Skip to content
Shop
Explore

INDEPENDENT BMW ENGINE EXPERTISE

Management of Lambda probes

N43 / N53 · NOXEM · DIAGNOSTICS

The oxygen sensors (Lambda probes) monitor the exhaust gases and are among the most important components of a petrol engine. Modern BMW engines use two, three or four of them. This article covers how they are arranged, what each type does, and how I check them when fuel mixture control is in doubt.

Sensor layout

4-cylinder engines use one of these arrangements:

  • A wide-band sensor for each bank (for example N43 in the 1, 3 and 5 series and X20i), or one wide-band probe for all cylinders (for example N43 in the 1 and 3 series, 116i, X18i).
  • A control (narrow-band) probe shared by both banks (for example all N43 series engines), or two control probes (N46 series engine).

6-cylinder engines have a wide-band sensor and a control (narrow-band) sensor on each bank.

Wide-band probes sit at the start of the exhaust system, before the catalytic converters. Control probes sit after the catalytic converters.

The exhaust system of the N43 series engine (one wide-band probe, one catalytic converter and one control probe):

[[[KEEP1]]](https://bimmerprofs.com/wp-content/uploads/2017/07/BMW_N43_exhaust_system.png)

The exhaust system of the N43 series engine (two wide-band probes, two catalytic converters and one control probe):

[[[KEEP2]]](https://bimmerprofs.com/wp-content/uploads/2017/07/N43_exhaust_3.png)

The exhaust system of the N46 series (4-cylinder) engine and the placement of the probes:

[[[KEEP3]]](https://bimmerprofs.com/wp-content/uploads/2017/07/BMW_N46_exhaust_system.png)

The exhaust system of the N52/N53 series (6-cylinder) engine and the placement of the probes:

[[[KEEP4]]](https://bimmerprofs.com/wp-content/uploads/2017/06/BMW_N53_exhaust_manifold.png)

Limited documentation

BMW’s technical documentation says almost nothing about the probe management algorithms. Simpler systems, such as cooling, come with block diagrams and descriptions of some components (for example the electric water pump), but the electronic control systems are not described. BMW evidently assumes the self-diagnostic system can handle probe diagnostics. Practice shows it is not reliable enough: in more than half of the cases where I suspect a problem with exhaust mixture control, there are no probe error messages, or the ones that exist are quite inadequate.

The usual solution is to replace all the probes. Four probes cost more than EUR 300, and in my opinion replacing parts before their performance has been evaluated is not professional.

Because of this, the material below may deviate from the real situation. Some nuances and functional algorithms can also differ between software releases.

Narrow-band probes

In modern petrol engines the narrow-band probe serves two purposes:

  • It determines the reference value of Lambda 1.00 for the wide-band probe (calibration of the wide-band probe).
  • It monitors the performance of the catalytic converters.

In a very simple explanation, the narrow-band probe generates voltage when no oxygen is present in the exhaust gases. If Lambda > 1.00 (lean mixture, oxygen present), the probe generates no voltage. If Lambda < 1.00 (rich mixture, no oxygen), it generates voltage. In a very narrow area the probe also works in a linear mode, where the voltage changes with Lambda.

Narrow-band probe voltage as a function of mixture (Lambda):

[[[KEEP5]]](https://bimmerprofs.com/wp-content/uploads/2017/07/o2sensor-output.png)

Wide-band probes

A wide-band probe measures Lambda (the oxygen/fuel ratio) over a wide range: from 0.65 (very rich mixture) up to the composition of atmospheric air (Lambda 30 .. 100). More information is available online, so here I mention only a few nuances.

The Lambda values reported by wide-band probes cannot be read with a multimeter or oscilloscope. Measuring the voltage between the Pump cell pins shows only whether the mixture is lean or rich, and only if the probe is connected to a correctly working ECU. An oscilloscope also shows dynamic changes (for example an imbalance in cylinder fuel mixture), but absolute values can be seen only with a special technical solution. That solution stabilizes the voltage of the Nernst cell at around 450 mV by changing the current through the Pump cell. Converted with the manufacturer’s data tables, this current gives the value we are interested in: Lambda.

Unfortunately, Lambda cannot be measured by several devices at once. You cannot connect in parallel to a working MSV/MSD unit and see what Lambda the wide-band probe actually measures (before the additional reference value corrections described below).

Catalytic converter monitoring

When the engine runs correctly in homogeneous mode (Lambda 1.00), the narrow-band probes are not held at the midpoint of the curve (0.45 .. 0.50 V), as one might expect. They are held at a generated voltage of around 0.7 .. 0.9 V. BMW gives no official explanation, but knowing how Lambda probes behave, I assume the following:

  • In the middle of the linear range the probe is very sensitive (voltage versus Lambda change) to both rich and lean mixture. Even with a perfect catalytic converter and well-balanced cylinders, the probe voltage would look like a rectangular wave.
  • Moving the working range to the top end of the curve makes the probe insensitive to enrichment, but its sensitivity to unburned oxygen is at maximum.
  • A slightly richer mixture (Lambda = 0.995) means that if a little unburned oxygen enters the exhaust system, the catalytic converter should burn it, because there is fuel to burn with it. If the converter does not function and does not burn the oxygen, a short-term oxygen excess appears after it. This creates a voltage drop in the signal sent to the DME.

The middle point of the narrow-band probe’s active range (corresponding to 450 .. 500 mV):

[[[KEEP6]]](https://bimmerprofs.com/wp-content/uploads/2017/07/o2sensor_output_midpoint.png)

The active range used to monitor the catalytic converter (corresponding to 0.7 .. 0.9 V) is marked:

[[[KEEP7]]](https://bimmerprofs.com/wp-content/uploads/2017/07/o2sensor_output_afetr_CO_converter.png)

With a good catalytic converter, the average DC value of the control probe signal is around 0.7 .. 0.9 V and the AC component is insignificant. With poor converter performance:

  • The amplitude of the AC component increases significantly and short voltage drops appear.
  • The control probe voltage drops quickly after an enriched mixture (the high-voltage “delay” disappears).

Note: MSV/MSD engine control units cannot distinguish quick Lambda pulsation caused by a poorly performing catalytic converter from, for example, pulsation caused by uneven cylinder performance. If the individual cylinder fuel adaptations are incorrect, or some cylinders misfire, MSV/MSD can register false error messages about insufficient catalytic converter performance.

Calibration of wide-band probes

Besides monitoring the catalytic converter, narrow-band probes perform another very important task: they calibrate the reference values of the wide-band probes, that is, they calibrate them at Lambda = 1.00.

The calibration is needed because of technical inaccuracies of the probes themselves, deviations in the electronics (the analog part of the DME), and so on. During calibration the DME adjusts the fuel mixture until the narrow-band probe reports 0.75 V. At that moment the reference voltage of the wide-band probe is modified so that the probe reports a Lambda close to 1.00.

Correct operation of an N-series engine in homogeneous mode (Lambda = 1.0). Both wide-band probes are calibrated and the narrow-band probes generate around 0.75 V:

[[[KEEP8]]](https://bimmerprofs.com/wp-content/uploads/2017/07/wide-band_trim_completed.png)

Diagnostic procedure

If incorrect fuel mixture control is suspected (misfire, trim errors, error messages about catalytic converters or probes), I perform the following checks.

1. Check the dynamic resistance of the Nernst cells. INPA shows it as analog values in the OBD 6th menu (../F5/F2/F6).

[[KEEP9]](https://bimmerprofs.com/wp-content/uploads/2018/02/INPA_diagnostics_Nernst.png)

The dynamic resistance of a Nernst cell reflects the electrochemical activity of the probe: the higher the activity, the lower the resistance. The resistance is also inversely proportional to temperature (in the newest engine generations it is used to stabilize the probe temperature). A correct dynamic resistance is several hundred Ohm. Unfortunately INPA displays values in steps of 256 Ohm, which is very rough and hides the true value. Correct values reported by INPA are 0, 256 and 512 Ohm (768 and 1024 Ohm are acceptable).

An additional check: on a cold (unheated) engine, the resistance between the Nernst and Pump cell connections, measured with a multimeter, is several MOhm. INPA should then display its maximum value: 65 KOhm for a narrow-band probe and 8 KOhm for wide-band probes. Within several dozen seconds of the start of probe heating, the resistance should drop to the 0 .. 512 Ohm mentioned above.

Possible problems:

  • Several dozen seconds after heating starts, the dynamic resistance is above 512 Ohm. The probe has aged and its electrochemical processes are inefficient, so it has to be replaced.
  • On a cold engine (just started, idling), a probe shows a dynamic resistance lower than the maximum values above. This indicates a current leak, or a short circuit if the resistance is several hundred Ohm. Check the electrical circuit first. If it is in order, measure the resistance of the probe’s sensor with a multimeter. If it is lowered, replace the probe.

2. Measure Lambda in the exhaust gases with an additional tool. For N43/N53 engines, the NOx sensor values at ../F5/Shift+F2/F1 can be used for this, although with lower reliability. Make sure the sensor is online, because a switched-off sensor shows a default value of 1.00. If the reference tool shows a Lambda different from 1.00 while all probes seem to show correct values, assume that some of the probes (or the only one) give incorrect mixture data and the wide-band probe is calibrated incorrectly. Of course, check that the exhaust system is airtight, but I believe any specialist can do that without difficulty.

3. Run the probe test. In INPA loader 2.023 it is in the diagnostics session (../F9/F3). During the test the injector opening time is changed by ±10% for several seconds. Observe:

  • The minimum and maximum voltages of the narrow-band probes. The minimum should be below 0.1 V and the maximum above 0.85 V. If either value is off (typically a maximum below 0.8 V), the probe is damaged and has to be replaced.
  • The minimum and maximum Lambda values of the wide-band probes. If the minimum differs significantly from 0.90 and/or the maximum differs significantly from 1.10, the probe is damaged and has to be replaced.

A sample of a successful Lambda probe test:

[[[KEEP10]]](https://bimmerprofs.com/wp-content/uploads/2017/07/BMW_INPA_Lambda_selftest_menu_2.png)

4. Test on the road if the test above is not possible. To check that a narrow-band probe can produce the necessary voltage, do a driving session. During aggressive acceleration (kick-down) the mixture should become rich and the narrow-band probe should generate more than 0.85 V.

5. Compare the fuel trims between banks (if the engine has two banks). Evaluate the multiplicative and offset STFT + LTFT. Large differences in these parameters (typically at idle, in the offset LTFT and integrators) give reasonable grounds to suspect incorrect functioning of one bank’s Lambda probes. Typical causes are leaking injectors and, as a result, incorrect individual cylinder adaptations and fuel mixture.

6. Check the engine temperature. If it is lower than required (+95 .. 115 °C), a richer mixture enters the exhaust. On N42/N53 engines, also check the exhaust gas temperature values and the wide-band probe setpoint at ../F9/Shift+F3/F1 (make sure it is 1.00).

7. Force homogeneous mode on N43/N53 engines. In the diagnostics menu (../F9/F1), choose homogeneous mode with Lambda 1.00 by pressing F4. Attention: it is very important not to forget to switch this mode off after the test. To do it safely, I recommend switching on and checking stratified charge: press F3, confirm that it turns on, then press F5 to exit the menu.

Limitations of the diagnostic tools

To make the situation even more complicated:

1. MSV70, MSV80 and MSD80 with loader 1.001 do not show correct offset and multiplicative STFT + LTFT (no sum is displayed, values are not equal or correct, and so on). 2. MSV70, MSV80 and MSD80 with loaders 1.001 and 2.023 do not show the wide-band probe setpoints when they differ from 1.00. For example, a not fully warmed engine can run an enriched mixture (say 0.98), but nothing shows that such a mixture is required. 3. MSD80 with loader 2.023 shows the multiplicative STFT + LTFT correctly, but does not show the impact of additional adaptations, which degrades the accuracy of these measurements. Wide-band probe setpoints are displayed, but there is no single menu showing the Lambda values of the wide-band, narrow-band and NOx sensors, which complicates viewing the data. 4. MSD80 with loader 2.023 does not show the offset STFT + LTFT (sum), and no menu shows the additional adjustments of each injector. This makes it impossible to compare the fuel mixture proportions of both banks at idle. 5. No engine shows the differences of the wide-band probe setpoints from their default values. This would help identify an incorrectly working narrow-band probe. 6. No single menu shows the Lambda values of the wide-band probes and the NOx sensors for N43/N53, and no differences between them are shown. That would ease the job. 7. No menu shows the status of the narrow-band probes: whether they work in the active range or regularly “fall out” of it, which would lead to incorrect operation of the whole system. 8. No engine shows status bits indicating whether the wide-band probes have been calibrated successfully, and there is no literature on the circumstances or modes in which these procedures run. For example, N53 starts the calibration of the wide-band probes at idle and keeps the narrow-band probes in the necessary mode only when the engine has reached nominal temperature and there are signs that the car has started moving (increased RPM). 9. No engine has control points where one could connect to the probes and check their live data.

This list could be continued for a long time.

Wide-band probe off 1.00 with no errors stored

A more interesting situation is one with no error messages in the ECU memory, where a wide-band probe shows a value different from 1.00 but the control probe or probes show correct values (0.7 .. 0.9 V). In this case, let the engine idle. If the mismatch does not disappear within a longer time (some 5 minutes), the engine is not performing the wide-band probe calibration for some reason.

For N46/N52 and other “ordinary” engines, delete the long-term trims and let the engine create new ones. If the problem persists (and the Lambda test shows seemingly correct results), swap the wide-band probes between positions. It is possible that the correction range of the reference point of the probe showing the incorrect value has reached its maximum deviation.

On N43/N53 engines, problems with creating long-term trims occur in several “interesting” cases:

Related entries

Wide-band Lambda probes

STFT and LTFT

INPA upgrade

BEYOND THE FAULT CODE

A clearer picture.
A better next step.

Get in touch
Enlarged technical image
Open original image