Skip to content
Shop
Explore

INDEPENDENT BMW ENGINE EXPERTISE

Intake manifold. Problems of air-tightness

N43 / N53 · NOXEM · DIAGNOSTICS

Select a technical image to view it in detail.

Air being sucked into the intake manifold is one of the more common faults I see. It is common, yet its symptoms vary and sometimes the fault hides. This article covers the typical symptoms, how to confirm the cause from fuel adaptation values, why the fault shows up mainly at idle, and how to find the leak.

Symptoms

  • Very uneven idle directly after starting the engine: the engine shakes and misfires, RPM “jumps” and so on. It can also appear suddenly on a warm engine.
  • Sometimes, while the idle is uneven, there is a loss of power at low torque (pulling away from a stop).
  • Fuel mixture error messages (lean or rich), and the EML symbol lights up in the KOMBI.
  • If the cause is a damaged CCVV, increased oil consumption.*
  • Often, and very often when there are lean mixture messages, cylinder misfire messages accompany the fault.

\* Increased oil consumption damages the Lambda probes and the CO catalytic converters (on N43/N53 also the NOx catalytic converter and the NOx sensor).

The symptoms vary, so typical scenarios are described in more detail below.

Typical behaviour of the fault

  • It can be sporadic: it appears and then disappears.
  • It shows up mostly at the moment of starting, but it can also appear suddenly on a warm engine.
  • It is most noticeable at idle and at low required torque, and does not affect the high-torque range.

Confirming the cause from fuel adaptations

Any of the following points to this fault:

  • The idle (offset LTFT) fuel adaptations are 0.00, meaning they are not confirmed. The DME does not confirm offset-type adaptations if the fuel correction required at idle is significant (exceeds the planned range) and/or changes significantly.
  • The idle (offset LTFT) fuel adaptations are strongly increased, close to the maximum value (1.0/1.5 mg/stk for N43/N53).
  • The idle (offset LTFT) adaptations change (are overwritten) or vary significantly with temperature, by more than ±0.2 .. 0.3 mg/stk per bank.

When lean mixture errors are recorded

Suppose additional air, which the air mass meter does not measure, is sucked into the intake manifold when the engine starts, and the leak gradually decreases or disappears as the engine warms up. Directly after starting, as soon as the wide-band Lambda probes are heated, the DME tries to correct the situation:

  • It raises the integrators to +30% (typical for MSV/MSD systems; other systems have maximum values of 20 .. 30%). Lambda rises above 1.00 on both banks, which confirms the lean mixture.
  • If that does not help, it modifies the offset-type (LTFT) adaptations upward, with a positive step.
  • It records a lean mixture error if, on reaching the 30% limit, it cannot reach a stoichiometric mixture (Lambda 1.00) within time X (usually 5 .. 10 seconds).

When rich mixture errors are recorded

These errors are somewhat rarer. They arise as follows.

If the fault appears gradually on a warm engine, the DME modifies (overwrites) the offset LTFT several times to significantly positive values (above +1.0 mg/stk). It injects additional fuel to compensate for the air that is sucked in and not measured by the air mass meter, and so keeps the mixture correct.

If the fault does not manifest on a cold engine, or on a warm engine at start, the DME detects that the mixture is very rich (because of the modified LTFT). Then:

  • It quickly reduces the integrators to -30% and modifies the offset-type LTFT downward, with a negative step.
  • Lambda drops below 1.00 on both banks, which confirms the rich mixture.
  • It records a rich mixture error if, on reaching the 30% limit, it cannot reach a stoichiometric mixture (Lambda 1.00) within time X (usually 5 .. 10 seconds).

Why the fault shows at idle and low torque

At idle the intake manifold has a very strong air shortage. This applies to N43/N53 in Homogeneous mode, and to the other N-series engines in “non-Valvetronic” mode, which is the state directly after starting. On M-series engines it applies without any additional condition. The throttle is practically closed and the engine tries to draw air into the cylinders “with all its power”. Under these conditions even a tiny leak, 1 .. 2 mm in diameter, can completely upset the fuel mixture stabilization. A small crack or hole is enough to make the DME unusable: misfires, shaking, loss of torque and fuel trim error messages.

At higher torque the throttle opens, so the air shortage in the intake manifold decreases and the throttle opening area increases. The effect of the leak drops very quickly and is close to zero by medium torque demand.

Most typical causes

  • A damaged CCVV membrane. Air is sucked in through the membrane’s ventilation holes, on the top of its body.
  • A defective CCVV pipe, usually close to the engine block in the hardest place to see.
  • Defective O-rings on the EGR valve or the DISA valve.

A leak upstream of the throttle, a damaged pipe to the intake manifold, has a similar effect. It causes problems both at idle and in the partial load range, with error messages for lean mixture and for air mass meter plausibility/data conformity. The good news is that this fault is easy to find by visually checking the pipe.

Locating the leak

To check the CCVV membrane for damage, read here.

To check other places for airtightness, a smoke test is normally used, because it identifies the leaking spots. How to perform it is described here.

Keep some caveats in mind:

  • During a smoke test the engine is not running, so it does not vibrate. If the cause is a pipe connection, the fault may appear only with the engine running and vibrating. When performing the smoke test, try to move all the connections of the intake manifold.
  • If the cause is a damaged CCVV membrane, the increased pressure during the smoke test may press the damaged spot shut, and the fault “hides”. The CCVV membrane therefore has to be checked separately, by checking for air suction through the membrane’s ventilation holes while the engine is running.
  • During a smoke test the pressure difference between the surroundings and the intake manifold is usually smaller than in real idle conditions, and it is reversed: the pressure in the intake manifold is raised, not lowered. These different conditions are sometimes why a smoke test fails to detect a tricky defect.

The smoke test helps to identify places where additional air is sucked in, but if it finds no damage, that does not guarantee that no additional air is sucked in. If the engine has the symptoms described above, check the CCVV very carefully. If it does not close, replace the unit.

If a smoke test is hard to perform yourself, go to the nearest workshop, which should have the equipment.

The image below shows the typical condition of a CCVV membrane after 5+ years of use.

Examples

Example No. 1

The MSD has not confirmed the offset adaptations (they are 0.00), and the mixture is very lean: even with +30% additional fuel, the MSD measured Lambda 1.32. The fault was observed when starting the engine, including a warm one.

Example No. 2

The MSD has confirmed the offset adaptations, but on the warm engine they grow and reach +1.5 .. 2.0 mg/stk. On the cold engine, however, the fault (air suction) does not manifest, and a rich mixture error is recorded.

Example No. 3

The additional air suction does not manifest directly after starting the engine, but, as in example No. 2, the engine had previously built up very positive offset-type adaptations. When the fault is not present, the mixture becomes very rich, with Lambda dropping significantly below 1.00 on both banks.

BEYOND THE FAULT CODE

A clearer picture.
A better next step.

Get in touch
Enlarged technical image
Open original image