Keep This House

energy

Load calculation (Manual J) in plain English: the number the bid is hiding

What an ACCA Manual J actually measures, why square-footage rules of thumb size the wrong machine, and what has to be on the report before you compare heat-pump bids.

A heat-pump bid without a load is a catalog page. The installer picked a tonnage the way people pick a truck: by square footage, by what came out of the closet, or by what the warehouse had on the truck. Manual J is the written heating and cooling load for this house, in this town, at the outdoor temperatures your code book actually uses. Until that number is on paper, you cannot tell whether Bid A and Bid B are the same job.

This is the math behind the heat pump quote checklist. Stay here when someone says “three ton” and you want to know what they measured. The weatherize-first guide is the order-of-work question. Ducted vs ductless is how the heat gets into the rooms after you know how much heat you need.

What is Manual J, in one paragraph?

Manual J is the Air Conditioning Contractors of America procedure for a residential load calculation. The current book is Manual J 8th Edition, published as ANSI/ACCA 2 Manual J - 2016. ACCA wrote it for single-family houses, townhouses, condos, and manufactured homes that have their own heating and cooling system. It is a room-by-room (or at least a whole-house) estimate of how many BTU per hour the house loses on a design-cold night and how many it gains on a design-hot afternoon. Those two numbers are the loads. They are not the model number, and they are not the square footage divided by 500.

ACCA: Manual J Residential Load Calculation

DOE Building America Solution Center: ACCA Manual J

The 2024 International Residential Code (IRC) M1401.3 tells the installer to size heating and cooling equipment with ACCA Manual S, or another approved method, using loads calculated with Manual J or another approved calculation. Most towns that adopted the IRC are asking for that paper with the mechanical permit. Local amendments win, so ask your building department what they want to see, not what a flyer promised.

ICC: 2024 IRC M1401.3 equipment and appliance sizing

What are you actually buying?

A same-size swap. The old sticker said 3 tons, so the new one says 3 tons. No one opened the attic, no one counted windows, no one used your town’s design temperature. This is the fastest bid and the one that most often leaves a bedroom cold or a system that short-cycles (on, off, on, off) in spring. It is not a load calculation. Treat it as a parts quote.

A block load. One heating number and one cooling number for the whole house. Useful as a sanity check. Many permit offices still want room-by-room numbers because ducts and mini-split heads are sized per room, not per roof. ACCA itself distinguishes block load from room-by-room load. If the report is one line that says “36,000 BTU cooling,” ask where the rooms went.

A room-by-room Manual J, then Manual S, then (if you have ducts) Manual D. Manual J is the load. Manual S is how the contractor picks a machine whose published capacity at your design outdoor temperature actually covers that load, without wandering off into a size that cannot dehumidify. Manual D is the duct design that delivers the air those rooms need. ACCA’s Manual S 3rd Edition (ANSI/ACCA 3 Manual S - 2023) is the equipment-selection book. Skipping S is how a correct J still becomes the wrong outdoor unit.

ACCA: Manual S Residential Equipment Selection

If Bid A is $12,000 with no load and Bid B is $17,000 with a J and an S attached, you are not looking at markup. You are looking at a catalog page next to a design.

What goes into the calculation?

The arithmetic is ugly in the software and simple in English. Heat moves through every surface that touches outdoor air. More area, worse insulation, bigger indoor-outdoor temperature difference, more load. Windows add sun. People and appliances add heat and moisture. Leaky ducts in an attic add a second house the machine has to condition.

A real report names, at least:

Design temperatures. Winter outdoor is the 99% dry-bulb temperature for your location (it is colder than that only about 1% of the hours in a year). Summer outdoor is the 1% dry-bulb. Indoor design in the ACCA/code-official training is 70 F for heating and 75 F for cooling. If the printout uses 95 F summer and 0 F winter because that is what the software defaulted to, it is not your town.

Orientation and glass. Which way the walls face, window area, U-factor, solar heat gain, overhangs. West glass in the afternoon is a cooling problem that square footage never sees.

Opaque surfaces. Wall, ceiling, and floor area with actual R-values, or honest defaults the report labels as defaults. “R-13 walls” on a 1958 house with no cavity insulation is fiction.

Infiltration and ventilation. How leaky the envelope is, plus planned outdoor air. A blower-door number beats a software default. If they never walked the attic, they guessed.

Ducts. Location (conditioned space vs attic vs crawl), insulation, and leakage. Ducts in a vented attic can be a large fraction of the load. That is why weatherizing first can drop a half-ton, and why a load run on today’s leaks is the wrong size after you air-seal.

Internal gains. Occupants, lights, appliances. These matter more for cooling and humidity than for a January night. If the report assumes a full house of people and you live alone, the cooling load is padded.

If any of those lines is missing, or every line is a round default, you have a marketing PDF. Ask them to rerun it with the house they stood in.

Why does square footage size the wrong machine?

The driveway rule of thumb is 400 to 500 square feet per ton of cooling. A ton is 12,000 BTU per hour, a size label, not the weight of the box. That rule ignores climate, glass, insulation, and leaks. ACCA’s low-load housing manual is blunt about how far new and tight houses have moved: a house built about ten years ago might land 800 to 1,200 square feet per ton of cooling; a house being built today might land 1,000 to 1,400; with current IRC and IECC provisions it is common to see cooling loads that exceed 2,000 square feet per ton. Heating in those tight houses can be on the order of 5 BTU per hour per square foot in a warm place (their Homestead, FL example) to 15 in a cold one (Duluth, MN).

ACCA: Manual LLH, low-load homes

So a 2,000-square-foot ranch is not “a 4-ton house.” In a leaky 1970s ranch with west glass and ducts in a hot attic, it might need more than the rule. In a tight 2018 house with good glass, it might need half. Putting the old 4-ton back in is how you buy short-cycles and a still-dry-feeling room that never quite hits setpoint because the coil never runs long enough to pull moisture.

Heat pumps make the mistake more expensive. Capacity falls as outdoor temperature falls. Sizing to cooling square footage in a cold climate undersizes winter. Sizing to a padded heating rule in a mixed climate oversizes summer. Manual S wants the manufacturer’s capacity table at your design outdoor temperature, not the 47 F heating sticker. The cold-climate spec sheet is that table. The SEER2 / HSPF2 decoder is the lab weather, which is not your design hour.

What should the report look like when it lands?

You should be able to read, without a decoder ring:

  1. The address and the outdoor design temperatures they used, with the source (Manual J Table 1A/1B or ASHRAE). If the city is two climate zones away, send it back.
  2. Whole-house heating load and whole-house cooling load, in BTU per hour, with cooling split into sensible (temperature) and latent (moisture) if they ran cooling.
  3. Room-by-room loads if you have ducts or more than one indoor unit. A four-head mini-split bid without room loads is four guesses.
  4. The inputs. Areas, R-values, window performance, infiltration assumption, duct location. Defaults should be labeled.
  5. Whether the attic, crawl, or addition was included. A load that stops at the 1960 footprint will not heat the 2004 sunroom.
  6. The Manual S pick, with the outdoor unit’s capacity at the same design temperature as the load, not only at 47 F / 95 F.
  7. Backup heat, if any: strip heat kilowatts, furnace lockout, or “none, heat pump covers design.” Dual-fuel is a different bid; see when keeping the furnace makes sense.

IRC M1401.3 has two exceptions that show up in real quotes. Multistage or variable-refrigerant equipment can sit anywhere in the manufacturer’s published capacity range for the calculated loads. If no published size hits both the total and sensible cooling loads, the contractor may step to the next larger standard size. Those are code valves, not a license to install a 5-ton on a 2-ton load. Ask which exception they are using, in writing.

How much does this change the install price?

EnergySage’s 2026 marketplace averages (before incentives) are $14,529 ducted, $14,353 hybrid (heat pump plus furnace), and $25,957 whole-home ductless, about $15,393 across the set they published. New ductwork, if you have none, is $2,000 for a small new house to more than $10,000 for a full retrofit. Those are install prices, not Manual J fees. The load is usually inside the bid, or it should be. A contractor who will not print the J is not saving you money. They are hiding the size.

EnergySage: heat pump vs mini-split

A tighter envelope can drop a zone or a half-ton. EnergySage’s insulation page puts professional work at $1,000 to $15,000, and practitioners they talked to often land $8,000 to $15,000. Run the load after you know whether the attic is in this year’s budget. If you weatherize after the condenser is in, you paid for capacity the house no longer needs.

EnergySage: do you need to insulate first?

Do not bake a dead federal credit into 2026 math. The IRS 25C window ended for qualifying property placed in service after December 31, 2025.

IRS: Energy Efficient Home Improvement Credit

What goes wrong when the J is fake?

Padded infiltration and “safety factors.” Every extra air-change and every 20% fudge moves you toward the next nominal ton. Oversized cooling short-cycles and leaves the house feeling damp. Oversized heat-pump heating in a mild climate does the same in summer.

Software with the wrong house. The address is yours. The windows are a library default. The ducts are “inside” when they are in a vented attic. Garbage in, 4-ton out.

Block load used to size six wall heads. The living room and the north bedroom do not have the same load. Manual D and a head-per-room plan both need the rooms.

Load run on today’s leaks, equipment bought, attic sealed in October. The machine is now too big. If the attic is happening this year, put it in the same week as the J.

Heating load in BTU, cooling in tons, winter capacity missing. You cannot check Manual S. Send it back.

“We don’t need a J, we’ve done this subdivision.” The IRC still wants a calculation on this house. Your addition, your replacement windows, and your spray-foam attic are not the house next door.

How do I decide this week?

  1. Ask every bidder for a written Manual J on this address, with design temperatures, whole-house loads, and the inputs they used.
  2. Walk the attic and the ducts yourself so you can tell whether “R-38 ceiling, ducts in conditioned space” matches what you saw.
  3. Decide whether envelope work happens before the equipment. If yes, put that scope on the load, not after the condenser is set.
  4. Require a Manual S pick that shows capacity at your design outdoor temperature, plus backup heat yes or no.
  5. Put the same load on three bids: same rooms, same ducts or heads, same electrical exclusions. The quote checklist is the markup sheet.
  6. If a bid is 40% cheaper and has no J, treat it as incomplete until the missing pages show up.

A load calculation is how you find out what the house needs. A model number is how someone sells you a box. Get the first on paper, then hire the second.