WITH TRANSIENT WASTEGATE CONTROL

This invention relates to turbocharger control systems for motor vehicle engines in general and more particularly to control systems for controlling the transient operations of the wastegate valve.

BACKGROUND OF THE INVENTION: Turbocharging of an internal combustion engine is accomplished by having the exhaust gases from the engine drive an exhaust turbine which is mounted coaxially with a compression turbine. The input to the compression turbine is the air intake to the engine and the output of the compression turbine, which is defined as the boost pressure output, is supplied to the intake valves of the engine generally via the throttle body and intake manifold. As the exhaust pressure increases, the exhaust turbine is driven faster and the compression turbine increases the pressure of its boost pressure output.

Without any controls to affect the exhaust turbine, the engine may be damaged due to overboosting the engine by excessive boost pressures. In addition, gasoline engine designers have found it necessary to limit boost pressures for several other reasons. Among these reasons are the likelihood of spark knock in a high BMEP (Brake Mean Effective Pressure) engine, driveability problems which are possible in a runaway boost, durability problems caused by boost application to a cold engine, and emission constraints imposed by the higher blow-by and blow-through (valve overlap) rates of a turbocharged engine.

In order to limit the exhaust pressure and thereby control the exhaust turbine, prior art systems have installed a valve, called a wastegate valve, in the exhaust system of the engine. The function of the wastegate valve is to divert the exhaust gas stream around the exhaust turbine thereby limiting or reducing the power available to drive the exhaust turbine and in doing so, limiting the boost pressure output of the compression turbine. In order to regulate or control the operation of the wastegate valve, an actuator or control valve is used which is basically constructed with a spring and diaphragm. The diaphragm has atmospheric pressure on one side and a control pressure from the intake manifold on the other side. The spring provides a predetermined preload pressure to allow the wastegate valve to be closed for a predetermined range of exhaust pressures. This system is quite successful in avoiding damage to the engine due to excessive boost pressures, but the control of the turbocharging system is not as effective as desired by the engine operator because among other reasons, the system controls about a single pressure setpoint.

A variation of this method of control utilizes a means to modulate the actuator pressure. This means is a microprocessor-controlled valve. A spark knock sensor and a controller cooperate to modulate the boost pressure in an integral fashion to limit detonation.

It is therefore a principle advantage to provide a turbocharging control system which is a closed loop system responding to a cartography of boost setpoints defined by the engine operating mode. It is another advantage to provide a closed loop system wherein the boost setpoints on the cartography are modified by

engine temperatures to avoid the problems of boost being applied to a cold engine. It is yet another advantage to provide a closed loop system additionally including a cartography of boost correction rates according to spark knock levels. It is another advantage to provide a closed loop turbocharge control system combining both cartographies of boost setpoints. Still a further advantage of the invention is to control the closed loop system to provide for enhanced turbocharged engine operation during transient conditions.

SUMMARY OF INVENTION

These and other advantages of the closed loop turbocharger control system are accomplished by a system having a turbocharger operatively connected to an internal combustion engine. The turbocharger has an exhaust turbine and a compression turbine coaxially mounted, the exhaust turbine has its input operatively connected to the exhaust system of .the engine and the compression turbine is connected to the air input system of the engine. A manifold pressure sensor is mounted in the intake manifold and responds to the pressure therein to generate an electrical signal representing such pressure. A wastegate means operates to provide a bypass path for exhaust gases around the exhaust turbine. The wastegate means has an actuator means responsive to variations in control pressure to control the operation of the wastegate valve according to a cartography of boost setpoints in a look-up table located in the memory means of a microprocessor based control unit. The look-up table responds to engine speed and throttle position for the determination of a boost setpoint.

Many other objects and purposes of the invention will be clear from the following description of the drawings taken in conjunction with the detailed description of a preferred embodiment.

DESCRIPTION OF THE DRAWINGS

In the drawings:

FIGURE 1 is schematic diagram of the turbocharger system. FIGURE 2 is a cartography of boost setpoints found in the memory look-up table for the turbocharger system.

FIGURES 3A and 3B are flow charts of the control strategy for wastegate control.

FIGURES 4A, 4B and 4C are flow charts of the control strategy for spark knock control.

DETAILED DESCRIPTION

An SAE Technical Paper 860,487 which is entitled "Closed Loop Turbocharger Control with Transient Wastegate Functions" by the inventors hereof, is included herein by reference. This paper was presented at the International Congress and Exposition in Detroit, Michigan, USA in February 1986.

Referring to FIGURE 1 there is illustrated in schematic form the mechanical hardware of the turbocharger system. As is well known, a turbocharger 10 is a device that has an exhaust turbine 12 mounted coaxially with a compression turbine 14. The input 16 to the exhaust turbine 12 receives the exhaust gases which operate to rotate the turbine 10. The exhaust gases are outputted from the exhaust turbine 10 to the exhaust system of the engine 18. The input 20 to the compression turbine 14 is typically ambient air and the compression turbine compresses the input air and

supplies it to the intake manifold 22. The compression turbine 14 is rotated by the rotation of the exhaust turbine 12.

The engine 18 is illustrated by an engine crankshaft 24 having a connecting rod 26 connecting the piston 28 to the crankshaft. The piston 28 is housed in a cylinder 30 having an intake valve 32 and an exhaust valve 34. The intake valve 32 is schematically positioned at one end of the intake manifold 22 which has at its other end the output of the compression turbine 14. Positioned between the compression turbine 14 and the intake valve 22 is a throttle valve 36 for controlling the flow of air from the compression turbine 14 to the intake valve. The intake manifold 22 has two pressure taps; one positioned upstream 38 and the other positioned downstream 40 respectively from the throttle valve 36.

The exhaust valve 34 of the cylinder 30 is schematically positioned at one end of an exhaust manifold 42 and at the other end of the exhaust manifold is the exhaust turbine 12. Mounted to the exhaust manifold 42 and upstream from the exhaust turbine 12 is a wastegate valve 44 which operates to provide a bypass passageway around the exhaust turbine 12 for the exhaust gases from the several cylinders 30. The wastegate valve 44 is operated by a control rod 46 from a vacuum motor which functions as a wastegate valve actuator 48.

Connected to the intake manifold pressure tap 40 which is positioned downstream from the throttle valve 36, is a manifold pressure sensor 50 for sensing the pressure in the intake manifold 22. This sensor 50 generates an electrical signal, MAP, in response to the magnitude of the pressure found in the intake manifold 22. Connected to the intake manifold pressure tap 38

which is positioned upstream from the throttle valve 36 and between the throttle valve and the compression turbine 14 is one input of a boost control solenoid valve 52. The boost control solenoid valve 52 has two inputs 54, 56 and one output terminal 58. The first input 54 is connected to the intake manifold 22 and the other input 56 is connected to a reference pressure 60 such as ambient pressure. The output terminal 58 is connected to the wastegate valve actuator 48. The boost control solenoid valve 52 is electrically controlled from an electronic control unit (ECU) 62 and operates to supply pressure to the wastegate valve actuator 48.

The wastegate valve actuator 48 is similar to the conventional spring/diaphragm device with the exception of the spring preload and rate. The preload in the present wastegate valve 44 is set to begin opening of the gate at a very low control pressure instead of the more conventional opening of the wastegate valve 44 at a maximum boost level. The very low control pressure which is applied to the wastegate valve actuator 48 is sourced from the output of the boost control solenoid valve 52.

The pressure from the pressure tap upstream 38 from the throttle valve 36 is inputted to the boost control solenoid valve 52 and modulated thereby. This input pressure is equal to or higher than the manifold pressure to allow the wastegate valve actuator 48 to operate even when the manifold pressure is below atmospheric. By so utilizing this pressure source, the control system controls the turbocharger system at all pressure ranges, even at relatively low part-throttle conditions.

Pneumatic restrictions, not shown, contained in the inputs 54, 56 of the boost control solenoid valve 52 determine the range of the control pressures produced by the valve 52 and available to the wastegate valve actuator 48. In addition a restriction, not shown, in the wastegate valve actuator input 64 damps the control pressure fluctuation without restricting the flow of the pressure to slow the response times of the wastegate valve 44. The system is completed by the electronic control unit 62 which is a microprocessor based control unit. Associated with the electronic control unit 62 are a plurality of various engine operation sensors 66, not all shown, such as engine temperature 68, engine speed 70, manifold pressure sensor 50, exhaust gas sensor 72, knock sensor 73, etc. Of primary importance for the present system are the manifold pressure sensor 50 and the throttle position sensor 74 which are used to close the loop of the turbocharger control system. The microprocessor control unit 62 controls the boost control solenoid valve 52 in response to a boost setpoint, as previously determined and stored in the memory of the control unit 62, between the source pressure and ambient pressure at a predetermined frequency. By varying the duty cycle of the solenoid control valve 52, the control pressure is changed and the operation of the wastegate valve actuator 48 changes the opening of the wastegate valve 44.

FIGURE 2 illustrates the cartography 76 of boost setpoints defined by the engine operating modes. Each of the boost setpoints is selected by the engine speed (RPM) and throttle position values (TPS) and is further modified by the current engine temperature. At manifold pressures above the desired setpoint, the wastegate

valve 44 is opened under control of the electronic control unit 62 to return the manifold pressure to the setpoint. This control is accomplished by using a proportional-integral-derivative (PID) gain strategy to assure stability and adequate speed of response.

The duty cycle to the boost control solenoid valve 52 is maintained by a software integrator which is modified, at a programmable rate, by proportional and derivative gain terms. The proportional gain provides a term as a function of the error from the desired setpoint, attempting to return the manifold pressure to the desired setpoint. The derivative gain provides a term as a function of the rate of the manifold pressure approaching the setpoint. This is the rate of change of the error term. The derivative gain is comparable to the damping function of an analog system. The proportional and derivative gain terms are summed and the resulting total gain is modified as a function of the available source pressure and the barometric pressure. In this manner the response from the wastegate valve actuator 48 is consistent.

This PID control is used, whenever the manifold pressure is above the desired setpoint and also for a small pressure or control band 78 below the setpoint cartography 76. This small pressure band is called the control band 78 because it allows small oscillations and overshoots about the setpoint to be managed without radical alterations in the control strategy including its destabilizing effects. Below the desired setpoint complete PID strategy is difficult as a drop of manifold pressure below the setpoint would demand corrective action to restore the boost. As an example, if the engine operator's input to close the throttle valve 36 is the cause of the pressure

drop, then restoring the boost pressure would be directionally incorrect. In order to handle this, the control system holds the wastegate valve 44 in position when the pressure drops below the setpoint. By so doing this keeps the wastegate valve close to the correct position for small pressure excursions back and forth across the control band 78.

For large pressure excursions back and forth across the control band 78, the control system reacts differently. If the engine operator's input closes the throttle valve 36 for a long time and the wastegate valve 44 is held in its last position, a rapid rebuild of boost is prevented. In such a condition the wastegate valve 44 should be closed directing all the exhaust gas to the exhaust turbine 12. Arbitrarily closing the wastegate valve 44 whenever the throttle valve 36 is closed by the operator will build power contrary to the engine operator's input.

To solve this problem, a compromise between the unwanted build of boost by closing the wastegate valve 44 and inability to build boost by holding the wastegate valve 44 open is to have the wastegate valve close very slowly at any time the manifold pressure is below the control range'. This compromise effectively holds the wastegate valve 44 open during a short excursion below the setpoint. However, during longer transients below the control band 78, the wastegate valve 44 is eventually closed for rapid throttle response by the engine operator. This closing is imperceptibly accomplished by the control of the wastegate valve actuator 48.

The imperceptible closing can be overridden when the engine operator requires a rapid build of boost before the wastegate valve 44 is actually closed. This

is accomplished by stepping the wastegate valve 44 completely closed upon the detection of a large throttle valve 36 opening motion by the throttle position sensor 74. When the control range is again reentered, the PID control takes over.

In turbocharged engines, preprogrammed boost levels are not enough to correct for spark knock caused by high manifold pressure, excessively warm inlet charge or poor fuel quality. Spark knock in a turbocharged engine can be reduced either by reducing boost pressure or by retarding the spark timing. However by traversing the operating surface of the cartography of the boost setpoints 76, the effects of each variable on fuel economy, power, emissions and knock vary widely and independently. This makes it imperative that there is combined in the control strategy the ability to select the optimum combination of spark timing retard and of boost reduction or wastegate control for each setpoint.

In the present embodiment, this is accomplished by creating a second cartography, not illustrated, of engine operating conditions containing rates of retard and of pressure reduction at each setpoint. If a large knock level is reached, the spark is retarded and the boost setpoint is reduced. Retardation and reduction of the boost setpoints are accomplished at independent rates until knock is acceptably corrected. These rates may_be programmed and controlled to achieve an all-spark strategy, an all-boost strategy or any weighted combination therebetween. In such situation, wherein a condition exists such as a fully opened wastegate valve which implies spark only strategy, system limits are provided for either parameter to switch control to the other parameter.

Further control of the system strategy includes a slower return of boost and spark setpoints to their normal calibration values once the knock level is acceptably low as determined by the knock sensor 73. Continuously combining both strategies of retard/boost reduction because of knock and advance/boost increase, balances the control outputs at the chosen weights for the engine operating point, yielding a programmable trace knock level.

OPERATION

WASTEGATE CONTROL LOGIC

Referring to FIGURES 3A and 3B, the operation of the closed loop turbocharger control system with transient wastegate valve control will be explained.

The purpose of the wastegate control logic is to limit the manifold pressure to a safe level on a turbocharged engine equipped with the wastegate type of bypass actuator. The wastegate valve 44 is controlled in a closed loop fashion with the MAP sensor 50 feedback. The PID control strategy provides gain as a function of rate of change of error to provide damping and an integral change in control valve output for wastegate stability. As the wastegate valve 44 is controlled by a microprocessor, the control laws are set into a program. The overall engine control program has a periodic interrupt 80, -fuel timing pulse, FTP, signifying the occurrence of an engine event. Upon the occurrence of an FTP, the engine control program operates to follow the control laws for the wastegate control logic system, WCLS. The microprocessor has a timer, BSTTMR, which operates the boost control solenoid 52 under the control of the wastegate control logic.

The FTP counts the timer. When the timer is addressed 82 and found to be running, the timer is decremented 84 by the FTP and the previously determined duty cycle, DO.T.C, to the boost solenoid driver continues to be applied 86. The WCLS is exited 88 until the next program interrupt.

If the timer is addressed and found to be timed out, the cartography of the boost setpoints which is stored in memory, is addressed 90 in accordance with the throttle position (TPS) and the speed (RPM) of the engine. Any corrections due to coolant temperature are made by addressing 92 a memory map in response to the coolant temperature. The purpose of this correction 94 is to protect the engine from operating the turbocharger when the engine is cold, if the engine is at the proper temperature, this factor is one. The resulting boost setpoint, which has also been adjusted for offset 96, is compared 98 against a minimum allowable setpoint which has been determined by the engine builder or control designer. The offset factor has been previously calibrated by the engine builder or control designer.

If the boost setpoint is less than the minimum allowable setpoint, the minimum setpoint is the default value 100. If the boost setpoint is greater than the maximum allowable setpoint determined by the engine builder, the maximum setpoint is the default value 102, 104. The resultant setpoint is the desired setpoint, BSTSETP, and it is stored 106 for use in the spark knock control logic. The boost setpoint may also be compared 108 with a knock setpoint that has been calculated by the spark knock strategy, hereinafter explained, to reduce spark knock. When the knock setpoint is less than the boost setpoint, the knock setpoint is used 110.

The WCLS then calculates 112 the difference, DELSETP, between the desired boost setpoint, BSTSETP, and the previous or old boost setpoint, OLDSETP, to determine the change. This change is used to inhibit correction if the error is due to a rapid setpoint change. The error between the BSTSETP and the value of the manifold pressure, MAP, is determined 114. This error term is used to determine the size of the proportional gain. The value of the BSTSETP is compared 116 with the value of MAP and if the BSTSETP is less than MAP, this indicates that PID control in the boost pressure is required 118. If the BSTSETP is greater than MAP, it must be determined 120 if the value of the boost setpoint is within the narrow control band 78 below the setpoint. This is accomplished by determining if MAP is below the value of the BSTSETP minus the value of the control band, BSTSETP-CNTLRNG. If MAP is below this value 122, then the PID strategy is replaced with gate closing logic. If it is in the control band 78, PID control in the boost pressure is required 118 so as to prevent oscillating about the boost setpoint.

If the manifold pressure is determined 122 to be below the control band 78 and if the change 124 in the throttle position, DELTPS, is greater than the wastegate threshold, GTTHRS, the duty cycle output to the boost control solenoid, DO.T.C, is reduced 126 by a fixed amount SHUTGT. However, if the change in DELTPS, is less than GTTHRS, the logic 128 closes the wastegate valve 44 by a decay step value DKSTEP.

In the case 118 that the change is a small amount, the control is PID and the logic looks up 130 the rate, BSTSTP, that the duty cycle is to be adjusted as a function of MAP error from the boost setpoint. As

illustrated on FIGURE 2, the cartography 76 of boost setpoints is divided into various cells 132 which are defined by values of TPS and RPM. The absolute value of the previous setpoint less the adjustment value is calculated 134 to determine if a full cell change, CLLCHNG, has been made. If the answer is yes, then the wastegate is under integrator control. If the answer is no, then a friction factor is calculated 136 to prevent overshoot and this factor is added 138, in a signed addition, to the boost integrator step change value, BSTSTP. This value is added 140 to the present duty cycle for operation of the wastegate valve 44. The changed value of the duty cycle, DO.T.C, is compared 142 with maximum 144 and minimum 146 values of the duty cycle to keep within the duty cycle design range.

The boost setpoint timer must now be adjusted for the new duty cycle. The value of the timer must be a function ^* 148 of the boost setpoint error which was previously determined. The timer is then reset 150 to this value and the the boost integrator is set to the new duty cycle value and supplied 152 to the boost control solenoid 52 and the WCLS returns 154 to the main control system.

SPARK KNOCK CONTROL LOGIC

Spark knock can occur by high manifold pressure, excessively warm inlet charge or poor fuel quality. In a turbocharged engine, spark knock can be reduced either by reducing boost pressure or by retarding the spark timing. It is a function of this control system to provide the optimum combination of spark timing retard and boost pressure reduction for each setpoint on the surface 76 of FIGURE 2. If a large knock level is reached, spark is retarded and the boost setpoint is

reduced at independent rates until knock is acceptably correct.

Figures 4A, 4B and 4C detail the spark knock control logic strategy to reduce spark knock to an acceptable level using a knock sensor value for each cylinder to retard the spark. The selection of spark retard rate and boost reduction rate is done at each individual engine event in the engine operating range by a TPS-RPM cartography of timer values. Both the spark and boost timers run all the time with time constants selected to give correction rates tailored to each particular engine event. Limits are provided on both controls so that when one control scheme runs out of range, the other is used as a default. These limits are set on minimum boost duty cycle, minimum manifold pressure and minimum spark advance.

Each time a return counter RETCTR, counts out, the respective spark and boost setpoints are returned one step toward the normal steady state spark and boost setpoints. This combination of step back in proportion to knock level and step forward on a time basis provides a balanced value for spark advance and manifold pressure that yields the desired knock level.

In the control strategy a spark correction flag is set to indicate that a change due to knock has been made to the spark advance during the particular program interrupt cycle. When the flag is set, only one spark correction can be made during a program interrupt cycle. Without this flag and if both spark and boost corrections are being made, more than one spark correction could be accomplished which is undesirable. Such a condition would occur if the spark correction was made first in the time sequence and the boost was later reduced to its limit. In this situation, the boost

would default to spark control and another spark correction would be made.

In a similar manner, a boost correction flag is used to indicate that a boost reduction due to spark knock has already been done during the present program interrupt cycle. Again, without the flag and if the spark correction runs out of range, the spark correction will default to the boost control and call for another boost reduction.

ADDRESSING THE PROGRAM:

Upon receipt of the fuel timing pulse interrupt 160, FTP, both the spark correction 162 and boost correction 164 flags are cleared. The boost knock counter 166 and the spark knock counters 172 are decremented and tested 168,174 to see if they are equal to or less than zero. If they are less than zero, both counters are set equal to zero 170,184.

CONTINUE PROGRAMMED CONTROL:

If the spark counter is greater than zero 176 and if the boost knock counter is not equal to zero 178, 208

(Fig 4B) , the return counter is decremented 209 (Fig

^' 4C) . If the return counter is not equal to zero 238, the program exits 240 and waits for the next FTP.

START NEW BOOST CORRECTION:

If the spark counter is greater than zero 176 and if the boost knock counter is equal to zero 178 (Fig 4B), the boost knock counter is set 180 from a cartography of boost time constants, BTCCRT, to a given time constant based on TPS and RPM. The timer cartography is a 9 X 13 matrix cartography which is stored in a ROM, wherein each cell contains a boost time

constant to which the boost knock counter is set each time it runs out. This time constant is used to determine the frequency of boost correction per program interrupt to eliminate spark knock. When the counter times out, the boost setpoint is modified in proportion to the amplitude of the knock. The boost correction flag, BSTCOR, is set 182 and the amount of boost reduction is determined 183 from a look-up table based on values of the knock sensor. The boost setpoint is changed by the subtraction of the boost reduction value 210 (Fig 4C) and the new boost setpoint is checked 214 to see if it is less than a minimum value. If it is 212, the boost setpoint is defaulted to the minimum value and the spark correction flag is checked 218. Since the boost setpoint is at the minimum value and the spark correction flag is reset, the program defaults to the spark knock control 220.

START SPARK ADVANCE CORRECTION: If the spark counter is less than or equal to zero 174, it is set equal to zero 184. The min spark flag is reset 186 indicating the end of min spark correction as a function of boost correction, spark correction flag is set 188 indicating that a spark correction by spark timing control will be done in the current interrupt and the program addresses the spark timing control.

SPARK KNOCK CONTROL:

The spark counter is set 190 with a value of spark time constant from the spark time constant cartography from a look-up table. This is to determine the frequency of spark retard correction to eliminate spark knock. The spark counter determines the frequency of the spark advance correction to compensate for knock.

When the spark advance counter times out, the spark advance command is modified in proportion to the amplitude of the knock. This counter is set from 9 X 13 matrix cartography of TPS-RPM wherein each cell contains a value to which the spark advance counter is reset 190 each time it runs out.

The control of each cylinder is individually checked 192-204 (Fig 4B) for the amount of spark advance angle which is a function of the knock sensor input per cylinder, K KADV,^^ 194. The size of the step in the retard direction that the spark advance is moved during each interrupt, FTP, is found 194 for each cylinder in a table in memory that is proportional to the knock input. The table is constructed to have a value of degrees of crank angle of advance for each entry. As in the wastegate control, a control band is provided to allow no knock correction below a threshold.

The knock advance angle per cylinder is corrected by a value RETSTP from the look-up table values of a predetermined size of the spark advance step 196. If the knock advance angle per cylinder is less than a minimum value MINSPK 198, the program defaults to a minimum value 199 and the min spark flag, MINSPK FLAG, is set 200. This indicates that the particular cylinder cannot be corrected with spark advance. Each cylinder in turn is checked until all of the clyinders have been processed 202.

MIN SPARK THEN GO TO BOOST CONTROL: After all cylinders have been checked 202, and if the min spark flag has been set 206, the boost knock counter BSTKCTR is set 180 from a cartography of boost time constants, BTCCRT, to a given time constant based on TPS and RPM. This time constant is used to determine

the frequency of boost correction to eliminate spark knock. The boost correction flag, BSTCOR, is set 182 and the amount of boost reduction is determined 183 from the knock sensor. The boost setpoint is changed by the subtraction 210 (Fig 4C) of the boost reduction value, BSTRED, and the new boost setpoint is checked 214 to see if it is less than a minimum value. The manifold pressure is another limitation to the control of spark advance strategy. Beyond a minimum pressure, there is no reduction in the boost control setpoint in response to spark knock. If the new boost setpoint is less than a minimun value 212, the boost setpoint is defaulted to the minimum value and the spark correction flag is checked 218.

BOOST GREATER THAN MIN BOOST:

If the new boost setpoint is greater than the minimum value 214, the duty cycle value, DO.T.C, 216 is check to see if it is at its maximum value indicating that the wastegate is fully opened. The duty cycle of the boost control solenoid is limited to cycle values, $EO, between a pair of limits which are a function of the wastegate valve. In the preferred embodiment these limits are 12% and 88%. The upper boundary is taken as an indication that the wastegate is fully open and no further correction of boost is possible. If it is, the spark correction flag is checked 218 and since it is set, the return counter is decremented 209. If the return counter is not equal to zero 238, the program exits 240 and waits for the next FTP.

SET BOOST SETPOINT ONE STEP TOWARDS NORMAL: If the return counter is equal to zero 239, a new value, RETTIN, is set into the return counter 221 and

the boost setpoint is returned 222 one step, BSTRTN, toward the normal steady state boost setpoint value. The value of the boost setpoint is compared with the desired boost setpoint, BST.CRT, 224 and if it is greater, the boost setpoint defaults 225 to the desired value.

SET SPARK ADVANCE BY ONE STEP ACCORDING

TO KNOCK SENSOR:

Next the spark advance angle, K KADV^ _j ., for each cylinder is returned 226-236 one step, SPKRTN, toward the normal steady state spark advance setpoint for each cylinder.

EXIT:

After all cylinders have been checked 236, the program exits 240 and waits for the next FTP.

Many changes and modifications in the above described embodiment of the invention can, of course, be carried out without departing from the scope thereof.

Accordingly, that scope is intended to be limited only by the scope of the appended claims.