Half of Air Conditioners Are Not Correctly Charged

Refrigerant charge is the one thing about an air conditioner a homeowner cannot check, cannot see, and has no way to verify was done correctly. It is also, according to the field measurements, wrong about half the time.

The measurement

The National Renewable Energy Laboratory pooled eleven field studies covering more than 2,600 residential systems. The split:

Note the tail on that: only around 28% were at the exact recommended charge. The “correctly charged” half includes systems inside an acceptable band rather than on the number.

A separate prevalence estimate, in NREL and Purdue's fault prioritization work, puts nonstandard charge at 72% of residential split systems against 42% of rooftop units, and condenser fouling at 44% of split systems against 4.8% of rooftops. The two studies use different definitions of “wrong”, which is why the numbers differ — but neither describes a rare condition.

Airflow is the same size problem

An audit of 4,168 air conditioners found improper airflow in 44% of them. This matters for a reason beyond the number: airflow faults and charge faults produce overlapping symptoms, and the standard field mistake is adding refrigerant to a system whose real problem is that air is not moving across the coil.

What it costs, nationally

NREL modelled those measured fault distributions against the U.S. housing stock. The result:

The same work finds that around 90% of homes pay some energy penalty from airflow and charge faults. Not a catastrophic penalty in most cases — a few percent — but paid every cooling hour of every summer, invisibly.

What a fault actually does to a system

NIST put numbers on the individual faults on an instrumented test stand. A 30% undercharge cost roughly 15% of capacity and 12% of COP. A 30% overcharge cost about 3% of COP while raising total power around 6%. Kim and Braun, quoted in DOE's 2018 review, measured 12–19% undercharge at about 7.6% lower efficiency.

Undercharge is the worse of the two by a clear margin — and it is the one that gets worse on its own, because a leak keeps leaking.

Why nobody notices

A system down 15% of capacity still cools. It runs longer to do it. On a mild day it keeps up perfectly well and nothing looks wrong at all. The shortfall only becomes visible on the hottest afternoons, when the reduced capacity finally cannot match the load — which is exactly when a service call is most expensive and least available.

That is the whole argument for measuring rather than waiting. Not that a fault is dramatic, but that it is quiet, common, and priced into every bill until somebody looks.