Furnace size calculator - heating load first, then the input bin.

How the heating load and furnace size are calculated.

The calculator runs the heating side of the whole-house Speed Sheet simplification of ACCA Manual J 8th Edition. It is the same code the BuildSolver assistant uses for a first sizing pass. Transmission, infiltration and duct losses are computed at the 99% heating design temperature and summed. Solar gain and internal gains are not credited: the house has to hold temperature on a cold night. The tool then picks a furnace input bin and a heat pump size from that total. Infiltration uses the air density at the ZIP's elevation where Census and USGS data pin that elevation down. Otherwise the result states that no altitude correction was applied.

Heating total

Q_heat = transmission + infiltration + ductΔT = 70°F indoor − 99% heating design dry-bulb for the ZIP. Every component uses the same ΔT.

Transmission

Q = Σ(U × A × ΔT)Walls, ceiling, windows, doors, and floors over a crawlspace or basement. U = 1/R from year-built × climate-zone defaults. BuildSolver assumes glass at 15% of floor area and 40 sq ft of doors at U 0.40. No CLTD multipliers.

Slab edge

Q = F2 × perimeter × ΔTSlab floors only (Manual J Table 5C). The page assumes no edge insulation, F2 = 1.0 BTU/h·ft·°F, and a square footprint: perimeter = 4 × √(floor area ÷ stories).

Infiltration

Q = 1.08 × CFM × ΔT; CFM = ACH × floor area × 8 ft ÷ 60BuildSolver's default winter ACH is 0.30 for tight, 0.55 for average and 1.0 for leaky. Scaled by the air-density ratio where Census and USGS data pin down the ZIP's elevation (about 0.83 at 5,000 ft); otherwise the result says no altitude correction was applied. Sensible only; humidification is not modeled.

Ducts in attic

Q_heat = (transmission + infiltration) × 1.20Applied only when ducts run through an unconditioned attic. Otherwise duct loss is zero.

Furnace input bin

smallest of 40K, 60K, 80K, 100K, 120K ≥ Q_heat ÷ AFUERounded up to the next input bin. AFUE is not on this form, so BuildSolver applies its own year-built default. It uses 80% before 2000 or when the year is unknown, 90% for 2000 to 2014 and 95% for 2015 and later. If the required input is above 120,000 BTU/h, no bin is returned.

Heat pump at 47°F

smallest of 18K, 24K, 30K, 36K, 48K, 60K ≥ Q_heat ÷ derateStandard split heat pump. Derate at the design temperature is interpolated from the tool's table: 1.00 at 47°F, 0.90 at 35°F, 0.72 at 17°F, 0.55 at 5°F, 0.45 at −5°F, 0.35 at −15°F and colder.

Capacity at design temp

Q_design = heat pump bin × derateWhat the selected size delivers at the design temperature. BuildSolver caps a single residential split at 60,000 BTU/h (5 tons). Above that at 47°F, it sets the dual-unit flag and returns no heat pump size.

Small loads

strip kW = Q_heat ÷ 3,412, rounded up to 0.5 kWBelow 12,000 BTU/h the tool returns an electric heat strip size instead of a furnace or heat pump bin. Above that, no aux heat kW is returned.

Dual-fuel flag

design dry-bulb < 17°F and zone 4A, 4B, 5, 6, 7 or 8Set from the ZIP's climate zone and design temperature only.

A worked example for a 2,000 sq ft two-story in Chicago.

2,000 sq ft, 2 stories over an unconditioned basement · ZIP 60614 · built 1995, average air tightness, 4 occupants, ducts inside conditioned space · window type from the year-built default. The old furnace is coming out and the homeowner wants a number today.

The load is the heat output the house needs. At 80% AFUE the tool rounds the required input up to the 80,000 BTU/hr bin. A higher-efficiency furnace can land in a smaller bin: give the real AFUE in chat. At -8°F a standard split heat pump keeps only a fraction of its 47°F rating, so no single residential unit covers this load. That is why the heat pump line is empty and the dual-unit flag is set. The dual-fuel flag is set because the design temperature is below 17°F in a cold zone.

Standards and data sources.

• ACCA Manual J 8th Edition - residential load procedure; the page uses its whole-house Speed Sheet simplification, which BuildSolver estimates lands within 15-20% of a full room-by-room calculation.
• NREL TMY3 weather data - outdoor design conditions as percentiles of typical-year hourly data (1% cooling dry-bulb with mean coincident wet-bulb, 99% heating dry-bulb) - the page applies the worst case of the ZIP's climate zone. Official city values are in ASHRAE Handbook - Fundamentals, Ch. 14.
• IECC 2021 climate zones (PNNL county map) - ZIP → county → climate zone; the zone sets design temperatures, solar factor and envelope defaults.
• ASHRAE Fundamentals Eq. 9.5 / 9.6 - sensible (1.08) and latent (4840) infiltration constants, with a standard-atmosphere density correction where Census and USGS data pin down the ZIP's elevation.
• USGS 3D Elevation Program (3DEP) - ground elevation at the 2020 census centers of population of the tracts overlapping the ZIP (block groups for tracts larger than about 10 sq mi), averaged by population in the ZIP's land area, for the air-density correction.
• ACCA Manual J 8th Edition, Table 5C - slab-on-grade heat loss by exposed perimeter (F2 coefficient).
• AHRI 210/240 and NEEP cold-climate air-source heat pump data - the 47°F and 17°F rating points and the capacity derate table the tool interpolates for a standard split heat pump.
• ACCA Manual S - equipment selection against manufacturer performance data. It is a separate step after the load; this page does not do it.

Limitations of the quick estimate.

• The load is a whole-house average, so it cannot size individual supply registers or find one hot bonus room.
• Design temperatures come from the worst case of the ZIP's IECC climate zone, so they can run more extreme than the ASHRAE values for the nearest station, hotter for cooling and colder for heating. BuildSolver estimates the gap at 2-8°F.
• Insulation levels and, unless you pick one, window type come from year-built × climate-zone defaults. Real upgrades since construction change the answer a lot.
• AFUE is not an input on this page. The tool uses 80%, 90% or 95% from year built. If the replacement furnace has a different rating, the input bin can change. The chat accepts an AFUE from 80% to 98%.
• Furnace bins are the 40K-120K series in 20K steps. Some manufacturers sell intermediate inputs that fit the load closer.
• The heat pump is sized as a standard split unit. Cold-climate heat pumps hold far more capacity at low temperatures. The chat can check one at your design temperature, and matching it to manufacturer data (Manual S) is on the Pro and Team plans.
• Slabs are assumed to have no edge insulation and a square footprint. Insulated slab edges or a long, narrow house change the floor loss.
• Heat pump backup is not sized. Aux heat kW is returned only for loads under 12,000 BTU/h.
• When the altitude correction applies, higher elevation lowers the infiltration loss for thinner air, and attic duct loss drops with it because it is figured as a share of that loss. The furnace's own high-altitude input derate is not applied: above roughly 2,000 ft, follow the manufacturer's installation instructions for the burner.
• The residential scope is 100-20,000 sq ft and 1-3 stories. Light-commercial, multifamily and VAV systems need a different procedure.
• Preliminary sales-phase estimate. Not ACCA-approved software output and not accepted for permit submission.
• Elevation is the population-weighted average for the ZIP, not the building's own. A house far up or down a slope in a hilly ZIP sits at a different elevation. BuildSolver treats the air-density correction as small below about 2,000 ft. Where the data cannot pin a ZIP's elevation down, no altitude correction is applied and the result says so.

Common furnace sizing mistakes.

1. Matching the old furnace nameplate. The furnace coming out was often picked by rule of thumb decades ago, before new windows, attic insulation or air sealing. A like-for-like swap carries that oversizing into the new furnace's whole service life.
2. Sizing by BTU per square foot. A per-square-foot rule ignores design temperature, insulation, glass and air leakage. Run through this calculator, a leaky 1952 crawlspace ranch in Cleveland with attic ducts comes out at 4.6 times the heating load per square foot of a tight 2022 slab house in Kansas City.
3. Input and output BTU mixed up. The load is output. Divide it by AFUE to get the input you order. An 80,000 BTU/h input furnace at 80% AFUE puts about 64,000 BTU/h into the ducts.
4. Ignoring AFUE on a change-out. Replacing an 80% furnace with a 95% unit of the same input adds heat output. Size the new input from the load and the new AFUE.
5. An extra size for safety. An oversized furnace short-cycles and puts extra starts on the heat exchanger. The larger blower can also push more air than the existing ducts handle.
6. Sizing a heat pump only at 47°F. The 47°F rating is the best case. Capacity falls as it gets colder, so check what the unit delivers at the design temperature and plan the backup heat for the difference.

Common questions about furnace sizing.

Related calculators.

• Manual J calculator The load method behind this page, with both cooling and heating framed side by side.
• Replacement AC sizing calculator Same load engine, framed for the cooling side of a change-out.
• HVAC design temperatures The 99% heating and 1% cooling design values by ZIP code.
• Duct sizing calculator Check whether the existing ducts carry the airflow of the new equipment.
• Manual S equipment selection Next step after the load: check a specific air conditioner or heat pump. Pro and Team plans.
• How to size a heat pump Balance point, capacity at design temperature and backup heat.
• Methodology Standards and data sources behind every BuildSolver tool.
• All tools Every public BuildSolver calculator in one list.