Insulation R-Values and Kneewalls
As part of my sealing and insulation project for my attic, I’ve been trying to determine what would be considered the proper r-value for insulating the knee walls. I’m assuming it’s somewhere between what would normally be recommended for an area’s walls and for an area’s attic, but that’s a lot of territory. This would be for a vented roof, and for my area of the country summer attic heat is as much a concern, if not more, as winter cold, at least for comfort’s purposes. Insulating the roof, rather than knee walls and attic’s floor, while usually preferable, is not a great option for this house for a number of reasons specific to this house.
So, for instance, I believe in my area, Zone 4, code would state you should R20 for walls and R49 for ceilings, what would you believe the R value should be for knee walls using an insulation goal comparable to the goal the code sought?
Of course, we can always go higher, but I’m just trying for purposes of comparison to get what people would consider comparable to that level of insulation. For others out there in warmer or colder climates that may have similar questions, would your advice vary? I ask because, I’m assuming normal insulation levels for walls would be sufficient in colder climates, but it’s really the extra heat in attics that differentiate knee walls from a normal exterior wall.
Thanks for your help!
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
Jeff,
Assuming that you are using the term "kneewall" correctly -- and you are referring to a short (vertical) wall that is between about 3 feet and 5 feet high -- then code requirements for exterior walls apply.
One side of the wall is interior conditioned space; and on the other side of the wall (if you are insulating the wall) the space is presumably unconditioned. So if you live somewhere where the code requires R-20 insulation for walls, that's the minimum for kneewalls, too.
For knee walls, why does code use the R-value for exterior walls and not attic floors? If they want R-42 on the floor in CZ-2, to keep hot attic air from migrating through the attic floor, why wouldn't they want that same R-42 on the knee walls, to keep that same hot attic air from migrating through the wall? To me, the knee wall is essentially a vertical floor. I must be missing something. What am I missing?
Ah, I guess I phrased the question incorrectly since I didn't really consider whether that was actually what the code would actually say. :) Assuming we should at least have R20 per the code, would you want to see more if its in an attic than just what you would put in an exterior wall given it might be 90 outside, but well above 100 in even the best ventilated attics on sunny days? I just assume that would warrant something more, despite the code.
The summertime delta-T of 100F in the attic, 75F in the conditioned space is only 25F.
In a zone 4A winter the difference can be 50F or more.
It's really about the wintertime performance, not summer.
And yes, R20 (or R13 + R5 continuous insulation) would still be the right amount for a code-min kneewall. If it's a retrofit with a 2x4 kneewall, R13- R15 in the stud bays with an inch of foil-faced polyiso on the attic side gets you over the R13 +R5 hurdle, but would also give you a bit of radiant-barrier boost from the foil facer.
Jeff,
Here at GBA, we have long advocated above-code levels of insulation. If you want to add more R-value, go to town.
As with any insulated wall, attention to airtightness is probably more important than R-value.
I am currently undertaking a similar kneewall project and can also vouch for the effectiveness of airtight rigid foam on the attic side (with batts in the cavities). The interior side heat reduction as I recently put up the foam board was increasingly noticeable, but as Dana mentions I am really hoping for winter time improvement when the temperature swings are even greater.
If you want to go beyond code min, consider adding 2X2s to the existing 2X4s to give you 5in of cavity space for R21 batts and then add the foam board (maybe 1.5in of foil faced with the R21).
Rather than furring out with 2x2s, furring out with foam insulation edge strips makes a much bigger difference. A 1.5" wide strip of 3/4" polyiso held in place with a 1.5" wide strip of 3/4" CDX (or 1x furring) doubles the R-value of 2x4 framing, and 5" deep cavity can fit a compressed R20, (which would deliver a true ~R18.3, slightly better than an R19 stuffed in a 5.5" deep 2x6 cavity.
Jeff,
There are a lot of issues here. In general, GBA does not recommend insulating kneewalls if there is an attic behind the kneewall. It's almost always a better idea to insulate the sloped roof assembly in the small triangular attic behind the kneewall, bringing the attic inside the home's conditioned space.
To learn why, read this article: “Two Ways to Insulate Attic Kneewalls.”
Dow Thermax polyiso has better fire ratings than most, and will usually pass muster with inspectors in a kneewall attic space.
If you insulate the kneewall without the foam board on the exterior side, it still needs SOME sort of exterior air barrier. (Kraft facers don't cut it for air tightness, and also run afoul of fire codes if left uncovered.) Even though wind-washing through rigid rock wool is low, stacked batts will always have some parasitic convection channels.
The foam edge-strips on framing trick really works. If you want to max it out, take 2" polyiso cut into 1.5" wide strips (tacked to the studs with dabs of foam board construction adhesive or cap-nailed in place) for a 5.5" deep cavity. Install R23 rock wool in the cavity, then long-scew the wallboard to the studs with 3" screws. With 2" polyiso edge strips the framing fraction comes in at R16+, and with R23 cavity fill it ends up at about R20 "whole-wall", which is better than code. It's also better than 2x4/R15 + 1" of continuous polyiso. But you give up a few square feet of interior space with that approach.
There's not much point to going higher than R20 whole-wall, which in a careful design is good enough to hit Net Zero Energy in a zone 4 location with a PV array that fits on the house, assuming other aspects of the house are similarly upgraded beyond code.
I see what you mean regarding the delta in temperature differences. I suppose what colored my thought on that was that, because of heat rising, keeping it warm in winter was never as much of a challenge as keeping it cool in summer. As much as I'd like to add rigid foam insulation, I think what just keeps holding me back is the code issue in this area that I might have to cover it with drywall as a fire retarding layer. My plan has been to insulate the wall cavities to the extent possible (in this case R-15), then build some extension to hold an additional rock wool batts either vertically or horizontally, then cover it with Tyvek as I've seen some insulation companies do. A backing barrier to hold it in place/prevent air movement is not required here, but I like the idea if even just for appearance. This left me a lot of choices on the batt sizes (an additional R-15 to R-30) I could be using, which led me to wonder what's just ridiculous overkill / waste of money. And would it be better saving that expense and just toss more on the attic floor.
I have a little different take on how you would decide on how thick to make the insulation. Code is based on some sense of whether the thicker insulation is worth the extra cost. Insulating an exterior wall to a higher R value costs more than insulating a knee wall to to a higher R value, because you have to deal with window and door openings and the details of making the exterior robust against rain and UV light. Maybe code should specify a higher R-value for knee walls for this reason. I would guess that the reason it doesn't is to keep things reasonably simple, both in the code and in the design and construction process. If you are doing this yourself, it's really a question of how much trouble you are will to go to to implement the suggestions given here--the material cost is pretty low. And the range of what would be reasonable R-values is as you originally suggested--greater than or equal to the code value for walls but less than the code value for the attic floor.