AutoLogicTools

OBD-II trouble code

P2073: Manifold Absolute Pressure / Mass Air Flow - Throttle Position Correlation at Idle

The words that matter in this title are the last two. At idle the engine is swallowing a few grams of air a second, so a leak that is trivial at speed is an enormous fraction of the total — which is why this code fires at idle and stays quiet everywhere else.

Medium severityPowertrainFuel & Air MeteringDrivable short-term

Quick facts

System
Powertrain
Category
Fuel & Air Metering
Severity
Medium severity
Drivable
Usually safe to drive short-term
Repair cost range
$90$700
DIY difficulty
Intermediate DIY

Browse every code in P2000–P20E8, or start from the full code library.

What does P2073 mean?

P2073 is set when the module compares the throttle opening it commanded against the airflow the mass air flow meter or manifold pressure sensor reports, finds they disagree by more than the calibration allows, and notes that the disagreement occurred at idle or close to it.

The operating point is the entire code. A typical four-cylinder engine at warm idle passes somewhere around 2 to 4 grams of air per second. The same engine at full throttle near its power peak passes 90 to 120. That is a range of roughly thirty to one, and it means an unmetered hole with a fixed flow behaves completely differently at the two ends. Suppose a split hose admits 1.5 grams per second of air that the meter never counted. At idle that is something like a third to a half of everything entering the engine — an error no plausibility monitor can miss. At full throttle it is barely one part in seventy, comfortably inside the noise. This is why an idle-specific correlation fault is a strong statement rather than a vague one: it says the extra air is roughly constant, and a roughly constant quantity of unmetered air is a hole, not a sensor.

Vacuum reinforces the same conclusion. The pressure differential across any leak is greatest exactly when the throttle is closed — twenty inches of mercury or so at idle, falling toward nothing as the throttle opens. So the leak flows its hardest at the one operating point where the engine can least afford it. A meter that had genuinely drifted, by contrast, would be wrong in proportion at every flow rate and would set the higher-load correlation code as well. The presence of P2073 with its higher-load counterpart absent is therefore the single most useful piece of evidence available before anybody opens the bonnet.

The two leak paths worth checking first are not the ones people picture, because neither of them is a vacuum hose. The first is the brake booster. It is a large sealed chamber permanently connected to the manifold through a big-bore hose, and its diaphragm can leak only under certain conditions — commonly when the pedal is pressed, sometimes only when it is released. That makes it the leak that comes and goes with the driver's foot, and the test for it takes seconds: watch the correlation data or the fuel trims while pumping the brake pedal at idle with the vehicle stationary. A booster fault also tends to arrive with a hard pedal or a hiss from behind the pedal box, which no other leak produces.

The second is the crankcase. The ventilation system connects the entire inside of the engine to the manifold, metered only by the calibrated orifice inside the PCV valve. If that valve sticks open, or its hose splits, or the oil filler cap seal has perished, or the dipstick tube is not seated, the engine gains an unmetered opening the size of that path. The result is indistinguishable from a vacuum leak, is invisible on a casual under-bonnet look, and is repaired for the price of a cap. It is also the version of this fault that will keep returning if only the symptom is treated.

What the driver notices is mild and easy to dismiss: an idle that wanders or sits higher than it should, an occasional stall coming to a stop, a light stumble pulling away, and long-term fuel trims sitting well positive at idle but returning to normal on the motorway. The trim pattern is the clue in miniature — a lean correction that exists only at idle and vanishes under load is the same fact this code is reporting, said a different way.

Common causes

  • Leaking brake booster diaphragm or check valve, often only with the pedal applied
  • Stuck-open or worn PCV valve admitting unmetered air through the crankcase
  • Split, hardened or disconnected crankcase ventilation hose
  • Loose oil filler cap or unseated dipstick tube
  • Cracked intake boot between the meter and the throttle body, opening only when the engine rocks
  • Failed intake manifold gasket, most often at one end of the manifold
  • Perished small vacuum lines to the purge valve, EGR or an accessory actuator
  • EGR valve leaking open at idle, admitting exhaust the meter never counted
  • Purge valve stuck open, connecting the manifold to the evaporative canister
  • Heavy carbon deposit narrowing the throttle bypass so the commanded opening no longer matches the flow

Symptoms

  • Unstable, hunting or higher-than-normal idle
  • Occasional stall when coming to a stop or when the air conditioning cuts in
  • Slight stumble or hesitation just off idle that smooths out as speed builds
  • Long-term fuel trim strongly positive at idle and near normal at cruise
  • Check engine light with no real complaint at road speed
  • Hiss audible under the bonnet with the engine idling
  • Hard or high brake pedal accompanying the engine complaint
  • Cold idle noticeably worse than warm idle
  • Additional lean codes such as P0171 or P0174 appearing alongside
  • Fuel economy essentially unchanged, because the fault disappears once the throttle opens

Diagnostic steps

  1. 1.Check first whether the higher-load correlation code is also stored. P2073 alone is evidence of a fixed unmetered leak; P2073 together with P2074 points instead at the meter itself or at something wrong with the model's inputs, and sends the diagnosis in a different direction entirely.
  2. 2.Read long-term fuel trim at idle and again at a steady 2,500 rpm. A large positive correction at idle that falls away at speed is the same finding as the code and confirms the leak is fixed in size rather than proportional to flow.
  3. 3.Test the brake booster before anything else, because it is quick. With the engine idling and the vehicle stationary, watch idle speed and fuel trim while pressing and releasing the brake pedal several times. A leak that responds to the pedal is the booster or its check valve.
  4. 4.Work through the crankcase path deliberately: PCV valve, both hoses, the oil filler cap seal and the dipstick seating. Removing the oil filler cap at idle should disturb the idle noticeably on a healthy engine; no reaction at all suggests the system is already leaking elsewhere.
  5. 5.Smoke-test the intake tract from the meter to the valves with the throttle closed. This is the operating point the code describes, so testing at closed throttle reproduces the real pressure differential.
  6. 6.Flex the intake boot by hand while watching live airflow data. A split that opens only when the engine rocks on its mounts is a common and easily missed idle-only leak.
  7. 7.Command the purge valve and the EGR valve closed with a scan tool, then look for the idle quality and the correlation error to improve. Either valve leaking at idle produces this code with a perfectly intact intake.
  8. 8.Compare the reported airflow at warm idle against what the engine should be drawing. A rough sanity figure is around 1 gram per second for each litre of displacement at idle; a value well above that with the throttle closed is the unmetered air showing up in the number.
  9. 9.Inspect the throttle bore and plate for deposit. Carbon narrowing the bypass path changes the relationship between commanded position and actual flow at idle, which is exactly what this monitor watches.
  10. 10.Only after the intake tract is proven sealed should the meter, the manifold pressure sensor or the throttle position sensor be suspected, and then confirm by data rather than by replacement.

Repair cost

$90$700

Diagnosis with a smoke test is $90 to $190. The cheap outcomes are genuinely cheap: an oil filler cap or a length of vacuum hose is $10 to $60 fitted, and a PCV valve and hose run $10 to $260 in parts with 0.3 to 3 hours of labour depending on access. An intake boot is $40 to $220 installed. A brake booster is the expensive end at $180 to $650 in parts with 1 to 3 hours of labour. An intake manifold gasket set runs $150 to $600 installed depending on how much has to come off first. Replacing the airflow meter should be the last resort on an idle-only correlation code, not the first move.

Estimate your repair

Run the numbers for your vehicle

Open the Repair Cost Estimator with vacuum leak repair preselected. Adjust labor rate and vehicle category to fit your situation.

DIY vs shop

This is an intermediate DIY job. It usually involves diagnostic steps, specialty parts, and some careful work in tight spaces. If you have the tools and a service manual or trustworthy video for your specific vehicle, it is achievable in a weekend. Otherwise, a competent independent shop will be faster.

Related codes

Frequently asked questions

Can I keep driving with P2073?

Yes, in ordinary use. The fault lives at idle, so the car generally drives normally once it is moving, and the risk is inconvenience rather than damage — a stall at a junction is the worst of it. The reason to deal with it fairly promptly is that unmetered air makes the engine run lean at idle, and a persistent lean condition eventually brings its own codes and, over a long enough period, its own consequences for the catalyst. If the stalling is frequent or happens in traffic, treat it as urgent for safety reasons rather than mechanical ones.

Why does the code only appear at idle if the leak is always there?

Because the engine's appetite for air changes by roughly thirty to one between idle and full throttle, while the leak stays about the same size. A hole passing one and a half grams of air a second is a third of everything the engine is breathing at idle and about one part in seventy at full power. On top of that, manifold vacuum — the pressure difference driving air through the hole — is at its maximum with the throttle shut and collapses as it opens. The leak is therefore both largest in proportion and strongest in absolute terms at exactly the operating point this code watches.

How is this different from P0068?

P0068 reports the same disagreement without saying when it happened. P2073 and P2074 split that judgement by operating point, and the split is what makes them worth having: an idle-only fault argues for a fixed leak, while a fault that appears under load argues for the meter reading low or for something restricting the intake. If you have P0068 on its own, you still have to establish the operating point yourself before the cause list narrows.

Could a bad brake booster really set an engine code?

It is one of the more common causes and one of the most frequently overlooked. The booster is a large chamber tied straight into the intake manifold, so a leaking diaphragm is effectively a big vacuum leak that the airflow meter never sees. What distinguishes it from the rest is that it often leaks only when the pedal moves, so the idle disturbance comes and goes with the driver's foot. Pumping the brake at idle while watching idle speed and fuel trim will usually find it in under a minute.

Editorial context

About This Diagnostic Information

AutoLogicTools diagnostic guides explain OBD-II trouble codes using recognized code definitions, standard automotive diagnostic principles, and practical automotive context. A trouble code records a condition detected by a control module. It does not automatically identify a failed part, and the right diagnostic procedure can vary by vehicle.

Manufacturer service information, technical service bulletins, wiring diagrams, and vehicle-specific procedures should take precedence when available.

AutoLogicTools was founded by Vincent Fisk, an automotive locksmith and shop owner in San Diego with hands-on experience in vehicle keys, immobilizer systems, electrical issues, modules, programming, and diagnostics.