Configurable interface circuit for exhaust gas oxygen sensors

Information

  • Patent Grant
  • 6831471
  • Patent Number
    6,831,471
  • Date Filed
    Thursday, November 14, 2002
    22 years ago
  • Date Issued
    Tuesday, December 14, 2004
    19 years ago
Abstract
An oxygen sensor interface circuit is configurable on the fly by an electronic controller such as an engine controller to support oxygen sensors having unique interface requirements, to reliably identify various oxygen sensor faults, and to enable rapid detection of a warmed up sensor. The interface circuit is configurable in a first respect to enable operation with any of a number of different sensors, and in a second respect to enable more reliable fault detection, including measurement of leakage to ground or battery.
Description




TECHNICAL FIELD




This invention relates to a circuit for interfacing an electronic engine control unit (ECU) to an exhaust gas oxygen sensor, and more particularly to a circuit that can be configured by the ECU to work with different types of oxygen sensors and to diagnose the operation thereof




BACKGROUND OF THE INVENTION




In a motor vehicle engine control system, engine fueling is controlled by a closed-loop control system that includes one or more exhaust gas oxygen sensors, and a corresponding number of circuits for interfacing the oxygen sensors with an electronic controller. Electrically, the oxygen sensor can be characterized as a two terminal analog device, and the interface circuit functions not only to develop a suitable input for the controller, but also to diagnose electrical faults such as shorts and open-circuits. See, for example, the U.S. Pat. No. 5,467,034 to Manlove et al., issued on Nov. 14, 1995, assigned to the assignee of the present invention, and incorporated herein by reference. As described in Manlove et al., the interface circuit can include RC filter elements to remove unwanted noise and a bias voltage to diagnose cold sensor and open-circuit fault conditions. However, oxygen sensors produced by different manufacturers require unique interface circuitry, and leakage paths between the sensor terminals and ground and/or battery make it difficult or impossible to reliably distinguish a normally operating rich-condition sensor from an open-circuit. Accordingly, what is needed is an improved interface circuit that is flexible enough to work with sensors produced by different manufacturers and to reliably diagnose sensor failures.




SUMMARY OF THE PRESENT INVENTION




The present invention is directed to an improved oxygen sensor interface circuit that is configurable on the fly by an electronic controller such as an engine controller to support oxygen sensors having unique interface requirements, to reliably identify various oxygen sensor faults, and to enable rapid detection of a warmed up sensor. According to the invention, the interface circuit is configurable in a first respect to enable operation with any of a number of different sensors, and in a second respect to enable more reliable fault detection, including measurement of leakage to ground or battery.











BRIEF DESCRIPTION OF THE DRAWINGS





FIG. 1

is a diagram of a motor vehicle powertrain, including pre-catalyst and post-catalyst oxygen sensors, an engine control module, and configurable oxygen sensor interface circuits according to this invention.





FIG. 2

is a diagram of a configurable oxygen sensor interface circuit according to this invention.





FIG. 3

is a diagram of a switched capacitor implementation of a configurable input impedance circuit for the oxygen sensor interface circuit of FIG.


2


.





FIG. 4

is a flow diagram of a software routine executed by the engine control module for configuring the oxygen sensor interface circuit of

FIG. 2

to diagnose sensor operation according to this invention.











DESCRIPTION OF THE PREFERRED EMBODIMENT




Referring to

FIG. 1

, the interface circuit of the present invention is disclosed in the context of a motor vehicle powertrain


10


including an internal combustion engine


12


having a throttle valve


14


for receiving intake air and an output shaft


15


connected to drive the vehicle. The intake air is combined with fuel supplied by a fuel controller (FC)


34


, and supplied to engine cylinders for combustion, with the exhaust gases being collected in an exhaust manifold


16


and passed through a catalytic converter (CC)


20


for emission control purposes. Feedback signals for air/fuel ratio control are developed by a first exhaust gas oxygen sensor


24


located upstream of the converter


20


and a second exhaust gas oxygen sensor


26


located downstream of the converter


20


. Oxygen sensor interface circuits (OSIC)


28


,


30


couple the respective oxygen sensors


24


,


26


to A/D input ports of a microprocessor-based engine control module (ECM)


32


, and the ECM


32


regulates the operation of fuel controller


34


based in part on the feedback signals. According to the present invention, the interface circuits


28


,


30


are configurable to support oxygen sensors having unique interface requirements and to reliably identify and distinguish various oxygen sensor faults, and the ECM


32


includes outputs


36


,


38


for individually configuring the interface circuits


28


,


30


as described below.




The interface circuits


28


,


30


of

FIG. 1

are identical, and

FIG. 2

depicts a top-level diagram of OSIC


28


. The input lines from oxygen sensor


24


are labeled


24




a


and


24




b,


as in

FIG. 1

, and additionally as Vin_high and Vin_low in FIG.


2


. Referring to

FIG. 2

, Vin_low is coupled to circuit ground


42


, while Vin_high is coupled to an interface circuit including a low-pass input filter comprising resistor


44


and capacitor


46


, and an amplifier comprising operational amplifier


48


and resistors


50


and


52


. The output voltage Vout of operational amplifier


48


on line


54


is applied as an input to an A/D input port of ECM


32


as shown in

FIG. 1

, and in the illustrated embodiment, the resistors


50


and


52


are selected to define an amplifier gain of 2.5.




The remaining circuit elements of OSIC


28


are selectively coupled to the above-described filter and/or amplifier by the switches


56


,


58


and


60


, depending on the operating characteristics of oxygen sensor


24


. If the oxygen sensor


24


is a so-called narrow-band sensor, the switch


56


is closed and the switches


58


and


60


are opened. If the oxygen sensor


24


is a so-called pumped-voltage reference sensor of the type manufactured by Robert Bosch. Corporation, the switch


58


is closed and the switches


56


and


60


are opened Finally, if the oxygen sensor


24


is a so-called pumped-current reference sensor of the type manufactured by Delphi Corporation, the switch


60


is closed and the switches


56


and


58


are opened. In practice, the switches


56


,


58


,


60


can be configured as semiconductor switches, the state of each such switch being controlled by the microprocessor of ECM


32


via a serial peripheral interface (SPI) circuit incorporated within ECM


32


.




In the case of a narrow-band sensor, the switch


56


couples input line


61


of operational amplifier


48


to a bias voltage source


64


through a configurable resistance (CR)


62


. The configurable resistance is remotely controlled by ECM


32


via line


58


, and acts like either a 480 kohm resistor or a 1.2 megohm resistor, depending on the signal Rsel impressed on line


58


by ECM


32


. At engine start-up when the sensor


24


is cold, its characteristic impedance is on the order of 5-10 megohms, and ECM


32


signals CR


62


to select the 1.2 megohm resistance to enable reliable cold sensor detection based on the output voltage Vout. Once the sensor


24


warms up and becomes operational, its characteristic impedance falls to a lower value (typically much less than 100 kohms), and ECM


32


signals CR


62


to select the 480 kohm resistance to enable reliable open-circuit detection accuracy, again based on the output voltage Vout. While the circuit may be implemented substantially as shown in

FIG. 2

, with switched discrete resistors in CR


62


, the combination of switch


56


, CR


62


and bias voltage source


64


is preferably implemented as a two-phase switched capacitor circuit coupled between a source voltage VCC and ground


42


, with ECM


32


providing a two-phase clock signal PH


1


, PH


2


, as shown FIG.


3


. Referring to

FIG. 3

, the capacitors


92


and


104


are joined at node


94


to form a capacitive divider, and the switch pairs


82


/


84


,


88


/


90


,


96


/


98


,


100


/


102


are turned on and off by the clock signals PH


1


, PH


2


as shown to alternately charge and discharge capacitors


92


and


104


at a switching frequency f


s


determined by ECM


32


. The switched capacitor circuit generates an equivalent bias voltage V


b


according to:








V




b




=VCC*[C




92


/(


C




92






30


C




104


)]






where C


92


and C


104


are the capacitances of capacitors


92


and


104


; in the illustrated embodiment, C


92


and C


104


are selected so that Vb is 0.45V. The equivalent resistance R


eq


of the circuit varies with the switching frequency f


s


, as follows:








R




eq


=1/[(


C




92




+C




104


)*


f




s


]






Given the capacitance values for achieving the desired bias voltage, the switching frequencies for achieving the alternate resistance values of 480 kohlms and 1.2 megohms are easily determined. When the sensor


24


is a pumped-voltage or pumped-current sensor, phase PH


2


can be maintained low to effectively duplicate the functionality of opening switch


56


.




In the case of a pumped-voltage reference sensor, the switch


58


couples Vin_high to a configurable high impedance reference voltage via the resistor


70


. The reference voltage is defined by an operational amplifier


66


configured as a voltage follower, the amplifier


66


receiving a reference voltage Vref from ECM


32


on line


68


. Although the manufacturer specifies a resistor value of 56.2 kohms and a reference voltage of 1.8V for open-circuit sensor detection, the reference voltage Vref is configurable by ECM


32


according to this invention to enable identification of sensor harness leakage, as explained below. When Vref is the specified value of 1.8V, an open-circuit or cold sensor


24


results in an output voltage Vout equal to the product (1.8V*2.5), or 4.5V.




In the case of a pumped-current reference sensor, the switch


60


couples Vin_high to a current source


74


via the resistor


70


. The current source


74


is powered by a 5V supply voltage as shown, and in the illustrated embodiment, sources a DC bias current of 8.5 uA. Should there be an open-circuit between input line


24




a


and sensor


24


, the output voltage Vout on line


54


becomes equal to the maximum output voltage of operational amplifier


46


(i.e., the positive source voltage for the circuit


28


), which indicates that the sensor


24


has either an open circuit or a short to battery.




Although the controlled switches


56


-


60


allow ECM


32


to configure the interface circuits


28


,


30


for use with any of a number of types of oxygen sensors as explained above, such configurability alone does not address the diagnostic ambiguity due to sensor leakage. For example, moisture or contamination in or near the sensor harness or circuit board, or even manufacturing variability, can result in a parasitic leakage path between the sensor line


24




a


and either ground or battery voltage (14V in a typical implementation). The leakage can cause an output voltage Vout that is indiscernible from the voltage one would expect with an operational sensor under rich fueling conditions, or the voltage one would expect with an open-circuit failure. Thus, the possibility of sensor leakage paths rules out reliable sensor diagnostic testing with the traditional diagnostic procedures.




The above-described problem of diagnostic ambiguity is addressed, according to this invention, by using the dynamic configurability of the interface circuits


28


,


30


in order to specifically access the sensor leakage. In general, the reference voltage Vref is configured so that the output Vout has a value of approximately one-half the dynamic range of operational amplifier


48


when the sensor


24


has an open-circuit failure. For example, if ECM


32


sets Vref to 1.1V using a D/A output port, Vout will have a value of (1.1V* 2.5), or 2.75V, when the sensor


24


is open-circuited. A sensor leakage path to battery will cause Vout to be higher than 2.75V, whereas a sensor leakage path to ground will cause Vout to be lower than 2.75V. In either case, ECM


32


can compute the effective resistance R


eff


of the leakage path based on Vref, Vout, and battery voltage Vbatt (which can be measured by ECM


32


) in the case of a detected leakage path to battery.

FIG. 4

depicts a simplified flow diagram for this procedure, wherein the block


120


initializes the lost by setting Vref to 1.1V, opening switches


56


and


60


, and closing switch


58


. Block


122


determines if Vout is within a calibrated value (CAL) of 2.75V. If so, the block


124


sets the OPEN SENSOR flag to TRUE, completing the routine. If not, there is significant sensor leakage, and block


126


determines whether the leakage path is to ground or battery. If the leakage path is to ground, block


126


will be answered in the affirmative; in this case, block


128


sets the LEAKAGE PATH TO GROUND flag to TRUE, and block


130


computes the effective leakage resistance Reff as a function of Vref and Vout. If R


eff


is low enough to prevent reliable sensor diagnostic detection, the block


132


will be answered in the affirmative, and block


134


sets a suitable diagnostic flag. If the leakage path is to battery, block


126


will be answered in the negative; in this case, block


136


sets the LEAKAGE PATH TO BATTERY flag to TRUE, reads the battery voltage Vbatt, and computes the R


eff


as a function of Vref, Vout and Vbatt. If Reff is low enough to prevent reliable sensor diagnostic detection, the block


138


will be answered in the affirmative, and block


140


sets a suitable diagnostic flag. The calibration value CAL takes into account variation in the reference voltage Vref, as it will typically be subject to a tolerance of ±5%; however, an additional configurable switch


150


as shown in phantom in

FIG. 2

may be added to enable ECM


32


to measure Rref via operational amplifier


48


and eliminate error due to tolerance variation, if desired.




In summary, dynamically configuring an oxygen sensor interface circuit as described herein allows a single interface circuit to be used with different types of oxygen sensors, and significantly improves diagnostic reliability by selecting circuit components that are well suited to the diagnostic test and by ruling out false diagnostic indications due to sensor harness leakage. In a typical implementation, the ECM


32


will configure the interface circuit for the specified oxygen sensor


24


, and conduct traditional cold-sensor diagnostic tests at engine turn-on to identify the moment at which the sensor warms up and becomes operational. When a presumed rich operational sensor is detected, the ECM


32


configures the interface circuit


38


,


40


to measure the sensor harness leakage as described in reference to FIG.


4


. If the leakage is too high to reliably diagnose proper sensor operation, a diagnostic flag is activated. If the leakage is insufficient to introduce diagnostic ambiguity, a warmed-up operational sensor is confirmed, and in the case of a narrow band sensor, CR


62


is configured to optimize open-circuit diagnostics. Since warmed-up sensor operation is detected quicker and more reliably, ECM


32


can begin closed-loop fuel control sooner than possible with conventional interface circuitry and diagnostic procedures.




While described in reference to the illustrated embodiment, it is expected that various modifications in addition to those mentioned above will occur to persons skilled in the art. For example, the invention is not limited to the illustrated types of oxygen sensors, the component and voltage values may differ in a given application, and soon. Accordingly, it Should be understood that interface circuits incorporating such modifications may fall within the scope of this invention, which is defined by the appended claims.



Claims
  • 1. An interface circuit coupled to an electronic control unit and an exhaust gas oxygen sensor, comprising:an amplifier having an input coupled to the oxygen sensor and an output coupled to the electronic control unit; a plurality of different bias circuits for diagnosing the oxygen sensor, said bias circuits being designed in accordance with diagnostic requirements of different types of oxygen sensors; and dynamically configurable means controlled by the electronic control unit for connecting a selected one of said different bias circuits to the amplifier input so that the selected bias circuit is designed in accordance with diagnostic requirements of the oxygen sensor that is coupled to said interface circuit, wherein at least one of said bias circuits receives a reference voltage from said electronic control unit that is adjustable for diagnostic purposes when such bias circuit is selected by said dynamically configurable means, and said electronic control unit includes diagnostic means for setting said reference voltage to a diagnostic value that produces a known output of said amplifier when there is no electrical leakage path between said sensor and another electrical potential, and for diagnosing the existence of an electrical leakage path when the output of said amplifier is different than said known value.
  • 2. The interface circuit of claim 1, wherein said electronic control unit is connected to a storage battery, and said diagnostic means diagnoses a leakage path between said sensor and said storage battery when the output of said amplifier is higher than said known value.
  • 3. The interface circuit of claim 2, wherein said diagnostic means determines an effective resistance of the diagnosed leakage path as a function of said diagnostic value, a measured voltage of said storage battery, and the output of said amplifier.
  • 4. The interface circuit of claim 1, wherein said electronic control unit is connected to a ground potential, and said diagnostic means diagnoses a leakage path between said sensor and said ground potential when the output of said amplifier is lower than said known value.
  • 5. The interface circuit of claim 4, wherein said diagnostic means determines an effective resistance of the diagnosed leakage path as a function of said diagnostic value and the output of said amplifier.
US Referenced Citations (2)
Number Name Date Kind
5405521 Nakamori et al. Apr 1995 A
6547955 Hada et al. Apr 2003 B1