P2279: intake air system leak — the code that names the fault

Air entering the engine after the mass airflow sensor has measured it is air the ECU does not know about. It fuels for the air it measured, the engine receives more, and the mixture goes lean.

P2279 is the ECU concluding exactly that.

Related: P0171 and P0174 lean mixture diagnostics and P2074 MAP/MAF and throttle correlation.

Where to look, in order of likelihood

1. The intake ducting between filter and throttle

The most common location, and the easiest to inspect. Rubber and plastic harden with heat cycles and crack — often on the underside where the crack is invisible without moving the duct.

Flex the ducting by hand with the engine off and look for cracks that open when it moves.

2. Vacuum hoses

Every hose connected to the intake manifold is a candidate: brake servo, PCV, purge valve, and any actuator that uses vacuum.

Old rubber goes hard and cracks at the ends, where it is stretched over a fitting.

3. The PCV system

Crankcase ventilation hoses and valves. On many engines the PCV valve itself is integrated into the valve cover, and a failed diaphragm inside it is an unmetered air path that no external inspection will find.

This is the leak people miss. If everything visible checks out, the PCV diaphragm is the next suspect: air filter and engine breathing.

4. Intake manifold gaskets

Particularly on engines with plastic manifolds, where the plastic distorts over time.

5. The brake servo

A failed servo diaphragm draws a large amount of unmetered air. The clue is a brake pedal that feels different — harder, or with a hiss when pressed.

Brake pedal feels spongy.

6. Injector seals

On direct injection engines, the seals where injectors enter the head can leak air, though this typically shows as a cylinder-specific problem.

7. Boost pipework on turbocharged engines

Same principle, higher pressure. A loose clamp or split hose after the compressor: turbocharger boost leak diagnostics.

The fuel trim signature

This is how you confirm a leak before crawling over the engine.

Log short-term and long-term fuel trim at idle, then at 2500 rpm.

Trims	Meaning
Strongly positive at idle, near zero at 2500	Vacuum leak
Positive at both	Fuel delivery shortfall or MAF under-reading
Negative at both	Too much fuel
Positive on one bank only	Leak on that bank’s manifold

The idle-versus-2500 difference is the whole test. A fixed-size leak is a large fraction of the small airflow at idle and a negligible fraction at higher RPM, so the trim correction shrinks as revs rise.

Fuel trim STFT and LTFT explained.

Finding it physically

A smoke test is the definitive method. Low-pressure smoke is introduced into the intake and escapes wherever the leak is, visibly. It is a workshop procedure and it finds in minutes what can take hours to find otherwise.

What not to do: spraying flammable liquids around the intake while the engine runs. It is a genuine fire risk in an engine bay with hot surfaces, and a smoke test does the job safely.

Symptoms

• Unstable or high idle: fluctuating idle RPM causes.
• Hesitation off idle.
• A stall when coming to a stop.
• Rough running when cold, improving warm: rough idle when cold.
• Increased fuel consumption as the ECU compensates.

Why to fix it rather than live with it

A lean mixture raises combustion temperature. Sustained, that promotes detonation and stresses valves and the catalyst.

The ECU compensates by adding fuel — up to its authority limit. Beyond that, it stops compensating and the engine runs genuinely lean.

Knock sensor diagnostics P0325.

Step by step

1. Record trims at idle and 2500 rpm — the signature confirms it.
2. Flex the intake ducting and inspect for cracks.
3. Check every vacuum hose end, where they harden first.
4. Check the PCV system, including the integrated valve.
5. Check the brake servo and its hose.
6. Smoke test if it is still not found.
7. Clear trims after the repair and confirm they settle near zero at idle.

Try it on the free tier of the app — pair an ELM327 adapter and compare fuel trims at idle against 2500 rpm. That single comparison distinguishes a leak from every other cause of a lean code.