Rescheck, Comcheck, and Manual J Reports
HVAC Heat Load Calculation Explained
A heat load calculation tells you how much heat a house loses in the winter and how much it picks up in the summer. Those two numbers are what you size the furnace and the air conditioner against. People call it a dozen different things. Load calculation, heat load calc, heat loss and heat gain, HVAC sizing calculation, and/or a Manual J. I've been doing these for over 25 years, and I'll be honest, most of what goes wrong in residential HVAC traces back to somebody skipping this step.
Two Numbers, Not One
Winter and summer are separate calculations, and that surprises people.
​
In winter, heat leaves the house. It goes out through the walls, the ceiling, the windows, the floor, and every gap somebody forgot to seal. The furnace exists to put it back as fast as the house sheds it.
​
For obvious reasons, summer is a different. Heat comes in through the envelope, sure, but it also pours in through the glass as sunlight, and a good chunk of it gets made right there in the house. Bodies, the oven, the lights, the fridge, the kid's gaming rig running twelve hours a day. All of that is load your air conditioner has to haul back out.
​
The reason you can't just run one calculation and scale it: these two don't move together. I've run tight new builds where the heating load came in modest and the cooling load was the bigger problem. Put a wall of west-facing glass on a house and the gap widens further.
So when somebody asks me how many BTUs their house needs, the answer is always two numbers.
Why Square-Footage Rules Don't Work
You've probably heard one ton per 400 to 600 square feet. It's still floating around out there. It's still wrong most of the time.
Square footage tells you how much air is in the house. It tells you nothing about how fast heat moves across the envelope, and that's the whole ballgame.
​Take two houses on the same street, both 2,000 square feet. One's got R-49 blown in the attic, the other has R-19 batts that settled to about R-13 fifteen years ago. One's got double-pane low-E, the other's still running clear glass from 1994. One sits with the long side facing south under a decent overhang. The other faces west with nothing shading it. One got a blower door test and real air sealing. The other leaks.
​Those two houses do not need the same equipment!
The rule of thumb sticks around because it's fast and because nobody calls to complain about an oversized unit. The house does get cold. It gets cold badly, in short bursts, without ever pulling the humidity down, while the compressor racks up cycles it was never going to survive. I wrote more about what that actually does to a house on the Manual J page.
How the Math Actually Works
The core of a heat loss calculation is one formula, and it's not complicated:
Q = U × A × ΔT
Q is the heat loss in BTU per hour. U is how easily heat passes through the assembly. A is the area in square feet. ΔT is the temperature difference between inside and outside.
U is just the inverse of R-value. R-21 wall? U is 1 ÷ 21, or 0.0476.
Run one wall through it. Say you've got 1,200 square feet of net wall area after you subtract the windows and doors, built with R-21 batts, and you're designing to 70 degrees indoors against a 5-degree outdoor design temperature. That's a 65-degree ΔT.
0.0476 × 1,200 × 65 = 3,714 BTU/h
That's the heat leaving through those walls on a design day.
Now here's where it stops being simple. That R-21 on your plans isn't the R-value of the wall. It's the R-value of the insulation sitting between the studs. Every stud is a thermal bridge running straight through the assembly, and wood is a lot less resistant than fiberglass. Once you account for framing at whatever spacing you're actually building, an R-21 wall usually performs somewhere around R-17 or R-18 whole-wall. Use 21 and you may undersize the load.
That's one assembly. A real house has exterior walls facing four directions, ceiling, floor or slab, foundation walls, rim joists, and every window and door as its own line item.
Air leakage is a separate formula:
Q = 1.08 × CFM × ΔT
CFM is how much outside air is coming into the house. That 1.08 is a constant that bakes in the density and specific heat of air. Take 150 CFM of infiltration against that same 65-degree ΔT and you're looking at another 10,530 BTU/h — often a bigger number than people expect, and the reason blower door numbers matter to the calculation.
Summer runs different. Heat loss and heat gain aren't mirror images. Cooling load has to account for solar gain through the glass, which depends on SHGC, orientation, shading, and time of day, plus internal gains from people and equipment, plus the latent load from moisture. There's no one clean formula for it. The solar piece alone is why a west-facing room and a north-facing room of identical size need different amounts of air.
So why not just do it yourself?
Honestly, you could. The formulas are public and the arithmetic is arithmetic. What takes the time isn't the math. It's knowing that the R-21 on the drawing isn't R-21. It's picking an infiltration rate that reflects how the house is actually built rather than accepting a software default. It's knowing which design temperature the plan reviewer in that county expects to see. It's getting the duct losses right when the trunk runs through an unconditioned attic.
​
Those judgment calls are where reports go wrong, and they're not in any formula. If you'd rather run the numbers yourself, my HVAC formulas page has more of them. If you'd rather have a report that gets stamped, that's what I do.
​
Sensible and Latent:
Why the House Can Feel Wrong at 72 Degrees
Cooling load splits into two pieces, and this is where a lot of the "my AC works but the house feels awful" calls come from.
Sensible is temperature. That's the part the thermostat reads.
​
Latent is moisture. Pulling water out of the air costs energy, and it shows up in the calculation as its own line.
Here's the thing people miss: an air conditioner only dehumidifies while it's actually running. A unit that's too big satisfies the thermostat in eight minutes and shuts off. It never runs long enough to do the moisture work. So you end up at 72 degrees and 65% relative humidity, which honestly feels worse than 75 and 45.
​
If you're building anywhere humid, the split between those two numbers matters as much as the total. It's right there in the report. It should be driving equipment selection, not just the tonnage on the front page.
Design Temperatures
A load calculation doesn't get run against the coldest day ever recorded. It gets run against design conditions — the 99% heating and 1% cooling temperatures for wherever the house is going.
​
Plain version: cold enough that only about 1% of the hours in a normal year are colder. Hot enough that only 1% are hotter.
If you sized to the all-time record low instead, you'd install equipment that's oversized 364 days a year and you'd be right back to short-cycling. Design conditions are the compromise, and it's a reasonable one.
​
Where this goes sideways is when somebody grabs the nearest big city instead of the actual project address. That's one of the more common reasons I see a load calculation come back rejected. Northern Ohio and southern Ohio are not running the same numbers, and a plan reviewer who works in that county knows it.
Is This the Same as a Manual J?
Basically, with one distinction worth knowing.
​
"Heat load calculation" is what the work is. Manual J is the standard that says how to do it — ACCA's Manual J8 Residential Load Calculation. That's what building departments recognize and what plan reviewers expect to see in the packet.
​
A load calculation run to Manual J standards is a Manual J. A load calculation somebody did on the back of an estimate sheet is not, and it won't get you a permit.
​
If you need one for a permit, you need a Manual J. Who needs one and how the requirements shift between jurisdictions is all on the
What Is a Manual J page.
​
Once the load is settled, Manual S is what proves your equipment actually delivers it.
What About the Free Online Calculators?
Fine for a gut check. Useless for a permit.
​
The free ones ask for square footage, your climate zone, and some vague insulation rating, then run it through a lookup table. That's a dressed-up rule of thumb. It isn't modeling your windows one at a time, it doesn't know which way the house faces, and it isn't giving you room-by-room output or anything a reviewer can verify.
​
If you want to know whether the 4-ton unit your contractor quoted is in the right neighborhood, go ahead and run one. If you need a report your building department will stamp, it won't get you there.
Need One Done?
I'm Brad Buzzell. RESNET-certified HERS rater, 25-plus years in residential energy compliance. Rescheck Review handles Manual J load calculations for builders, contractors, architects, and homeowners in all 50 states.
​
Send me your plans and I'll tell you what you need and what it runs. If you're not even sure whether your permit office wants a load calculation, call me at (800) 671-1895 x702 and I'll find out for you. Costs you nothing either way.
Common Questions
Is a heat load calculation the same as a Manual J?
Close enough day to day. "Heat load calculation" is the work. Manual J is the ACCA standard that says how it's supposed to be done.
A calculation run to Manual J standards is what building departments accept.
Why is my contractor's number different from the load calculation?
Usually because he used a square-footage rule and the calculation modeled the actual house. When the two disagree, go with the one that has the windows and insulation in it.
Do heat loss and heat gain have to be calculated separately?
Yes. Different physics, and they don't scale together. A house can have a modest heating load and a big cooling load, or the other way around, depending on climate, glass, and which way it faces.
What are design temperatures?
The 99% heating and 1% cooling temperatures for your specific location. Cold enough that only about 1% of hours in a normal year are colder, hot enough that only 1% are hotter. Equipment gets sized against those, not against record extremes.
Can I use a free online HVAC load calculator?
For a rough sanity check, sure. For a permit, no. Those calculators run lookup tables instead of modeling your building, and they won't give you room-by-room numbers or documentation a plan reviewer can check.
How do you calculate heat loss?
The core formula is Q = U × A × ΔT. Q is heat loss in BTU per hour, U is the inverse of the assembly's R-value, A is the area in square feet, and ΔT is the difference between your indoor design temperature and the outdoor design temperature. Run it for every wall, ceiling, floor, window, and door, add air leakage separately, and the total is your heating load.
Why doesn't R-21 insulation give me an R-21 wall?
Because the studs run straight through the assembly and wood is far less resistant than fiberglass. That R-21 is the value of the insulation between the framing, not the wall as built. Once you account for framing, an R-21 wall typically performs closer to R-17 or R-18 whole-wall. Using the label number may undersize your load.