Method, sources and limits
Every number on this site comes from one tested module, and the pages are generated from it at build time. The worked examples in the prose and the answers the calculator gives are produced by the same code, so they cannot drift apart.
The calculation, step by step
- Pick the per-square-foot figure. Each DOE climate zone has a published band for cooling and for heating. The insulation setting selects a point inside that band: a poor envelope sits at the top, a good one at the bottom. It selects within the band rather than multiplying on top of it, so the result can never leave the published range.
- Multiply by floor area. Linear.
- Scale by ceiling volume. The bands assume 8 ft. A 10 ft ceiling multiplies the load by 10/8.
- Apply solar exposure. Up to 15% on cooling, and a small credit on heating.
- Add internal gains. 600 BTU per occupant past the first two, and 4,000 BTU for a kitchen in the conditioned area. Cooling only.
- Select equipment. The smallest standard size at or above the load, with the percentage overshoot reported rather than hidden.
Sources
- Climate zones. DOE and IECC climate zone definitions, zones 1 to 8, assigned by county in the published map. This site aggregates them to state level and says where a state spans more than one.
- BTU per square foot bands. The conventional published ranges used in US residential practice, organised by zone. No proprietary or copyrighted table is reproduced.
- Adjustment factors. The standard set: ceiling volume, envelope quality, solar exposure, occupancy and kitchen gain.
- Equipment sizes. The nominal capacities ductless, ducted and furnace equipment is actually sold in.
- A ton of cooling. Exactly 12,000 BTU per hour, by definition.
What the method cannot see
These are the limits, stated plainly rather than buried:
- Your actual construction. Three insulation settings stand in for every wall, window and roof assembly there is.
- Orientation and window specifics. One sun setting stands in for glazing area, orientation, shading coefficient and overhangs. On a house with a lot of west glass this is the largest single source of error.
- Air leakage. Assumed by envelope quality rather than measured. A blower door test is the only way to know.
- Duct losses. Not modelled at all. Ducts in an unconditioned attic can add a large fraction to both loads.
- Your design temperature. Climate zones are broad, and a house at the edge of a zone behaves like its neighbour.
- Equipment capacity at temperature. Nominal ratings are at standard conditions. A heat pump delivers less than its nameplate on a cold night.
This is a square-foot estimate, not an ACCA Manual J load calculation. A Manual J uses your building’s real construction, orientation and air leakage, and on a well-sealed house it usually comes out smaller than this. Use this to sanity-check a quote, not to replace one. What a Manual J measures.
Corrections
If a figure here is wrong, it is a defect worth fixing. The arithmetic is covered by a test suite that asserts known answers, band monotonicity across zones, and that furnace output always covers the load it was sized for.
Frequently asked questions
How accurate is a square-foot HVAC sizing estimate?
It is a planning figure. Against a real Manual J it typically reads high on a well-sealed modern house, sometimes by 20 to 40 percent, and can read low on an older house with ducts in an unconditioned attic. It is close enough to question a quote and not close enough to buy equipment on.
Where do the BTU per square foot figures come from?
They are the conventional published ranges used in US residential practice, organised by DOE and IECC climate zone. Each zone has a band rather than a single number, and the insulation setting selects a point inside that band rather than multiplying on top of it, so a result can never leave the published range.
Does this account for duct losses?
No. Duct losses are not modelled at all, and ducts running through an unconditioned attic or crawlspace can add a large fraction to both the heating and the cooling load. This is one of the clearest reasons a real load calculation is worth having.
Why do the calculator and the worked examples always agree?
Because both are produced by the same tested module. The examples in the text are computed when the page is built, not typed in, so the page a search engine reads and the answer a visitor gets cannot drift apart.