How Cold Should AC Air Be? The Delta T Test

Why You Measure Close To The Coil

A register in the far bedroom, at the end of forty feet of flex duct baking in a 135 degree Sunnyvale attic, reads warmer than a register six feet from the air handler. That is not a cooling problem, that is duct heat gain, and it is a different conversation.

Measuring at the nearest register gets you as close to true coil output as possible without cutting into the plenum. If the near register shows a 19 degree split and the back bedroom shows 9, congratulations, you just diagnosed your own duct insulation problem without lifting a panel.

Conditions That Make The Test Valid

Run it wrong and the number is worse than useless, because you will act on it.

• Clean filter. Non-negotiable. A restricted filter drives the split up and makes a sick system look healthy.
• Fifteen minutes of continuous runtime minimum. No short cycling, no starting the timer at switch-on.
• Do not test during a heat wave with a hot house. A house sitting at 88 all day feeds the coil abnormally hot, humid air and the split scatters. Test in the morning, or on a moderate day with the house near setpoint.
• Do not test when it is cool outside. Below roughly 70 degrees F outdoor, system pressures fall out of their normal range and none of the numbers mean what they usually mean.
• Measure in the return airstream and at the nearest supply register. Not at the thermostat, not against a wall.
• Humidity counts. With the marine layer parked over the Peninsula, more of the coil’s capacity goes into wringing water out of the air instead of dropping its temperature. That legitimately shrinks the split a couple degrees with nothing wrong.

What A Low Split Means

Under 16 degrees F and the system is not removing enough heat. Common causes, roughly in the order I find them:

• Low refrigerant charge, which means a leak, because refrigerant does not evaporate out of a sealed system on its own.
• Overcharge, which sounds backwards but is real. Too much refrigerant floods the condenser, drives head pressure up, and costs capacity.
• A filthy outdoor condenser coil that cannot reject heat, a whole genre of problem in the dusty East Foothills.
• A weak compressor no longer pumping to spec.
• Air or nitrogen mixed into the charge from a sloppy prior repair.
• Return duct leakage pulling 130 degree attic air into the system, which raises return temperature and shrinks the apparent split.

Most of those require gauges to separate. A low split tells you to call somebody. It does not tell you what to buy.

What A High Split Means, And Why It Is Not Good News

This is the counterintuitive part, and the most useful thing on this page.

A split above 22 degrees F is usually not a powerful air conditioner. It is usually low airflow.

The coil removes a certain amount of heat per minute. Push half the normal volume of air across it and each cubic foot that makes it through gets colder, because the same cooling power is spread over less air. The register spits out 48 degree air, feels fantastic on your hand, and the house still never reaches setpoint. Total heat removal went down even though the air feels colder.

Worse, low airflow drops the refrigerant’s boiling temperature inside the coil. Once that saturation temperature falls below 32 degrees F, condensate stops draining and starts freezing.

So a 26 degree split is not a bragging right. It is an early warning. Dirty filter, crushed duct, closed registers, dust caked in the blower wheel, or a coil nobody has cleaned since the Obama administration. The full breakdown lives in the guide to static pressure and HVAC airflow problems.

Be Honest About What This Test Is

Delta T is a screening test. A smoke alarm, not a fire investigation.

It tells you something is wrong. It does not tell you what. A 12 degree split could be undercharge, overcharge, a fouled condenser, a dying compressor, or leaky return duct. Those five share one symptom and nothing else, and the parts, labor and price are completely different.

It has real blind spots too. It cannot see an overcharge on its own, and it cannot see whether coil saturation temperature has slipped below freezing. A system can post a picture perfect 19 degree split while carrying a charge that is quietly grinding the compressor down.

The real diagnosis is a stack of measurements taken in order. Filter, then measured airflow, then the temperature split, then gauges after 10 to 15 minutes of stable runtime, then superheat and subcooling read against the actual metering device in that coil box. Superheat reads what the evaporator is doing. Subcooling reads what the condenser is doing. Together they name the fault. Alone, neither does, and neither does delta T.

What delta T buys you is leverage. Walk into the conversation with a measured 13 degree split and a clean filter and you are a customer with data instead of a customer with a feeling.