GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter X Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Plus Icon Minus Icon Picture icon Hamburger Icon Close Icon Sorted

Community and Q&A

Block foundation wall in zone 5 – when do I reach diminishing returns?

baxt1412 | Posted in Energy Efficiency and Durability on

I’m in zone 5 and have a concrete block foundation in a 100 year old house.  I’m wanting to insulate the walls on the inside of the **unconditioned** basement and plan on using rigid foam insulation.  I do not plan on finishing the basement but know that for code for most rigid foams I’ll have to cover it with something fire rated.  I’ll look into options that will be cheapest and that I can paint white to help improve brightness.   

My question is: when do I reach the point of diminishing returns in terms of r-value per dollar? I’m sure that’s going to be somewhat subjective but I guess to put into reference of an old house that the majority of has nearly nothing in the exterior walls.  I’ll be addressing the attic and rim joists in the basement as well. 

For what it’s worth, If I’m reading our code right here, it is r-13 for basement walls.

GBA Prime

Join the leading community of building science experts

Become a GBA Prime member and get instant access to the latest developments in green building, research, and reports from the field.


  1. baxt1412 | | #1

    Bump - happy to clarify if there's anything that my ramblings don't make sense of

  2. Expert Member
    NICK KEENAN | | #2

    If you think of a house as a six-sided cube and calculate the heat loss for each side, it's easy to see that the insulation level of one side doesn't affect the heat loss of the other ones. The same is true of real houses, the insulation level of one part of the house doesn't affect the heat loss of other parts. So the fact that part of the house is poorly insulated doesn't affect the break-even point for other parts of the house. I'd say the fact that your basement is unconditioned is irrelevant, you're heating that basement whether you want to or not if it's connected to the rest of the house. Calculating the exact break even point requires knowing a lot more about your exact situation, but I'll say this: modern codes are made with a lot of thought toward cost-effectiveness, so R15 is probably pretty close to where you want to be if that's your code.

    The one exception to this rule is air infiltration, air leaking into one part of a house can make heat loss worse in others. You really want to try and make your house as tight as you can. In the basement the area to focus on is the rim joist and the area between where the house meets the foundation to the bottom of the floor. A block foundation may be somewhat air leaky as well, but if you apply board foam and then tape between the pieces of foam that should provide a good seal.

    1. Expert Member
      MALCOLM TAYLOR | | #3


      While the heat loss from one part of a house may not affect the heat loss the rest of it, an un-insulated component may significantly reduce the overall R-value of the house.

      What would the overall R value of a cube be where one face was R-2 and the remaining five faces were R-50?

      1. Expert Member
        NICK KEENAN | | #6

        Let's say you have two cubes of the same size. One has five faces that are R-50 and one that is R-2, the other has six faces that are R-2. Increasing one R-2 face on either cube to R-50 saves the same amount of energy. Yes, proportionately the two cubes are very different, but when you're looking at paybacks it's absolute amounts that matter not percentage changes.

        Think of how you do a Manual J: you add up the wall area and divide by R-values. One poorly-insulated wall doesn't make the well-insulated walls perform worse. We usually focus on the poorly-insulated parts of a house, because that's where the highest return is, but that doesn't mean there aren't returns to be had in the well-insulated parts.

        1. Expert Member
          MALCOLM TAYLOR | | #7

          Right but if you had a cube with one face at R-2 and the rest at R-50, there would be little point in messing with anything but the R-2 one. Increasing the other faces to R-100 would still do very little to improve the overall R-value of the assembly. You are simply losing too much heat through that one weak point.

          1. Expert Member
            NICK KEENAN | | #8

            I'm afraid you're missing my point.

            Let's say the heat lost through the R-2 wall costs $50 per month to replace. So for the six sides it's $300/month to heat the building. At R50, that same wall costs 1/25 as much, or $2 a month, and at R100 half of that, or $1 a month. So for any given wall the decision on whether to go from R2 to R50 is whether the cost of doing so is worth the long-term value of $48 a month, and the decision to go from R50 to R100 is whether the cost of doing so is worth $1 per month. So yes, adding insulation suffers from diminishing returns, adding the same amount of insulation to an R2 wall has 48 times the return as adding it to a an R50 wall. And if the cost of adding insulation to every surface is the same, then it always makes sense to add it to the poorest-insulated surfaces first.

            But the point is that each wall is independent of each other. The "overall R-value of the assembly" isn't an important measure at all. If going from R50 to R100 is cost effective, it doesn't matter that another wall is at R2. And in real buildings it often happens that all of the surfaces aren't the same. The most common example is adding pour-in insulation to an attic. Once you've gone to the trouble of getting the insulation delivered and figured out a way to get it up into the attic it costs almost nothing to slice open a few more bales and spread them around. Conversely, a situation we run into here in DC all the time is that the front façade of a building is protected by historic preservation and can't be modified. OK, leave it alone. But it still makes sense to insulate the attic, and if you're throwing in a few bales of loose insulation, it makes sense to throw in a few more.

          2. Expert Member
            BILL WICHERS | | #9

            You're both right (Malcolm and DC). DC is correct that the insulation in one area doesn't have an affect on the insulation in another area. Malcolm is correct that there is a lot more to gain by insulating the weak spots with low R value than the ones with more.

            R value is not an additive scale for absolute energy loss. R value is additive for the R value itself, but is essentially a logarithmic scale for absolute energy loss. This is the same basic way the decibel scale works if you're familiar with that.

            The easy way to explain this is that if R3 cuts your heat loss by half, and you add another R3 to that for a total of R6, the first R3 cuts your heat loss by half to 50% relative to what you started with, then the second R3 cuts that remaining 50% by half again to 25% for a total reduction of 75% compared to what you started with. I used those numbers for a reason since in the decibel (dB scale), 3dB is half, 6dB is 1/4 (remaining), 9dB is 1/8, etc. The "what you started with" is defined as 0dB, so +3dB is twice what you started with, -3dB is half what you started with.

            The nice thing about a log scale is that you can add the units of whatever you are measuring with, just keep in mind that the energy flow IS NOT LINEAR and drops off essentially proportionally with each additional step, which would essentially be with each additional layer of insulation in this case.

            In the R value example, adding an additional R3 to a wall with only R3 makes a much bigger difference (25% additional reduction in energy loss) compared to adding R3 to a wall that already has R10 (in which case the extra R3 only makes a 5% additional reduction in energy loss). "10" in the dB scale is 10 times, so it's another "magic" number. If you start with R30 (1000x), that extra R3 only makes a 0.05% difference in total energy loss through the wall.

            As you increase the R value, each additional increase results in less and less reduction, as a percentage, in additional energy savings on a project. This is how you get into that "diminishing returns" area. This is also why building science people always talk about air leaks loosing more energy than low R value insulation -- an air leak BYPASSES all the insulation, so it's like a loophole in those R value calculations!


    2. baxt1412 | | #4

      Definitely addressing air tightness in the basement and attic when I get to it. I plan on doing rigid foam with one part foam around it for the rim joists. And will likely even caulk between the rigid foam boards for the wall and a high quality tape as well. Figure if I'm going to do it once, do it right. I will caulk where the rigid foam touches the floor as well. Any benefit in doing a small strip of foam on the face of the floor with adhesive below it and on the face that faces the wall (Almost like a furring strip made of EPS on the face of the floor) rather than just extending the wall foam down to the floor and caulking below it?

  3. Expert Member
    BILL WICHERS | | #5

    Code now is R15 for basement walls in CZ5. I saw a graph some time ago, I believe it was from an agency or university in Minnesota, that pretty much showed the point of diminishing returns for basement wall insulation is right around R10, which used to be code.

    If you don't have to pass inspection, R10 is a good amount of basement insulation, and it's easy to install (2" of XPS). If you want to do a little beter, use 2" of polyiso for about R13. If you use Dow Thermax or equivalent, you don't need to worry about covering the interior since it's already rated to be left exposed. If you install polyiso like that using the "pound in" plastic insulation anchors, it's a relatively quick and easy installation.

    If you need to meet code, you need 3" of XPS, about 4" of EPS, or 2-1/2" of polyiso. If you use XPS or EPS, you need a thermal barrier on the interior side.


    1. baxt1412 | | #12

      I was going to shy away from polyiso because of it's poorer performance in colder temperatures. Is it not that big of a difference?

  4. user-723121 | | #10

    Heat loss for a basement wall is calculated using above grade and below grade dimensions. The portion of above grade foundation has a straight heat loss like walls and the below grade area loses more heat near the soil surface and less farther down in the soil. ASHRAE has quite a detailed explanation on determining foundation heat loss and it includes soil types in the calculation. The MN Code task force did a lot of study on foundation walls and they bumped the R value up to R-15. R-10 is not much insulation in a cold climate. There was time in MN you could cut the foundation insulation down to R-5 if you installed a high efficiency heating system. One of the trade offs wrangled by the production home builders at the time.

  5. norm_farwell | | #11

    The $ ROI for above grade masonry walls in a cold climate is very good. For below grade it is very poor, because the delta T between interior temp and ambient ground is minimal.

    Fwiw, when I did the energy modeling on my own basement (climate zone 6) some years ago, I was surprised to find that insulating the uninsulated slab floor would save me the grand sum of $13/yr. But beyond energy benefits, there’s the issue of condensation which insulation will take care of, so there’s potentially benefit in comfort/health/ durability.

    1. baxt1412 | | #13

      Sounds like I should be focused on my attic more. Still hoping I get some at least noticeable comfort changes in the basement. I think the air leakage improvement from walls and rim joists will be huge I hope. I have a finished conditioned bathroom in there with an undercut door and when I close it, I can feel significant drafts coming under it. The regular basement door from my kitchen to my basement I have tightened up with weatherstripping and a tight door sweep so that has helped there.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |