Why Square-Footage HVAC Sizing Rules Are Wrong — and What Manual J Actually Does — Blog

Ask the internet what size air conditioner a 2,000 square foot house needs and you will get a confident answer in about four seconds: 4 tons. Divide the floor area by 500, or multiply it by 20 or 25 Btu per square foot, and there it is. The arithmetic is clean, every calculator on the first page of results does it, and it is wrong often enough — and by enough — that the industry's own standard exists specifically to replace it.

Where the rule of thumb comes from

The square-footage rule is not a formula anybody derived. It is a back-figure: someone looked at a lot of houses in one climate, noticed that installed equipment averaged roughly one ton per 400 to 600 square feet, and wrote the average down. That is a description of what got installed, not a calculation of what those houses needed — and much of what got installed was oversized to begin with, so the rule partly encodes the very error it is used to make.

The tell is that no two sources agree on the number. Depending on where you look, a ton covers 400, 500, 600, or 700 square feet, and the Btu-per-square-foot figures run from about 18 to 30. Those spreads are not rounding. A 2,000 square foot house is a 2.9 ton job at 700 square feet per ton and a 5 ton job at 400 — the same house, the same rule, nearly double the equipment. When a method's published range spans a factor of two, the method is not producing an answer. It is producing a starting point that somebody else has to check.

The error is bigger than most people assume

Published criticism of rule-of-thumb sizing puts the resulting oversizing in the range of 40 to 100 percent. That is not a marginal safety factor. It means a house that genuinely needs 2.5 tons routinely gets 4 or 5.

The reason the error runs almost entirely in one direction is that floor area is the one variable the rule uses, and it is nowhere near the one that matters most. Two 2,000 square foot houses can have wildly different cooling loads:

Windows. Glass area, orientation and glazing type can swing solar gain by several tons on their own. A wall of west-facing single-pane glass and a wall of north-facing low-E glass are not the same building.
Insulation and air sealing. A 1968 house with R-11 walls and a leaky envelope against a 2020 house with R-21 walls and a blower-door result are different loads by a wide margin.
Climate. The design temperature difference in Minneapolis is not the one in Houston, and the humidity load is a different problem again.
Internal gains. Occupants, appliances, lighting and equipment all add heat.
Duct location. Ductwork in a 130°F attic loses capacity that ductwork inside the envelope does not.

None of that is in the square footage. A rule that ignores all of it and gets within 40 percent is doing well.

Oversizing is not the safe direction to be wrong

The instinct behind rounding up is understandable: nobody wants a callback on the hottest day of the year. But an oversized system is not a system with margin. It is a system that behaves worse.

It short-cycles. An oversized unit satisfies the thermostat quickly and shuts off. Then it starts again. Compressors draw hard at startup and are least efficient in the first minutes of a cycle, so a system that never reaches steady state runs at a fraction of the efficiency on its own label.

It does not dehumidify. This is the one homeowners actually feel. Moisture removal happens once the coil has been cold and wet for a sustained run. A short cycle drops the dry-bulb temperature and stops before it has pulled meaningful moisture out of the air. The result is a house that hits 72°F and still feels clammy — the classic cold-and-damp complaint that follows an oversized replacement.

It wears out sooner. Starts, not run hours, are what age a compressor and a contactor.

It costs more up front. Bigger equipment, sometimes bigger ductwork and a bigger circuit. The customer pays for capacity that then degrades their comfort.

So the trade is not "a bit of extra safety for a bit of extra cost." It is a real premium for a system that dehumidifies worse and dies earlier.

What Manual J actually does

ACCA Manual J, Residential Load Calculation, is the recognised method, and it is referenced by the International Residential Code and by most energy codes and utility programs. What makes it different is not that it is more complicated for its own sake. It is that it calculates the load instead of recalling it.

A Manual J calculation works component by component through the actual building:

Design conditions for the specific location — the outdoor temperature and humidity the equipment is expected to hold against, taken from published weather data rather than from the worst day anyone remembers.
The envelope, surface by surface: wall area and its assembly R-value, ceiling, floor, and every window and door with its own U-factor, solar heat gain coefficient, orientation and any external shading.
Infiltration, from the tightness of the house rather than an assumed air-change rate.
Internal gains from occupants, lighting and appliances.
Duct losses, which depend on where the ducts run and how well they are sealed.
Sensible and latent loads separately — this is the part the rule of thumb cannot do at all. Sensible load is temperature; latent load is moisture. Equipment is rated for both, and a system sized only on sensible load is exactly how the clammy-house problem gets built in.

The output is a heating load and a cooling load in Btu/h for that house, split sensible and latent. Manual S then selects equipment against those numbers, and Manual D sizes the duct system to deliver the airflow it needs.

What this means in practice

The honest summary is short. Floor area is one input to a calculation with roughly a dozen, and it is not the dominant one. A number derived from floor area alone is a rough order of magnitude — useful for a ballpark conversation on the phone, not for selecting equipment.

If you are a contractor, a Manual J is what separates a proposal you can defend from a guess with a price on it, and it is increasingly what code officials and rebate programs ask to see. If you are a homeowner and a contractor sizes your replacement by walking the perimeter and quoting the same tonnage you already have, you are getting the average of a lot of other houses — including, very likely, the oversizing already sitting in your basement.

Certified Manual J calculations are performed with ACCA-approved software, because the method is a published standard with a specific procedure rather than an equation you can reproduce in a spreadsheet. That is also why we do not offer a square-footage sizing calculator here: there is no authoritative table behind Btu-per-square-foot to build one on, and a number with a confident interface around it would only make an unreliable rule look reliable.

What we do offer is the question that does have a verifiable answer. If you need to know what size the equipment already installed is — as the starting point for a replacement conversation, or to check that a quoted model is the size it was sold as — the AC tonnage calculator decodes it from the model number on the data plate, including the published capacity codes where the usual divide-by-twelve shortcut returns a size no manufacturer builds.