In 1990 I was provided generous funding by CertainTeed Corporation to study attic ventilation. The study continued for about 10 years. We learned a lot from that study, which I will report on, later. I just want to share with you what was the major finding from the study. We began with a long building extending east-west in Champaign IL which contained eight study bays, each with a different attic construction. Each roof had a 5/12 pitch with OSB roof sheathing. We had white shingles and black, organic felt shingles and fiberglass, cathedral ceiling and truss-framed flat ceiling, vent devices and no vent devices, fiberglass batt, fiberglass blown-in and cellulose insulation, hole in the ceiling and airtight ceiling. The interior was maintained at 70 degF in winter and 75 in summer. Normally, in winter, we kept the indoor humidity at 50%–a pretty high humidity. Basically we wanted to know which factors, or combination of factors would lead to damage, which we considered to be mold discoloration of the underside of the sheathing. And what did we find?
No mold. None. Zero. OSB was clean as can be.
Not in 10 years, not until the building was abandoned and roof leaks let the rain in. So. If you live in Champaign IL and have OSB on your roof, there’s no combination of the factors listed above that will lead to mold damage. So. Do whatever you want. Vent, don’t vent, put holes in the ceiling or not, use any insulation, use any framing, do anything you want. This research says you’ll be fine and mold-free, for the first 10 years anyhow.
Well, that’s pretty poor research design. You’ll notice there was no null hypothesis. A good research design might have gone like this:
- Anticipate an unsatisfactory outcome, say with a hole in the ceiling and no venting and 50% indoor RH.
- Set up instrumentation that has a good chance of measuring the unsatisfactory outcome.
- Posit the null hypothesis, that the combination does not lead to the unsatisfactory outcome.
- Wait for the data, and for the example of unsatisfactory outcome. If it takes more than 10 years for the problem to arise, find out why.
- When it happens, then reject the null, and conclude that the combination does indeed lead to the bad outcome.
An appropriate outcome of the research would have been a showing that doing A leads to no problem while doing B leads to a problem. In that case you can recommend A and not recommend B. But in the absence of a performance threshold, we learn that A is drier than B (or cooler than B, or better than B by some factor), which would lead people to prefer A, when B remains as a perfectly fine option, with no black marks against it. Takeaway for future researchers: in the research plan, begin with a well-defined performance threshold and measure to that. Simple relative values are not, and should not be, actionable. They are fun, I can vouch for that. They may not be allowed to affect policy.
But we did learn a lot, which I’ll be describing in future posts.
- Fiberglass attic floor insulation can have a density so low that the warm air at the bottom of the insulation mass bubbles upward just from natural buoyancy of warm air. You may imagine the impact of that on the thermal insulating value.
- You can’t cool the top of a cathedral ceiling with venting. It’s the warmth of the air in the slot which carries it up and out by buoyancy. The air gets warmer as it nears the top of the vent slot. You can’t cool a warm roof with warm air, the top anyway.
- Venting cools shingles, a little. How much? Adding venting to a roof has the same cooling effect as moving that roof about 100 miles to the north. People from Chicago ask me if they need venting to cool the shingles. I tell them to order their shingles in Champaign IL, 100 miles to the south, and any shingle which works fine on a vented roof in Champaign will work just fine in an unvented roof in Chicago.
- Infrared radiant exchange with the sky is very important. Roofs can be much colder than the surrounding air. Using only outdoor temperature, indoor temperature and total solar radiation measurements, we can get to only 92% agreement between predicted and actual temperature measurements. Add in infrared radiation, and the agreement jumps to 98%.
- Does attic venting save on cooling energy? Don’t know. Our results showed that if there were benefits they stayed in the noise range, not in the signal range. Later research by others suggests attic ventilation carries an energy penalty in the South (Zone 2).
- Organic fiber shingles suck. Good riddance.
- The resistance to air flow of a vent device is quite independent of its rated net free area. You might think that the resistance to air flow of a vent device is the same as the resistance offered by a hole of the net free area size. Think again.
- Roof surfaces go from quite hot, thanks to the sun, to quite cold, thanks to radiant cooling to outer space. Cold things tend to be wet while hot things tend to be dry. The daily wet-dry cycle of a roof is wide. And that daily swing tends toward the dry side. Attics are hot so it’s pretty easy to keep them dry.
- We had three cathedral ceiling types, stuffed, vented and slotted. Vented roofs had slots, with a vent at the low and high end; slotted roofs had the slot without the vents. Slotted roofs performed almost like vented, and unlike stuffed. There’s a buoyancy effect at work here, but it is not stratification.
- Duff sensors are little matchstick ends that are charged with voltage, and they operate as variable resistors depending on their moisture content. Thermocouples are temperature sensors that register the microvoltage produced with two different metals in contact. Don’t put Duff sensors and thermocouples together with salty wet cellulose. I’ll let you imagine the resulting circuitry.
- Vented attics are 30 degrees F cooler than unvented attics on a sunny afternoon. Otherwise, they’re roughly the same temperature.
- In a vented attic, wind is a stronger driving force than buoyancy, by about 20 to 1. Makes you wonder about the calls for high and low vents.
- Of course there were no ice dams. The only heat available to melt snow was heat coming from below through R-30 insulation. That can make, at most, a couple of gallons of water from ice, over the whole project, in 10 years. And besides, there were no valleys which act as funnels, channeling the water, and ice, and icicles to the mouth of the valley.
- Cool roof? White shingles. Hot roof? Dark shingles. Duh.
Leave a Reply