Air is a heat carrier, and airflow is how much you send
A supply register does not cool a room. It delivers air at a temperature different from the room's, and the room is conditioned by the heat that air absorbs or gives up before it leaves. How much heat that is depends on two things only: how much air arrives, and how far its temperature has to move to reach room temperature.
Putting that in numbers gives the whole calculation. One cubic foot per minute of standard air weighs 0.075 lb per cubic foot and there are 60 minutes in an hour, so one cfm carries 4.5 lb of air past a point every hour. Air's specific heat is 0.24 BTU per pound per degree Fahrenheit. Multiply: 4.5 × 0.24 = 1.08 BTU per hour for every cfm and every degree of difference. That single number is the sensible air equation, and it appears everywhere in this trade: in load calculations, in coil capacity checks, in furnace temperature rise limits.
Rearranged for design, the airflow a room needs is its sensible load divided by 1.08 times the temperature difference you intend to supply at. A 6,000 BTU/h room with 55 °F supply air into a 75 °F space needs 6,000 ÷ (1.08 × 20) = 278 cfm.
Note that only the sensible load appears. Latent load, the moisture the coil condenses, does not change the air temperature, so it does not change how much air you have to move. Latent load is handled by the coil's condition, not by the duct's size, which is why a room-by-room airflow schedule is built from sensible loads and the system's latent capacity is checked separately.
The direct method and the proportional method, and when they disagree
There are two ways to arrive at a room's airflow and the calculator shows both, because the gap between them is diagnostic.
The direct method is the equation above: room load divided by 1.08ΔT. It answers what the room needs, independently of anything else in the house.
The proportional method takes the total system airflow, already fixed by the equipment selection, and splits it in proportion to each room's share of the total load: CFMroom = CFMtotal × (loadroom ÷ loadtotal). It answers how the air you actually have should be divided.
Write the two out and the condition for agreement falls out directly. They are equal when
CFMtotal × loadroom ÷ loadtotal = loadroom ÷ (1.08ΔT)
which cancels to CFMtotal = loadtotal ÷ (1.08ΔT). In other words, the two methods agree exactly when the system's total airflow is what the whole-house sensible load would need at that same supply temperature difference, and they disagree by whatever ratio those two figures differ by. The disagreement is not a rounding error; it is information. If the system moves more air than the whole-house load requires at your assumed ΔT, then the real ΔT will be smaller than assumed, and the proportional answer is the one that matches what the system will actually do.
In practice the proportional method governs, because the equipment is chosen before the ducts and its airflow is set by the coil's requirement, which is the subject of the CFM per ton calculator. Use the direct method to sanity-check the supply temperature difference you assumed.
Worked example: a 6,000 BTU/h bedroom in a three-ton house
A bedroom of 200 sq ft with an 8 ft ceiling has a sensible cooling load of 6,000 BTU/h. The system is a three-ton unit moving 1,200 cfm, the whole-house sensible load is 30,000 BTU/h, supply air is 55 °F, the room holds 75 °F, and the registers selected are good for 100 cfm each. The site is at sea level.
- Temperature difference. |55 − 75| = 20 °F.
- Direct airflow. 6,000 ÷ (1.08 × 20) = 277.8 cfm.
- Proportional airflow. 1,200 × 6,000 ÷ 30,000 = 240 cfm.
- Why they differ. The whole-house load at 20 °F would need 30,000 ÷ 21.6 = 1,389 cfm, but the system moves 1,200. The ratio 1,200 ÷ 1,389 = 0.864 is exactly the ratio of the two answers, 240 ÷ 277.8. The system will therefore run a larger temperature difference than 20 °F: 30,000 ÷ (1.08 × 1,200) = 23.1 °F, giving supply air near 51.9 °F.
- Registers. 240 cfm at 100 cfm each needs three outlets; at 277.8 cfm it also needs three. Either way, three.
- Air change rate. The room volume is 200 × 8 = 1,600 cu ft, so 240 cfm gives 240 × 60 ÷ 1,600 = 9.0 ACH and 277.8 cfm gives 10.4 ACH.
The design number here is 240 cfm, because that is the air the system has. The useful thing the direct method has told you is that the real supply temperature will be about 52 °F rather than 55 °F, which matters for register selection: colder air needs a diffuser that mixes well, or occupants will feel a draught.
Choosing a supply temperature difference
The temperature difference is a design choice with real consequences, because airflow is inversely proportional to it. Doubling ΔT halves the airflow, halves the duct cross-section needed, and roughly halves the fan energy for that branch. It also makes the air colder or hotter at the register.
In cooling, residential supply air normally lands between about 52 and 58 °F, giving a difference of 17 to 23 °F against a 75 °F room. That is not arbitrary: it is what a coil produces at 350 to 450 CFM per ton, which is the range the equipment is rated in. You cannot independently choose both the CFM per ton and the supply temperature difference; the coil links them.
In heating, a furnace's temperature rise is limited by its rating plate, commonly a band such as 30 to 60 °F or 40 to 70 °F. Below the minimum the heat exchanger runs cold enough to condense flue gas in a non-condensing appliance; above the maximum the limit switch trips. Heating differences are therefore much larger than cooling ones, which is why a heating-only system needs far less air than a cooling system of similar capacity, and why a dual-purpose system is almost always sized on the cooling airflow.
Do not choose a difference that the equipment cannot deliver. If the calculator's supply temperature implies a coil colder than 50 °F or a furnace rise outside its plate, the airflow that follows is fictional. Check the equipment side first, then compute the branch airflows, then size the ducts with the duct size calculator.
Airflow per 1,000 BTU/h of sensible load
| Supply dT (F) | CFM per 1,000 BTU/h | CFM for a 6,000 BTU/h room |
|---|---|---|
| 15 | 61.7 | 370.4 |
| 18 | 51.4 | 308.6 |
| 20 | 46.3 | 277.8 |
| 22 | 42.1 | 252.5 |
| 25 | 37.0 | 222.2 |
| 30 | 30.9 | 185.2 |
| 40 | 23.1 | 138.9 |
| 50 | 18.5 | 111.1 |
| 60 | 15.4 | 92.6 |
Rows from 15 to 25 F are the cooling range; rows from 30 to 60 F are the heating range a furnace rating plate typically permits. Divide every figure by the air density ratio at altitude.
Getting the inputs right
- Use sensible load, not total load. Latent load does not change air temperature and therefore does not change airflow. A room-by-room schedule built from total loads oversizes every branch by the latent fraction.
- Do not assume a supply temperature. It is a consequence of the coil and the system airflow, not a free choice. Compute it as total load divided by 1.08 times total airflow and check it is achievable.
- Correct for altitude above about 2,000 ft. At 5,000 ft the air is 17% less dense, so it carries 17% less heat per cfm and the room needs proportionally more air. The same correction applies to every 1.08 in your calculations.
- Register count is a throw problem, not just a division. Two registers at 120 cfm each is not equivalent to one at 240 cfm if the room is long: throw, spread and drop all matter, and manufacturers publish them. ACCA Manual T covers selection and placement.
- Check the return path. Air delivered to a room with a closed door and no return path or transfer grille will pressurise the room and reduce its own supply airflow. Every room with a significant supply needs a way back.
- Verify by measurement. The design airflow and the delivered airflow are different numbers until someone has put a flow hood on the register. Compare the result against the air changes per hour the room should be seeing.
Where room airflow sits in the design
Room airflow is the hinge between the load calculation and the duct design. Upstream, the load calculation produces a sensible load for each room; ACCA Manual J does this room by room precisely so that these numbers exist. Downstream, each branch is sized to carry its room's airflow at the system friction rate, which comes from the available static pressure calculator and feeds the duct size calculator. Manual D is the standard that ties those steps together for residential systems, and Manual T covers the register selection at the end.
Two failures come from skipping this step. The first is sizing branches by room area, which allocates air by floor plan rather than by load and starves the rooms with the most glass and the most exposed wall, which are exactly the rooms that will be uncomfortable. The second is sizing branches by register count, which allocates air by how many holes were cut in the ceiling.
The corollary is worth stating: two rooms of the same size routinely need very different airflows. A north-facing interior bedroom and a south-west corner bedroom with the same floor area can differ by a factor of two in sensible load, and therefore by a factor of two in cfm. That difference is invisible on the plan and obvious to whoever sleeps in the corner room. The whole point of computing airflow from load is to catch it before the metal goes in.