Cooling and heating load calculation field guide with ACCA Manual J

The load calculation, and what it actually sizes

A Manual J load calculation is the room-by-room accounting of how much heat a building gains on the design summer day and loses on the design winter day, expressed in Btu per hour. That number is what the equipment has to remove or add to hold the indoor setpoint. Size the equipment to it and the system runs long, even cycles that hold temperature and pull moisture. Skip it and you are guessing with a compressor.

The load is not the square footage. Two houses the same size carry wildly different loads depending on the windows, the insulation, the orientation, the leakage, and how many people and appliances are inside. A tight new build with good glass can carry half the load of an old house the same size with single-pane windows facing west. The calculation is how you tell them apart instead of pretending they are the same.

On the job the math is not the hard part. Software does the arithmetic. The trouble is the inputs: the wrong design temperature, a window U-factor pulled from memory, infiltration guessed instead of estimated, and a load done whole-house when the duct guy needed it room by room. Garbage in, oversized out. Then nobody writes down which assumptions drove the tonnage, so when the house runs clammy in July, there is no record to check.

Right size, not bigger

The whole reason to run the load is to get the size right, and right means matched to the building, not padded for safety. Bigger is the instinct, and bigger is the mistake. An oversized cooling system cools the air fast, satisfies the thermostat, and shuts off before it has run long enough to wring moisture out of the air. You end up cold and damp, which is worse comfort than warm and dry.

Three things ride on the size. Comfort, because a right-sized unit runs longer and holds temperature and humidity steadier across the house. Efficiency, because short, hard cycles are the least efficient way a compressor can run and they waste the part of the cycle where the equipment was getting good. And the moisture, because dehumidification only happens while the coil stays cold and wet, and an oversized unit never stays on long enough to get there.

Padding the size to cover for a sloppy calculation does not buy insurance. It buys short cycling, higher bills, more wear, and a humidity complaint you will get called back for. The fix is not a bigger box. It is a real load and equipment selected to it.

Manual J, then S, then D

ACCA splits residential design into a sequence, and the order is not optional. Manual J is the load, the Btu per hour the building gains and loses. Manual S is the equipment, where you select a unit whose published capacity at your design conditions matches the load, and where you check that the sensible and latent split of the equipment fits the sensible and latent split of the house. Manual S also sets the actionable ceiling: cooling capacity should land no higher than about 115 percent of the cooling load, while a heat pump can run higher, often near 125 percent in a heating-dominated climate, to cover more of the heating before the strips take over. Manual D is the duct, sizing every run so the blower delivers each room its share within the available static pressure.

Each step feeds the next. The load from J is the input to S. The room loads from J set the airflow targets that D sizes the ducts to hit. Do them out of order, or skip J and jump to picking equipment off the old nameplate, and everything downstream inherits the error. There is also Manual T for sizing the registers and grilles, which sits between equipment selection and final duct layout in ACCA's set.

The Manual D guide on this site carries the duct side in full: friction rate, available static pressure, total effective length, and sizing the trunk and branches. This guide stays on J, the load that starts the chain. Get the load wrong and no amount of good duct design saves the job, because you are moving the wrong amount of air to meet the wrong capacity.

What design temperature do you use?

You size to the 1 percent cooling and 99 percent heating design conditions, not the record extreme. The 1 percent summer design temperature is the dry-bulb the location exceeds only about 1 percent of the annual hours, roughly 88 hours a year, on a multi-year average. The 99 percent winter design temperature is the one the location stays above 99 percent of the hours. These come from ASHRAE and are tabulated in Manual J, and they are not the hottest or coldest day on record.

This trips people up, so it is worth being blunt. You do not size for the 105 degree afternoon that hits once a decade. Size for that and the equipment is oversized for the 8,700 other hours of the year, and it short-cycles through all of them to cover a handful you could ride out a degree or two warm. The percentile conditions are the honest design target, and they already build in nearly all of the real weather.

Cooling also needs the design wet-bulb or grains, not just the dry-bulb, because that is what sets the latent load. The indoor side is the setpoint you design to, commonly around 75 degrees and a target relative humidity for cooling and around 70 degrees for heating, with the project or the energy code controlling the actual numbers. The difference between indoor and outdoor, the design delta-T, is what the envelope conduction is calculated across.

How do you calculate a heat load?

At the bottom of every line in a load calc are two kinds of arithmetic. Conduction through a surface is the area times the U-factor times the design temperature difference. That is the wall, the roof, the window glass, the door, every opaque and glazed surface losing or gaining heat by the temperature gradient across it. Solar gain through glass adds the window area times the solar heat gain coefficient times the solar factor for that orientation, which is what makes a west window a summer problem the math has to capture.

The airflow side uses the standard-air sensible and latent equations. Sensible heat carried by an airstream is 1.08 times the airflow in CFM times the temperature difference. Latent heat is about 0.68 times the CFM times the difference in grains of moisture, and total is roughly 4.5 times the CFM times the difference in enthalpy. Those constants assume standard air near sea level, so at altitude they need correction, which the software handles when you tell it the elevation.

Nobody runs the whole house by hand anymore, and that is fine. What matters is knowing what each term is doing, so when the software spits out a number that smells wrong you can find the input that did it. A roof load that looks low is usually a U-factor that does not match the actual assembly. A latent load that looks high is usually an infiltration or ventilation rate set too generous.

U

Conductance of the assembly in Btu per hour per square foot per degree F, the inverse of R-value

A

Area of the surface or glass in square feet

SHGC

Solar heat gain coefficient, the fraction of solar energy a window admits

grains

Grains of water vapor per pound of dry air, the unit the latent load is figured in

What is the difference between sensible and latent load?

Sensible load is the heat that changes air temperature, the kind a thermometer reads. Latent load is the heat tied up in moisture, the energy it takes to condense water vapor out of the air, which a thermometer does not see at all. Manual J figures them separately because the equipment removes them separately, and a unit that handles one well can handle the other poorly.

Sensible load comes from the envelope and the temperature difference: conduction through walls and roof, solar gain through glass, the temperature part of infiltration and ventilation, and the heat off lights and equipment. Latent load comes from moisture: people give off water vapor, infiltration and outdoor ventilation air drag humidity in, and cooking and bathing add to it. In a humid climate the latent load is a real fraction of the total, and ignoring it is how you end up with a cold, clammy house.

The ratio matters when you pick equipment. Sensible heat ratio is the sensible load divided by the total, and a typical comfort cooling job lands somewhere around 0.70 to 0.80 sensible, with the rest latent. The equipment has its own sensible heat ratio at your conditions, published in the expanded performance data. Manual S is where you match the two. A high-sensible unit on a high-latent house dries nothing, and that is the most common humidity complaint there is.

Envelope loads: walls, roof, windows, orientation

The envelope is the biggest sensible load on most houses, and the windows usually run it. Opaque surfaces, the walls and the roof and the floor, lose and gain heat by conduction: area times U-factor times the temperature difference, plus an adjustment on sunlit surfaces for the solar heating of the surface itself. The U-factor has to match the real assembly. A 2x6 wall with continuous exterior insulation is a different number than a 2x4 wall with batts, and pulling the wrong one throws the whole wall load off.

Glass is where the calculation earns its keep, because a window does two things at once. It conducts heat by its U-factor like any surface, and it admits solar energy by its solar heat gain coefficient, which depends heavily on orientation. A west-facing window in the late afternoon of the design day is a different load than a north window of the same size, and Manual J applies the orientation and shading to capture it. This is why a builder who flips a floor plan east-to-west changes the room loads even though nothing about the house got bigger.

Shading, overhangs, and interior coverings all change the glass number, and the software has fields for them. The rookie move is to take the manufacturer's center-of-glass numbers and skip the orientation and shading, which lands the cooling load high on the sunny rooms and low on the shaded ones. Get the glass right room by room, because that is what sizes the registers in the rooms people actually complain about.

Infiltration and ventilation: the air you did not mean to condition

Infiltration is the uncontrolled air leaking in through the envelope, and ventilation is the outdoor air you bring in on purpose. Both carry a sensible load, because you have to heat or cool that air to setpoint, and both carry a latent load in a humid climate, because that air drags moisture in with it. On a leaky house in a humid zone, infiltration and ventilation together can be a surprising share of the total cooling load, most of it latent.

Infiltration is estimated, not measured, in a design-stage Manual J, from the construction tightness, the climate, and the house geometry, often expressed as an air change rate. A blower-door number from a tested house is better when you have it, and on tight new construction the measured ACH50 turned into a natural rate gives a far more honest infiltration load than the old assumption that every house leaks the same. Tight houses changed this. A modern sealed envelope leaks a fraction of what a 1980s house did.

Mechanical ventilation is a deliberate load you add because the house needs fresh air, and the rate commonly follows the ventilation standard the energy code or the spec invokes. Outdoor air in a humid climate is where the latent load lives, which is why a tight house with a ventilation system still needs its latent load taken seriously even though the infiltration dropped. The air you bring in on purpose has to be dehumidified just like the air that leaks in by accident.

Internal gains: people, lights, and plug loads

Everything inside the house that gives off heat is an internal gain, and it counts toward cooling and works against you, while it helps heating and is deliberately not credited in the heating design. People are both: each occupant throws off sensible heat and latent heat, the moisture from breathing and perspiration, and Manual J counts a number of occupants based on the bedrooms rather than a packed-house worst case. The default basis is the number of bedrooms plus one, counted as one person per bedroom with two in the primary bedroom, and the load is spread to the rooms people actually occupy.

Lighting and appliances add sensible heat, and the kitchen is the big one. A range, an oven, and the refrigerator all dump heat into the space, and cooking adds latent load on top. Plug loads, the televisions and computers and chargers, are a smaller but real sensible gain that modern houses carry more of than the old assumptions accounted for. The software has default values, and for most houses the defaults are reasonable, but a home with a heavy kitchen or a server closet deserves a second look.

The honest move is to use realistic internal gains, not inflated ones. Padding the occupant count or the appliance load pushes the cooling size up the same way every other padded input does. Count what is really there.

Block load or room by room?

You need both, and they answer different questions. A block load is the whole conditioned zone added up at once, accounting for the fact that the peak gains do not all happen at the same hour, so the block load is a little less than the sum of the room peaks. The block load sizes the equipment, because the equipment serves the whole zone. That is the Manual J number that feeds Manual S.

The room-by-room load is what sizes the air distribution. Each room's load sets how much airflow that room needs, which sets the register and the branch duct the Manual D design has to deliver. A block-only load tells you the tonnage but tells the duct designer nothing about how to split the air, and a system sized right at the equipment can still starve the bonus room over the garage if nobody ran that room's load. Do the block for the equipment and the room-by-room for the ducts, off the same calculation.

The room loads also tell you where zoning earns its place. When parts of a house peak at different times, a west wing in the afternoon and east bedrooms in the morning, or when an addition or a finished attic swings far from the rest, the room-by-room load is the evidence for a zoned system or a second piece of equipment instead of one oversized unit trying to satisfy a thermostat in one hallway. The load drives the zoning decision, not the other way around.

Why is oversizing an AC bad?

An oversized air conditioner satisfies the thermostat on temperature before it has run long enough to remove moisture, so it shuts off, and that short cycle is where the damage lives. Dehumidification only happens while the coil stays cold and wet long enough for water vapor to condense and drain, commonly cited at on the order of 10 to 15 minutes of run time, and an oversized unit reaches setpoint and quits before it gets there. The result is the cold, clammy house: at temperature but above a comfortable humidity, which feels worse than a warmer dry room and grows mold given time.

Short cycling costs you more than comfort. The hardest, least efficient part of a cooling cycle is the start, and a unit that starts and stops constantly spends its life in that worst phase, burning energy and wearing the compressor and contactor faster. The equipment that was supposed to last fifteen years gets tired early. None of this shows up on the day of startup, which is why it gets installed and signed off and only surfaces as a humidity complaint and a high bill the next summer.

The rule-of-thumb sizing that causes most of this is the square-feet-per-ton habit, somewhere around 400 to 600 square feet per ton depending on who you ask, applied without a load. It is fast and it is usually wrong, almost always wrong on the high side, because it ignores the windows, the insulation, the orientation, and the leakage that actually set the load. A widely cited figure is that the large majority of central systems are sized incorrectly, and oversizing is the usual direction. The calculation exists precisely to replace that habit. Size to the Manual J, not the floor plan.

The heating load: conduction and infiltration, no free credit

The heating load is simpler than the cooling load, and the simplicity is a design choice. It is the heat the building loses on the 99 percent winter design night, which is conduction through the envelope plus the heat to warm the infiltration and ventilation air, calculated across the heating design delta-T. There is no solar gain in the heating design, because you size for the cold night when the sun is down, and there is no credit for internal gains, because you cannot count on the lights and the people being there when the load peaks.

That no-credit rule is deliberate and worth understanding. The internal gains and the sun do reduce real heating demand most of the time, and that is good for the energy bill, but you do not size the furnace assuming they will be there at 4 a.m. on the coldest night. The design load is the worst credible case, and the equipment has to cover it without help from gains that may not show up.

In a heating-dominated climate the load calc still matters, just for the opposite reason from cooling. An oversized furnace short-cycles too, blows cold-feeling air at the start of each short burst, and swings the temperature, though it does not carry the moisture penalty that oversized cooling does. The right move is the same: size the heating to the calculated loss, then let Manual S confirm the equipment covers both the heating and cooling loads at your design conditions.

Latent load and humidity in a humid climate

In a humid climate the latent load is not a rounding error, and treating it like one is how you build a house that is cold and wet. The moisture comes from the outdoor air, both the infiltration and the ventilation, plus the people and the cooking inside. On a Gulf Coast or humid-continental job, the latent fraction of the cooling load is large enough that the equipment's ability to dehumidify, not just to cool, is the thing that makes or breaks comfort.

This is where right-sizing pays off twice. A right-sized unit runs longer cycles, and long cycles are what keep the coil cold and wet so it condenses moisture out. An oversized unit in a humid climate is the worst combination there is, because it satisfies the temperature fast and leaves the humidity behind. When the load calc shows a high latent fraction, that is the signal to select equipment by its sensible heat ratio and, on some jobs, to add dedicated dehumidification rather than hoping a bigger AC will handle the water.

Get the latent inputs honest, because they swing the answer. The infiltration rate, the ventilation rate, and the occupant count are the three that move the latent load most, and inflating any of them oversizes the latent side the same way padding the sensible inputs oversizes the cooling. In a dry climate the latent load is small and this matters less. In a humid one it is half the job.

Load calculation software, not a guess

A Manual J done right is run in software that ACCA has approved as conforming to the procedure, because the calculation has hundreds of inputs and the approval is what tells a plan reviewer the method behind the number is the real one. The current residential procedure is published as ANSI/ACCA Manual J, the 8th edition, often written Manual J8 or MJ8, and approved programs are checked against it. Several exist, including Wrightsoft Right-J, Elite RHVAC, and Cool Calc among others, and the energy code or the AHJ may require an approved one for permit.

Approved software does not make the answer right. It makes the method right. The inputs are still yours, and a conforming program fed lazy inputs produces a conforming, oversized, wrong answer that looks official on the report. The default-everything load, where the designer accepts every preset without checking the windows or the tightness against the actual house, is the quiet way a Manual J becomes a rule of thumb wearing a report cover.

Treat the software as the arithmetic and yourself as the judgment. Pull real U-factors and SHGC values from the window schedule, set the infiltration from the construction or a blower-door number, count the real occupants and appliances, and confirm the design conditions match the location. The program will do the rest correctly. Garbage in still oversizes, approved or not.

Commercial loads and Manual N

Manual J is the residential and light-commercial load procedure. For full commercial buildings the ACCA procedure is Manual N, which carries the same physics but adds what commercial loads bring: high and variable occupancy, large internal gains from lighting and equipment, big ventilation requirements driven by the occupancy, and diversity across spaces and hours that a house does not have. A conference room packed at 10 a.m. and empty at 2 p.m. is a load profile a residential method was never built for.

The bigger difference on commercial work is that the loads can be internal-dominated rather than envelope-dominated. A deep-plan office or a retail box has so much lighting, plug, and people load in the core that the envelope is a minor part of the cooling load, and the building can need cooling in the dead of winter while the perimeter needs heat. That simultaneous heating and cooling is normal on commercial jobs and foreign to most residential work, and the load method and the system selection both have to account for it.

ASHRAE Fundamentals carries the underlying load calculation methods that the commercial world leans on, and on engineered commercial projects the mechanical engineer's load model and the project specification govern. The point that carries across from residential is the same one: size to a real, room-by-room or zone-by-zone load, not to a per-square-foot rule, because the consequences of oversizing scale up with the building.

Why a data center load is a different animal

A data center is the extreme case of an internal-dominated load, and a residential or even a normal commercial load method does not fit it. The cooling load is set almost entirely by the IT equipment, the racks of servers turning electrical power into heat around the clock, and the envelope is nearly irrelevant by comparison. The load is essentially the electrical draw of the IT gear plus the support systems, and it runs at full tilt regardless of the weather outside.

That changes everything about the design. The load is steady, not weather-driven, so the 1 percent design-day thinking that anchors Manual J does not drive it. The load is almost purely sensible, because servers add heat without adding moisture, so the latent concern flips from removing humidity to not over-drying or condensing. And the consequence of getting cooling wrong is not a comfort complaint, it is downtime, so the redundancy and the airflow management dominate the design in a way comfort cooling never sees.

If you are working a critical-cooling project, the methods come from the data center cooling discipline and the thermal guidelines for that space, not from Manual J. The cooling pillar content on this site is the place for the IT-load-dominant approach. Naming it here only to be clear: do not run a server room as if it were a bedroom. The load is a different shape entirely.

Verify the calc against the building you built

A load calculation is a prediction, and the building is the test of it. The honest close on a job is to confirm that the system the load designed actually holds the setpoint and the humidity once it is running, which is where the load work and the balancing work meet. The calc said the equipment would carry the load and the rooms would get their airflow. The field is where you find out.

The airflow side gets verified by measurement, not by faith. The room loads set design airflow, and the air balancing work confirms each outlet delivers its share within tolerance, which is the proof that the room-by-room load and the duct design actually landed. The air balancing guide on this site carries that procedure in full. When the far bedroom never gets comfortable, it is usually the airflow to that room, not the equipment tonnage, and the balance report is what tells you which.

The equipment side gets verified by how it behaves. A right-sized system runs long, steady cycles and holds humidity. A system that short-cycles on mild days, or holds temperature but never pulls the humidity down, is telling you the load was wrong or the equipment was selected past it. Catch that at commissioning, with the data, not a year later when the homeowner calls about mold. The calc is only as good as the building proves it to be.

What to document

Equipment size is the first thing questioned when a house runs hot or muggy, and a load calc you can pull up shows the tonnage came from real inputs, not a hunch. The record is what answers, a year out, whether the tonnage was ever right or whether it was a guess dressed up in a report. It is also what a plan reviewer or the energy code wants to see, room by room.

Capture the design conditions used, the indoor setpoint, the room-by-room sensible and latent loads, the block load that sized the equipment, the airflow each room calls for, the major assumptions on windows and infiltration, the equipment selected against the load through Manual S, and who ran and checked it. The table below is the spine of a defensible record: room, sensible, latent, design CFM, and the equipment the loads led to.

Common mistakes

• Sizing by square-feet-per-ton or the old nameplate instead of running a load, which almost always oversizes.
• Using the record high or low for the design temperature instead of the 1 percent and 99 percent conditions.
• Ignoring the latent load in a humid climate, so the system cools but leaves the house clammy.
• Running a whole-house block load only and giving the duct designer nothing to size the rooms to.
• Accepting the software defaults on windows and infiltration without checking them against the real house.
• Stopping at Manual J and skipping Manual S, so the equipment is never matched to the load's sensible and latent split.
• Padding inputs for safety, which compounds into an oversized system the same as a bad number would.

Field checklist

Standards and references

ACCA Manual J is the residential load calculation procedure, published as ANSI/ACCA Manual J in its 8th edition, and it is the recognized method for sizing residential equipment in most adopting jurisdictions. It pairs with Manual S for equipment selection, Manual D for duct design, and Manual T for register and grille selection, with Manual N covering commercial load calculations. The residential code commonly requires that heating and cooling equipment be sized from a Manual J load and a Manual S selection, but the adopted code edition and local amendments control whether and how that is enforced.

The design conditions trace to ASHRAE, whose tabulated 1 percent and 99 percent values Manual J uses, and ASHRAE Fundamentals carries the underlying load calculation methods that both the residential and commercial procedures rest on. Ventilation rates commonly follow the ventilation standard the energy code invokes, which on many jobs is an ASHRAE standard, but the project specification and the adopted energy code set the actual rate.

Edition numbers and the specifics of code adoption shift between cycles and between jurisdictions, so confirm the procedure edition and the adopted code against what the AHJ actually enforces before you cite a requirement on a submittal. Where the project specification or the equipment manufacturer's data imposes something tighter than the rule of thumb, that governs the design.

Units, terms, and conversions

Load work moves between a few units and a few names for the same thing, so the same load can read differently across a report, a spec, and an equipment sheet. Loads are figured in Btu per hour and equipment is sold in tons, where one ton of cooling is 12,000 Btu per hour. Airflow is in cubic feet per minute, CFM. Temperature difference is the design delta-T, indoor minus outdoor.

Moisture shows up as grains of water vapor per pound of dry air, or as relative humidity, or on the psychrometric chart as wet-bulb and enthalpy, and the latent load is figured in the grains form. Conductance is the U-factor in Btu per hour per square foot per degree F, the inverse of the R-value people quote for insulation. Solar admission through glass is the SHGC, a dimensionless fraction from 0 to 1.

Btu/h and ton

Load is in Btu per hour; one ton of cooling equals 12,000 Btu per hour

Sensible / latent

Heat that changes temperature versus heat tied up in moisture

SHR

Sensible heat ratio, sensible load divided by total load

U-factor / R-value

Assembly conductance and its inverse, how the envelope conducts heat

SHGC

Solar heat gain coefficient, the fraction of solar energy a window admits

Design conditions

The 1 percent cooling and 99 percent heating outdoor values from ASHRAE

Related tools

Calculators and readiness checks for this work

Compare your options

FAQ

People also ask

• What is a friction rate in duct design?
• What is air balancing?
• What is an electrical load calculation?
• How do you size a data center generator?
• What is the difference between DX and chilled water cooling?