This invention relates to a method of measuring the air fuel ratio in spark ignition engines.
This method can be applied to any spark ignition engine which may have the following physical features but is not limited to them:
1 .Two stroke, four stroke, rotary (Ankle)
2. Single or multi-cylinder
3. Inline, V-type, Opposing or Radial
4. Air cooled, or liquid cooled
5. Naturally-aspirated, Super charged or turbo charged
6. Engine speed ranging from 500 rpm to 20000 rpm
The proposed method can be applied to any spark ignition engine in which the fuel metering

is done by the following types but is not limited to them:
1. Fuel Injection of type:
a. Port Fuel Injection (PFI)
b.Throttle Body Injection (TBI)
c. Multi-point Fuel Injection (MPFI)
d. Direct fuel Injection (Dl)
2. Injector of type
a.Single hole
b. Multi-hole
c. Pintle type
3.Injector located in the manifold or throttle
body or cylinder head at any orientation 4.Carburetor controlled (Elebtronically or
Mechanically) 5.Any other fuel metering system by which
fuel supplied to the engine can be
changed

.This method can be applied to Engines running on the following type of Fuels but is not limited to them:
1. Liquid fuels: a. Gasoline b.Ethanol C. Methanol d. Mixture above fuels of any ratio (by
volume or mass) e.Any other Hydro Carbon fuels in the
liquid form f. Hydrogen in liquid form
2.Gaseous fuels:
a. Liquefied Petroleum Gas (LPG)
b. Hydrogen
c.CNG

d.Biogas
e. Producer gas
This method can be applied for the Exhaust Gas Oxygen (EGO) sensor of type given below but is not limited to them:
Binary or switching type EGO sensor which can detect rich and lean air-fuel ratios
The proposed method of air fuel ratio estimation using a 'EGO sensor uses the variations in the • sensor switching response time when the air fuel ratio is changed from slightly richer than stoichiometric values to various lean conditions suddenly. The response time at a given operating speed and throttle position is shown to be related to the final air fuel ratio (lean in this case).
Due to ever-increasing fuel costs; there is a demand for fuel-efficient automobiles. The concern to reduce emissions of CO, HO and NOx, on the other hand is driving research

owards the realization of cleaner engines. In Asian countries two-wheelers of small capacity SI Engines) of about 100 - 150cc have become ;he most popular modes of transportation in urban areas. These are mainly used by the cost consious middle income group of the population. Thus it becomes important to find low cost solutions to achieve reduction in emissions and fuel consumption. Reduction of Green house gas (CO2) emissions is related to improvements in fuel efficiency which can be attained by operating the engine at lean mixture conditions while using a veduction , cataJyst to take care of NO Proper control of the air fuel ratio is needed with lean mixtures as irregularities can lead to cyclic variations in combustion and even misfire. Thus in the case of small two wheeler engines a mapped air fuel ratio has to be followed closely with low cost systems. The air fuel ratio has to be maintained close to a predetermined value. This may require atleast a low cost oxygen sensor even for this cost sensitive two-wheeler application. Proper air-fuel ratio control will also lead to good drivability.
Presently, there are two types of oxygen sensors used for measuring the exhaust gas oxygen level namely: Binary type Exhaust Gas

I Oxygen ( EGO) sensor and Universal Exhaust • ^ Gas Oxygen (UEGO) sensor. The first one is also known as switching type EGO, as its output . signal switches between a low (e.g 0.1)when the mixture is lean and a high level (e.g 0.7)for a rich mixture. The switching takes place even with a small amount of change in the air-fuel ratio around the stoichiometric value of 14.7:1 in the case of gasoline fuel. Hence, the EGO sensor is normally used only for detecting the lean or rich air-fuel ratio condition and for the closed loop control of stoichiometric air-fuel ratio. The second type of oxygen sensor known, as the amperometric type oxygen sensor is capable of detecting air-fuel ratio over a wide range (rich to lean condition). These sensors -need additional circuits for pumping / measuring the current into sensor which increases their cost. Due to cost considerations these lean air-fuel ratio measuring sensors are not suitable for cost effective two wheeler engine controllers which are generally built around an 8-bit embedded controller. Thus there is a need for new air-fuel ratio estimation preferably using the existing low cost. EGO sensor.
The proposed method of air fuel ratio estimation with a EGO sensor uses the variations in the

sensor switching response time when the air fuel ratio is changed from slightly richer than stoichiometric values to various lean conditions suddenly. The response time at a given operating speed and throttle position is shown to be related to the final air fuel ratio (lean in this case). In order to implement this method for example two wheeler engine, first at a given throttle position and engine speed, when the engine is operating at steady state conditions, the fuel injection pulse width is adjusted such that the output signal of the IEGO sensor oscillates ' ' between high level (e.g0.7.) and low level (e.g o.1) volts). This oscillatory response indicates thaf the air fuel ratio is at stoichiometric value. Then the fuel injection pulse width is increased slightly so that, the air-fuel mixture becomes rich and the sensor output signal reaches the high level (e.g 0.7)' This new fuel injection pulse width becomes the' reference pulse width for this speed and throttle position.
From this reference value, the, injection pulse width is reduced to a predetermined lower value (below the stoichiometric injection pulse width) so that the air-fuel ratio becomes leaner suddenly. As the air-fuel ratio is changed from the rich to lean condition, the EGO output '

signal switches between high (e.g.0.7V) to low (e.g.O.IV) level. The sensor output signal is recorded and the switching time (transition time) taken by the EGO sensor to detect the change in the air-fuel is measured and recorded. To compute the actual air-fuel ratio of the mixture at this operating condition, the fuel and air mass flow rates are also recorded.
This procedure is repeated for different lean air-fuel ratio conditions by switching the air-fuel ratio value from the reference (slightly rich) condition to a new lean condition: by reducing, the injection pulse width progressively in small steps (e.g.SOps). The EGO sensor switching time for • each experiment is measured and recorded. The ^whole set of experiments for the particular throttle and speed condition is repeated many times, and the averaged sensor switching time is stored in a look-up table. With this look-up table, it is possible to estimate any lean air-fuel ratio value for the same throttle and speed condition. The actual implementation using software is shown later in Fig.6 The new lean-air fuel ratio estimating technique using the EGO sensor can be used only when the engine operates at steady state ie. when the throttle position remain

signal switches between high (e.g.0.7V) to low (e.g.O.1v) level. The sensor output signal is recorded and the switching time (transition time) taken by the EGO sensor to detect the change in the air-fuel is measured and recorded. To compute the actual air-fuel ratio of the mixture at this operating condition, the fuel and air mass flow rates are also recorded.
This procedure is repeated for different lean air-fuel ratio conditions by switching the air-fuel ratio value from the reference (slightly rich) condition to a new lean condition: by reducing, the injection pulse width progressively in small steps (e.g.50|js). The EGO sensor switching time for • each experiment is measured and recorded. The *whole set of experiments for the particular throttle and speed condition is repeated many times, and the averaged sensor switching time is stored in a look-up table. With this look-up table, it is possible to estimate any lean air-fuel ratio value for the same throttle and speed condition. The actual implementation using software is shown later in Fig.6 The new lean-air fuel ratio estimating technique using the EGO sensor can be used only when the engine operates at steady state ie. when the throttle position remain

at constant value with small dead-band. Trials were conducted on a 125 cc scooter engine to determine the suitability of the present method for detection of air fuel ratio.
According to this invention, the method of measuring the air-fuel mixture ratio of a spark ignition engine, at a given throttle position and engine speed, when the engine is operating under steady state conditions, comprises the steps of adjusting the fuel injection pulse width such that the output signal of the EGO sensor -oscillates between a predetermined high level value and a predetermined low level value, the oscillatory response indicating that the air fuel ratio is at stoichiometric value; increasing the fuel injection pulse width slightly so that, the air-fuel mixture becomes rich and the sensor output signal reaches the said high level, this new fuel injection pulse width being the reference pulse width for this speed and throttle position; reducing the injection pulse width from this reference value, to a predetermined lower value (below the stoichiometric injection pulse width) so that the air-fuel ratio becomes leaner suddenly; changing the air-fuel ratio from the rich to lean condition, causing the EG0 output

F 1 ; , '■
signal to switch between the said high level value to the said low level value; measuring and recording the sensor output signal and the switching time (transition time) taken by the "EGO sensor to detect the change in the air-fuel; recording the fuel and air mass flow rates to compute the actual air-fuel ratio of the mixture at this operating cqndition; repeating this procedure for different lean air-fuel ratio conditions by switching the air-fuel ratio value from the reference (slightly rich) condition to a new lean condition: by reducing, the injection pulse width progressively in predetermined small steps; measuring and recording the EGO sensor switching time for each such procedure; repeating the procedure for the particular throttle . and speed condition several times and storing the averaged sensor switching time in a look-up table, to enable any lean air-fuel ratio value for the same throttle and speed condition to be estimated.
The method proposed herein using the EGO sensor can be used only when the engine operates at steady state, that is, when the throttle position remains at constant value with

small dead-band. Trials were done on a 125 cc scooter engine to determine the suitability of the present method for detection of air fuel ratio.
The stoichiometric air fuel ratio is detected in the following manner depending upon the type EGO sensor used:
In case of two state, switching type or Binary type EGO sensor the output signal will be switching between low level (depending on the type of sensor) and high level (depending on the type of sensor). The switching time will depend on the materials used in the sensor, its, -construction, other catalysts used and on other factors.
The engine is switched from a current lean air fuel ratio to a condition that is slightly richer than stoichiometric at the said throttle position, speed, and ambient condition by increasing the pulse width of the injector of electrical / mechanical signal sent to the fuel delivery system.
First the quantity is increased such that the stoichiometric air fuel is set and this condition is

detected by identifying oscillations of the sensor output.
The rich condition is obtained by further increasing the injection quantity slightly between 1 and 25% of the current value.
Thereafter the operation is switched to previous lean condition suddenly and the sensor output signal is filtered using a low pass filter with a cutoff frequency ranging from 50 HZ to 500Hz or used as it is and then sampled on the time basis.
The time rate of change of the sensor signal from a high value in the range defined by the sensor that is used to a low value depending on the type of sensor that is used and the sampled signal is recorded.
The data is compared with previously recorded values from several trials as a look up table and the present air fuel ratio is thus determined.
The look up table is created by conducting many such trials between different rich and lean air fuel ratios, different throttles, different speeds, and different ambient conditions, different fuels

and sensors.. The lookup table indicates the air fuel ratio values vs.the time rate of change of the signal.
From the look up table data, the switching time in any particular case would indicate the corresponding air fuel ratio.
This is used to control the desired air fuel ratio. Using this information corrections can be made in the engine management system and rechecked. This corrected air-fuel ratio value could be used for updating the stored lean air-fuel ratio look-up table. Also this corrected value could be used as a new reference air-fuel ratio set point to achieve the adaptive closed loop control of lean air-fuel ratio, and there by achieve fuel economy and reduction of exhaust emissions.
The accompanying drawings illustrate in
Fig.1 EGO sensor response for rich-lean air fuel ratios (300rpm, 25% throttle)
Fig.2 EGO sensor response for rich-lean AF (throttle =25%, initial AFR =13.0)

Fig.3 EGO sensor response for rich-lean AF ratio (throttle =40%, Initial AFR = 13.0)
Fig.4 EGO sensor switching time stored as Look-up table for estimating air-fuel ratio for specific speed and throttle value.
Fig.5 measured vs. EGO predicted air-fuel ratio with 5% error banned.
Fig.6 EGO signal based lean air-fuel estimator implementation.
From Fig.1, clear differences are noticed between the time needed for the sensor to change its state when the air fuel ratio is changed from a given initial slightly rich value to different final lean values (rich to lean). The sensor switching time has been defined in Fig.1. The transition or switching time can thus be used to estimate the final air fuel ratio if the initial air fuel ratio is fixed.
Further, trials were done with the initial mixure being set at a slightly richer than stoichiometric condition. This condition was noted by gradually increasing the pulse width of the injector till the output of the EGO sensor switches to a high

factors. The controller implemented in simulink is also seen in Fig.6.
The estimated air fuel ratio values, obtained by the method proposed, are within 5% of the actual measured air fuel ratio. The new method may be used as a cost effective air fuel ratio estimating technique using a commerically available low cost 'EGO sensor for the cost sensitve two-wheeler application.

We Claim:
1.A method of measuring the air-fuel mixture ratio of a spark ignition engine, at a given throttle position and engine speed, when the engine is operating under steady state conditions, comprising the steps of adjusting the fuel injection pulse width such that the output signal of the EGO sensor oscillate.s between a predetermined high level value and a predetermined low level value, the oscillatory response indicating that the air fuel ratio is at stoichiometric value; increasing the fuel injection pulse width slightly so that, the air-fuel mixture becomes rich and the sensor output signal reaches the said high level, this new fuel injection pulse width being the reference pulse width for this speed and throttle position; reducing the injection pulse width from this reference value, to a predetermined lower value (below the stoichiometric injection pulse width) so that the air-fuel ratio becomes leaner suddenly; changing the air-fuel ratio from the rich to lean condition, causing the EGO output signal to switch between the said hiah level value to the said

low level value; measuring and recording the sensor output signal and the switching time (transition time) taken by the EGO sensor to detect the change in the air-fuel; recording the fuel and air mass flow rates to compute the actual air-fuel ratio of the mixture at this operating condition; repeating this procedure for different lean air-fuel ratio conditions by switching the air-fuel ratio value from the reference (slightly rich) condition to a new lean condition by reducing the injection pulse width progressively in predetermined small steps; measuring and recording the EGO sensor switching time for each such procedure; repeating the procedure for the . particular throttle and speed condition several times and storing the averaged sensor switching time; repeating the said procedure for different engine speeds and throttle positions; storing the EGO sensor switching time values in a lookup table, comparing the measured lean air-fuel ratio with the desired air-fuel ratio and determining correction factor for the injector pulse width and storing; using the corrected air-fuel ratio value to update the stored lean-air fuel ratio look up table; also using this corrected value as a new reference air-fuel

ratio set point to achieve the closed loop control of lean air fuel ratio and thereby achieving fuel economy and reduction of exhaust emissions, to enable any lean air-fuel ratio value for any throttle and speed condition to be estimated using the look up table.
2.A method as 'Claimed in Claim 1 wherein the high level value is exemplified by 0.7 V.
3. A method as claimed in Claim 1 or Claim 2 wherein the low level value is exemplified by 0.1V.
4.A method of measuring the air-fuel , mixture ratio of a spark ignition engine substantially as herein described and illustrated with reference to the accompanying drawings.