Measuring the return-to-supply temperature split
The difference between return and supply air moves with humidity and airflow, so a single target number is misleading. How to take a reading worth reporting.
What this means
Subtract the supply-air temperature from the return-air temperature and you have a return-to-supply split. ACCA rejects a universal target because the expected result depends on OEM sensible capacity, measured airflow, indoor dry- and wet-bulb conditions, and outdoor conditions. A register-level reading is useful context, not a diagnosis or a direct measurement across the coil.
Equipment and model context
Cross-brand explanation; product figures come from the model documentation
- Residential ducted air conditioners and heat pumps in cooling mode
This describes a screening measurement an owner can take at a return grille and supply register on residential ducted equipment in cooling. Duct heat gain, leakage, and register location separate that result from a technician's entering- and leaving-air readings at the equipment. OEM data and measured airflow govern diagnosis.
What this guide covers
- The house cools slowly and the air from the registers does not feel as cold as expected.
- A technician has asked for a temperature reading before attending.
- Cooling performance has changed and there is no fault code to work from.
What changes the answer
- ACCA states that the expected split changes with sensible capacity, airflow, and indoor and outdoor conditions.
- NREL identifies filters, ducts, blowers, coils, and zoning as airflow variables and relates temperature difference to sensible load divided by airflow.
- Latent moisture removal does not appear in a dry-bulb temperature difference, so indoor humidity has to accompany the reading.
What the measurement is and what it leaves out
Air arrives at the indoor coil at one temperature and leaves at another. A thermometer at a return grille and supply register can estimate the change seen by the occupied space, but ducts between those points and the coil can add heat or leak air.
What the reading captures is only the sensible part of the work. When the coil removes moisture from the air, it is doing cooling that shows up as condensate in the drain pan rather than as a lower supply temperature.
So the same dry-bulb split can accompany different total cooling, and different splits can be correct at different sensible heat ratios. The measurement captures one component of performance rather than the whole of it.
Why a large split can be bad news
The intuition is that a bigger difference means better cooling. ACCA's relationship between sensible capacity, airflow, and temperature difference shows why that conclusion is incomplete: the expected split cannot be judged without airflow.
NREL lists a restricted filter, undersized or constricted ducts, blower settings, coil restrictions, and incorrectly sized zoned ducts among the conditions that alter airflow. A large split therefore cannot prove that the system is delivering the correct capacity.
Taken far enough the same mechanism produces ice. The coil runs below freezing, condensate accumulates as frost rather than draining, and airflow collapses further. Trane routes a frozen evaporator coil to a professional rather than to further measurement.
How to take a reading that is worth reporting
Let the system run in cooling for at least ten to fifteen minutes before measuring, so the coil and the ductwork have reached a steady condition. A reading taken in the first minutes of a cycle describes the warm-up rather than the operation.
Take the return reading at a return grille and the supply reading at a nearby representative register, avoiding direct sun. Note both figures and their locations rather than only the difference, because the result includes the duct path between the equipment and each grille.
Then record what the reading cannot see: the outdoor temperature, roughly how humid the house feels, when the filter was last changed, and whether the system had been running continuously. Those four notes turn a number into evidence.
Where the reading points, and where it stops
A smaller-than-expected split can accompany a refrigeration problem, a different sensible heat ratio, or an airflow condition. A register reading cannot separate those without OEM performance data and technician measurements.
A large split with weak register airflow makes the air side relevant: the filter, accessible registers, ductwork, blower, and coil. An owner can check the filter and unobstructed grilles; airflow verification at the equipment belongs to a technician.
Neither pattern is a diagnosis. Refrigerant charge is established with pressures and temperatures measured at the equipment by someone licensed to connect to the system, and a split taken at a register is a screening measurement that decides which conversation to have.
| Split across the coil | Airflow at the registers | What the pair is consistent with |
|---|---|---|
| Larger than expected | Weak | Airflow needs measurement; check accessible filter and grilles |
| Smaller than expected | Normal | Compare OEM data, indoor conditions, and refrigerant measurements |
| Smaller than expected | Weak | Take the reading again after correcting airflow |
| Matches OEM expectation | Verified | One performance check agrees at the recorded conditions |
| Falling during the cycle | Falling | Ice forming on the coil, which is a stop-and-service condition |
| Cannot get a steady reading | Varying | The system is cycling; time the cycles before measuring |
| Condition | Effect on the split | Effect on total capacity |
|---|---|---|
| Indoor humidity changes | Changes sensible heat ratio | Some capacity appears as moisture removal |
| Reduced airflow | Raises the calculated split at a given sensible capacity | System performance must be checked against OEM data |
| Increased airflow | Lowers the calculated split at a given sensible capacity | Only valid within the equipment's specified airflow |
| Outdoor condition changes | Changes the OEM expected capacity | Use the performance table for that condition |
Only one of them shows up on a thermometer. That is why a temperature split describes part of the picture and why the conditions have to be recorded with it.
- Sensible coolingLowers air temperature, and this is what the split measures
- Latent coolingRemoves moisture, and appears as condensate rather than temperature
- Airflow rateDivides the same capacity across more or less air
- The reading at the registerOne number produced by all three of the above
Questions people ask about this
What should the temperature difference across my AC be?
There is no universal figure. ACCA says the expected result depends on manufacturer sensible-capacity data, airflow, indoor wet- and dry-bulb conditions, and outdoor conditions.
Is a bigger temperature split better?
Not on its own. Temperature difference is related to sensible capacity and airflow, so a larger result cannot be judged without measured airflow and the manufacturer's expected performance at the recorded conditions.
Why is my split small when the house is humid?
Moisture removal is latent cooling and does not appear directly in a dry-bulb temperature difference. Indoor wet-bulb or humidity conditions are therefore part of judging the expected split.
How long should the system run before I measure?
Give it at least ten to fifteen minutes of continuous cooling so the coil and the ductwork have settled. A reading taken in the first minutes of a cycle describes the warm-up rather than steady operation.
Does a small split mean low refrigerant?
It is consistent with a refrigeration problem, and Trane lists low refrigerant among the reasons a system stops blowing cold air. It is equally consistent with a humid house, which is why the reading screens rather than diagnoses.
Can I use this reading instead of calling a technician?
It decides which conversation to have rather than replacing it. Charge is established with pressures and temperatures measured at the equipment by someone licensed to connect to the refrigeration circuit.
Evidence recordHow this page was checked
standards body publication, government guidance, official manufacturer support article · checked 2026-09-08
Every technical claim above was written from primary documentation held in the HVAC Bench evidence record: Air Conditioning Contractors of America, National Renewable Energy Laboratory and Trane technical literature. Where a source limits a definition to certain models, test conditions, or product classes, that limit is repeated here rather than generalised.
- Documentation class
- standards body publication, government guidance, official manufacturer support article
- Scope of the definition
- Confirm against the exact model manual
- Last checked
- 2026-09-08