Calculator
Heat Loss Calculator
Two houses with the same 1,500 square feet of floor can need furnaces that differ by 10,000 BTU/hr, and the square footage tells you nothing about which one that is. What sets the heating load is the area of every surface that touches the outside — walls, ceiling, floor over a crawlspace, windows — weighted by how poorly each one resists heat, plus how much outside air leaks in and has to be warmed from scratch. A house with an undersized furnace and a house with a 30% oversized one can look identical from the street.
This calculator does the envelope arithmetic surface by surface: it divides each area by its R-value (or multiplies by U-factor for windows), multiplies by the temperature difference between inside and the coldest day you design for, adds a separate term for air leaking in through gaps rather than through the assembly, and sums the five numbers. That is different from a flat rule of thumb like 30 to 60 BTU/hr per square foot, which assumes an average envelope and cannot tell you that your windows, specifically, are half your loss.
Calculate your quantity
Exterior walls, minus windows and doors.
From the NFRC label. Lower is better.
cubic feet
Design heat loss
22,772 BTU/hr
Equipment size
28,465 BTU/hr
- Temperature difference
- 65 deg F
- Walls
- 6,000 BTU/hr
- Ceiling
- 2,566 BTU/hr
- Floor
- 5,132 BTU/hr
- Windows
- 4,160 BTU/hr
- Air infiltration
- 4,914 BTU/hr
- Loss per square foot
- 15.18 BTU/hr/sq ft
- Walls are the largest single loss at 26.3% of the total.
Shopping summary
- • A furnace or heat pump delivering about 29 MBH output at design conditions
This is an estimate — confirm structural work with a professional.
How this calculation works
Conduction loss through an opaque assembly is area divided by R-value, times the temperature difference: a wall's R-value describes how much it resists heat flow per square foot, so more area or a lower R-value both push more BTU/hr through it. Windows and doors are rated the opposite way, in U-factor, which is conductance rather than resistance — multiply by area directly instead of dividing. That is why the window field asks for a U-factor from the NFRC label rather than an R-value: entering one where the other belongs silently reverses the effect of a bigger number.
Air leakage is not conduction through a surface at all — it is outside air physically entering the house and needing to be heated from the outdoor temperature up to room temperature. The constant 0.018 is air's heat capacity: roughly 0.075 pounds of air per cubic foot times 0.24 BTU per pound per degree, which is the energy it takes to warm one cubic foot of air by one degree. Multiply by the house's volume, its air changes per hour, and the temperature difference, and you have the fourth loss term, one that a tight modern house and a leaky older one can differ on by a factor of four even with identical insulation.
- This is a design-day number, not an annual fuel estimate. It answers 'how many BTU/hr on the coldest morning I design for,' not 'how many therms will I burn this winter' — those are different calculations built on different inputs.
- The outdoor design temperature should be a 99% or 99.6% design-day figure for your location, not an average January low. Using the average understates the load on the mornings the furnace actually has to keep up.
- This sums conduction and infiltration across the whole building envelope. It does not model duct losses through an unconditioned attic, solar gain through south glass, internal gains from people and appliances, or room-by-room distribution — a Manual J calculation does that, and this is not one.
- The sizing margin is a single input, applied once. Adding it here and then buying the next size up from a contractor's own margin compounds into equipment that short-cycles instead of running steady.
The formula
Q = (wallArea/wallR + ceilingArea/ceilingR + floorArea/floorR) x deltaT + windowArea x windowU x deltaT + 0.018 x volume x ach x deltaT
- wallArea / wallR
- Net exterior wall area in square feet divided by the assembly's R-value. This is conductance: the higher the R-value, the smaller this term gets for the same area.
- ceilingArea / ceilingR, floorArea / floorR
- The same conductance calculation applied to the attic-side ceiling and any floor over an unheated crawlspace or garage. Ceiling R-value is usually the highest number on the page because attic insulation is the cheapest R-value to add.
- windowArea x windowU
- Window and door loss, using U-factor directly rather than R-value, because that is how glazing is rated on its NFRC label. A U-0.32 window is roughly R-3 — far below any wall it sits in — which is why glazing punches above its square footage.
- 0.018 x volume x ach
- The infiltration term. 0.018 BTU per cubic foot per degree F is air's volumetric heat capacity; volume is the heated interior space in cubic feet, and ACH is how many times that volume of outside air enters per hour through cracks, not through the assemblies above.
- deltaT
- Indoor design temperature minus outdoor design temperature. Every term in the sum is multiplied by this once, at the end — it is the single number that turns five conductances into a BTU/hr rate.
- oversizePct
- A margin applied to the summed design loss, not to any single term, to leave headroom for a cold snap beyond the design day without buying a furnace sized for the coldest morning in a decade.
Where these numbers come from
- 0.018 BTU / cu ft / deg F (air's volumetric heat capacity)
- Physics, not a code figure: air's density (about 0.075 lb/cu ft) times its specific heat (about 0.24 BTU per lb per deg F). It is the standard multiplier used throughout residential air-change load calculations, not something this page derived on its own.
- U-0.32 default window
- Manufacturer-published: a mid-range NFRC-labeled U-factor for a double-pane low-E window, the number printed on the sticker in the corner of the glass. Older single-pane windows run closer to U-1.0; high-performance triple-pane can be under U-0.20.
- R-13 wall / R-38 ceiling defaults
- Trade convention: roughly what a 2x4 wall with batt insulation and a well-insulated attic deliver in much of the country, used here as a starting point the field is meant to be overwritten with your own construction's actual values.
- 25% sizing margin
- Trade convention. Contractors commonly carry 15-25% above a calculated design load; more than that risks short-cycling equipment that never runs long enough to reach steady, efficient operation.
Worked examples
A code-minimum house on a cold-climate design day
| Wall area | 1200 sq ft |
|---|---|
| Wall R-value | 13 |
| Ceiling area | 1500 sq ft |
| Ceiling R-value | 38 |
| Floor area over unheated space | 1500 sq ft |
| Floor R-value | 19 |
| Window and door area | 200 sq ft |
| Window U-factor | 0.32 |
| Heated volume | 12000 |
| Air changes per hour | 0.35 |
| Indoor design temperature | 70 |
| Outdoor design temperature | 5 |
| Sizing margin | 25 % |
| Design heat loss | 22772 BTU/hr |
|---|---|
| Equipment size | 28465 BTU/hr |
| Temperature difference | 65 deg F |
| Walls | 6000 BTU/hr |
| Ceiling | 2566 BTU/hr |
| Floor | 5132 BTU/hr |
| Windows | 4160 BTU/hr |
| Air infiltration | 4914 BTU/hr |
| Loss per square foot | 15.18 BTU/hr/sq ft |
At a 65 degree design difference, the five terms land within a fairly narrow band of each other: walls lose 6,000 BTU/hr, floor 5,132, air infiltration 4,914, windows 4,160, ceiling 2,566. No single surface dominates here, which is what a reasonably balanced envelope looks like — the fix for a house shaped like this is a smaller improvement spread across several surfaces, not one big obvious target.
28,465 BTU/hr with the margin applied lands this house in furnace nameplate territory around 30,000 to 36,000 BTU/hr input, once you account for a modern condensing furnace's efficiency rating rather than shopping on the design number directly. That step — output design load to input nameplate rating — belongs to whoever is selling the equipment, not to this page.
A glassy, well-insulated house in a milder climate
| Wall area | 1000 sq ft |
|---|---|
| Wall R-value | 21 |
| Ceiling area | 1200 sq ft |
| Ceiling R-value | 60 |
| Floor area over unheated space | 1200 sq ft |
| Floor R-value | 30 |
| Window and door area | 350 sq ft |
| Window U-factor | 0.5 |
| Heated volume | 10000 |
| Air changes per hour | 0.25 |
| Indoor design temperature | 70 |
| Outdoor design temperature | 20 |
| Sizing margin | 20 % |
| Design heat loss | 16381 BTU/hr |
|---|---|
| Equipment size | 19658 BTU/hr |
| Temperature difference | 50 deg F |
| Walls | 2381 BTU/hr |
| Ceiling | 1000 BTU/hr |
| Floor | 2000 BTU/hr |
| Windows | 8750 BTU/hr |
| Air infiltration | 2250 BTU/hr |
| Loss per square foot | 13.65 BTU/hr/sq ft |
Everything about this envelope is better than the first example except the glass: R-21 walls, R-60 ceiling, a tighter 0.25 ACH, and a 45 degree design difference instead of 65. Despite that, the total loss is only 28% lower, because 350 square feet of U-0.5 windows loses 8,750 BTU/hr on its own — more than the walls, ceiling and floor combined, and more than double what the first house's smaller, better-rated windows lost.
That is the case where upgrading insulation everywhere else buys little until the windows are addressed. A homeowner who reinsulates the attic here without touching the glazing is spending money against a surface that was never the problem — the design loss barely moves, because ceiling loss was already small relative to the windows next to it.
Common mistakes
- Entering an average winter low instead of a design-day temperature. The furnace has to keep up on the coldest reasonable morning, not on an average day — using the average understates deltaT and undersizes the equipment for the day it matters.
- Putting an R-value into the window U-factor field, or vice versa. Windows are rated in U-factor on their own label; a bigger U-factor number means more loss, the opposite of a bigger R-value. Reversing the two silently flips whether a change makes the number better or worse.
- Applying the sizing margin twice — once in the oversizing field here, and again when a contractor rounds up to their own next equipment size. Two margins compound into equipment too large to run a full efficient cycle.
- Using this design-day BTU/hr figure to estimate an annual heating bill. Fuel cost depends on heating degree-days over a whole season, a completely different calculation with different inputs — this number describes one cold morning, not a winter.
- Skipping the infiltration term because it feels like the hardest input to know precisely. A rough ACH estimate from a blower-door-test rule of thumb is far closer to reality than leaving it at a default that assumes a tightness your house may not have.
- Treating this as a substitute for a room-by-room Manual J when buying equipment. It sums the whole building; it says nothing about which room is starved for supply air, which is a duct-design question this page does not answer.
Shopping summary
There is nothing to put in a cart here — the output is a design number to bring to whoever sizes and quotes your heating equipment, and to use as the starting BTU/hr figure for the duct size calculator, which needs a heat delivery target per room rather than a whole-house total.
If the worked examples above make windows or a specific R-value look like the dominant loss in your own house, the fix is a real product change, not a bigger furnace: upgraded glazing, or more insulation sized against the insulation calculator for the surface that is actually losing the most. Chasing a bigger furnace instead of the envelope just moves the same problem to a higher monthly bill.
If your goal is a comfortable room rather than a whole-house heating system — sizing a window AC or a single mini-split head for one space — the room AC BTU calculator is the right tool instead of this one; it works from room area and adjustments rather than summing an envelope, and it is built for cooling a single space, not heating a whole house.
FAQ
Why is this number so different from the BTU rating already printed on my existing furnace?
Existing equipment is very often oversized from the day it was installed, sometimes deliberately as a margin, sometimes because a past contractor rounded up without running a load calculation at all. A furnace nameplate tells you what was bought, not what the house actually loses on a design day — this calculator answers the second question.
How is this different from the room air conditioner calculator on this site?
They diverge on purpose, and it matters if you're specifying multiple ductless mini-split heads for one house rather than a single central system: a whole-house furnace or heat pump sized on this page shares its capacity across every room at once, while several independent room units, each sized on the BTU/sq ft page with its own margin and its own worst-case room, commonly sum to a higher total capacity than one system sized to the whole envelope together. If you're specifying multiple heads, this page's whole-building number is worth running anyway as a sanity check against the sum of the individual room figures.
What outdoor design temperature should I actually use?
A 99% or 99.6% heating design temperature for your specific location — available from ASHRAE climate data or a local energy code table — not your area's average January low. The design temperature is deliberately colder than average because the furnace has to keep up on the cold mornings, not on a typical one.
Does a heat pump need a different design number than a furnace?
The design heat loss itself is the same physical number regardless of what heats the house. What changes is what you do with it: a heat pump's output falls as outdoor temperature drops, so its capacity at your design temperature — not its rated capacity at a mild test condition — is what has to meet this number, and backup heat covers the gap on the coldest days. That balance-point sizing is equipment-specific work for whoever specs the heat pump.
I doubled my ceiling R-value from 30 to 60 and the total barely moved — is the calculator wrong?
That is the arithmetic working correctly, not a bug. R-value has diminishing returns: going from R-13 to R-26 cuts a surface's loss in half, but R-30 to R-60 also halves it, off a much smaller starting number. If ceiling loss was already a small share of the total, halving it again barely moves the sum — the surface worth upgrading is whichever one is still large in absolute BTU/hr, not the one with the lowest R-value.
Is this accurate enough to buy a furnace against on its own?
No, and it says so above the tool. This is an envelope calculation — conduction and infiltration, summed — not a full Manual J, so it leaves out duct losses through unconditioned space, solar and internal gains, and room-by-room airflow. Use it to sanity-check a contractor's number or to see which surface dominates your loss; equipment selection and final sizing is a professional's calculation.
Where to go next
The projects this number is a step of, the guides that explain the method behind it, and the rest of its trade group.