The big picture:

The cooling trade has no single target for that gap. One training site notes that blower type and heat load shift the math, calling the old 20-degree rule too simple to be a strict standard [1]. A smart thermostat maker lists the best range at 16 to 22 degrees, while admitting the exact target changes depending on who you ask [2]. An owner holding a report with just one number is looking at a reading with no firm baseline.

Both sources assume the tech put the probe right across the coil. The training site clearly warns against taking this reading at the room vents, because heat is gained and lost in the ducts before the air ever reaches a room [1]. Vents are the wrong place to look. Leaky ducts alone add 15 to 20 percent to the energy a system burns, even in a new house [3], and a vent reading includes whatever the duct lost or gained on the way. A split above that 22-degree ceiling can point to a lack of air, from a dirty filter or a fouled coil [2].

A split below 16 degrees opens a list of eight suspects, and a bad compressor sits way down at number six [2]. What strikes me is that the compressor is the costliest item on the list. It is the only suspect that turns a basic service visit into a giant sales quote.

By the numbers

  • 14 to 23 degrees — The split band: A trade training site gives this broad range for healthy cooling, warning that fixed rules fail because blower type and load move the number [1].
  • 350 cfm per ton — Air before charge: The inspection standard requires indoor coil airflow to clear 350 cubic feet per minute (cfm) per ton before a tech can check the refrigerant charge [3].
  • 15 percent — Low charge frequency: Field studies find both new and working systems run low on refrigerant by about this much, and units often arrive that way on day one [5].
  • $72 to $144 — Cost of a weak system: That is what a 10% to 20% efficiency loss adds to a summer of cooling, priced against the 2024 average household bill of $719 [5].

What I’d watch:

What I am watching is the paper that comes with a verdict. The numbers a tech writes down say much more than the final price at the bottom. A split logged with exact probe spots, an airflow figure, and a refrigerant charge reading turns a sales call into a real checkup. A single temperature on a work order does neither.

  • The coil-side number: A split read right before and after the coil, far enough out for the air to mix, is the only version that compares well with published bands [1].
  • The airflow proof: Refrigerant charge cannot be judged until airflow clears the required baseline, which is a step a rushed service call tends to skip [3].
  • The charge figure: The ticket needs to show superheat for a fixed metering device or subcooling for a thermostatic expansion valve (TXV), written down rather than just spoken [3].
  • The replacement floor: A new split-system condenser manufactured on or after January 1, 2023 must clear 13.4 Seasonal Energy Efficiency Ratio 2 (SEER2), and a heat pump needs 14.3 SEER2 with 7.5 Heating Seasonal Performance Factor 2 (HSPF2) [4].
  • The wasted kilowatt-hour: At the 2025 average home price of 17.30 cents per kilowatt-hour, a system missing its rated output pays that gap back every month it runs [6].

The catch

Wet air is the honest snag in these checks. When indoor air holds more moisture, the coil spends its power pulling that water out. The cooling split then drops for a reason that has nothing to do with a broken part [1]. That is physics, not a fault.

One summer reading is just a quick look at one afternoon’s heat load. That is exactly why the thermostat maker suggests tracking the number over time rather than trusting a single glance [2]. A split that lands right inside the band is not a clean bill of health either. A mild day can easily hide a charge fault that a hot afternoon would expose.

Where I have landed is that the split is a sorting tool rather than a final ruling. It narrows eight suspects down to two or three, and it makes a compressor pitch something you can check. The case for a new compressor needs a set of numbers read together, spanning airflow, gas charge, and the split across the coil. An owner must demand that the reading came from the right place, and that the airflow and charge numbers exist at all. The same caution applies to the tools themselves: a thermal image is a pointer, not an answer, and a reading taken at a distance deserves the same suspicion.

At a glance

  • The Big Shift: The reading that decides between a basic fix and a total compressor swap is a temp split taken across the coil, making the probe spot more vital than the number itself.
  • Why It Matters: A compressor is the most costly part on the low-split suspect list, meaning a reading taken at a vent can turn a small valve repair into a huge quote without anyone lying.
  • What I’d Watch: Whether a service ticket carries the three numbers a true checkup needs, and how the federal output floor shows up on a new quote.
  • Temperature split: The gap between the air going into the cooling coil and the air coming out, read right at the coil rather than at a room vent.
  • Superheat and subcooling: The refrigerant charge readings taken at the lines, one metric for a fixed device and another for a valve.
  • Airflow floor: The minimum air volume per ton of cooling that must move through the system before a charge reading means anything.
  • The Catch: Wet indoor air lowers the split on its own, making a single reading just a snapshot of one afternoon that cannot clear a system of a charge fault.

Sources

[1] HVAC School, Bryan Orr, “What should the Air Delta T be? (Air Temperature Split),” 2019-10-16. https://www.hvacrschool.com/videos/what-should-the-air-delta-t-be-air-temperature-split [2] Resideo, “Why Delta T is the magic number for proactive diagnostics and problem-solving,” 2021-05-03. https://www.resideo.com/us/en/corporate/newsroom/featured-articles/why-delta-t-is-the-magic-number-for-proactive-diagnostics [3] InterNACHI, “HVAC” (inspection article and standards commentary). https://www.nachi.org/hvac.htm [4] Electronic Code of Federal Regulations, 10 CFR 430.32, Energy conservation standards (edition current 2026-10-29). https://www.ecfr.gov/current/title-10/section-430.32 [5] RMI, “Why Refrigerant Matters More Than You Think for AC (or Heat Pump) Efficiency.” https://rmi.org/resources/why-refrigerant-matters-more-than-you-think [6] U.S. Energy Information Administration, “Electricity explained: Prices and factors affecting prices” (2025 annual averages). https://www.eia.gov/energyexplained/electricity/prices-and-factors-affecting-prices.php