What delta-T tells you and what it does not
Delta-T is the temperature change of the air as it crosses the equipment. In cooling it is the split: return dry bulb minus supply dry bulb. In heating it is the temperature rise: supply minus return. It is the fastest whole-system check a technician has, because it needs two thermometers and no gauges.
What it measures is the ratio of sensible capacity to airflow. Rearranging the sensible heat equation, ΔT = Q_sensible ÷ (1.08 × CFM), so a wide split means high capacity or low airflow, and a narrow split means the reverse. That is why the split on its own never tells you which. It tells you that something is off and points you at the next measurement.
The trap is the single number. Technicians quote "20 degrees" as if it applied everywhere, but the correct split depends heavily on how humid the return air is. A coil pulling 12,000 BTU/h out of dry 30%-RH air spends nearly all of it lowering temperature, and the split can exceed 28 °F. The same coil pulling the same 12,000 BTU/h out of 70%-RH air spends much of it condensing water, and the split falls to about 15 °F. Both systems are working correctly.
Why the target moves with humidity
An air conditioner does not remove degrees; it removes enthalpy, the total heat content of the air including the heat locked up in water vapour. Total capacity across a coil is:
Q_total = 4.5 × CFM × Δh
where Δh is the enthalpy drop in BTU per pound of dry air and 4.5 is 60 × 0.075, the pounds of dry air moved per hour per CFM. Turn that around: if you know the capacity the coil delivers and the airflow through it, the enthalpy drop is fixed. At 11,000 BTU/h per ton and 400 CFM per ton, Δh = 11,000 ÷ (4.5 × 400) = 6.11 BTU/lb, whatever the humidity.
What is not fixed is how that 6.11 splits between temperature and moisture. Air leaving a wet evaporator sits close to saturation, around 90–95% relative humidity. So the calculation is: take the return air state, drop its enthalpy by Δh, find the temperature at which air at 95% RH has that enthalpy, and subtract. The wetter the return air, the more of the enthalpy drop goes into condensing water and the smaller the temperature change.
Furnaces are simpler, because combustion heat is entirely sensible. The rise is straight out of the sensible heat equation with no psychrometrics involved. That is also why the rise is such a good airflow measurement: with the output known from the nameplate, CFM = Q_output ÷ (1.08 × rise) gives you real airflow from two thermometers.
Worked example: a 3-ton system on a 75 °F, 50% RH day
You measure 75.0 °F and 50% RH in the return, and 55.0 °F in the supply plenum. The system is a 3-ton split running 400 CFM per ton.
- Measured split. 75.0 − 55.0 = 20.0 °F.
- Return humidity ratio. Saturation pressure at 75 °F is 0.43016 psia, so pv = 0.215 psia and W = 0.621945 × 0.215 ÷ (14.696 − 0.215) = 0.009237 lb/lb (64.7 gr/lb). That corresponds to a wet bulb of 62.6 °F.
- Return enthalpy. h = 0.240 × 75 + 0.009237 × (1,061 + 0.444 × 75) = 18.00 + 10.11 = 28.11 BTU/lb.
- Enthalpy drop. Taking 11,000 BTU/h delivered per ton at 400 CFM per ton: 11,000 ÷ (4.5 × 400) = 6.11 BTU/lb.
- Leaving enthalpy. 28.11 − 6.11 = 22.00 BTU/lb.
- Leaving temperature. Air at 95% RH reaches 22.00 BTU/lb at about 53.8 °F.
- Target split. 75.0 − 53.8 = 21.2 °F.
The measured 20.0 °F is 1.2 °F under target — inside normal measurement uncertainty, so the system passes this check and you stop here. Had it come out at 27 °F, the next step would be a static pressure reading; had it come out at 13 °F, the next step would be superheat and subcooling.
Reading the deviation
Treat ±3 °F as the noise band. Two field thermometers disagreeing by a degree, a probe seeing radiant heat, a slightly wrong humidity reading and a coil that is not quite at steady state will move the number by that much on a healthy system. Run the equipment for at least 15 minutes before you believe any reading.
Split wider than target. The usual cause is airflow below design: a loaded filter, a fouled evaporator, crushed flex duct, a closed damper, or a return that was never sized for the equipment. Measure total external static pressure and look the airflow up in the blower table before you touch refrigerant. A wide split can also appear on a legitimately dry day, which is exactly why this page asks for return humidity rather than assuming it.
Split narrower than target. Low refrigerant charge is the first suspect, followed by a compressor that is not pumping, airflow above design, and a room load the coil simply cannot overtake — a house at 85 °F pulling down after a hot day shows a narrow split for the first hour because the return air is hot and the coil is saturated. Take pressures and temperatures before concluding anything.
Furnace rise above the nameplate range. This is the one delta-T fault that is a safety issue rather than an efficiency one. Too little airflow across a heat exchanger overheats the metal, cycles the high-limit switch, and cracks heat exchangers over time. Fix the airflow — do not turn the gas down to hide it.
Furnace rise below the nameplate range. Either the blower is moving more air than the entered figure, or the furnace is firing below its rated input. Clock the gas meter and check manifold pressure before changing blower taps.
Target cooling split by return humidity
| Return RH | Return wet bulb (°F) | Target split (°F) | Supply air (°F) |
|---|---|---|---|
| 30% | 56.7 | 28.5 | 46.5 |
| 40% | 59.7 | 25.0 | 50.0 |
| 45% | 61.2 | 23.0 | 52.0 |
| 50% | 62.6 | 21.0 | 54.0 |
| 55% | 63.9 | 19.5 | 55.5 |
| 60% | 65.2 | 17.5 | 57.5 |
| 65% | 66.5 | 16.0 | 59.0 |
| 70% | 67.7 | 14.5 | 60.5 |
Change the capacity, airflow or return temperature above and the calculator rebuilds this table for your own assumptions. Manufacturer charging and performance data always take precedence over any generic table.
Things that corrupt a delta-T measurement
- Reading too soon. A system needs 10–15 minutes of run time to settle. A reading taken at three minutes shows a wide split because the coil is still cold and the house is still hot.
- Duct leakage between the probe and the equipment. A return that pulls 130 °F attic air through a leaking joint reads high, which narrows the apparent split and makes a healthy system look undercharged.
- Radiant heat on the supply probe. On a furnace, a probe in line of sight of the heat exchanger reads well above the air temperature. Get past the first elbow.
- Measuring a register instead of the plenum. Ducts in unconditioned space gain or lose heat on the way. Register temperature is what the room gets; plenum temperature is what the equipment made.
- Assuming humidity instead of measuring it. The target moves by roughly 14 °F between 30% and 70% return RH in the table above. Guessing the humidity is the same as guessing the answer.
- Diagnosing charge from delta-T alone. Delta-T is a symptom, not a diagnosis. It tells you to go measure something else — static pressure, superheat, subcooling — not what to adjust.
Never adjust refrigerant charge from delta-T
A wide or narrow split has at least four common causes, and only one of them is refrigerant charge. Adding refrigerant to fix a split that is actually caused by a dirty filter leaves you with a dirty filter and an overcharged system, which is worse than what you started with. Confirm airflow first, then charge by the manufacturer's method — subcooling for TXV and EEV systems, target superheat for fixed-orifice systems.
Where delta-T sits in a service call
Delta-T is a screening test. It is fast, non-invasive, and it costs nothing to take, so it belongs at the front of a diagnostic sequence rather than at the end of one.
A workable order on a cooling call: measure return dry bulb and humidity, measure supply dry bulb, and compare against the target here. If the split is off, measure total external static pressure and convert to airflow with the blower table — that separates the airflow causes from the refrigerant causes immediately. Only then connect gauges and take superheat and subcooling, which tell you what the refrigerant circuit is doing.
On a heating call the same logic runs in reverse: the temperature rise is the airflow measurement, provided you trust the nameplate output. Rearranging gives real CFM directly, which is often more reliable than a blower-table lookup because it does not depend on knowing the static pressure accurately.
If you want to work the airflow side of the same equation, the sensible heat calculator solves Q = 1.08 × CFM × ΔT in any direction, and the CFM per ton calculator checks the result against design airflow. For the humidity input, the wet bulb calculator converts between dry bulb, wet bulb and relative humidity.