One small observation: When I had no insulation in my attic, the attic temperatures remained consistently close (within 20-25F) of the room temperatures underneath (separated by R-1 gypsum board). When I had R-38-ish insulation, that delta was 40+ F. So more insulation slowing/rejecting the heat transfer downwards could increase the average attic temperature (because it can’t escape in the other directions any faster) could maybe indirectly make your rooms underneath the R-10 warmer. Maybe. But R-10 is still R-10. My other observation was that attic air temperatures can be quite variable depending on where in the attic you measure, and the surface temperature of the insulation at any point may be more susceptible to radiant heat from stuff above and around it than said average air temperature. And then there’s attic ventilation, and HVAC duct leakage that is highly localized.

From a practical perspective, I have R-45 in some of my attic and R-1 in the rest. The HVAC mixes the air inside my house so having the partial R-45 is still a net benefit and doesn’t seem to make the uninsulated rooms “worse” if that’s the root of your question..

@Jonathon
I will speculate that although fluid dynamics apply to air just as much as water, the heat transfer in question here is not so much by convection, but by conduction (and radiation). Also, the air pressure is not building up in the attic. So the heat transfer through the R-10 area is likely related only to the air temperature above it and any radiant heat hitting it if you ignore air leaks for a moment. Is it possible that having R-40 elsewhere increases the temperature of the attic air and the roof deck because less of that heat is transferred to the conditioned area? Possibly although who knows if this effect is significant.

There are some other great blog posts explaining how the heat flows. The heat will flow through the R-40 section, and the heat will flow through the R-10 section (at a faster rate because by definition of R-10). But the heat will also continue to “flow” in all the other directions. Up back to the roof. Out the roof /gable vents. Onto the ducts. Onto the HVAC equipment and other stuff that’s in the attic. Etc.

PS I am no HVAC expert nor trained. Just participating in your thought experiment.

@Jonathon, maybe this will help: your garden hose analogy doesn’t hold water 🙂

Restricting the hose with your thumb does in fact change the flow rate. That’s how the faucet valve works, and why low-flow shower heads work.

In the case of thermal equilibrium, surface temperatures largely determine the heat flow through an assembly of a given R-value (resistance), so as long as the rooms are maintained at the same temperature, the R-40 ceilings have no effect on the heat flowing through the R-10 ceilings (ignoring the effects of 3-D heat flux at the boundaries).

My brain has two conflicting conclusions:
1. Insulation that has u=0.1 is u=0.1 insulation, whatever the insulation around that area.
2. the thumb-on-hose analogy is not perfect but contains a lot of truth, particularly that with the same [[pressure]] of heat wanting to expand/equilibriate/rise, if you add R-40 to part of the attic, then more heat will rise thru the R-10 area than when the whole attic was R-10.

Help?

@Deniz, regarding light can covers… The poly needs to be cut. Technically, it needs to be sealed back to cover but the integrity of a ceiling vapor barrier isn’t nearly as important as the air barrier. My recommendation was to caulk the cover TO the ceiling, as an air seal, and secondarily as an adhesive to prevent shifting (and thus air barrier failure). Foaming over the flange accomplishes the same thing.

My XPS box reference was intended as a DIY solution. In fact, many air sealing details ultimately become DIY projects, unless you’re fortunate enough to live in an area where multiple expert air sealing subs are competing for business.

Regarding LED refrofit fixtures… caulking them (instead of gasket) sounds like it would work.

Regarding roof vents… couldn’t they be sealed from the top? Keep in mind, you don’t have to “air seal” the attic to the outside, only reduce the leakage enough so that the fan can reach 50 pascals. And if the fan doesn’t hit 50, there’s a formula or table for extrapolating the results (the house blower door fan must be adjusted to match the max pressure achieved by the attic fan). You’re primarily interested in the ratio between ceiling leakage and total house leakage. BTW, be sure to unseal the roof vents before measuring the reference total house leakage.

Regarding sprinklers… Here’s something more recent: Air Sealing Fire Sprinkler Systems.

@David,

My installer sloppily set the Tenmats (not flush) on the plastic sheets that cover all the gypsum board attic floor, made big slits for wiring, and then sloppily sprayfoamed the circumference. Their sorry explanation was that it was for insulation purposes, not air sealing. Ha!

You mentioned caulking – do you mean caulk the circumference to the gypsum (I should have cut away the plastic sheeting)? Or do you mean using caulk underneath the flange as an adhesive sealant?
Energystar’s Rule Your Attic video show expanding spray foam being used on the perimeter. Tenmat’s own installation video and install instructions do not show or mention caulking or foam sealant, they seem to have more of an “It’s the thought that counts” attitude.

I can assume mine aren’t airtight because some of my cans (while off) are barely visible whereas some have obvious hot areas when viewed from the ceiling side via IR camera.

energystar video
https://www.youtube.com/watch?v=UTODEUGgKbs

tenmat video
https://youtu.be/yZXHw20q8bo

What was upsetting is that my installer asked for $250 per XPS box, times 34 cans. I assume each must be DaVinci-grade artwork for that price.

I still think the $7.50 LED retrofit kit with foam seal might be airtight “enough” and a viable DIY alternative. Or perhaps I’ll just caulk the trim to the ceiling – there you go, why not? – those things last 50,000 hours anyway and even if you replace them, the trim from the next one will hide the ugly.

Using blower door/duct tester to test ceiling airtightness is a great idea but impossible in an existing home like mine. You can’t possibly reach every O’Hagin vent to seal it because crawling around is near impossible with all the ductwork and low pitch. Would love new construction to require this though.

I found a reference to Jeff Hugo’s article from the 2012 Nov/Dec edition of the National Fire Sprinkler Magazine titled, “Caulking Fire Sprinklers for Energy Code Compliance,” but the article doesn’t seem to be available online.

> My question about higher rock wool mass was – can it actually be DETRIMENTAL, because the higher mass absorbs more heat (?) during the day and then gives it off at night.

In general, thermal mass in a building enclosure is beneficial, not detrimental. Not all the heat absorbed by the buidling’s mass ends up being released into conditioned space. A lot gets released back into the attic, more than half if the diurnal swing is favorable. Also, I think you’d be surprised at how little heat is actually absorbed. Also, even a higher density product like rock wool is still mostly air.

> The Tenmat covers are actually not as easy to really install airtight as I thought.
I assume you’re calking them to the ceiling? I find that DIY 1-inch XPS boxes are easy to make and install, and they work really well. You can size the box to each fixture so as to minimize the amount of displaced insulation, and the boxes themselves have an R-value of 5, which makes up a little for of what’s lost due to displacement.

> I wonder if a ceiling-side LED retrofit kit with a foam gasket wouldn’t have done the same air sealing job much easier.
I haven’t tested that but I’m not sure the clips would be strong enough to create a true seal. OTOH, with surface mount fixtures, wherever possible, you can easily seal off the original hole.

> I’m curious if there is a good way to test / measure the all-important ceiling airtightness…
This can be done with a blower door and duct tester. To get the attic to 50 pascals with the small fan, you’d probably need to seal the O’Hagin vents. Also, let the tech know the size of your attic access hatch so he can make a mounting surface for the fan (typically cardboard). If your attic access is a regular door, it would be easier to use 2 blower doors, but it could still be done with a duct tester fan.

> some things like fire sprinkler heads and bathroom fan dampers/flaps can’t be easily made airtight after the fact.
I recall reading an article about air sealing sprinkler heads maybe 10 years ago. You can probably find with with a Google search. Some bath fans are easier to air seal than others.

@David You are right, I misstated when I wrote ’11″ thickness of R-45 insulation all the way.’ instead of ‘as far as it will go before tapering’. With no foil at the end, I can get maybe 5″ thick insulation at the bitter end, which extends slightly past the wall plate, so by the time I’m over the wall plate I may be averaging 7″ of thickness already. I understand now that maxing out insulation here is much better than just having 3.5″ for the last few feet.

In fact this was one of the reasons for choosing rock wool: higher R value per inch of thickness helps at the edges of my low pitch roof (I need 10.75″ of rock wool vs 13″ fiberglass to achieve R-45). Plus I can taper the rock wool nicely to fully fill the shape of the eave vs fiberglass I’d have to stuff (I’m avoiding blown-in because of IAQ issues plus it wouldn’t stay put nice and flat). From my DIY experience, rock wool seems easy to work with, to cut to shape for even coverage, and the spongelike material makes it easy to layer in alternating directions to fully cover the joists. Probably more labor intensive than fiberglass batts but it feels like it would result in better distribution of insulation. I don’t remember the math for lost R-value due to the joists, but this way I’ve got guaranteed R-30 still above them to make me feel good. Except those darn eaves.

My question about higher rock wool mass was – can it actually be DETRIMENTAL, because the higher mass absorbs more heat (?) during the day and then gives it off at night. I certainly wish it weren’t as heavy as I’m hitting the limit of how much weight gypsum board can carry.

I had considered adding soffit vents but unsure about the cost, and their potential value in absence of a ridge vent. In another blog post, I read that a certain balance between high/low vents is desirable, maybe 50/50 or 40/60, so I’ve shied away from randomly opening up holes for gable vents as well.

The Tenmat covers are actually not as easy to really install airtight as I thought. I wonder if a ceiling-side LED retrofit kit with a foam gasket wouldn’t have done the same air sealing job much easier. I’m curious if there is a good way to test / measure the all-important ceiling airtightness, since blower doors don’t isolate ceiling leakage, and some things like fire sprinkler heads and bathroom fan dampers/flaps can’t be easily made airtight after the fact.

Thank you for this discussion.

@Deniz, there’s no harm in having insulation contact with the already-installed RB at the edges, but you should remove (or simply slice) the radiant barrier in those areas as you definitely don’t want to give up the additional insulation depth where the foil would prevent that. Again, there’s no reason not to allow insulation to contact roof deck over the plates. Of course, if you later decide to add soffit vents, you would need to insert baffles.

Regarding the additional mass of rock wool… I’ve seen no studies that demonstrate any additional thermal benefits due to the higher mass of rock wool. I would be skeptical of any claims to that affect, and in any case, I can’t see the rationale for the considerable cost of rock wool in this application.

Regarding light cans and other interference…. aside from displacing insulation, light cans are never air tight even if they’re “AT” rated, and getting a tight ceiling is probably more important than improving your insulation coverage. The best way to achieve that is to cover and seal the fixtures with air-tight caps (e.g., CanCap, Tenmat Recessed Light Cover, diy XPS box, etc.). Or you may want to consider replacing some of your light cans with surface-mount LED fixtures.

BTW, the reason I thought you might have raised heel trusses is that you indicated you could get 11″ depth over the wall plate. I don’t see how that’s possible if rafters bear on outer edge of top plate, which is how regular trusses are typically designed.