Study: Vented vs Unvented Roofs

evan Filed Under: Tags: , , , , ,

Things are moving along at 4600 Augusta Ave: all of the windows and doors are in, and we’re quickly approaching the time to insulate. Pinning down the best method for insulation has been a surprisingly difficult process… while building science reports have increasingly advocated tight full-home conditioned envelopes using unvented roofs and conditioned attics, we’ve found contractors to be generally skeptical of the spray-foam technique, citing the lack of ventilation along the sheathing as cause for concern about shingle temperature and degradation. Shingle manufacturers themselves have expressed confidence in their products in the assembly; however, given the degree/prevalence of contractor pushback we decided to throw this one over to the R&D department here at R. Michael Cross | Design Group, and enlisted the invaluable help of Guy DuBois of Commonwealth Building Sciences for a sort of case study to examine the effect of underside ventilation (or lack thereof) on roof surface temperature.

To break this down a little, I should explain that vented attics were mandated by code for years as a means of preventing roof damage resulting from trapped moisture. In this method, insulation gets put down between the second floor ceiling joists, vents are installed in the eaves and gable ends, and the attic is left essentially unconditioned.  There are a few main thoughts behind the original mandate: preventing the buildup of moisture during the colder months by allowing it to escape, and preventing ice damming at the eaves in the winter by keeping the entire roof cold. Dominion Power further explains:

At some point, the original purpose for attic ventilation was forgotten and/or replaced, in cooling climates, with the belief that it was to reduce roof and attic temperatures, thus lowering cooling expenses and increasing shingle life. This was further compounded by the leap of faith that increased or powered ventilation would be even better. While sounding logical, there is simply no research to validate it. In fact, scientific testing has shown that attic ventilation has almost no effect on roof surface/ shingle temperatures and very little effect on attic temperatures.

 While Dominion’s research would seem to answer our question… we decided to move forward anyway. Because we’re scientists, dammit, and we need some of that good empirical stuff.

The key to unvented roof assemblies and conditioned attics is managing the temperature of the first condensing surface either with rigid insulation on top of the sheathing, or with an impermeable barrier of spray foam insulation adhered directly to the underside of the sheathing. In the case of rigid insulation, this keeps the condensing surface at the roof sheathing sufficiently warm (45°F) to minimize or prevent condensation. In the case of applied spray-foam, this moves the condensing surface from outside of the insulation (the roof sheathing) to the inside of the spray-foam, which is essentially self-insulated by the rest of the foam behind it.

This method of keeping the condensing surface below significant insulation also allows the exterior roof surface to remain relatively cool, preventing ice dams. Furthermore, the absence of direct attic ventilation helps minimize uplift pressure in high wind events.

To study the surface temperature performance of unvented roof assemblies compared to unvented roof assemblies, we constructed two 2’-0” x 4’-0” mockup sections of roof, one with two inches of spray-foam applied directly to the sheathing, one without.

From there we set the roof panels up at a 12:12 slope and set up matching heat lamps (250 watts). Guy spent some time calibrating the whole set-up, making certain both panels received equal heat exposure. After heating both panels for about 15 minutes, the heating lamps were turned off and initial surface temperature readings were taken.

We found that both panels had a surface temperature within 1°F of 181°F. We chose to continue measuring at intervals over the next half hour to see if the panels rejected surface heat at the same rate. 

Ultimately, the panels performed almost identically, with the unvented panel cooling slightly slower, presumably because of the foam’s thermal mass. Measurements taken behind the panel showed an expected but marked difference in hot spots, with the insulated panel performing significantly better. 

We admit that this investigation could have been more controlled, and that the panels may not exactly reflect the performance of an actual enclosed roof assembly; however this was a helpful exercise in essentially pointing towards a confirmation that unvented roofs won’t melt your shingles off. At the end of the day, we believe the positives of an unvented spray-foam assembly present a better solution for residential roofs in most situations.

We wish we had somewhat more conclusive results, and may still pursue further tests on these panels. In fact, we’d love to hear any of your suggestions and ideas for testing methods, or even general questions that you don’t think we have addressed. You can also email us at



| edit post

The Augusta Project @ VCA

R. Michael Cross Filed Under:

Quick news
The Virginia Center for Architecture has invited the R Michael Cross | Design Group to speak about the Augusta Project: Richmond's First LEED Registered Home as part of their program series supporting their current exhibition The Green House: New Directions in Sustainable Architecture and Design. We're flattered by the invitation and very excited to participate in this series.
The presentation is scheduled as a lunchtime lecture on November 24th from noon until 1pm and it is our great pleasure to announce that lunch will be provided by the fine people at Ferguson. Mark your calendars!


| edit post