Temperature Split (Delta-T) Explained
The temperature split is the fastest gut-check you own. Two probes, thirty seconds, and you know whether a cooling system is moving air and pulling heat the way it should — or hiding a problem.
Turn your split into CFM
Enter delta-T and capacity to back into actual airflow and see if the blower is delivering.
In This Guide
What Temperature Split Actually Measures
Temperature split — also called delta-T or the evaporator split — is simply the difference between the air going into the coil and the air coming off it in cooling mode. It is the most direct evidence you have that the system is removing heat.
The Formula
Both measured dry-bulb, at the plenums or registers, with the same thermometer to cancel out probe error.
The reason this one number tells you so much is that it sits at the intersection of the two things that have to be right for a coil to work: airflow and refrigerant heat transfer. The sensible heat equation ties them together:
The constant 1.08 = 0.075 lb/ft³ (air density) × 60 min/hr × 0.24 BTU/lb/°F (specific heat of air).
Rearrange that equation and delta-T becomes a mirror. For a fixed cooling capacity, if airflow drops the split rises, and if airflow climbs the split falls. That inverse relationship is the whole reason a two-second temperature reading can point you straight at a duct restriction, a slipping blower, or a charge problem.
The Target Range (and Why It Moves)
For most residential and light-commercial cooling, a healthy delta-T lands between 14°F and 22°F, with the sweet spot around 18–20°F on a properly charged, properly ducted system running near design conditions.
Usually low airflow — dirty filter, closed dampers, undersized returns.
Normal window. Trend toward the high end in dry climates, low end in humid ones.
Suspect low charge, high airflow, or a coil that is not pulling capacity.
The range moves because delta-T is sensible only — it does not see the latent (moisture) heat the coil is also removing. On a humid day, a large share of the coil's capacity goes into condensing water vapor rather than dropping air temperature, so the split reads lower even though the system is working hard. The same equipment in Phoenix in July will show a noticeably higher split than in Houston.
Rule of thumb: Cross-reference the return-air wet bulb against the entering dry bulb. Higher indoor humidity (higher wet bulb) shifts your expected split lower. A 15°F split at 67°F wet bulb can be perfectly normal, while the same 15°F at 57°F wet bulb is a red flag.
How to Measure It Correctly
The number is only as good as the measurement. Follow these steps in order:
Step 1 — Run it first. Let the system cool for a solid 10–15 minutes. Readings taken on a fresh startup are meaningless because the coil and plenum haven't stabilized.
Step 2 — Return air. Probe the return plenum or a return grille well away from any door, window, or the coil itself. You want the true mixed-air temperature entering the equipment.
Step 3 — Supply air. Probe the supply plenum past the first elbow. Measuring in line-of-sight of the cold coil lets the probe "see" the coil radiantly and reads artificially low.
Step 4 — Subtract. Return minus supply is your split. Use the same instrument for both readings so any calibration offset cancels out.
Best Practice
Drill dedicated 3/8" test ports in the supply and return plenums and plug them with a snap grommet. Consistent probe placement trip after trip is what makes delta-T a repeatable diagnostic instead of a guess.
When the Split Is Too High
A split above 22°F almost always means not enough air is moving across the coil. The refrigerant has plenty of time to pull heat out of the small amount of air passing through, so that air comes off very cold — but the total heat removed from the house is actually low. Common causes:
- Dirty air filter — the number-one call. A loaded filter can choke airflow 20–40%.
- Dirty evaporator coil — a matted coil face restricts airflow and insulates the fins.
- Undersized or crushed return ductwork — the blower can't pull enough air back.
- Closed dampers, shut registers, or a collapsed flex return.
- Blower running too slow — wrong tap on a PSC motor, a slipping belt, or a bad ECM program.
- Iced coil — low airflow and low charge both cause icing, which then further blocks airflow.
Watch for freeze-up: A high split combined with a low suction pressure is the classic set-up for a coil that will ice over. Starving the coil of airflow drops coil temperature below 32°F. Fix the airflow before you touch a gauge on the charge.
When the Split Is Too Low
A split under 14°F means the coil is not pulling its rated heat out of the air. Two families of causes:
Refrigerant side
- Low charge / undercharge
- Restricted metering device or filter drier
- Failing compressor not pumping
- Non-condensables in the system
Airflow side
- Blower running too fast (over-airflow)
- Return-air leakage pulling in hot attic air
- Wrong blower tap for the tonnage
- Very high indoor humidity masking the split
This is exactly why delta-T is a screening tool, not a verdict. A low split tells you something is off; superheat, subcooling, and static pressure readings tell you which something. For the refrigerant-side confirmation, pair this reading with a static pressure check and a look at your superheat and subcooling values.
Ventora — Your AI HVAC Companion
The AI HVAC assistant in your pocket: airflow and charge calculators, code lookups, and instant diagnostics from a photo of your gauges. Free on iOS.
From Delta-T to Actual Airflow
The real power of the split shows up when you combine it with the system's known capacity to back into actual airflow. Rearrange the sensible heat equation to solve for CFM:
Use the sensible portion of capacity (roughly 75% of total for a typical AC), not the nameplate total.
Design airflow for cooling is about 400 CFM per ton — so a 3-ton system should be moving roughly 1,200 CFM. If your delta-T math says you're only pushing 900 CFM, that's a 25% airflow deficit you can now go chase in the ductwork instead of guessing.
| Measured ΔT | Likely Airflow (3-ton, ~27k sensible) | Read as |
|---|---|---|
| 12°F | ~2,080 CFM | Over-airflow or low charge |
| 15°F | ~1,665 CFM | High side of normal |
| 18°F | ~1,390 CFM | On target |
| 20°F | ~1,250 CFM | Near 400 CFM/ton design |
| 24°F | ~1,040 CFM | Low airflow — investigate |
Values assume ~27,000 BTU/hr sensible capacity. Real sensible capacity shifts with indoor and outdoor conditions, so treat these as directional, not exact.
Worked Example: 3-Ton System
Scenario
A homeowner complains the house won't cool below 78°F on hot afternoons. You're at a 3-ton (36,000 BTU/hr) straight-cool split system. Return air reads 78°F, supply reads 54°F.
Step 1 — Find the split.
ΔT = 78°F − 54°F = 24°F
Step 2 — Estimate sensible capacity. ~75% of total.
36,000 × 0.75 = 27,000 BTU/hr sensible
Step 3 — Back into airflow.
CFM = 27,000 ÷ (1.08 × 24) = 27,000 ÷ 25.92 ≈ 1,042 CFM
Step 4 — Compare to design. A 3-ton needs ~1,200 CFM.
1,042 ÷ 1,200 ≈ 87% of target — about a 13% airflow shortfall
Diagnosis
The 24°F split and the ~1,040 CFM estimate both point the same direction: airflow restriction. Before condemning charge, pull the filter, inspect the coil face, and take a total external static pressure reading. The system is likely making cold air but not enough of it to satisfy the load on a design day.
Field Tips and Common Mistakes
- Delta-T is sensible only. On humid days a low split can be normal — check wet bulb before you condemn anything.
- Never diagnose charge on split alone. Confirm with superheat/subcooling on the refrigerant side.
- Mind probe placement. A supply probe that sees the coil reads several degrees low and fakes a high split.
- Give it time. Ten minutes minimum; a system that just cycled on hasn't stabilized.
- Airflow first, then refrigerant. Restricted airflow skews every pressure and temperature you take afterward.
- Heating rise is a different target. For a gas furnace, temperature rise runs 35–75°F per the nameplate — don't apply cooling numbers to a furnace.
Bottom line
Treat delta-T as the opening question on every cooling call, not the closing argument. It costs you thirty seconds and instantly narrows a no-cool or weak-cool complaint to either an airflow problem or a refrigerant problem. From there, static pressure, superheat, and subcooling finish the story.
Your HVAC Reference, Powered by AI
Ventora — the AI HVAC assistant in your pocket. Run airflow and charge calculations, snap a photo of your gauges for instant analysis, and get code answers in seconds. Free 7-day trial, then $19.99/month.