JPH0267658A - Method for diagnosing fault in non-volatile memory - Google Patents

Abstract

PURPOSE:To improve the fault detecting accuracy of a memory by initializing only a controlling storage area when abnormality is generated in the storage format of a memory and holding the stored contents of a fault diagnosing storage area in that condition. CONSTITUTION:In a controller 1, fault diagnosing bits A5 to A7 and B5 to B7 for the memory 10 are respectively arranged in the storage areas A, B of the memory 10. At the time of generating abnormality in the storage format of the memory 10, only the residual control bits A0 to A4 and B0 to B4 and are set up and the contents of fault diagnosing bits A5 to A7 and B5 to B7 are held in their condition. Consequently, history of fault generation such as RAM formation can be left in the memory 10 without troubling the control operation of the controller 1 and a fault in the memory 10 can be rapidly detected.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of the Invention The present invention relates to a method for diagnosing a failure in a nonvolatile memory suitably used in a control device for an internal combustion engine of an automobile.

BACKGROUND OF THE INVENTION An example of the use of non-volatile memory is, for example, a control device for an internal combustion engine of an automobile. This control device is
The fuel injection amount and ignition timing are controlled based on intake pressure, throttle valve opening, internal combustion engine speed, oxygen concentration in exhaust gas, etc. It changes from moment to moment due to factors such as environmental conditions, individual differences in internal combustion engines, and changes over time. For this reason, a non-volatile memory is provided in the control device, and the optimal fuel ↑;) control of injection amount, etc., is stored in the non-volatile memory by learning. written.

A typical prior art storage mode in the memory is 512th.
It is indicated by 1. That is, a store area A for normal storage and a storage area A for inversion storage are provided as a pair, and the detection output of the detector that measures the intake pressure, throttle valve opening, etc. When it is normal, the fifth [2! (
As shown in 1), the space area 21. 0 is stored in mutually corresponding bits of b, for example bit a1 of store area a, and
The opposite "1" is stored in "1".

On the other hand, when an abnormality is found in the output of each of the detectors, the control device performs a control operation using, for example, hardware, without performing calculation operations based on the outputs from the detectors. At this time, in the store area a of the memory, as shown by the 50th (2>), "1" is stored in the bit corresponding to the detector in which the abnormality was discovered, for example, bit ct 2 of the store area a. , "0" is stored in the inverted S[・A visible area SNOPI)-b2. In this way, by using the store W4 area a for directly storing data and the store area S for storing data in an inverted manner, a self-examination function called so-called data storage is realized.

On the other hand, when J% has an abnormality in the memory itself, Figure 5 (
3), the storage areas it
, b, for example, the stored contents of rt_lbl become the same, and at this point, as mentioned above, the arithmetic operation corresponding to the faulty bit 71 is stopped, and the hardware ') control is performed.

When the driver detects such an abnormality through a warning display and repairs are carried out at a repair shop, the memory is disconnected from the backup power source such as the battery, and the 51st
As shown by 1Z(1), all store contents in the storage area Et are initialized to "0", and all store contents in the store area Et are initialized to "1".

That is, as shown in FIG. 6, in step rll, the store contents of the face area a and the store contents of the store area 1-) are stored.
The contents of a are added and assigned to the store domain a. As mentioned above, when there is no abnormality in each detector, the stored contents of the store area a are all "0", and the contents of the storage area a are all "0".

It is. At this time, when the addition result in step n1 is expressed in hexadecimal using the upper 11 bits and lower 4 bits, it is ¥FF (representing hexadecimal number), and in step n2, the addition result is ¥FF. If so, it is assumed that the memory itself is operating normally and the process moves on to the next operation.

Also, in this step rr2, the addition result is ¥F
If not F, that is, if the stored contents of the store area t1 and the inverted value of the stored contents of the store area \FF is assigned to the area S, and the store area εi,b is thus initialized as described above.

Problems to be Solved by the Invention In the prior art as described above, when the stored contents of store area a and the inverted value of the stored contents of store area S are not equal, both store areas a and b are initialized. Therefore, the learned values obtained up to that point are also initialized, and the abnormality in the storage mode of the memory (RAM) is erased by the above-mentioned initialization operation. As a result, it is difficult to discover defects in the memory itself, such as the above-mentioned Fj AM corruption, even during inspection at a repair shop.

An object of the present invention is to provide a storage area for failure diagnosis in addition to a storage area for control [1] in the memory, and to initialize only the storage area for control when an abnormality occurs in the memory f3 of the memory. To provide a method for investigating failures in a non-volatile memory, in which a history of failures of the memory itself can be left by retaining the stored contents of a storage area for failure 3 as they are.

Means for Solving the Problems The present invention includes a control storage area used for control and a fault snooping area used for snooping memory failures, and the control storage area includes: [Control] Store a predetermined logical value when the bidirectional fY is abnormal, compare whether the logical value in the store area for failure corresponds with the preset logical value, and check whether it differs from the preset logical value. The non-volatile memory failure diagnosis method is characterized in that, at times, the logical value of the control storage area is initialized and the logical value of the failure diagnosis storage area is maintained as it is.

Further, in the present invention, when the logical value of the storage area for serious crime diagnosis differs from the predetermined logical value at the time of failure of the memory, a code corresponding to the logical value of the storage area for failure jt is displayed. This is a nonvolatile memory failure diagnosis method characterized by outputting.

According to the present invention, the memory is provided with a control storage area used for control and a failure storage area used to investigate a failure of the memory. A predetermined logic value is stored in the control storage area when an abnormality occurs in the control operation, and therefore, an abnormality such as a wave control object can be detected in accordance with the logic value of the control storage area.

In addition, when the logical value of the required storage area matches the preset logical value for failure 3, the memory is operating normally, and when they do not match, an abnormality has occurred in the storage mode of the memory. As a result, the logical value of the storage area for serious crime control n is initialized, and the logical value of the fault diagnosis storage area is maintained as is. Therefore, even if an abnormality occurs in the storage mode as described above, the logical values in the control storage area are initialized, and this type of storage will not cause any major problems with the control (taku). The abnormality in the aspect is left as it is within the obvious area for failure diagnosis.
Therefore, the history of failures that have occurred in the memory itself is maintained, and the accuracy of memory failure detection can be improved.

Further, according to the present invention, when the logical value of the storage area for failure diagnosis differs from the predetermined logical value when diagnosing a memory failure, when the logical value of the storage area for failure diagnosis is different from the predetermined logical value, the storage area Display output is performed using a code corresponding to the logical value. Therefore, it is easy to see abnormalities such as so-called RAM sink marks that occur in the memory history.

Embodiment 1 [2I is a block diagram showing the electrical configuration of a control device 1 of an internal combustion engine according to an embodiment of the present invention. This control device 1 includes an atmospheric pressure detector 2, a rotation speed Based on the detection results from the detector 3, oxygen concentration detector 4, cooling water temperature detector 5, etc., the processing circuit 6 calculates the optimal fuel injection amount, ignition timing, etc. The processing circuit 6 also responds to the output from the coolant temperature detector 5, etc., and turns on a warning light when the coolant temperature rises abnormally, for example, and alerts the driver. will be notified.

In this processing circuit 6, there is 3g of Beck Arami source from the battery 11! A non-volatile memory 10 is provided which can retain its stored contents even if the power is disconnected, and this memory 10 stores, for example, the optimal air-fuel ratio based on the output from each of the detectors 2-5. The learning control amount of the fuel injection amount, the state of each of the detectors 2 to 5, etc. are stored.

Close to the processing circuit 6, there is also a read-on memory (hereinafter referred to as
(abbreviated as ROM) 12 is provided, and a processing circuit 6
As will be described later, when detecting a location where an abnormality has occurred, code data corresponding to the location is successively output, and the warning light 9 is driven to blink according to the read code, thus notifying the operator of the location where the abnormality has occurred.

FIG. 2 is a diagram showing the manner of storage in the memory 10. FIG. This memory 10 is composed of a storage area A for main storage and a storage area B for inverted storage.
, B's lower bits AO to A4; B OB 4 is used as a control storage area to store abnormalities occurring in each of the detectors 2 to 5. When no abnormality is detected, bits AO-A4 of store area A are reset to "0", as shown in FIG. 2 <1>.
Correspondingly, the inverted rl is set in bit Bo□-B4 of store area B. Each of the bits AO to A4; BO to B4 in such a pair corresponds to each of the detectors 2 to 5, etc., and if an abnormality occurs in the detector, for example, as shown in FIG. 2 (2), When bit AO of store area A is “1”, the
・Bit BO of A visible area B becomes “0”, and the warning light 9
lights up.

In this way, when an abnormality occurs in each of the detectors 2 to 5, a warning light is turned on to notify the driver, and the so-called software control operation by the processing circuit 6 is canceled, and the hardware Switching to the control operation is performed by If such an abnormal condition is displayed, the defective part (V) must be replaced at a repair shop or the like, and power must be supplied from the battery 11! It is reset by disconnecting.

On the other hand, upper bits A5 to A7 of store area A and
The upper bits 85 to B7 of the external area B are used as a storage area for fault diagnosis, and bits A5 to A7 store "1", and bits 85 to B7 store "0". .

Therefore, the storage areas A and B are set to lower bits A+1.
−． A4; BO~B4 and upper bits A5~A7:
When divided into B5 to B7 and expressed in hexadecimal, when the output from each of the detectors 2 to 5 and this memory 10 is normal, store area A represents ¥EO, and store area B represents ¥IF. .

In the control device 1 configured as described above, when an abnormality occurs in each of the detectors 2 to 5, the corresponding bit in the storage area A, for example, as shown in FIG. Bit AO becomes "1" and correspondingly, bit B1 of store area B becomes "0" and warning light 9 lights up. In this way, when an abnormal condition occurs, the processing circuit 6 performs so-called hardware control without depending on the output of the detector.

When an abnormality occurs in this way and the terminal C provided in the processing circuit 6 is grounded at a repair shop, for example, the processing circuit 6 will output the code data corresponding to the location where the abnormality has occurred, for example, ¥11. is successively read from the ROM 12, and a warning light is driven to blink in response to the code thus read out, thereby notifying the fv-company of the location where the abnormality has occurred.

On the other hand, when the stored contents of the 1118 disable bits A5 to A7; B5 to B7 of the storage areas A and B change, that is, as shown in the 21st Elj (3), the so-called RAM When a monster occurs, the warning light 9 lights up, and by grounding the terminal C of the processing circuit 6 as described above, code data that is completely different from that at the time of normal abnormality detection, such as ¥99, is read out from the ROM 12. The warning light will start blinking.

FIG. 3 shows bit AO□ of store area AB of memory 10.
-A4; It is a flowchart for explaining the initialization operation of BO to B4. In step "ri 1, the store contents of store area A and the store contents of store area B are added and input to store area A. Step r"-2
Then, it is determined whether the stored content of the store area A where the above addition result is stored is ¥FF, and if so, that is, the stored content of the store area A is ¥EO, and the stored content of the store area B is When it is ¥IF, that is, when it is in a normal state, it moves to a direct operation.

In step rn2, if the stored contents of store area A are not ¥FF, the process moves to step m3, where the stored contents of store area A are multiplied by ¥EO, and the multiplication result is inputted into store area A again. In this way, without changing the contents of the failure diagnosis store area made up of bits A5 to A7, the stored contents of the control store area made up of bits A O- and A4 are changed to "0".
”. In step m4, the store contents of store area B and ¥IF are summed,
As a result, the stored contents of the control store area made up of bits BO to B4 can be initialized to "1" without changing the stored contents of the fault diagnosis store area made up of bits 85\B7. .

FIG. 4 is a flowchart showing the operation when a failure diagnosis of each of the detectors 2 to 5 is performed in a repair shop or the like, and shows the detection operation for the store area A.

In step [011, it is determined whether the terminal (k) of the processing port n6 is grounded, that is, whether the test mode is selected, and if so, the process proceeds to step rnl2. 0,
1, -, ? ) is initialized to "0" and the process moves to step m13.

In the 17711 port 13, it is determined whether the variable 1&rI exceeds "4", that is, whether it is a storage area for failure diagnosis, and if not, the Stella 11 port 14
Then, it is determined whether the bit A rI is "1" or not. If so, that is, if the detector corresponding to the bit A n is abnormal, the process moves to step "n15. In the 17711 port 15, the ROM 12 From this bit A
The code data of the item corresponding to rI is read out, the warning light 9 is driven to blink as described above, and the process moves to step m16. If the bit A n is not "1" in step m14, the process moves directly to step to 16. Change 1 at step rn l 3! When Lr+ exceeds 4, that is, when the storage area is for fault diagnosis,
In step rn 17 it is determined whether the bit A rI is "0" and if so, the step m
If not, the process moves to step m16.

In step [n16, "1" is added to the variable n and updated, and in step m18, it is determined whether the variable number exceeds 7 or not.
That is, it is determined whether the contents of all bits AO to A7 have been read. If not, the process returns to step m13, and if so, the process moves to other operations. In step m11, if the test mode is not selected, the process moves directly to another operation.

In this way, in the present control device 1, fault diagnosis bits A5 to A7 and B5 to B7 of the memory 10 are provided in the storage areas A and B of the memory 10, respectively, and the memory 10 is
1! In cases where an abnormality has occurred, so-called RAM conversion, the remaining control bits AO~A4: BO~
Reset only B4 and set bits A5 to A7 for the faulty hu.
．． Since the contents of B5 to B7 are retained as they are, they can be stored in the memory 10 in RAM (a history of failures such as sinks) without interfering with the control operations of the control device 1. Therefore, a failure of the memory 10 can be quickly detected.

In addition, if a failure occurs in the memory 10 as described above, the output code stored in the ROM 12
), the warning light is made to blink, which also makes it easier to detect a failure of the memory 10.

In the embodiment described above, the code data read from l(OM l 2) is not displayed by the flashing action of the warning light, but in other embodiments of the present invention, it may be displayed by sound or the like.

Effects of the Invention As described above, according to the present invention, there is a control storage area in the memory used for control, and a failure area 1 used in the event of an nh disconnection of the memory.
When the logical value of the fault diagnosis storage area does not match the preset logical value, it is assumed that an abnormality has occurred in the storage mode of the memory, and the logical value of the control storage area is set. The logic values in the storage area for failure diagnosis are maintained as they are after initialization, so if an abnormality occurs in the storage format due to so-called RAM corruption, the logic values in the storage area for control The values are initialized, and such abnormalities in the storage format can be left in the fault diagnosis storage area without causing any major problems with control, thus preserving the history of faults that have occurred in the memory itself. Therefore, the accuracy of memory failure detection can be improved.

Further, according to the present invention, when the logical value of the storage area for failure diagnosis differs from the predetermined logical value when diagnosing the failure of the memory, the logical value of the storage area for failure diagnosis when diagnosing the memory. Since the display output is performed using a code corresponding to the above, it is possible to easily know the state of an abnormality such as so-called RAM corruption that has occurred during the memory history.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the electrical configuration of a control device 1 for an internal combustion engine according to an embodiment of the present invention, FIG. 2 is a diagram showing how storage areas A and B are stored in the memory 10, and FIG. The figure shows memory 1
0 bit - AO~A = 1. , 70-chart for explaining the initialization operation of B O to B-1, the 40th chart shows the failure detection operation f? of each detector 2 to 5 and the memory 10. FIG. 5 is a diagram showing the storage mode of the storage area εt,1 of the memory of the prior art, and FIG. 6 is a flowchart for explaining the initialization operation of the storage area it,b of the memory of the prior art. This is a flowchart for 1... Control device, 2... Atmospheric pressure detector, 3... Rotation speed detector, 4... Oxygen concentration detector, 5... Cooling water temperature detector, 6... Processing circuit , 7... fuel injection valve,
8... Ignition circuit, 9... Warning light, 10... Memory, 11... Battery, 12... ROM, A, B...
- Store area, AO to A4; BO to B4... Control bits, A5 to A7: B5 to B7... Fault snooping bits

Claims (2)

[Claims] (1) A control storage area used for control and a failure diagnosis storage area used for memory failure diagnosis, the control storage area stores a predetermined logical value in the event of an abnormality in a control operation, and the failure It is compared whether the logical value of the diagnostic storage area matches a preset logical value, and if it is different from the preset logical value, the logical value of the control storage area is initialized and the fault diagnosis is performed. A method for diagnosing a failure of a non-volatile memory, characterized in that logical values in a storage area are maintained as they are. (2) When diagnosing a memory failure, if the logical value of the failure diagnosis storage area differs from the predetermined logical value,
2. The nonvolatile memory failure diagnosis method according to claim 1, wherein the display output is performed using a code corresponding to the logical value of the failure diagnosis storage area.