JP2751488B2 - Pulse input processing method - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a pulse input processing method for processing a pulse signal having a repetition frequency related to a vehicle speed, a rotation speed, and the like.

[Conventional technology]

As a pulse input processing method for detecting a vehicle speed or the like, for example, a rotational speed detecting device for an internal combustion engine disclosed in Japanese Patent Application Laid-Open No. 49-27722 and an automatic engine speed controlling device disclosed in Japanese Patent Publication No. 61-20274 are disclosed. and so on. In these methods, a pulse interval is calculated with a constant clock pulse, and input processing is performed.

By the way, reed switches are mainly used in vehicle speed sensors and the like.
An illegal signal (hereinafter referred to as chattering) occurs. Such chattering cannot be sufficiently removed (chattering removal) depending on the hardware (electric circuit). Therefore, a chattering signal is generated for both edges of the pulse signal, and a malfunction occurs when the chattering signal enters the edge input port of the microprocessor.

That is, as shown in FIG. 14 and FIG.
A magnet 91 whose N and S poles are magnetized in the circumferential direction is rotated by an automobile speed cable according to the vehicle speed, and the on / off state of a reed switch 92 close to the magnet 91 is obtained as a pulse signal 94. . Then, the sensor using the reed switch 92 generates a chattering waveform as shown by the SPD signal in FIG. 15 when the on / off signal is switched with the rotation of the magnet 91.

In order to prevent this, a chattering removing circuit shown in FIG. 14 is provided. According to the circuit, the resistors R1 and R2 are R
1 << R2 is set, and the threshold level voltage of the transistor Tr is "5V-about 0.6V".
Chattering of the hardware circuit can be removed by the time constant of the resistor R2 and the capacitor C1.

[Problem to be solved]

However, in the chattering removal by the above hardware, at the time of chattering, all are determined to be “L” level (signal 93). Also, considering the on / off time ratio of the vehicle speed sensor, the "H" level may disappear at high speed. Therefore, chattering by hardware is not enough.

In view of the above, Japanese Patent Publication No. 41944/1994 proposes a pulse input processing method for removing the chattering. In this method, when the level of a signal corresponding to an edge input within an interrupt is the same for a certain period of time or more, the input is regarded as correct and the chattering removal processing is performed.
Therefore, there is a disadvantage that other processing cannot be performed while monitoring the level change for a certain period of time.

As a general pulse input process, an interrupt is generated in an edge direction where chattering of the sensor is likely to occur, and chattering is removed. The processing shown in FIG. 16 is performed. In this method, a rising edge interrupt of SPD (vehicle speed) pulse input is performed in step 9-0, a rising time interval is calculated in 9-1, and 1 ms is calculated in 9-2.
The re-interruption within is ignored for chattering. Then, in step 9-3, the SPD calculation is performed based on the time interval, and the current time is stored for the next calculation.

And for the time chart of this waveform,
As shown in the figure, first, the SPD input is rising edge 11, 12, 1
3, 14, between which there are falling edges 31, 32, 33.
In the same figure, between the rising edge 12 and the falling edge 32, chattering edges 310 and 120 between the falling edge and the rising edge occur. Further, between the falling edge 33 and the rising edge 14, chattering 130 and 330 between the rising edge and the falling edge occur.

Therefore, as described above, the chattering time T1 (1m
When s) is provided, the rise time interval becomes as shown in FIG. That is, the chattering 310, 120 in T1 after the rising edge 12 is interrupted, whereas the chattering 130, 330 after the falling edge 33 is calculated at a time interval based on the rising edge 130. Therefore, the former case is normal, while the latter case is erroneously determined to be high speed.

That is, in the above method, the rising edges 11, 12, 13, 14
Although the chattering input on the side can be removed, the chattering input on the falling edges 31, 32, and 33 cannot be removed, resulting in the above erroneous determination.

To cope with the above problem, a pulse input processing method shown in FIGS. 18 and 19 can be considered.

That is, in FIG. 18, the aforementioned steps 9-0 to 9-
In step 9-7 after step 2, it is determined whether or not the edge is a rising edge. If the edge is a rising edge, in step 9-8, the RAM contents stored in the previous interrupt (falling) 9-10 (time from the previous rising edge to the falling edge time) ) And the time obtained by adding the difference between the previous interrupt time (falling edge) and the current interrupt time (interval between the last rise and the current rise time)
Performs SPD operation from. In 9-9, the current time is stored for time calculation and the next interrupt edge switching is performed.
If there is no rising edge in 9-7, the time interval from the previous interrupt (rising) to the current interrupt (falling) is stored in 9-10, and the process proceeds to 9-9.

Then, in the above method, the time chart shown in FIG. 19 is obtained. That is, in this method, the rising edge
Interrupt generation chattering is performed for both edges 11, 12, 13, 14 and the falling edges 31, 32, 33, and the SPD calculation is performed based on the total time interval of the rise-to-fall time 20 and the fall-to-rise time 25. Done.

However, this method also has the following problem. That is, S
PD signal on / off duty is not 50%, 0.8 ≦ ON
/OFF≦3.5. Therefore, at high speeds including hard delay, the on-time or off-time may be shorter than the chattering elimination time (1 ms).

Therefore, as shown in FIG. 20, when the "H" level (on time) 20 between the rising edge 11 and the falling edge 31 is shorter than the chattering elimination time T1, this time T1
If the chattering edges 310 and 120 occur during the period, it is erroneously determined that the speed is low as shown in FIG.

As shown in FIG. 21, the “L” level (off time) 25 between the falling edge 31 and the rising edge 12 is
If the chattering edges 310 and 120 occur during the time T1 when the chattering time T1 is smaller than the chattering removal time T1,
It is erroneously determined to be low speed.

If the “L” or “H” level time is shorter than the chattering time, an erroneous judgment is made as described in the prior art.
Can also be said.

SUMMARY OF THE INVENTION In view of the conventional problems, the present invention provides a pulse input processing method capable of ignoring a pulse input time interval during chattering, obtaining an accurate pulse input time interval, and performing an accurate calculation on vehicle speed and the like. It is something to offer.

[Solutions to solve the problem]

The present invention relates to a pulse input processing method for processing a pulse signal of a repetition frequency sent from a sensor, wherein a first interrupt mode in which an interrupt is generated at a rising edge of the pulse signal and a second interrupt mode in which an interrupt is generated at a falling edge of the pulse signal In the first interrupt mode, the previous first interrupt and the current first interrupt
The first time interval between the interruption is calculated as the input pulse period, and the first time interval calculated in the first interrupt mode within the second fixed time from the second interruption is not set as the input pulse period. , A first time interval calculated in the first interrupt mode after the second fixed time is added to the first time interval to obtain an input pulse period.

The most remarkable point in the present invention is that a chattering process is performed on the falling edge to accurately calculate the input pulse period in consideration of the input time interval of the chattering width. In addition, the rising edge (first interrupt mode) And the falling edge (second interrupt mode) has different input edge timing, thereby distributing the interrupt load.

Further, in the present invention, the first interrupt generated within the first fixed time from the first interrupt mode can be ignored.

Further, the arithmetic processing using the input pulse period can be performed in the second interrupt mode.

Also, comparing the present invention with Japanese Patent Publication No. 41944/1994, the former always checks whether or not the level of the port corresponding to the edge within the interrupt is correct, and corrects the level for more than a fixed time. Occasionally, a normal interrupt is determined and processed. In this case, the last interrupt time in the chattering is the calculation time.

On the other hand, according to the present invention, a normal interrupt is determined when a previous same-direction edge interrupt has passed a predetermined time or more with respect to the current interrupt. That is, the first interruption time in the chattering is the calculation time.

As the chattering removal time for the same-direction edge input, a value close to a time equal to or less than the minimum time interval of the pulse input frequency range is adopted on the first interrupt mode side for calculating the pulse time interval. On the other hand, on the side of the second interrupt mode in which the calculation is performed based on the time interval, a value close to the chattering time that is longer than the chattering time that can occur from the sensor used is adopted. Both chattering times may be the same or different.

In the present invention, the above technical terms mean the following.

 First interrupt; interrupt at rising edge.

 Second interrupt; interrupt that occurs at the falling edge.

1st interrupt mode: the processing state performed by the 1st interrupt, here, the chattering occurring on the rising edge side and the time interval between the rising edges (ignoring chattering if present) The calculation and the calculation of the pulse period using the calculated time interval are performed.

In the calculation of the pulse period, when chattering is performed in the second interrupt mode, the time interval calculated this time in the next first interrupt mode and the time interval calculated next time are added, and the pulse period is calculated. I do.

When chattering is not performed in the second interrupt mode, the time interval calculated in the next first interrupt mode is set as the pulse cycle.

The second interrupt mode; the processing state performed by the second interrupt, in this case, the chattering removal occurring on the falling edge side and the first state (when chattering removal is not performed)
The calculation of the vehicle speed or the like is performed using the pulse cycle calculated in the interrupt mode.

 First fixed time; T1 (see FIG. 2).

 Second fixed time; T2 (see FIG. 2).

(Operation)

In the present invention, as shown in the block diagram of FIG. 3, the pulse signal sent from the sensor ignores the first interrupt generated within the first fixed time from the first interrupt mode. In the first interrupt mode, the previous first interrupt and the current
A first time interval between the interruption and the interruption is calculated to be an input pulse period.

However, the first time interval calculated in the first interrupt mode within the second fixed time from the second interrupt is not set as the input pulse period, and the first time interval generated after the second fixed time elapses in the first time interval. The first time interval calculated in the embedded mode is added to obtain an input pulse period.

According to the second aspect of the present invention, the first interrupt generated within the first fixed time from the first interrupt mode is ignored.

According to the third aspect of the present invention, the arithmetic processing using the input pulse period is performed in the second interrupt mode.

〔Example〕

First Embodiment Regarding a pulse input processing method according to an embodiment of the present invention,
This will be described with reference to FIGS.

In the method of this embodiment, as shown in the block diagram of FIG. 3, a pulse signal generated from the sensor 40 is processed in a first interrupt mode 41 and a second interrupt mode 42.

That is, in the first interrupt mode 41, the rising edge detection 411 of the input pulse signal is performed, then the chattering 412 of the time T1 is performed, and the time interval calculation 413 is performed. On the other hand, in the second interrupt mode 42, the falling edge detection 421 of the input pulse signal is performed, then the chattering 422 of the time T2 is performed, and the time interval calculated in the first interrupt mode 41 is input. The calculation 423 of the vehicle speed and the like is performed. These processes are performed using a microcomputer.

Next, a flowchart of the above process will be described with reference to FIG. If there is an interrupt input, it is determined in step 1-1 whether or not the edge input is a rising edge. If the edge input is a rising edge, it is determined in 1-2 whether the port level is "H" (high level, reference numeral 20 in FIG. 2). Determine whether or not. If it is “H” level, 1−
In 3, the value obtained by subtracting ZTiNH (rising interrupt time) from TiN (current interrupt occurrence time) is set to D (rising interrupt time interval).

Then, in 1-4, it is determined whether or not D is smaller than the rising chattering time T1 in the first interrupt mode. If D <T1, chattering is determined to be 1-9.
To switch the input interrupt edge. On the other hand,
When D> T1 in step 4, the TiN is set to ZTiNH in step 1-5, and the operation flag set is checked in step 1-6. If NO, that is, if the operation was not performed by the previous interrupt (falling edge) (standby) If it is the second or subsequent rising edge interrupt of falling chattering, the rising pulse time interval T ₀ (addition of D to the previous time T ₀ ) is calculated in 1-8, and the process proceeds to 1-9. Further, when the flag set at 1-6, the D as the time interval T ₀ at 1-7, the process proceeds to the 1-9 clears the operation flag.

On the other hand, when it is determined in the above 1-1 that the edge is not a rising edge but a falling edge, 1-10
It is determined whether the port level is at "L" (low level, reference numeral 25 in FIG. 2). If “L”, at 1-11
D is the value obtained by subtracting ZTiNL (falling interrupt time) from TiN, and ZTiNL is TiN.

Then, in 1-12, it is determined whether or not D is shorter than the chattering time T2 in the second interrupt mode. If D <T2, the process proceeds to 1-9. On the other hand, when D <T2, 1-13
In step 1-8 or 1- of the first interrupt mode,
Than the time interval T ₀ calculated at 7 performs calculation of vehicle speed, 1-
At 14, the operation flag is set, and the routine proceeds to 1-9.

Next, FIG. 2 shows a time chart obtained based on the above flowchart. As shown in the figure, for the rising and falling input signals, the time T1, T
The chattering of 2 is performed, and a time interval T ₀ based on TiN−ZTiNH is calculated and the calculation is performed.

For example, between the rising edge 12 and the falling edge 32, chattering of the edges 310 and 120 occurs.
However, the D at the rising edge 120 is shorter than the chattering time T1. Therefore, this rising edge 12
Time intervals are calculated ignoring 0. Regarding the falling edge 310, since TiN−ZTiNL (falling 31) is larger than T2, the calculation is performed irrespective of chattering. At this time, the calculated time interval is the rising edge 12
Is the time interval from 11 to 12 obtained in.

Further, between the edges 32 and 13, chattering of the edges 121 and 320 occurs after the falling edge 32. In this case, at edge 121, the chattering removal time from 12
Since the operation flag after a and calculating T1 or later is set, T ₀ is set the time from 12 to 121. Edge 32
0 is within the chattering removal time T2 from 32 and is ignored. In the next edge 13, because after the last T ₀ set operations has not been made, the operation flag is cleared, thus T ₀ is T ₀ (12
12 to 13 which is the time of 121 to 13)
Is calculated at the next fall 330, 33 as the time interval of.

Next, FIG. 4 to FIG. 6 show that the “H” level 20 between the rising edge 11 and the falling edge 31
When the former is smaller than the latter and the time of the former is smaller than the chattering time T1 (FIG. 5), the former is smaller than the latter, as shown in FIG. Is larger and the latter time is chattering time
The case where it is smaller than T2 (FIG. 6) is shown.

And, as is known from the time charts in the above figures, in any case, regardless of the occurrence of chattering,
After confirming the correct pulse input time interval, the calculation based on the time interval can be performed. In the prior art, such processing cannot be performed, and an erroneous determination has occurred (see FIGS. 19 to 21).

Second Embodiment Next, a specific example will be described with reference to FIGS.

7 and 8 are flowcharts showing a specific embodiment of the present invention.

Fig. 7 shows SPD (vehicle speed) interrupt, Fig. 8 shows 4.096ms
This is SPD processing with a periodic interrupt that occurs every time. For example, if the SPD signal is generated at 637 rpm × 4 pulses at a vehicle speed of 60 km / h and the LSB of the interrupt occurrence time is calculated as 4 μS and the SPD time interval is calculated as 8 bits and 2 bytes, the maximum time interval width is 0.26214 S
(4 [mu] S × 2 ^16), also the vehicle speed at this time (minimum detection speed)
Is about 5.4km / h [(1 / 0.2621) x (60 x 60/637 x 4)]
It is.

The CSPDi (4.096 ms counter) 1 sets the vehicle speed to 0 km / h when the SPD signal is not 0.7 S or more (vehicle speed <2 km / h) and detects the vehicle speed up to 2 km / h below the minimum detected vehicle speed by the interrupt. , 5-7, 5-9. In addition, the smoothing process of averaging the previous SPD calculation time interval and the current SPD calculation time interval is performed in steps 5-8, 5-9, 5-10, 5-18, and 5-
It is configured to be performed in 19,5-20.

Step (hereinafter abbreviated) 5-1 is interrupt edge judgment, 5-
Switching of next interrupt edge in 3,5−14, 5−4, 5−15
Performs chattering at the same direction interrupt edge. In addition, the rising edge of 5-8, or 5-9 and 5-20
The SPD calculation time interval is calculated, and the SPD is calculated at 5-21 which is the timing of the falling edge.

 Next, the operation will be described.

9 to 12 are time charts corresponding to FIGS. 7 and 8 for specific embodiments for various SPD signals.

First, in the regular interrupt processing of 4 ms (4.096 ms) when there is no SPD signal input, CSPDi is incremented by 1 in 6-1 in FIG.
In 2, it is determined that CSPDi> 172 (SPD pulse interval> 0.7s, SPD <2km / h), SPD = 0km / h in 6-3, and SPD operation time interval TSPD
Is set to the time measured by CSPDi1.

Here, it is assumed that CSPDi1 incremented by 1 is guarded so as not to overflow. At this time, the SPD interrupt edge is set to rise.

Next, when the SPD signal comes in, 5-1 in FIG.
After the level judgment [whether the port level is H (high side) or L (low side)] corresponding to the interrupt edge of 5-2 and 5-13, the time is judged at 5-3 and 5-14. Switching the interrupt edge. Reference numerals 5-4 and 5-15 denote chattering removal processing, which is the rising side which is the time interval measurement side (11, 12, 13, 1 in FIG. 9).
In 4), the first interrupt time Z after the previous chattering
From the SPDH, when the difference between the next rising interrupt time TiN and the rising chattering time T1 is shorter (12 to 120 in FIG. 9), it is ignored that there is no interrupt.

On the other hand, the falling side of the SPD calculation side (31, 32, 33, 34 in FIG. 9)
In this example, when the difference between the falling interrupt occurrence time TiN and the falling chattering time T2 is smaller than the falling chattering time T2 from the time ZSPDL for each interrupt occurrence (33 to 330 in FIG. 9), it is ignored that there is no interrupt. . The reason for this will be described later.

In addition, the vehicle speed is determined to be 5-6 <2 km / h in FIG. More specifically, it is determined by the counter CSPDi1 which is incremented by 1 at 6-1 in FIG. 8 of the 4.096 ms interrupt process, which is a regular interrupt, and is cleared (5-12) at the rising edge of the SPD interrupt in FIG. Also, the SPD <5.7km / h determination in 5-7 is performed using the same counter, and the SPD operation time interval TSPD processing (5-8,5
Perform -9).

In the TSPD processes 5-8 and 5-9, the rise time interval measured this time is added to the SPD operation time TSPD one after another while adjusting the LSB (measurement resolution). TSPD, LSB = 32μS, TiN, ZSPD
H, LSB = 4 μS, CSPDi1, LSB = 4.096 μS, and the ZSPDH used here is ZSP before TiN data is changed to ZSPDH in 5-5.
DH data. Also, at the time of the SPD calculation after the addition, a smoothing processing request flag that should be halved is set.

When it is determined that chattering occurs on the falling edge according to 5-15, the rise time included in the chattering is added to the TSPD, but after chattering is removed, the SPD operation is performed except at the first falling edge interrupt. Is not performed. In 5-18, 5
According to the TSPD smoothing request flag set in -10, TSPD is set to 1
/ 2, and after clearing the flag, calculate the SPD based on the rise time interval TSPD. In 5-18, when the smoothing request flag is not set, the next process may be performed without performing the SPD operation.

9 to 11 show time charts for various SPD signals.

First, FIG. 9 will be described. Rising signal (11)
Then, a new time interval is added to the TSPD, and when the smoothing request flag is set, the SPD is calculated after halving the TSPD at the falling edge (31) of the signal.

When chattering (320,120) occurs at the rise, TS
After adding the rise time interval to PD this time, fall (32
In 0), TSPD is halved and SPD is calculated. Subsequent interruptions in chattering are ignored. In the next normal falling (32) signal, the smoothing request flag is cleared, so that TSPD is not halved.

Next, if chattering occurs at the fall 33,
At the first fall (33), the already added TSPD is halved.
After annealing, SPD calculation is performed. And rise (130)
Approximately 1/2 of the time interval of the original rise (13-14) in TSPD
Is added, and subsequent interrupts in chattering are ignored. At the next regular rise (14), the original rise time interval is reached.
D is calculated.

FIG. 10 is a time chart when the "H" level of the SPD input signal is less than the chattering removal time T1, and FIG. 11 is a time chart when the "L" level is less than the chattering removal time T2.

In FIG. 11, when chattering occurs at the falling edge (33), the first rising signal (14
Since 0) is valid as the rise time interval and the other is ignored as chattering, the SPD is calculated temporarily at short time intervals.

However, the next time interval is measured to be longer by that amount, and "H" level time >>"L" level time is assumed, and there is no problem with the smoothing process.

In addition, the SPD signal originally has four pulses per rotation of the crankshaft, and the same is a reed switch, and the same N or S pole magnet is provided every fourth. For this reason, the time interval requires a smoothing process or a calculation based on a 4-pulse time interval.

Further, as shown in FIGS. 14 and 15, the chattering removal by the hardware becomes “L” during the chattering period.
With such a configuration, it is possible to use a signal in which the "L" level is secured, instead of the SPD input as shown in FIG.

Here, according to FIG. 12, at the time of the falling SPD calculation,
Set the chattering removal time ZSPDL to fall interrupt (6
1, 62, 63,...) Will be described.

That is, if chattering is performed based on the first time as in the case of rising (51, 52 ...) as described above, an erroneous determination is made when an SPD signal as shown in FIG. There is.

However, in the present invention, as shown in FIGS. 9 to 11, the chattering removal time T1 when chattering occurs.
Is always based on the rise, and the fall is later than that, so there is no misjudgment as shown in FIG. Also,
If the "L" level by the hardware processing (FIGS. 14 and 15) can secure the falling chattering taking time T2 or more, the falling chattering time is necessarily updated every time an interrupt occurs. No need.

Third Embodiment This embodiment will be described with reference to FIG.

In the second embodiment, a smoothing process is performed in which the current measurement time interval is added to the previous SPD calculation time interval and halved.

However, in the method of calculating the SPD from each time interval in FIG. 1, the calculation in 1-13 can be replaced as shown in FIG.

That is, SPD calculated at time intervals of every four pulses in 11-1 to 11-5 using four pulses of one rotation of the crankshaft.
And SPDs 0 to 3 at 11-6 to 11-9, respectively.
At -10, the average value may be used as SPD.

In this case, the time interval may be set to TSPD0 to TSPD3, and the SPD may be calculated from the average time interval.

Next, in the above embodiment, the SPD calculation is performed in the interrupt processing. However, in order to shorten the interrupt processing time, only the pulse time interval measurement is performed in the interrupt, and the SPD calculation request flag is set in the main routine. , SPD may be calculated from the time interval.

In the above embodiment, the vehicle speed sensor is mainly described. However, the present invention can be applied to a sensor input including the same type of chattering signal such as various kinds of rotation speed measurement.

Furthermore, even if there is no chattering of the sensor itself, it can also be used as a pulse input processing circuit for an illegal signal after a normal signal level change due to fluctuations in power supply, noise above hysteresis near the threshold, etc. .

[Brief description of the drawings]

1 to 6 show a first embodiment, FIG. 1 is a flowchart thereof, FIG. 2 is a time table, FIG. 3 is a block diagram, and FIGS. 4 to 6 are diagrams of H level and L level. FIGS. 7 to 12 show the second embodiment, in which the ratio is different, FIGS. 7 and 8 are flowcharts thereof, FIGS. 9 to 11 are various time tables, and FIG. 13 is a timetable for comparative explanation, FIG. 13 is a flowchart of the third embodiment, FIGS. 14 and 15 are circuit diagrams and timetables for chattering removal by conventional hardware, and FIGS.
The drawings are a flowchart and a time table of chattering removal by conventional software, and FIGS. 18 to 21 are a flowchart and a time table of chattering removal by other conventional software.

Claims (3)

(57) [Claims] 1. A pulse input processing method for processing a pulse signal of a repetition frequency sent from a sensor, comprising: a first interrupt mode in which an interrupt is generated at a rising edge of the pulse signal; and a second interrupt mode in which an interrupt is generated at a falling edge of the pulse signal. In the first interrupt mode, a first time interval between the previous first interrupt and the present first interrupt is calculated and used as an input pulse period. The first time interval calculated in the first interrupt mode within the fixed time is not set to the input pulse period, but the first time interval calculated in the first interrupt mode after the second fixed time in the first time interval. A pulse input processing method comprising adding an interval to an input pulse period. 2. The pulse input processing method according to claim 1, wherein a first interrupt generated within a first fixed time from the first interrupt mode is ignored. 3. The pulse input processing method according to claim 1, wherein the arithmetic processing using the input pulse cycle is performed in the second interrupt mode.