What Size Furnace Do I Need? Input and Output BTU
By Danish Javaid, founder & editor · Published · Updated
Pick a furnace whose rated output is between 100% and 140% of the home's design heating load. Output is not the input BTU/h on the nameplate: it is input times AFUE. For a house with a 41,660 BTU/h heating load, the target output is 41,660 to 58,324 BTU/h, which is 43,400 to 60,800 BTU/h of input at 96% AFUE. Square footage does not enter into it.
The conversion that causes most mistakes
Screening estimate only. Use the manufacturer's published rated output for the model, fuel, and altitude.
The U.S. Department of Energy's Building America sizing guide, which cites ACCA Manual S, states the limit this way: the furnace must meet the Manual J heating load without exceeding it by more than 40%. It also gives a conversion example: a 60,000 BTU/h furnace at 92.5% efficiency yields around 55,500 BTU/h of output. So input labels, loads, and efficiency have to be reconciled before anything is compared.
| Design heating load | Target output (100% to 140%) | Input needed at 96% AFUE | Input needed at 80% AFUE |
|---|---|---|---|
| 30,000 BTU/h | 30,000 to 42,000 | 31,250 to 43,750 | 37,500 to 52,500 |
| 41,660 BTU/h (DOE Chicago house) | 41,660 to 58,324 | 43,396 to 60,754 | 52,075 to 72,905 |
| 60,000 BTU/h | 60,000 to 84,000 | 62,500 to 87,500 | 75,000 to 105,000 |
Worked example: the DOE Chicago house
The guide models a one-story, 2,223 ft² house in Chicago with a conditioned basement, built to the 2009 IECC prescriptive path. Its Manual J heating load is 41,660 BTU/h, which is 18.7 BTU/h per ft² (41,660 ÷ 2,223). The maximum target output is 41,660 × 1.4 = 58,324 BTU/h. The guide selects a furnace rated 60,000 BTU/h input at 92.5%, giving 55,500 BTU/h output, 55,500 ÷ 41,660 = 133% of the load. It passes, with about 13,800 BTU/h above the load and about 2,800 BTU/h under the cap. The guide also notes that furnace sizes span a wide range (a line might run from 60,000 to 120,000 BTU/h input) so the smallest available furnace can already sit near the limit, and that a two-stage unit can drop to a lower stage, 39,000 BTU/h in its example, to cut the gap in mild weather.
The same house, six replacement choices
Suppose the owner of a house with that 41,660 BTU/h load is replacing a furnace rated 100,000 BTU/h input at 80% AFUE. The old unit puts out about 80,000 BTU/h, which is 80,000 ÷ 41,660 = 192% of the load before any change. Here is how replacements compare.
| Option | Approximate output | Output ÷ 41,660 | Verdict |
|---|---|---|---|
| 100,000 input at 80% (like for like) | 80,000 BTU/h | 192% | Above the 140% cap; repeats the oversizing |
| 100,000 input at 96% | 96,000 BTU/h | 230% | Higher efficiency raises output at the same input |
| 75,000 input at 96% | 72,000 BTU/h | 173% | Above the cap |
| 60,000 input at 96% | 57,600 BTU/h | 138% | Inside the window, near the top |
| 45,000 input at 96% | 43,200 BTU/h | 104% | Inside the window with about 4% headroom |
| 40,000 input at 96% | 38,400 BTU/h | 92% | Below the load |
Efficiency cuts both ways. Going from 80% to 96% at the same input raises output by 20% (96,000 versus 80,000), so a like-for-like swap gets larger. If ducts in an unheated attic lose heat, put that loss in the load calculation, not in a bigger burner. And if the home will get air sealing or insulation first, run the calculation for the improved house, because a smaller load changes every row.
Check the blower and temperature rise
Temperature rise across the furnace is output ÷ (1.08 × CFM). With 55,500 BTU/h of output at 1,000 CFM the rise is 55,500 ÷ 1,080 = 51.4°F; at 1,200 CFM it is 42.8°F. The rise has to land inside the range printed on the furnace nameplate, which is why the installer sets the blower speed from the manufacturer's table at measured static pressure. In the DOE example, the blower must also meet the required cooling airflow within 10% when the same blower serves an air conditioner, and the guide assumes 0.50 in. w.c. of external static pressure for that selection.
- Low-stage output: ask how it compares with mild-weather loads. A two-stage or modulating furnace still cycles if its minimum output is too high, and a label alone does not tell you the minimum.
- Cold rooms: check returns, filters, and duct runs before blaming burner size. More burner capacity does not repair missing airflow.
- Short cycling: it can also come from thermostat settings or limit-switch trips caused by poor airflow, so diagnose before replacing anything.
Get the load first
A floor-area chart cannot tell a sheltered, well-insulated house from one with a lot of glass and leakage. A load calculation needs measured envelope areas, construction, the winter design temperature, the indoor target, infiltration and ventilation, and duct location. The design temperature is a frequency-based value, not the coldest night on record; see design temperature explained. The heating load calculator gives an early estimate, and the furnace size calculator converts a load into a target output and input.
Questions to settle before installation
- Does the load describe the house after any planned air sealing, insulation, or window work?
- Does the proposal show both input and rated output, with the altitude adjustment for your elevation?
- What output does each stage deliver, and how will the thermostat use the stages?
- Can the selected blower meet both heating temperature rise and cooling airflow against the actual duct resistance?
- Will the contractor document airflow, static pressure, temperature rise, venting, and combustion checks?
In many climates the question comes before size: see heat pump versus furnace. Keep the final load, model data, and commissioning record together, and ask for the output and airflow checks to be redone if a different furnace or coil is substituted. The J, S, and D walkthrough shows the full handoff.
Sources: DOE Building America, Strategy Guideline: HVAC Equipment Sizing (Manual S furnace limit, Chicago house, furnace selection example); ACCA Manual S.
Calculation guidance is reviewed against our published methodology. To report an error or request a correction, use the contact page.
Written and maintained by Danish Javaid, founder & editor. About this site.
Revision history
- : Rebuilt around the DOE Building America Chicago house (41,660 BTU/h load, 58,324 BTU/h cap, 60,000 BTU/h input at 92.5% giving 55,500 output) and recomputed a six-option replacement table and the temperature-rise check.