JPH0631564B2 - Knock control device for internal combustion engine - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Field of Application] The present invention is for an internal combustion engine that detects knock conditions occurring in an internal combustion engine (engine) and controls knock control factors such as ignition timing, air-fuel ratio, intake pressure, and the like. Regarding the knock control value.

[Conventional technology]

Conventionally, knock detection means for detecting knock generated in an internal combustion engine, amplification means for amplifying the output of the knock detection means, and amplification factor varying means for changing the amplification factor of the amplification means according to the rotation speed of the internal combustion engine And a knock discrimination means for discriminating the knock state by the output of the amplification means at the amplification rate determined by the amplification rate varying means, and an ignition timing, an air-fuel ratio, an intake pressure according to the discrimination result of the knock discrimination means. There is known a device including a knock control factor control means for controlling a knock control factor such as (for example, Japanese Patent Laid-Open No. Sho 60-3).
5238). There is also known a device which generates a wail output when the output signal of the knock sensor is lower than the fail judgment level (for example, Japanese Patent Laid-Open No. 60-17336).
Issue).

[Problems to be solved by the invention]

However, in the case of simply combining the former one with the latter one described above, even when the knock sensor is normal, the sensor output in the low speed and light load range of the internal combustion engine is very small. It is very difficult to distinguish from sensor abnormalities such as deterioration. For this reason, a special amplifier for high-magnification fail detection is added, safety is sacrificed to some extent to detect a failure only in a higher rotation range, or abnormal detection of sensor output deterioration is abandoned. There is a problem in that cost and reliability cannot be achieved at the same time because detection of only the electrical connection state such as disconnection or short circuit of the sensor signal line is stopped.

Therefore, an object of the present invention is to balance cost and reliability.

[Means for solving problems]

Therefore, according to the present invention, as shown in FIG. 1, knock detection means for detecting knock generated in the internal combustion engine, amplification means for amplifying the output of the knock detection means, and Amplification factor varying means for gradually changing the amplification factor of the amplifying means in a plurality of stages, knock determining means for determining knock by the output of the amplifying means at the amplification factor determined by the amplification factor varying means, and the knocking factor. Knock control factor control means for controlling knock control factors such as ignition timing, air-fuel ratio, intake pressure, etc. according to the determination result of the determination means, and fail determination for generating a fail output when the output of the amplification means is lower than the fail determination level. Means, and the fail judgment of the fail judgment means substantially when the amplification factor of the amplification device is set to a small amplification factor by the amplification factor variable device. And it provides a knock control device for an internal combustion engine and a failure determination disabling means for disabling.

[Action]

As a result, the fail determination of the fail determination means when the amplification factor varying means sets the amplification factor of the amplification means to the smaller side is substantially invalidated by the fail determination invalidation means.

〔Example〕

 The present invention will be described below with reference to embodiments shown in the drawings.

In FIG. 2, 1 is a knock sensor that serves as knock detection means for detecting engine knock, 2 is an engine control electronic control unit (ECU) for controlling engine ignition timing, air-fuel ratio and the like, and 3 is An igniter 4 which receives an ignition control signal from the ECU 2 to energize / disconnect an ignition coil, and an injector 4 which receives an air-fuel ratio control signal from the ECU 2 to inject fuel into an intake system of the engine. Further, in the ECU 2, 21 is a filter circuit that passes only the frequency component specific to the knock of the knock sensor output signal, and 22 is a part of amplification means for amplifying or attenuating the signal after passing through the filter with a predetermined amplification factor. , 23 is an amplifier forming a part of an amplifying means for further amplifying or attenuating the output of the amplifier 22 at a predetermined amplification factor, 24 is an A / D converter built-in microcomputer, and 25 is various types. Ignition timing in response to knock detection results from the sensor and the microcomputer 24,
It is a host computer for controlling the air-fuel ratio and the like.

The knock sensor 1 is a piezoelectric knock sensor and converts vibration of the engine block caused by knocking into an electric signal. A noise component is removed from this electric signal by passing through the bandpass filter 21, and only the frequency component (eg, 8 KHz) peculiar to knock is input to the amplifier 22. The amplification factor of the amplifier 22 is determined in consideration of the output characteristics of the knock sensor used, the dynamic lens of the A / D converter 24a described later, and the like. If the amplification factor of the amplifier 22 is C ₁ , G ₁ may be 1 or more and 1 or less.

The output of the amplifier 22 is branched into two, one of which is an analog input port A of an A / D converter 24a incorporated in a microcomputer 24 (in this embodiment, a MB88413 manufactured by Fujitsu is used as a one-chip microcomputer).
It is input to N1 and the other is input to the amplifier 23. The amplifier 23 is for further amplifying the output of the amplifier 22 at the previous stage, and its amplification factor is, for example, 4 times. The output of the amplifier 23 is input to the other analog input port AN2 of the A / D converter 24a. By doing so, sensor signals differing only by a factor of 4 are input to the two ports AN1 and AN2 of the A / D converter 24a at the same time (AN1 is amplification factor G ₁ and AN2 is amplification factor G ₂ = 4 × G ₁ ). The amplification factor of the amplifier 23 can be any value, but is 2, 4, 8, ... Or 1 /
It is desirable to set to "power of 2" such as 2, 1/4, 1/8, .... Because the microcomputer 24
Controls the input gain of the knock sensor 1 by appropriately selecting and using the sensor input ports AN1 and AN2. At that time, it is necessary to correct the calculated value, and if the relative ratio of the powers of 2 is reached, the calculation is performed. This is because a shift calculation process that is simple and has a short execution time can be used. Further, by setting the power of 2, it becomes possible to widely apply to a microcomputer without any division or multiplication function.

The amplifier 23 is set after the amplifier 22 for the purpose of keeping the relative signal ratio (four times in this example) with high accuracy. That is, even if the output of the filter 21 is branched into two and input in parallel to the amplifier 22 and the amplifier 23 and the output thereof is input to the ports AN1 and AN2, the function is the same. However, in this case, in order to maintain the accuracy of the relative signal ratio (for example, 4 times), it is necessary to increase the accuracy of both the amplification factor G ₁ of the amplifier 22 and the amplification factor G ₂ of the amplifier 23 (in this case, since G ₂ / _{G 1} is relative signal ratio).

On the other hand, in the case of the embodiment, the error can be reduced because the relative signal ratio is determined only by the amplifier 23.

Now, the microcomputer 24 has an A / D
The converter 24a and a central processing unit (CP not shown)
U), storage device, (ROM, RAM), input / output device (I
/ 0) etc., analog input port AN for A / D conversion
Knock determination and knock sensor fail detection are executed based on the knock sensor signal values fetched from 1 and AN2. The knock determination result and the knock sensor fail result are output from the output port of the microcomputer 24, and the engine control host computer 2 is output at an appropriate timing.
5 through its input port. The host computer 25 is also a microcomputer, and it is preferable to use a relatively high-grade microcomputer such as 8-bit. As is well known, the engine control microcomputer 25 calculates a basic ignition timing and a basic injection time based on sensor signals from a rotation angle sensor, an intake air amount sensor (air flow meter), etc., which are not shown. The air-fuel ratio is controlled by adding a correction value from various sensors (for example, a water temperature sensor) to the basic injection time, or by feedback-controlling the injection time by a rich / lean signal from an oxygen concentration (O ₂ ) sensor. More achieved.

On the other hand, the ignition timing is always controlled near the knock limit by advancing and retarding the ignition timing according to the knock determination result of the microcomputer 24 with reference to the basic ignition timing. Further, when the microcomputer 24 detects a failure of the knock sensor 1, the engine control host computer 25 also executes measures such as maximizing the ignition timing. The host computer 25 also manages the reset signal of the microcomputer 24. That is, when the power of the host computer 25 is turned on, the host computer 25 first initializes itself and then sends a reset signal to the knock detection microcomputer 24 to initialize it.

Next, the operation of this embodiment will be described. FIG. 3 shows a timing chart of the microcomputer 24 for knocking detection, determination, sensor fail detection, and delay angle calculation output of this embodiment. FIG. 4 shows a microcomputer 24
3 is a flowchart showing a basic program flow of the above. In the main routine starting from step 220 of the start, the ignition period T180 is calculated by the built-in timer (step 221), and the delay time (masking time) T1 until the start of A / D conversion and A / D conversion are performed based on the calculated value. Time (judgment time) T2, calculation of a predetermined experimentally determined magnification (K value) for multiplying the average value of the knock sensor signal in order to obtain a judgment level to be compared with the A / D converted value (step 222), and knock The calculation of the fail determination level (V _F ) for sensor fail detection (step 223) is executed. At the falling edge of the interrupt start (IRQ) signal output from the host computer 25 (211 in FIG. 3), the microcomputer 24 is interrupted (step 211-1 in FIG. 4). In this embodiment, the host computer 25 causes the IRQ signal to fall at the timing of 10 ° CA before the top dead center (BTDC) of each cylinder of the engine. After the interruption routine is started, the microcomputer 24 is operated during the masking time T1 (FIG. 213) following the interruption process (212 in FIG. 3).
Has A / D conversion (step 213-1 in FIG. 4).
After the masking time ends, the knock sensor signal input to the analog input port AN1 or AN2 is repeatedly A / D converted for the period of the value of time T2 (FIG. 3 214).
(Step 214-1 in FIG. 4), after the A-D conversion value is smoothed (step 214-2 in FIG. 4), the A / D conversion value V _AD for each time and the cylinder-by-cylinder determination of the previous calculation Level V
_{When LE} is compared with V _LEV <V _AD , the knock pulse is counted up (step 214-3 in FIG. 4).
The A / D converted value smoothing process in step 214-2 is performed as follows.

Here, V _ADi is the A / D converted value of this time, V _MADi-1 is the result of the previous _smoothing , and V _MADi is the result of the current _smoothing .

Next, to perform the data substitution for the maximum value V _peak calculated in step 214-4. That is, when the current A / D converted value V _ADi is larger than the maximum value V _peak up to the previous time, V _peak is replaced with the current V _ADi , and when V _ADi is smaller than V _peak, up to the previous time. The V _peak value of is stored as it is.
This is repeated during the T ₂ loop (step 214-5).
Therefore, the maximum peak value V of the knock determination section (T ₂ )
_Peak can be obtained (see FIG. 8).

After the A / D conversion is completed, the average value of the A / D converter 24a is calculated at the timing of 215 in FIG. 3, and further the cylinder-by-cylinder determination level for the next determination section is calculated (step 2 in FIG. 4).
15-1, 215-2).

Here, the average value calculation in step 215-1 is performed as follows.

Here, V _MAD is the final _smoothed result calculated in step 214-2, V _MEANi-1 is the average value up to the previous cycle, and V _MEAN is the average value up to this time.

The determination level V _LEV in step 215-2 is calculated as follows.

V _LEV = K × (V _MEAN + V _MOS ) Here, V _MOS is a constant for correcting an A / D conversion error, an arithmetic processing error, and the like, and is obtained in advance by an experiment.

Knock determination level V obtained as described above
_LEV changes according to engine conditions and also changes for each cylinder. That is, the amount V on which V _LEV is calculated
_MEAN is an amount corresponding to the sensor output average value of only that cylinder. Therefore, even if the K value is constant in all cylinders, V _LEV changes from cylinder to cylinder.

The microcomputer 24 stores / updates the determination level (for example, four cylinders have four cylinders) in the RAM area for each cylinder.

Next, within the timing of FIG. 215 of FIG. 3, the fail output is calculated as follows in step 215-3 of FIG.

Where V _Faili is the result of this fail output, V
_Faili-1 is the fail output up to the previous time, and _Vpeak is the maximum _peak value in the knock determination section as described above. This V _Faili
Is calculated for each cylinder similarly to V _MEAN .

Next, the selection of the A / D conversion port in step 215-4 of FIG. 4 within the timing of FIG. 3 215 is performed as follows according to the flowchart shown in FIG.

First, when the microcomputer 24 is initialized, the A / D conversion port is preset to the port with a relatively large amplification factor, that is, AN2. When the engine condition changes and the engine speed increases, the sensor output increases and V _MEAN increases. As a result, knock determination level V _LEV = K × (V _MEAN +
There are cylinders whose V _MOS ) exceeds a predetermined value V _MAX . This will be supplementarily described with reference to FIG. For example, the knock determination level V of a cylinder having a relatively large output (j-cylinder)
_LEV exceeds a predetermined value V _MAX at a certain rotation speed as the engine rotation speed is increased.

At this time, the microcomputer 24 switches the A / D conversion port at the timing of the cylinder to the AN1 side with a low amplification factor (steps 261 to 263). For another cylinder, the knock determination level V _{LEV for} that cylinder is the predetermined V _MAX.
When it exceeds, the A / D conversion port is switched from the AN2 side to the AN1 side (see #k cylinder in FIG. 9). Then, for the cylinder whose port is switched from the AN2 side to the AN1 side, V _MEAN , V _LEV and V _Fail corresponding to the cylinder are corrected by the relative signal ratio (step 264). In this example, it is 1/4. At this time, if the ratio is set to a power of 2, high speed processing can be performed by shift calculation. That is, in this case, by shifting V _MEAN, V _LEV , and V _Fail to the right by 2 bits, the value can be corrected to 1/4 at high speed. Now, on the contrary, when the engine speed decreases and V _LEV falls below the predetermined value V _MIN , the A / D conversion port is switched to the side with the higher amplification factor (that is, the AN1 to AN2 side is switched:
Together with steps 261, 265 and 266), V _LEV , V _MEAN and V _Fail of that cylinder are shifted to the left by 2 bits (ie 4 times: step 267). Therefore, for example, under the condition that the engine speed is constant, the #j cylinder may be selected as the port on the AN1 side and the #k cylinder may be selected as the port on the AN2 side. To explain this a little more, for example, in a 4-cylinder engine, the ignition order is # 1, # 3, # 4, #.
If it is the order of 2 cylinders, the A / D conversion value from the port on the AN2 side is taken in at the knock detection timing of the # 1 cylinder, and the A / D conversion value is taken at the next knock detection timing of the # 3 cylinder.
The A / D conversion value from the port on the N1 side is taken in. By such an operation, the amplification factor can be switched for each cylinder and according to the output level of the sensor signal.

If the ratio between the upper and lower switching threshold values V _MAX and V _MIN is set with a margin greater than the relative signal ratio (4 times in this example), the switching characteristics of the port can have hysteresis, and the switching characteristics can be changed. It is possible to prevent motion hunting (see FIG. 9 (a)).

Next, a detailed operation of the A / D conversion portion and the knock determination output will be described with reference to FIGS.

Reference numeral 301 in FIG. 5 denotes one cycle of the knocking detection signal. As described above, the detection signal has a sine wave of 8 KHz. Referring to FIG. 6, when the A-D conversion program is started (step 320), the mode of the comparator which is the comparison function in the one-chip microcomputer 24 is designated and the threshold level Th is changed from 0 level. It is set (step 321). Then, the comparator is started (step 322), and the falling time TDO of the sine wave 301 from Th is detected (step 323-).
1) Then, the Th level is further confirmed for the purpose of preventing malfunction (step 323-2). After the Th level is detected, the routine proceeds to the routine for detecting the rise from the Th level (step 325), the rise detection time TZ1 is delayed by a predetermined time ΔD (step 326), and the time TS1 is reached. In the rising edge detection routine (step 325), the judgments are arranged in series so that even if the rising edge is not detected, the final step 3
By shifting to 26, the arithmetic processing is prevented from falling into an infinite loop. Then, the time when the A / D conversion is started from the time TS1 is set around the peak 310-1 of the sine wave 301. At that time, the slope near the peak of the sine wave 301 is considered to be almost constant as compared with the rising and falling slopes of the sine wave, and even if A / D conversion is performed by delaying the delay time ΔD to near the peak ( (Step 327) The peak value V _P of the sine wave 301 is obtained. That is, both the value V _S1 given at the conversion start time T _S1 and the value V _fi given at the end time T _fi are much smaller on the left side than the peak value V _P, and in this example, A / D conversion is performed successively. Since the comparison A / D conversion is used, the A / D conversion value obtained is closer to V _P than V _S1 and V _fi . The obtained conversion value is placed in the read memory (step 32).
8) Then, move to the next operation step 329.

FIG. 7 is a more detailed flowchart of the pulse count portion (step 214-3 in FIG. 4) for knock intensity determination.

The A / D converted value V _AD is compared with the cylinder-by-cylinder determination level V _LEV (step 215-2 in FIG. 4) obtained up to the previous time (step 331), and the A / D converted value is determined by the cylinder. If it is lower than the judgment level V _LEV , the next operation step 3
Move to 33. When it is determined that the value is greater than the cylinder-specific determination level V _LEV , 1 is added to the knock pulse counter (step 332).

By repeating during the pulse count of the knock determination period _{(T 2} loop step 414-5 of FIG. 4) 1
The total number of pulses during the ignition cycle is determined. This total number of pulses is the amount corresponding to the knock intensity. For example, the knock intensity can be classified and determined such that the total number of pulses 0-1 is no knock, 2-5 is small knock, 6-9 is medium knock, and more is large knock. The inclusion of the number of pulses of 1 without knocking is a device for preventing the electric sharp noise from being erroneously judged as knocking.

The result of the knock strength thus determined is transmitted from the output port of the microcomputer 24 to the host computer 25 in step 216-1 of FIG. For example, if there are four types of knock intensity, none, small, medium, and large, this can be transmitted to the host computer 25 by a combination of 0 and 1 of two signal lines. The host computer 25 reads this at an appropriate timing, executes advance / retard calculation of the ignition timing based on this information, and finally sends an ignition timing control signal to the igniter 3 to perform knock control.

Next, the fail judgment output of step 216- of FIG. 4, which is the main object of the present invention, will be described in detail with reference to the flowchart of FIG.

As described above, the detected signal of the knock sensor fail is the smoothed value V of the maximum _peak value V _peak in the knock determination section.
_You are using _Faili . This V _Faili is calculated for each cylinder (step 215-3), and this is compared with the fail determination level V _{F in} step 216-2. This is shown in Fig. 9 (b).
The failure determination level V _F changes according to the engine condition (for example, engine speed). The fail judgment level V _F may be changed for each cylinder, but the fail judgment level common to all the cylinders is not a problem in practical use. This fail judgment level V
_F and the detected signal V _Faili are compared for each ignition (step 2)
71). This V _Faili is the fail judgment level V _F
Only when it is smaller and the amplification factor of the cylinder is larger (determined in step 272), the fail counter is incremented by 1 (step 27).
3). Then, V _Faili may ignition cycle has come to exceed V _F, or the case where the cylinder in which the amplification factor is set to a small side have brought this clears the fail counter to 0 (step 274). In this way, when the count value of the fail counter reaches a predetermined value, for example, when the fail counter is incremented continuously for 30 ignitions (determined in step 275), it is determined that a sensor failure has occurred and the host is output from the digital output port. A fail signal is sent to the computer 25 (step 276). Upon receipt of this signal, the host computer 25 side takes safety measures such as making the ignition timing the most retarded.

Although the fail determination level V _F is an engine adaptation constant, a too small value cannot be used due to noise margin or the like. In FIG. 9 (b), this lower limit is set to V
_{When Fmin} is set, the conventional system that detects a failure without switching the amplification factor can detect a failure only in the rotation range of N ₁ or more. In the present invention, however, this is a low rotation (N ₂ ).
Fail detection is possible up to. Therefore, in this embodiment, it is determined in step 270 whether the engine speed is N ₂ or higher, and the rotation range of N ₂ or higher is set as the fail detection execution condition.

Further, as in the above-described embodiment, the fail detection accuracy is further improved by setting the amount related to the maximum peak value in the combustion section (knock determination section) such that the sensor signal becomes relatively large as the fail determination signal. To do.

Although the amplification factor is switched for each cylinder in the above-described embodiment, it may be switched for all cylinders (simultaneously). However, according to the present embodiment, high-performance cylinder-by-cylinder gain switching can be achieved only by the software of the microcomputer, so this embodiment is preferable.

Further, in the above-mentioned embodiment, the amplification factor is switched to two stages, but it is of course possible to switch to three or more stages. At this time, for example, G ₁ , G ₂ ,
If G ₃ ..., Fail detection is performed only for G ₁ and G ₂ , (fail detection is prohibited only for G ₃ ), and the fail determination level V _F itself is switched between G ₁ and G ₂ . It is also possible.

In the above-described embodiment, the outputs of the amplifiers 22 and 23 are directly and simultaneously input ports AN1 and A of the A / D converter 24a.
Although input to N2, it is also possible to selectively input to the A / D converter 24a via the analog switch 26 as shown in FIG. At this time, the switching control of the analog switch 26 can be executed by the digital port of the microcomputer 24.

Further, in the above-described embodiment, the 16 ignition _smoothing value V _Faili of V _peak is used as the fail detection signal.
_Vpeak can also be used directly as the fail detected signal. However, the use of the _smoothed value V _{Faili improves} the accuracy of fail detection, so that the present embodiment is preferable. Also, V _Faili is created for each cylinder,
It is also possible to create this for all cylinders.

Further, in the present embodiment, the fail detection is performed by using the smoothed value of V _peak , but the fail detection may be performed by using V _MEAN . Also, the median distribution of V _peak (cumulative 50%
It is also possible to make a fail determination by using (dot). Distribution median value V ₅₀ of the V _peak is, V ₅₀ = V ₅₀ large if _{V 50 = V 50 + ΔV 50} , if inverse to V _peak is smaller than V ₅₀ than V _peak is V ₅₀ for each ignition cycle -ΔV It can be obtained by updating it sequentially like ₅₀ .

Further, although the amplification factor is switched by the knock determination level V _LEV in the above-described embodiment, various amounts of change depending on the knock sensor output such as V _MEAN can be considered.

Further, in the above-described embodiment, the fail counter is cleared to 0 when switching to the side where the amplification factor is small, and the failure detection is cancelled. However, when the amplification factor is small and small, the count value of the fail counter is held. It is possible to leave it.

〔The invention's effect〕

As described above, the present invention has the following excellent effects.

Since the fail detection is performed when the amplification factor that changes stepwise in a plurality of stages is relatively high, the fail detection accuracy is improved, and the fail detection can be performed even in a lower rotation range and a lower load range.

Since it is not necessary to add a special fail detection amplifier,
There is not much cost increase.

Since the amplification factor is changed according to the output of the sensor, the relative signal ratio can be increased without being influenced by the tolerance of the sensor and the engine, and therefore, both knock detection accuracy and fail detection accuracy can be improved.

[Brief description of drawings]

FIG. 1 is a diagram corresponding to the scope of claims of the present invention, FIG. 2 is a block diagram showing an embodiment of the present invention, and FIG. 3 is a timing chart of arithmetic outputs used for explaining the operation of the apparatus shown in FIG.
Figures 4, 6, 7, 7, and 11 are second
FIG. 5 is a flow chart showing various processing procedures in the microcomputer of the apparatus shown in FIG. 5, FIG. 5 is a diagram showing A / D conversion timing according to the flow chart shown in FIG. 6, and FIG. 8 is for calculating the maximum value according to the flow chart shown in FIG. FIG. 9 is a diagram provided for explanation, FIG. 9 is a characteristic diagram showing the cylinder-specific discrimination output signal and the detected signal for fail determination obtained by the knock sensor output in the apparatus shown in FIG. It is a block diagram which shows the principal part structure of other Example of an invention apparatus. 1 ... Knock sensor as knock detection means, 2 ... Engine control electronic control unit, 3 ... Igniter, 4 ...
... injectors, 22, 23 ... amplifiers constituting amplification means, 24 ... microcomputers.

Claims (2)

[Claims] 1. A knock detecting means for detecting knock occurring in an internal combustion engine, an amplifying means for amplifying an output of the knock detecting means, and a plurality of amplification factors of the amplifying means according to an output level of the amplifying means. In accordance with the discrimination result of this knock discrimination means, and a knock discrimination means for discriminating a knock by the output of the amplification means at the amplification factor determined by the amplification factor variable means, Knock control factor control means for controlling knock control factors such as ignition timing, air-fuel ratio, intake pressure, and the like, fail determination means for generating a fail output when the output of the amplification means is lower than a fail determination level, and the amplification factor varying means. Therefore, there is no fail judgment that substantially invalidates the fail judgment of the fail judgment means when the amplification factor of the amplification means is set to the small side. Knock control system for the internal combustion engine and means. 2. The knock control for an internal combustion engine according to claim 1, wherein a value related to a maximum peak value in a knock determination section of the output from the amplifying means is used as a signal to be determined by the fail determining means. apparatus.