R-values add up. U-factors don't.
Heat crossing a wall goes through every layer in turn, so the layers act like resistors in series. Total thermal resistance is the plain sum of each layer's R-value. Siding, sheathing, cavity fill, drywall and the thin films of still air on each face all count, and the order doesn't change the total.
The U-factor is the reciprocal of that finished sum, U = 1 ÷ R, computed once,
at the end. It is the rate of heat flow per unit area per degree, and energy codes outside
the United States are usually written in it.
The trap is averaging or adding U-factors layer by layer. U is a reciprocal, so summing each layer's U gives a number bigger than the U of any single layer. That says adding insulation makes the wall worse, and no reading of that result means anything. Sum the R-values. Invert once.
The R-value printed on the batt is not the R-value of the wall
A roll marked R-19 gives you R-19 through the fiberglass. Through the stud next to it, you get nothing close. Softwood framing lumber is about R-1.25 per inch, so a 3.5-inch stud is R-4.4, a quarter of the cavity beside it, and that stud runs unbroken from the drywall to the sheathing.
Framing also takes up more of the wall than people picture. Count it properly, with plates, corners, headers and the framing around openings, and a wall built at 16 inches on center is typically 25% framing by area. Advanced framing at 24 inches with insulated headers brings that down to roughly 18%. Ceiling joists are nearer 11%.
Handle it with the parallel-path, or area-weighted, method. Compute the U of the cavity path, compute the U of the framing path, weight the two U-factors by their share of the area, then invert. If you area-weight the R-values instead, you get a higher answer, and a wrong one, because what holds per unit of area is heat flow, and heat flow tracks U.
A standard 2×4 wall with vinyl siding, OSB, an R-13 batt, drywall and both air films sums to a nominal R-15.5. The framing path is R-6.9. Weighted at 25%, the wall performs at R-11.8. That 24% is far bigger than a rounding error, and almost nothing on the first page of search results accounts for it.
Insulation R per inch varies more between products than between categories
Insulation is sold by R-value but manufactured to a conductivity, and the R per inch that comes out of it differs a lot:
| Material | R per inch | RSI per inch (25.4 mm) |
|---|---|---|
| Blown fiberglass, attic | 2.2 to 2.7 | 0.39 to 0.48 |
| Fiberglass batt | 3.1 to 3.4 | 0.55 to 0.60 |
| Cellulose, loose or dense-pack | 3.5 to 3.7 | 0.62 to 0.65 |
| Open-cell spray foam | 3.6 to 3.8 | 0.63 to 0.67 |
| EPS board | 3.6 to 4.2 | 0.63 to 0.74 |
| Mineral wool batt | 3.9 to 4.3 | 0.69 to 0.76 |
| XPS board | 4.7 to 5.0 | 0.83 to 0.88 |
| Polyisocyanurate board | 5.6 to 6.5 | 0.99 to 1.14 |
| Closed-cell spray polyurethane | 6.0 to 6.8 | 1.06 to 1.20 |
| Softwood framing lumber | 1.25 | 0.22 |
| Dense concrete | 0.08 | 0.014 |
The ranges come from the U.S. Department of Energy insulation fact sheet and the ASHRAE Handbook of Fundamentals, chapter 26. Treat them as starting points. That's why every R per inch in the calculator is an editable field and not a hidden constant. Density, aging, mean temperature and the manufacturer all move the number. Polyiso in particular loses R at low temperature, which is the wrong direction on a cold night.
R-value, RSI and the factor of 5.678
R-value is imperial: h·ft²·°F/BTU. RSI is metric: m²·K/W. Both describe the same physical property, and they differ by 5.678. RSI 1.0 is R-5.678. R-30 is RSI 5.28.
Because of that factor, a Canadian label reading RSI 3.5 and an American label reading R-20 are the same product, and reading an RSI figure as an R-value understates a wall by nearly six times. This tool shows both at once, plus the metric U-value in W/m²·K, so you can't mix them up. The conversion comes from the exact definitions of the BTU, the foot and the degree Fahrenheit, so the round trip closes.
Air films and cavities are free R-value that people leave out
A thin layer of still air clings to every surface and resists heat on its own. ASHRAE puts the inside film at about R-0.68 on a wall, R-0.61 on a ceiling with heat flowing up and R-0.92 on a floor with heat flowing down. Direction matters because convection helps a rising flow and fights a falling one. The outside film is about R-0.17 in wind and R-0.25 in still air.
An unvented cavity between 20 and 100 mm adds roughly R-1.0. Give one face a reflective foil and it can get close to R-2.8, because most of the heat crossing an air gap moves as radiation and foil won't emit it. So the foil does nothing by itself. A radiant barrier pressed flat against a surface has no air space to work with and adds almost nothing.
On a well-insulated wall these items are a small share of the total. On a bare wall they can be most of it, and that is exactly when leaving them out skews the comparison you're trying to make.
What the IECC climate zones ask for in R-value
The prescriptive R-values in the 2021 IECC climb with the zone, and they climb much faster for ceilings than for walls:
| Zone | Ceiling | Wood-frame wall | Floor |
|---|---|---|---|
| 1, Miami, Honolulu | R-30 | R-13 | R-13 |
| 2, Houston, Orlando, Phoenix | R-49 | R-13 | R-13 |
| 3, Atlanta, Los Angeles, Dallas | R-49 | R-20 | R-19 |
| 4, Baltimore, Seattle, St. Louis | R-60 | R-30 or R-20+5ci | R-19 |
| 5, Chicago, Denver, Boston | R-60 | R-30 or R-20+5ci | R-30 |
| 6, Minneapolis, Burlington | R-60 | R-30 or R-20+5ci | R-30 |
| 7 and 8, Duluth, Fairbanks | R-60 | R-30 or R-20+5ci | R-38 |
Ceilings jump to R-49 in zone 2 while the wall stays at R-13. Geometry explains it better than climate does: an attic floor is flat, cheap to reach and has unlimited depth, so the code puts the R where the R is easy to get. The wall is expensive per unit of R.
Your jurisdiction may still enforce an older edition (many are on the 2018 or 2015 IECC), so read these as the direction of travel and confirm what your building department has adopted. The target field in the calculator stays editable for that reason.
Insulate the attic first, and seal it before you insulate
In a house with an uninsulated or under-insulated attic, the attic is almost always the best dollar you'll spend. Going from R-11 to R-49 in a ceiling removes about 78% of the conducted loss through that surface. Getting the same fractional improvement from a wall means opening it up.
Attic work is only as good as the air sealing under it, though. Warm air leaks upward through can lights, top plates, plumbing chases and the attic hatch, and air-permeable insulation, loose fill and batts alike, doesn't stop that flow. It just filters it. Blow R-49 over an unsealed ceiling and you get much less than the R-value suggests. The dirty streaks you see in old fiberglass are that air passing through.
Doubling the R-value does not double the saving
Heat loss through an assembly is proportional to U, and U is 1 ÷ R. So every added increment of R takes a smaller slice of what's left:
- R-5 to R-10 removes half the loss.
- R-10 to R-20 removes another quarter of the original.
- R-20 to R-40 removes another eighth.
- R-40 to R-80 removes another sixteenth, a huge amount of material for almost nothing.
Each step is a doubling. Each one costs at least as much as the last and returns half as much. The calculator shows the next identical step next to the current one so you can see the curve flatten. At some point on it, your money does more in air sealing, in the windows or in a better heating system, and that point arrives far sooner than the marketing suggests.
What a series R-value calculation can't tell you
- Air leakage. In an older house it is often a comparable share of the heating load, and an R-value can't see it at all. R measures conduction only.
- Installation quality. A batt that is compressed, gapped or cut short around wiring loses a substantial share of its rated performance. Grade I installation is an assumption, never a given.
- Steel framing. Steel conducts hundreds of times better than wood, and simple parallel-path badly overstates it. You need ASHRAE's zone method or published correction factors.
- Thermal mass and solar gain. A steady-state R-value says nothing about how a heavy wall delays a heat pulse, or about sun hitting a roof.
- Moisture. Where the dew point lands inside your stack decides whether the assembly dries or rots, and that is a different calculation entirely.
Everything stays in your browser
Every figure is plain arithmetic done on your device. Nothing is uploaded, nothing is stored, and there's no account. The math lives in a separate module that is checked against hand-worked examples outside the browser, including the parallel-path wall above, so you can verify the numbers instead of trusting them.
Frequently asked questions
Do you add R-values or U-factors when you stack insulation?
Add the R-values. Heat crosses the layers one after another, so the resistances add in series like resistors: total R is the sum of every layer. The U-factor is the reciprocal of that finished total, U = 1 ÷ R, and you calculate it once, at the end. Adding up the U-factors of the individual layers is the classic mistake. It gives a U larger than any single layer's, which would mean adding insulation makes a wall worse. That can't happen physically, and it still turns up in spreadsheets and in more than one calculator.
Why is my R-19 wall not actually R-19?
The studs aren't insulation. R-19 describes the batt between the framing, but softwood lumber is about R-1.25 per inch, and framing takes up roughly 25% of a typical wall built at 16 inches on center once you count studs, plates, corners and headers. Heat takes the easy path through the wood. Weighted by area, a nominal R-15.5 assembly with R-13 batts delivers about R-11.8 in practice, a 24% loss. The calculator works out that parallel path for you instead of just mentioning it.
What is the difference between R-value and RSI?
Same property, different units. R is imperial, in h·ft²·°F/BTU. RSI is metric, in m²·K/W. One R equals 0.1761 RSI, so RSI 1.0 equals R-5.678. R-30 attic insulation is RSI 5.28, not RSI 30. Read one as the other and you are off by a factor of 5.7. That is an order of magnitude, far past rounding, and it is why a product sold as RSI 2.5 in Canada and R-14 in the United States is the same product.
How much R-value do air films and air gaps add?
More than most people expect, and they cost nothing. The still-air film on an inside wall surface is about R-0.68, on a ceiling with heat flowing up about R-0.61, and on a floor with heat flowing down about R-0.92. A windy exterior surface is about R-0.17. An unvented cavity of 20 to 100 mm adds about R-1.0, and the same cavity with one reflective foil face can reach about R-2.8. On an uninsulated wall these make up most of the total resistance, so leaving them out can understate a bare assembly by half.
If I double the R-value, do I halve my heating bill?
You halve the heat lost through that one assembly. The bill is a different number. Heat loss is proportional to U, not to R, so doubling R from 13 to 26 cuts the loss through that surface in half. Doubling again from 26 to 52 removes only half of what was left, a quarter of the original. The same added R buys half the result the second time around. And that assembly is only one of several losses: windows, air leakage and ventilation don't change at all.
Does this work for steel studs?
Not accurately, and you should know that up front. Steel conducts roughly 400 times better than wood, so the simple parallel-path method overstates a steel-framed wall by a wide margin. A cavity filled with R-13 can perform closer to R-6 or R-7 overall. ASHRAE uses a modified zone method or published correction factors for steel framing. Use the tool for wood framing, masonry and continuous assemblies, and treat any steel-stud number it gives you as an optimistic ceiling, never the answer.