ridge
beam, which the engineer clearly thought was a sound approach
and met code as he interpreted it.
Note that we also added bearing knee walls on both sides of
the roof, with connections to the rafters that exceeded code.
Both of these knee walls were directly over — or close
enough to — interior walls below, which carried loads to
the ground in the basement.
The engineer and I spent considerable time verifying the load
path for the knee walls and the support posts. We even opened
up part of the downstairs wall framing to visually verify a
solid, continuous load path. And as the article stated, we
added structural elements at both gable ends to spread the load
over the walls below.
Had we used rafter-to-ridge connectors, we would have used
Simpson H1s or H11s. Because of the narrowness of the space, no
other connector could be securely fastened — and even the
H1 would have presented access problems.
We also could have invented a new tie, similar to a collar
tie, to straddle the ridge beam and connect the opposing
rafters, with a flange bent down on either side of the ridge
beam to allow fastening. This would have required separate
engineering costs because it would not have been covered by the
code. Ultimately, the engineer did not specify additional
connections at the ridge.
We gave all of this considerable thought; the design was the
result of collaboration between builder and engineer and our
collective knowledge and skill.
It was Sunday evening, and as I opened my February issue of
JLC, it was if some strange power led me directly to the
picture leading off the article on 16-gauge finish
It was as if someone were trying to remind me of my foolish
behavior from only a day earlier, when, unlike the carpenter in
the picture — who uses a block of wood to align the
material being nailed — I had ignored my better judgment
and instead chose to align a corner using my thumb while
nailing. The photo shows the result. Kudos for showing the
right way to do it.
Glen Allen, Va.
Regarding Jason Seltin's otherwise clear article "Using an
Airless Paint Sprayer" (3/07), I think he's misinformed on
back-rolling. While 2,000-psi delivery pressures "should" embed
paint in the work surface adequately, all surfaces are not
The question is which surfaces and conditions warrant
back-rolling. Particularly with new rough-sawn siding or
shingles, the paint is often too thick to soak in properly
before it dries. This can lead to a paint job that looks great
when it's fresh — but two years later any pro can see the
shadowy effects of inadequate coverage. Repainting a house that
has a good original paint job can often be done without
back-rolling, but I'd be careful.
Ben Lomond, Calif.
Leave a Gap in Subfloor
The article "Framing the First-Floor Deck" (12/06) contains
good tips for floor layout and framing. We applaud the use of a
glued-nailed floor system and the care that this crew shows for
One photo shows a worker with a sledgehammer beating the wood
structural panels together "so that the T&G joints will
close." Because it's not clear from the text what "close"
means, I wanted to remind readers that it's important to leave
a 1/8-inch gap along the top edges to allow the panels to
expand to reach their equilibrium moisture content.
Sometimes it's necessary to coax panels together, but forcing
the tongue too deep can damage the panels. Plus, panels that
are tightly butted together may buckle and telegraph through to
the finished flooring, especially with vinyl, thin carpet, or
APA/Engineered Wood Association