Wideband O2 Sensor & Tuning Strategy

The two methods for tuning an engine are with wideband o2 sensors and with narrowband o2 sensors.     Each method have benefits and limitations, let's review the technology behind each and how to apply this to our tuning of the Bosch ECM.

When using wideband sensors, direct Air Fuel Ratio (AFR) measurements are performed by the sensors and integrated (via PCVX) with the other ECM engine data to provide a complete picture of the current state of tune for the engine.  This is the preferred method to tune your engine since all fuel cells that the ECM uses to control the fuel delivery can be safely adjusted.  The wideband o2 sensors typically are accurate from 10 AFR to 19 AFR, which is more than enough for us to tune our engine.

AFR is a ratio, so for technical correctness the number for AFR should be written as 10:1 to 19:1 AFR in the above paragraph.  This is 10 parts air to 1 part fuel to 19 parts air to 1 part air.  For simplicity, I will write it at a whole number with the :1 understood.

Here is a typical transfer function for a wideband sensor.  Stoic is 2.5V in this example.

(Click on image to expand)

A word about AFR measurements.  What the sensor actually measures is the presence or absence of oxygen molecules.  It has no knowledge of fuel ratios, it is properly called a Lambda sensor since it measures the oxygen molecules present in the gas it is analyzing.

Let's discuss how we make adjustments using wideband sensors.

The first step is to set our Target AFR to open-loop numbers and consistent over a larger range.  13.2 to 13.8 are good numbers to start with, with Hi-load & RPM regions richer at 12.7 to 13.0.  Here is the stock AFR table modified for tuning with widebands.  

(Click on image to expand)

Since you are monitoring the actual AFR from each cylinder, be sure to maintain safe operating parameters at all times.  If you see certain areas going lean, stop, make VE corrections and run again.

Note: The starting AFR table is very rich.   I prefer more reasonable numbers from 13 to 13.8 AFR, but it is a safe starting point.  Here is the table I like:

(Click on image to expand)

We have one VE table and two IPW Comp tables (Front and Rear), I take the average of the Front & Rear AFR and find the error to apply to correct the VE table.  What is the Volumetric Efficiency (VE) table?

In it's simplest form, it is the measure of volume of air in the cylinders at any given engine speed versus the physical volume of the engine (piston displacement).

Notice I stated cylinders.  This is why the average method I describe below makes since for correcting the VE table.

This looks like: Average AFR = (AFR1 + AFR2)/2.  The AFR error is calculated by Average AFR/Requested AFR.  The resulting error is then multiplied to the VE cell that the ECM used to calculate the requested AFR at that load/rpm operating condition. 

Example: Average AFR = 13.8 with a target AFR of 13.5.  The error is 1.0222, which is then multiplied to the VE cell that called for the target AFR of 13.5, say it is 50.5 VE.  The new average VE will be 51.62 for that cell.  This makes since because we are leaner than what is called for so we need to increase the VE cell to represent more air flow through the engine at that point.  This will cause additional fuel to be added to the mixture when the engine hits that VE cell.

Once we get our average AFR consistent throughout the load and RPM range that we are tuning, we are now ready to move on to tuning the individual cylinders to achieve accurate AFR readings throughout the same load and RPM range.

Note: Since the VE is an average you must look at both cylinder actuals with that in mine.  If your target AFR is 13.5 as our example, you can have one cylinder at 13.2 AFR and the other cylinder at 13.8 AFR.  This is acceptable and expected (13.2 AFR1+13.8 AFR2)/2 = 13.5 AFR average.

One thing to keep in mind is the VE values change smoothly throughout the VE table.  You will not see large swings causing peaks and valleys.

(Click on image to expand)

Moving on to the individual cylinders, which final fueling injector pulse width is corrected by the IPW Comp Front & Rear tables so we will use these tables to get accurate AFR measurements from each exhaust.

So to find the correction to be applied to the front cylinder,  take the actual AFR1/requested AFR.  This is the error to be applied to the IPW Comp Front table cell that this corresponds to.

Example:  AFR1 = 13.8 with a target AFR of 13.5.  Error is 1.022, which will be multiplied to the IPW Comp Front cell that corresponds to the load and RPM at that operating condition.  So if the IPW Comp. Front cell is 1.056ms, the new value will be 1.079ms.  This makes sense because we need to richen the mixture by that percent, so we are increasing the front injector time to increase the fuel mixture at that operating condition.

The same calculation is done for the rear cylinder.

The IPW Comp tables are not smooth, since they are affected by injector latency, air flow imbalance between cylinders, cams, etc.  We will look more fully into these tables as we tune with narrowbands.

Note: If I start seeing extremes in my IPW Comp Front or Rear table, I go back and correct the VE tables to shift the average up or down toward the problem table.  Then go back and correct the IPW Comp tables.

To properly tune your engine requires time.  It is an iterative process of recording logs, make corrections to tune, flash tune, record new logs, make corrections to tune, etc.  One pull on the dyno or one run around the block is not sufficient to develop a good tune.