* § * § * § *

FIELD OF THE INVENTION

The present invention concerns an electronic control system for the preparation, dosing and ignition of the mixture of air and fuel in internal combustion engines. In particular, it concerns such a system for carburettor-supplied internal combustion engines .

BACKGROUND ART

As known, at the time being the majority of small internal combustion and controller-ignition engines (typically motor bicycles, lawn-mowers, snowploughs, engine-driven cultivators, ... ) , with both a 2-stage and 4-stroke cycle, employs the tradi- tional carburettor supply and ignition systems with a fixed value of the spark advance. These means for controlling engine operation, however, are unsuitable for an effective improvement of fuel consumption, of performances, as well as of drivability in difficult climate conditions.

In order to improve the control of the carburettors employed in these types of engines, also with the purpose of reducing polluting emissions, control systems have already been suggested employing programmable electronic central units (ECU) .

These units, by reading and analysing a certain number of operation parameters, are designed to adjust, in an optimised manner, the control of a series of actuation devices which determine the dosing of the fuel mixture by means of adjustable carburettors (for example with bypass valves) and combustion conditions (acting also on the adjustment of the spark advance) .

Adjustable carburettors for such applications comprise an auxiliary air conduit (bypass air) the flow of which is controllable in various ways by means of electric device, such as solenoid valves .

A solution which up until today has proved highly effective is the one disclosed in EP 1.835.154 in the name of the same Ap ^¬ plicant. In this system, due to an effective control strategy, advantageous effects are obtained in terms of consumptions and reduction of polluting emissions.

In the past, on similar systems, it has already been suggested to use also an oxygen sensor (the so-called "lambda" probe) , integrated in the exhaust system. The probe analyses ex- haust gases and supplies to the electronic control unit a series of detected values, which allow the same unit to assess the operative conditions of the engine and to act in a corrective manner on the carburettor. Examples of such systems are disclosed in US 3.759.232 and US 5.575.268.

However, such a solution is very costly, because it necessarily resorts to complex probes, provided with a plurality of control cables suitable for transferring a number of electric signals to and from the probe.

SUMMARY OF · HE INVENTION

The object of the present invention is hence that of providing a control system of the combustion of controller-ignition engines, which overcomes the mentioned drawbacks and which is even more effective than known systems, being able to employ also a feed-back signal relative to the composition of exhaust gases.

Such objects are achieved by means of the system and corre ^¬ sponding control method which are disclosed in their essential features, in the enclosed claims.

In particular, according to a first aspect of the inven- tion, it is provided an electronic control system of the opera ^¬ tion of carburettor-supplied internal combustion engines, of the type comprising a control unit suited to drive at least one electrovalve device for the adjustment of the emulsion air of the carburettor which determines the air/fuel ratio (AFR) of the fuel mixture and further comprising an 0 ₂ sensor probe arranged for reading the composition of exhaust gases and for sending a signal of the amount of 0 ₂ towards said control unit, wherein said 0 ₂ sensor operates switching between two states of 0 ₂ (Lean02Threshold, Rich02Threshold) and the control of said elec- trovalve device for the carburettor adjustment is a closed loop control based on said two 0 ₂ states.

According to a further aspect, said closed loop control (CLC) is built as the sum of a step term, which determines a control jump, with a plurality of increment terms, which deter ^¬ mine a control ramp, where the step term is greater than the increment term.

Preferably, different step terms (StepL, StepR) are provided on the lean side and on the rich side of the control and also different increment terms (IncrL, IncrR) are provided on the lean side and rich side of the control.

According to an additional aspect of the invention, the control of said electrovalve device for carburettor adjustment is performed with PWM technique and the drive PWM signal issued by the control unit (ECU) consists of

PWMoutput=PWMbase+CLC

where PWMbase is the signal generated in conditions in which the signal coming from the 0 ₂ sensor is disregarded.

Preferably, the CLC signal is used to trim the AFR offset compared to the stoichiometric conditions of a preset amount.

A further aspect of the invention is that said 0 ₂ sensor probe is of the single-wire, non-preheated type, mounted close to the exhaust port of the engine.

According to a different aspect of the invention a control method of the operation of internal combustion, carburettor- supplied engines, is provided, of the type comprising a device for the adjustment of the emulsion air of the carburettor which determines the air/ fuel ratio (AFR) of the fuel mixture as well as an 0 ₂ sensor probe arranged to issue a signal (02State) of the amount of 0 ₂ in the exhaust gases, said 0 ₂ sensor switching between two 0 ₂ states (Lean02Threshold, Rich02Threshold) , wherein

the control factor (CLC) is calculated in a closed loop based on said two 0 ₂ states,

and is built as the sum of a step term, which determines a control jump, with a plurality of increment terms, which deter ^¬ mine a control ramp, where the step term is greater than the in- crement term, and wherein

If the 02State changes from Rich to Lean

Then upon the subsequent cycle CLC = CLC - StepL If the 02State is Lean

Then upon the subsequent cycle CLC = CLC - IncrL,

If the 02State changes from Lean to Rich

Then upon the subsequent cycle CLC = CLC + StepR

If the 02State is Rich

Then upon the subsequent cycle CLC = CLC + IncrR

BRIEF DESCRIPTION OF THE DRAWINGS

Further features and advantages of the invention will be in any case more evident from the following detailed description of some preferred embodiments, given by way of a non-limiting example and shown in the accompanying drawings, wherein:

fig. 1 is a diagram of interactions in the closed loop system according to the invention,- and

figs. 2 and 3 are plots of 0 ₂ probe voltage signal showing two different operating conditions, whose comparison gives an idea of good and bad drivability;

fig. 4 is a plot of some monitored values according to the prior art; and

fig. 5 is a similar plot as in fig. 4 but using the system of the invention.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

As shown in fig. 1, a control system of a carburettor vehicle substantially comprises a central electronic control unit (ECU) - provided with a logical unit wherein the control strate- gies are implemented (through suitable software and/or firmware) - a series of detection sensors of the fundamental operation parameters and a series of actuators which are diagrammatically clearly identified in the drawing.

An electronically controllable carburettor (known per se) typically has one or more solenoid actuators for controlling the A/F (air/fuel) ration of the mixture, acting on the supplemental (emulsion) air circuit in addition to the main (emulsion) air adjusted through a traditional throttle.

Moreover, the system is preferably provided with capacitive discharge ignition (CDI) with variable and controllable spark advance, although it is possible to employ also simpler inductive discharge ignitions . In the central electronic unit there is implemented the control strategy, which must take into account the specific operation conditions, i.e. the low operation frequency of the electrovalve on the carburettor, the control mode of the elec- trovalve in PWM and the need to prepare an A/F mixture suited to a carburettor.

According to the invention, the control strategy of the electrovalve on the carburettor, implemented in the electronic unit, takes into account the availability of a closed loop sig- nal, i.e. a signal coming from an oxygen probe arranged in the exhaust pipe of the engine. In particular, for the sake of economy, simplicity and efficiency, the probe which one wants to employ must be suited to operate simply and to quickly achieve full functionality, to avoid complicating the system to take into account transient and complex operation conditions.

For such purpose, the oxygen probe (so-called lambda probe) is of the single-wire type, hence not electrically heated. It has a reduced size and is mounted directly at engine exhaust, i.e. downstream of the exhaust valve. This mounting, made possi- ble by the presence of a single wire (the ground pole is obtained by mounting locally the probe in contact with the engine and/or exhaust metal parts) and by the small size of the probe, ensures that the probe warms up quickly (even without having conductors which provide to the preheating thereof) , can quickly achieve full functionality and can operate also without the need to install a battery connected to the system.

Conversely, with a single wire the transmittable signal to the control unit is simplified, with respect to what exists in the prior art. In particular, the probe is capable of operating between two extreme switching conditions, i.e. between a condition in which a signal "rich mixture" is issued and a condition in which a signal "lean mixture" is issued, based on the exceeding of thresholds of detected 0 ₂ .

Consequently, the closed loop control provided for the sys- tern of the invention is based on a switching control between rich and lean conditions, detected by an unheated switching 0 ₂ sensor. The closed loop control is enabled only when the engine is warm, the 0 ₂ sensor is ready and the driving style is not aggressive .

In the following the various aspects peculiar of the system of the invention will be analysed in detail.

Q ₂ feedback Control

As known, the advantage of using an 0 ₂ probe is to be able to obtain an optimal adjustment of the fuel mixture (in the controlled carburettor) , as well as to be able to employ catalytic converters which cut down polluting emissions.

Another significant advantage inherent in the use of a feedback signal, simplified as it may be, is that of also being able to compensate production engine variability, engine ageing, carburettor ageing, 0 ₂ sensor ageing, altitude conditions, air- box filter clogging, fuel quality, etc.

The feedback signal coming from the 0 ₂ sensor, according to the invention is hence entered in the ECU to change the basic control strategy of the carburettor, in particular of the elec- trovalve of the air by-pass duct (and hence adjusting as desired the air/fuel ratio) .

In order to optimise the conversion efficiency of a three- way catalytic converter, the AFR (air/fuel ratio) is controlled in a narrow window around the stoichiometric value. With an oxygen sensor ' feedback control it is possible to keep the AFR close to the stoichiometric value, provided that the electrovalve AFR correction range covers the stoichiometric value.

0 ₂ lambda sensor

The unheated switching oxygen sensor voltage output signal is used for AFR compensation in closed loop control. An oxygen sensor in the exhaust gas is used to detect the actual rich or lean condition of the AFR. Depending on whether a rich or lean condition of the exhaust gases is detected, the sensor switches between two different voltage values (Clc_Lean02 and Clc_Rich02) which are supplied to the control unit (ECU) . Then, said 0 ₂ sensor feedback is used by the logic implemented in the ECU to trim the AFR back to stoichiometric conditions or outside the stoichiometric condition of a preset value, depending on the situation. When the sensor output is high, compared to a preset threshold, it means the fuel mixture is rich and vice versa.

In order to improve the response time and the warming-up of feedback control, the 0 ₂ sensor is arranged close to the exhaust port of the engine. In any case, the 0 ₂ sensor probe must not affect the flow capacity of the exhaust system, i.e. limiting the cross section of the exhaust pipe. Furthermore, the exhaust pipe shall be preferably free from air leaks.

0 ₂ input in the unit

The 0 ₂ sensor is properly grounded via the exhaust pipe, then only one wire and one ECU input pin is used. The input stage is implemented on a board within the ECU and provides, among other things, also to A/D convert the signal, through a specific sampling rate, and to filter said signal.

ClosedLoopControl Enabling Condition

The closed loop control (CLC) is enabled and its flag is set to 1 when the following conditions are met:

1. Engine is running (RPM other than 0 , and out of cranking)

2. Temperature oil > preset parameter (CLCEnToil) , and

3. No error is detected by the 0 ₂ sensor, and

4. Engine Revolution number must be higher than a preset pa- rameter (CLCEnRevN) , and

5. 0 ₂ sensor is ready, i.e. 02Ready signal=l, and

6. Variation in time of the throttle position sensor (Δ TPS) <preset parameter CLCDeltaTPS

7. TPS < CLCTPSEnLo

where CLCTPSEnLo is the lower enablement operation theresold (TPS enable low) .

The CLC mode is preferably disabled and its flag is set to 0 when one of the following conditions is met:

1. One of the conditions above in 1-5 is not true, or

2. Δ TPS > CLCDeltaTPS (keep CLC disabled for a certain time, for example 3 sec, and then check again)

3. TPS > CLCTPSEnHi where CLCTPSEnHi is the higher enablement operation theresold (TPS enable high)

4. TPS = 0 and RPM > of a preset threshold (ThresHoldRPM) , which represents a required cut-off condition

5. RPM > of a preset threshold (RPMthCLCOFF)

6. if the lambda sensor voltage is between Clc_Lean02 and Clc_Rich02 for more than 2 seconds (this is to be understood as the sensor is not warm and the CLC contribution must be disabled) .

Preferably, the RPM (revolution per minute) and Temperature threshold values are defined with a certain tolerance margin (hysteresis) , for example 5°C for temperature and 300rpm for RPMs. Parameter Evaluation Timings

The calibration parameters (as described above) to be used in the control strategy of the carburettor electrovalve sole ^¬ noid, based on the values transmitted by the 0 ₂ sensor, are to be assessed also based on the experimental results of the spe- cific system. As a matter of fact the system is affected by a series of typical reaction times, which are typically a) closed loop behaviour of the electrovalve due to its P M period; b) physical delay between electrovalve air metering, mixture preparation, intake port transportation, in-cylinder processing and exhaust port transportation; c) 0 ₂ sensor response time. All these typical times are also linked to the specific system ther ^¬ modynamics ' and can hence be taken into account, acting on a series of calibration parameters determined experimentally in the real application.

In order not to affect in an instable way on the closed loop of the control, the correction introduced on the basis of the value received by the oxygen probe can be applied preferably every two (or more) engine cycles. Engine Running Condition

The CLC strategy is enabled while the engine is running (i.e. key-on and engine speed is detected and is above the cranking threshold) and an assigned time has elapsed. The latter is defined by the calibration constant CLCEnRevN.

TPS Sensor Enabling Condition

The CLCTPSEnLo, CLCTPSEnHi and CLCDeltaTPS are calibration parameters used for the throttle thresholds of the CLC working range. Thereby the CLC is not active in the presence of an aggressive driving style - as detected by the throttle position sensor (TPS) , i.e. at high loads or during sudden acceleration. In order to improve drivability, a hysteresis range is provided, by adopting different High (Hi) and Low (Lo) threshold values.

If the TPS fails - i.e. no coherent signal arrives at the ECU from the throttle position sensor - then TPS is assumed = 100%, so that the CLC is disabled.

0 ₂ Sensor Ready

02ReadyHigh and 02ReadyLow are two fixed calibration parameters (with a defined sensor these two values will be fixed) that enable lambda sensor use.

If the 0 ₂ sensor output is greater than 02ReadyHigh, or lower than 02ReadyLow values, the 0 ₂ sensor is considered ready for working for the closed loop control . Then the 02 sensor is ready and its status flag 02Ready is set to 1. On the other hand, whenever the 0 ₂ sensor voltage lies outside of the above mentioned threshold, the 02 Ready flag is set to 0.

Lean and Rich 0 ₂ Sensor

Combustion is considered rich if the output from the 0 ₂ sensor is above rich voltage (Rich02Threshold) . Combustion is considered lean if the output from the oxygen sensor (lambda voltage) falls below 0 ₂ lean voltage (Lean02Threshold) .

1. 02 output < Lean02Threshold - AFR Lean - 02State=Lean

2. 02 output > Rich02Threshold -> AFR Rich - 02State=Rich Lean02Threshold and Rich02Threshold are calibration parameters.

CLC (Closed Loop Control) factor calculation

According to the invention, it has been detected that supe- rior results can be obtained when the closed loop control factor is calculated with a control technique having a jump and ramp between rich and lean condition. This technique involves determining two operating parameters: a first, relatively high, step term (StepL, StepR) to be used for the jump portion, and a second, lower, increment term (IncrL, IncrR) to be used in the ramp portion. After which, the logic of the closed loop control is as follows. The voltage value of the oxygen probe (02State) is read upon each control cycle CLCStepCycle (which is a calibration constant, equal to or multiple of an engine cycle) and, based on the two switching conditions the simplified probe can take on: If: 02State changes from Rich to Lean

Then upon the subsequent cycle: CLC = CLC - CLCStepL

If then: 02State is Lean

Then upon the subsequent cycle: CLC = CLC - CLCIncrL,

If: 02State changes from Lean to Rich

Then upon the subsequent cycle: CLC - CLC + CLCStepR

If then: 02State is Rich

Then upon the subsequent cycle: CLC = CLC + CLCIncrR

While entering the CLC control, the CLC factor is set equal to 0.

CLCStepR, CLCIncrR, CLCStepL, CLCIncrL are calculated as follows :

CLCStepL = isoStepL-deltaStepL

CLCIncrL = isoIncrL-deltalncrL

CLCStepR = isoStepR+deltaStepR

CLCIncrR = isolncrR+deltalncrR

Being iso-X and delta-X values to be founds on 3-D maps as a function of RPM and TPS. In this way it is possible, working in real time, to set the stoichiometric condition on one iso-X map, and then offset the AFR still using only one real time delta-X map .

The actual step and increment term values to be applied shall be interpolated from the maps .

The values of the step terms can be obtained as follow. A first calibration phase employing a "iso-step" reference map, where stepR=stepL (i.e the same values on the lean or rich side of the control) , so as to center the control on stoichiometric condition. Than a subsequent calibration phase employing a "delta-step" map, where deltaStepR=-deltaStepL, so as to offset the control compared to the above stoichiometric. The same tec- nique can be used for the increment ("iso-incr" and "delta- incr") terms.

CLC Limits

The CLC compensation value must remain within the range of

MaxCLC (+100) and MinCLC (-100). MaxCLC and MinCLC are calibration parameters . When the calculated CLC is above or below the specified range, the CLC factor is assumed to be equal to MaxCLC or MinCLC, respectively.

EOL CO Setting

Since, according to a preferred embodiment of the invention, the system employes a carburettor with pilot circuit setting screw, a specific End Of Line (EOL) setting procedure has to be implemented. Improper setting of this screw may result in lambda control troubles or inefficiency. Due to this issue, the above mentioned setting screw shall have a stroke limiter, like i.e. +/- ½ turn, or a anti-tampering cap. The pilot screw has to be set on the engine and/or on the carburettor production line, while measuring the pilot circuit flow.

Moreover, if the combination of engine, carburettor and air box variability is wide, it should be necessary to introduce via software (acting on the setting parameters of the operating method) a CO setting, that alter the PWM output from the map ta- bles. Such a setting shall be done at the EOL station (TPS adjustment, MAP loading, etc.). At the EOL stage the engine should be fully warmed-up and the operating point for CO correction shall be fast idle (i.e. no load, 1000 RPM above idle). The amount of the PWM EOL correction factor EOLCF is then stored in the unit memory and applied to the PWM values stored in the calibration map. Electrovalve PWM Driving Signal

The electrovalve driving the A/F ratio on the carburettor, acting on the air bypass conduit, is controlled with the PWM technique. The PWM control calculated in a conventional way (PWM base), for example as shown in EP1835154, is corrected by the CLC factor and, if used, AL (Auto Learning) factor, as follows:

PWMoutput=PWMbase+CLC

Advantageously, due to the feedback correction given by the lambda probe, the PWM factor can be managed also in the non- linear ranges of 0-15% and 85-100%, which range in the conventional control (i.e. without closed loop control) had instead to be cut .

If, due to the CLC contribution, the calculated PWM becomes negative or exceeds 100%, the actual driving PWM sent to the so- lenoid valve of the carburettor must be kept at 0 or 100%, respectively.

Since the control electrovalve is placed on the pilot circuit, its AFR (air/fuel ratio) variation potential could be limited in some working conditions. In these cases, the electro- valve is driven by the correction factor to remain fully closed whenever the AFR is lean, or completely open whenever the AFR is rich, even if the stoichiometric condition cannot be approached.

Additionally, according to a preferred embodiment of the control method of the invention, the calibration parameters of the closed loop control are set in order to bias the AFR value on the lean or rich side of the stoichiometric condition. In other words, the calibration parameters are set so that the operation conditions are offset compared to the stoichiometric conditions . This has the advantage that better results in terms of maximizing catalytic reduction of NOx or CO e HC cab be obtained.

The closed loop control may be further enhanced by using a self-teaching strategy.

With the control methodology offered here it has been de- tected that it is possible to obtain excellent results in terms of consumptions and pollution, which are maintained also over time, despite using an overall simple and economic system. The system of the invention, using a non-preheated probe and a very simple operation logic, can operate also without batteries, which makes it suitable to the specific application disclosed in the preamble, i.e. small motorcycles and gardening equipment, for example.

Fig. 5, compared with fig. 4, shows clearly that the A/F value can be kept more regular and constant with the system according to the invention.

However, it is understood that the invention is not limited to the particular configurations illustrated above, which represent only non-limiting examples of the scope of the invention, but that a number of variants are possible, all within the reach of a person skilled in the field, without departing from the scope of the invention.