JPS6053189B2 - Crank angle pulse generation circuit - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a crank angle pulse generation circuit for generating a pulse corresponding to a rotation angle (crank angle) of a crankshaft rotationally driven in an internal combustion engine with respect to a reference point.

A crankshaft that is rotationally driven in an internal combustion engine is set at a certain reference point, such as top dead center (TDC) or bottom dead center (BDC).
) It is important to know at what rotational angle the crank angle is currently located, and various proposals have been made in the past as crank angle detectors.

In particular, in the ignition system in an internal combustion engine, it is required to strictly set the ignition timing of the gas mixture injected into the cylinder to a predetermined crank angle (advance angle). If the crank angle (advance angle) deviates from a predetermined crank angle, various problems will occur, such as hindering purification of exhaust gas, reducing output torque, or excessively increasing fuel consumption. Generally, the above-mentioned predetermined crank angle (advance angle) is not a constant value, but must change according to a certain characteristic, with the engine speed of the internal combustion engine as one parameter, and generally increases or decreases in the engine speed. Accordingly, the predetermined crank angle (advance angle) for ignition must be set larger or smaller, respectively. Conventionally, the crank angle has been set based on a mechanical method such as a governor advance device.
However, with this mechanical method, ignition is performed at different crank angles depending on the engine speed. It was effective only for warning, and it was not required to detect the crank angle itself. However, in recent years, electronic control within internal combustion engines has progressed, and for example, electronic fuel injection (EFI) controls the crank angle at which fuel should be injected and the angle at which it should be injected. The range of rank angles must be set accurately,
Furthermore, with the addition of special equipment including computers, there is a need to selectively ignite at different crank angles only within a certain range of engine speeds. The advent of precision and rank angle detectors has become awaited.The most commonly thought of crank angle detector is a rotating disc that rotates in synchronization with the crankshaft, arranged at equal intervals. This is a method of detecting the current crank angle by adding a large number of optical or electrical detection means and counting the detection pulses obtained from these detection means.

However, in this method, if the crank angle is 10
If the following resolution is to be obtained, the detection means must be 36
This would require more than just a hat, which would impose considerable strain in terms of manufacturing technology and economy. Taking this into consideration further, the number of the above-mentioned detection means is at most several, for example, J2, and between the detection pulses output every 1/2 rotation from each detection means arranged at equal intervals, other oscillations occur. If a prescribed oscillation pulse is accommodated by the means, each oscillation pulse can be made to correspond to each crank angle. Since the number of oscillation pulses can be freely and extremely selected using electrical circuit means, it is easy to obtain a high resolution that far exceeds that of one crank angle. Such a rank angle detector is described, for example, in U.S. Pat. No. 3,832, No. The present invention is constructed based on this kind of idea.

However, this type of crank angle detector can effectively detect the crank angle only while the engine speed remains constant.

Assuming that at a certain engine speed R, a specified number of r oscillation pulses are included between the two detection pulses, each of the r oscillation pulses will be exactly as long as R remains constant. Corresponds to crank angle. However, if the engine is in an acceleration or deceleration state and the engine speed becomes R+ or R-, the time TD between the generation of the two detection pulses will correspond to the rate of acceleration or deceleration. It becomes progressively shorter or longer depending on the situation. Therefore, in the acceleration state, the prescribed r oscillation pulses do not fit within the shortening time TD and overflow occurs, while in the deceleration state, the prescribed r oscillation pulses become longer. The time Tc. Even if it fits inside, there is still more,
cause a shortage. As a result, the crank angle to which the r oscillation pulses should correspond gradually deviates from the original crank angle depending on the degree of overflow or shortage, making it impossible to detect the crank angle with high precision. will occur. Therefore, an object of the present invention is to eliminate the above problems, generate a pulse corresponding to a normal crank angle, and detect the crank angle with high precision regardless of whether the engine is in an acceleration state or a deceleration state. A crank angle pulse generation circuit has been proposed. In accordance with the above object, the present invention synchronizes with the crankshaft and rotates under the force of the crankshaft (N is 1, 2, 3,...
・N detection means that generate a reference detection pulse every time a predetermined constant value) is provided, and the reference detection pulses Pl, P2, . . . P are sequentially sent out from the N detection means.
Any reference detection pulse P1 among O, 1, ), PN, . . . and PN, it has an oscillation circuit for generating L crank angle pulses (L=21, 7, 23, ...) of predetermined constant values corresponding to each crank angle, and further The optional reference detection pulse P(N-1)
If the time required from when PN is detected until the next reference detection pulse PN is detected is T(N-1)1, then the required time T(N-2)1(N -1), the unknown required time TN→(N-1) immediately after that is expressed as T
N3(N+1)0 Liver(N-1)+NT(N-2)3(
Predicted time TN from the formula N-1) →
, N-1), and between the next reference detection pulse PN+, and the reference detection pulse PN, there are immediately L pulses corresponding to each crank angle. The invention is characterized in that the crank angle pulse is included.

The present invention will be explained below with reference to the drawings. FIG. 1 is a diagram showing an example of the reference detection pulse detection means used in the present invention.

In this figure, the detection means 10 is arranged on a disk 11 and on the circumference of the disk 11. The electromagnetic pickup includes a magnetic body (hereinafter abbreviated as magnetic body) 12 that causes the electromagnetic pickup to generate an electromotive force, and an electromagnetic pickup 13 that is fixedly arranged in the vicinity of the disk 11. The disc 11 is interlocked with and rotates in the direction of the arrow in synchronization with a crankshaft (not shown). Therefore, when the upper magnetic body 12 in the figure comes to a position facing the electromagnetic pickup 13, assuming that the piston in the cylinder that is linked to the crankshaft is at top dead center (TDC), this magnetic body 12 is at the lower position, the piston is at bottom dead center (BDC)
), and when it returns to its original position, the piston has moved one stroke. As already mentioned, in an internal combustion engine, normally, when igniting the gas mixture in the cylinder, the ignition timing is set to a predetermined crank angle (advanced angle) O from the top dead center (TDC) of the piston. This crank angle θ is -7
It has a range of about 00 to +200, and must be changed according to certain characteristics with respect to the rotational speed of the engine, that is, the rotational speed of the crankshaft. The degree of this change is extremely subtle, so it is very important to know which crank angle the piston is currently at.
Alternatively, a computer-controlled ignition system requires detection of the current crank angle with high resolution. Therefore, if you want to obtain the crank angle detection pulse P from the electromagnetic pickup 13 with a high resolution of 1 or less, the disc 11
This means that 360 or more magnetic bodies 12 must be arranged at equal intervals on the circumference. However, in this case, disk 11
It is clear that not only will the size be considerably large, but it is not a good idea from the viewpoint of manufacturing technology and cost. Therefore, a separate reference oscillator is provided, and only one magnetic body 12 is provided, and after the first reference detection pulse P'' is obtained from the electromagnetic pickup 13 by the magnetic body 12, the second reference detection pulse P is
If more than 360 oscillation output pulses are output from the oscillator before obtaining ``, then by using only one magnetic body 12, more than 360 detection pulses can be obtained in appearance, and each individual With the detection pulse, a crank angle of 1 or less can be obtained with high resolution.However, since the engine rotation is accompanied by some rotational unevenness during one rotation, N pieces of magnetic material 12 (N = 2, 3 ,...) are arranged at equal intervals, and N reference detection pulses Pl, P2,... are placed during one rotation.
...If PN is obtained, a more accurate crank angle pulse can be obtained through the oscillator. In other words, a prescribed L number of crank angle pulses are obtained each time the crankshaft force is rotated. In one embodiment of the present invention, two
Since two magnetic bodies 12 and 12'' are arranged, the N is 2
Thus, one reference detection pulse P can be obtained every 1/2 revolution of the crankshaft. Further, the number L of the crank angle pulses was set to 23 (=256) in this 1/2 rotation. The reason why the number is a power of 2 is to facilitate the digital processing described later.

In this case, the seven crank angle pulses Pl, P2,...PK...P256 generated in 1/2 rotation are crank angles of 0.7P, 1.4■,...0, respectively. .7
The crank angle is detected with a high resolution of 11 or less, which corresponds to xKOl 79.2 angles. Note that below 0.7 is 180 noh (1/2
rotation) divided by 28 (256). Now, assuming that the engine rotation is constant (constant speed state), the reference detection pulses Pl, P2, Pl″, P2″, P1″,
... is a fixed time interval (
It is outputted from the electromagnetic pickup 13 (Fig. 1) with a cycle) T, and is outputted from the oscillator every 1/2 rotation (Pl
9l9Pl? 29P19k91P19256)9(P2
9l9P2929P29k91OOP29256))(
P'1919P'1929p''1,k,...P''1
, 256)..., these crank angle pulses are equal to the normal crank angle of 0.7
Therefore, 1.4k and 2.1 correspond one-to-one to...

However, when the engine rotation is in an accelerated state, the reference detection pulses Pl, P2, P″1, P″2, P″1, . . .
As shown in FIG. 2 (■), the output is output with gradually decreasing time intervals (periods) T1-, T2-, D1-, D1-, T''''1-, and if the oscillator If the frequency of the oscillation output pulse is constant, for example, the crank angle pulse Pl92549Pl92559P at the end of 1/2 rotation.
l92569 no longer occurs during 1/2 rotation) Pl
9l9Pl929Pl93..., Pl, 253 is the regular crank angle 0.7l, 1.42, 2.1P...1:1
, and the crank angle cannot be detected with high accuracy. Conversely, when the engine rotation is in a deceleration state, the reference detection pulses Pl, P2, P″1, P″2, P″1,
As shown in Fig. 2 (■), ... is output with gradually increasing time intervals (periods) T1+, T2+, D,+, D,+, for example, 2/2 rotation. A space is created between the final crank angle pulse P2, 256 and the crank angle pulse p''1, 1 at the beginning of 3/2 rotation,
Again, these crank angle pulses are equal to the normal crank angle.
There is no longer a one-to-one correspondence, and the crank angle cannot be detected with high accuracy. This is an inconvenient problem caused by keeping the frequency of the oscillation output pulse from the oscillator constant. In order to solve this inconvenient problem, if the frequency of the oscillation output pulse is changed to an apparently higher or lower value depending on the acceleration or deceleration state of the engine, and the period of the crank angle pulse is changed to be shorter or longer, Second
There should be no overlap or gaps in the crank angle pulses shown in the figures (■) or (■), and it should be possible to detect the normal crank angle with high accuracy.

However, this is effective only when the acceleration or deceleration is performed at a constant rate (constant positive or negative acceleration), and in internal combustion engines such as automobiles, the acceleration or deceleration state is Considering that this occurs randomly, it is not possible to detect the rank angle with high accuracy by simply controlling the period of the crank angle pulse as described above. In other words, if the acceleration is switched at random, the overlap or gap of the crank angle pulses will be further expanded at the time of the switch. This is due to the essential reason that future changes in engine rotation that are to occur next cannot be set in the present or in the past. Therefore, the present invention aims to overcome the above-mentioned essential reason and predict the future change in engine speed that will occur next.

That is, the period of the crank angle pulse is determined to be the optimum value in the future based on the predicted future change in engine speed. Here, future changes in engine speed are predicted as follows. Figure 3 (1) is a graph showing an arbitrary example of how the engine rotation speed changes over time, with the elapsed time 1 on the horizontal axis and the engine rotation speed ■ on the vertical axis. To show. As shown in this example, the engine rotational speed (2) changes randomly with any positive, negative, or zero acceleration. It is of course impossible to predict future changes in the engine rotational speed from a macroscopic perspective. However, in Figure 3 (1
) If we look microscopically at the characteristic curve of Figure 3 (■), we can see that no matter which minute part of this characteristic curve is taken, the curve of Figure 3 (■) + α
, -α or C, which is observed to change approximately linearly. However, the horizontal axis of Fig. 3 (■) shows the enlarged elapsed time 1, the vertical axis shows the enlarged engine rotational speed v, and the curves +α, -α and C represent the acceleration/state, respectively. Shows changes in deceleration state and constant speed state. Therefore, Figure 1, Figure 2 (1), (■), (■)
The reference detection pulses Pl, P2, P″1, P
″2, P219be0 period (T9T-19T-29″1
T+19T+2, etc.) also changes almost linearly when viewed microscopically. That is, if we now consider the acceleration state curve +α in FIG. 3 (■), the period of the reference detection pulse changes almost linearly as shown in the curve +αT in FIG. 3 (■). However, the horizontal axis of FIG. 3 (■) is the same as that of FIG. 3 (■), and the vertical axis 4 (taken in the opposite direction for the purpose of explanation) indicates the period T of the reference detection pulse.
In other words, the period T is past (time 1-1), present (time 1)
and increases almost linearly as Ti-1, Ti and Ti+1 along the future (time 1+1). If the curve α in the deceleration state shown in Fig. 3 (■) is obtained, its periods T, -1, Ti, and Ti+1 decrease almost linearly, and the curve C in the constant speed state is obtained. If so, the slope is almost 0
It changes in a straight line. In any case, the period T,−1,T
Since i and Ti+1 change almost linearly when viewed microscopically, the period T of the reference detection pulse that will occur in the future
1+, is predicted by the linear approximation of the following equation. This calculation requires at least two of the past known period T, and the currently known period T. Next, find the future period Ti+1 of the reference detection pulse according to the above formula,
Furthermore, within this determined period Ti+1, there is exactly 28 (=
FIG. 4 shows a basic block diagram of the present invention for generating 256) crank angle pulses at equal intervals.

In this figure, the crank angle pulse generator 40 of the present invention has an input terminal 411n and an output terminal 410ut,
The input terminal 411n receives a reference detection pulse P() which is output from the electromagnetic pickup 13 shown in FIG. Pl, P2, P″1, P″2, P
``1,...) is 1/2, 2/2, 3 of the crankshaft
/2, 4/2, 5/2... are applied every rotation. On the other hand, the output end. From the child 410ut, a crank angle pulse P(Pl, P2...
The prescribed 7 (256) crank angle pulses are generated every four unspeeded rotations of the crankshaft, and therefore the individual crank angle pulses Pl, P2
・・・・・・ is the regular crank angle 0.7 pa, 1.4,・
There is almost a one-to-one correspondence with... First, input terminal 41
The first reference detection pulse P1 applied from 1n is
The binary counter 43 is reset and its contents are cleared through the first and second delay elements 42-1 and 42-2. After this, the binary counter 43 starts counting the clock pulses CLK output from the reference oscillator 44. Subsequently, the second reference detection pulse P2 is applied to the input terminal 411n after a period Ti-1, and when it passes through the first delay element 42-1, it is applied as a clock pulse CLKI to the second latch circuit 45-■, This clock pulse CLKI
The count contents of the binary counter 43 (accumulated count value of the clock pulse CLK) at the time when CLK is applied are taken in parallel into the second latch circuit 45-2 as a binary bit string. Since this capture operation requires several tens of seconds, the second reference detection pulse P2 is delayed by the second delay element 42-■ until the capture operation is completely completed, and the second reference detection pulse P2 is outputted again to the binary counter 43. and clear its contents. Here, the second latch circuit 4
The clock pulse CLK cumulative count value stored as a binary bit string in 5-■ is the period from the generation of the first reference detection pulse P1 to the generation of the second reference detection pulse P2. T, -1 (which will later become the time required for 1/2 rotation in the past).
Subsequently, the third reference detection pulse P3 is generated after a time Ti (
Ti-1) is applied to the input terminal 411n and is applied to the first latch circuit 4 as a clock pulse CLK■.
5-1. At this time, the second latch circuit 45-
■It is stored in. A binary bit string corresponding to period T, -1 is taken in parallel into the first latch circuit 45-1. Therefore, the binary bit string corresponding to the period T, -1 is now stored in the first latch circuit 45-1.
Since this capture operation also requires several tens of seconds, the third reference detection pulse P3 is delayed in the first delay element 42-2 until this capture operation is completed. The delayed reference detection pulse P3 is applied as a clock pulse CLKI to the second latch circuit 45-■,
The count contents of the binary counter 43 at the time when this clock pulse CLKI is applied (clock pulse CLKI)
(accumulated count value of K) is taken into this second latch circuit 45-2 in parallel as a binary bit string. Since this capture operation also requires several tens of r1s, the third reference detection pulse P3 is delayed by the second delay element 42-■ until this capture operation is completely completed, and the binary counter is started again. It is reset, its contents are cleared, and the next clock pulse CLK is counted. Here, the second
The cumulative count value of the clock pulse CLK stored as a binary bit string in the latch circuit 45-■ is calculated from the generation of the second reference detection pulse P2 to the generation of the third reference detection pulse P3. period T, (current 1
/2 revolutions). Thus, at the current end point, the first latch circuit 45-1 has the past required time T,-, and the second latch circuit 45-■ has the current required time T,-. T,
have been stored, so at the start of the next future, the required time T,+1 for the future 1/2 rotation can be calculated using the above formula Ti+1=
While calculating by liver 1-T,-1, the future period T, +
Output terminal 4 outputs crank angle pulses Pl, P2, etc. that accommodate '7 (=256) pulse trains per day.
Must be sent from 10ut.

The calculation of the above equation is performed in the calculation circuit 46, which takes in the value of the past cycle T1-1 from the first latch circuit 45-1 through the route R1, and
2 from the second latch circuit 45-■ through the current period T,
The value of is taken in and the calculation 1-T, 1 is performed. The result of this calculation is the period T,+1 that will occur in the future.
is applied to one input of the comparison circuit 47 through route R3. The role of this comparison circuit 47 is to ensure that exactly 7 (=256) pulse trains fit within the future cycle Ti+1, so that the rank angle pulses Pl9P29 reach OPk0.
OOP 256) and is sent from the output terminal 410ut in synchronization with the clock pulse CLK from the reference oscillator 44. In other words, if the binary bit string from the arithmetic circuit 46 indicating the value of the future period T,+1 is, for example, 51200 in decimal notation, the binary bit string will be displayed on the binary counter 43 in the future period Ti+1. But lol
200 x 1, 200 x 2, 200 x 3 in hexadecimal
Just output a pulse every time 7
(=256) When the pulse is output, its 51200 (
= 200 x 256), and each of these pulses corresponds to a crank angle pulse Pl, P2,...P256, so that even during the future 1/2 revolution, the crank angle will be the same as the normal crank angle. A prescribed 7 (=256) crank angle pulses corresponding to each other on a one-to-one basis are obtained. In the above example, the value of 1 pheasant 200 (which is the value in the above example) means that in the future period Ti+1 the crankshaft will be 1/200.
It remains fixed for two rotations. In this case, it is clear that the value of 200 in w-adic is the value 51200/2B (=256) obtained by dividing the decimal 51200 by 7 (=256). At the beginning of explaining this example, the future period T
Indicates the value of I+1. Although it is assumed that the binary bit string from the arithmetic circuit 46 is 51200 in W-adic, for example, the bit processing of the first and second latch circuits 45-I and 45-■ is actually performed by the upper group of the binary counter 43. It is done in digits, specifically 8+1, 8+2, +3...8+m digits (this 8 is exponent 8 of 7 (=256)), so from the arithmetic circuit 46 The value displayed by the binary bit string is not 51200 in decimal, but it is already 7
The value divided by (=256), that is, 200 in the w-adic above
(51200/23 (=256)). For example, 99999 is 1 Italian (n=1, 2, ・
・・・Dividing by ri means rounding down the lower digits of 99999 from the lowest digit, 999g, 999...
The reason for dividing a binary bit string by 7 is that it is sufficient to truncate only the lowest 8 digits. In this way, the comparator circuit 47 is provided with one input through the route R3 as a bit pattern obtained by dividing the binary bit string that should be present in the predicted future period TO by 7.

Furthermore, the other side of the comparison circuit 47 is a counter (hereinafter abbreviated as R4O) through R4, as shown in detail in FIG. The same bit pattern as this binary bit string (T1+1) is R4. When it appears, one pulse is output from 410ut and R4. Clear. This periodically synchronizes with the clock pulse CLK of the oscillation circuit 44,
= 256) times. If one pulse is sent each time this same bit pattern appears on the counter R4O, these 28 (=256) pulses become individual crank angle pulses. j Now, suppose that the counting of the future period Ti+1 actually starts in the binary counter 43, then the counter R,.

The bit pattern currently given by the arithmetic circuit 46 appears seven times. However, if the counted value of the actually started future period Ti+1 deviates from the predicted value of the period Ti+1, seven crank angle pulses may not be obtained accurately. However, since the value of this predicted period Ti+1 is accurate, deviations hardly occur, and even if deviations occur, the values are so small that they do not pose a problem. A simple illustration of the above-mentioned operation can be expressed in a time chart as shown in Fig.
This is done continuously every two revolutions.

In this figure, first, 2
The count value BC,-1 of the binary counter (Binary COUNTer) 43 is stored as T,-1, and then 1/2
The count value B-Ci counted during the rotation is stored as Ti, and during the next (1+1)/2 rotation, from the already stored Ti-1 and Ti, at 4 in the figure, Ti+
Calculate 1. Since this is a predicted value, it is shown with a bar below Ti+1. The count value B-Ci of the binary counter 43 actually counted during this (1+1)/2 rotation is stored as the actual TI+1, and (1+2)/2
During rotation, the already stored T and Ti+1
Accordingly, in step 4, T,+2 is calculated and the required period for (1+2)/2 rotations is predicted. Repeat the same operation below,
Generates a crank angle pulse that should always occur after 1/2 revolution. FIG. 6 is a block diagram showing an embodiment in which the basic blocks of the present invention shown in FIG. 4 are further embodied.

In this figure, parts corresponding to the basic blocks in FIG. 4 are given the same reference numerals. First, the input terminal 411n is connected to an electromagnetic pickup (
A reference detection pulse P output from 13) in FIG. 1 is applied. The first reference detection pulse P1 becomes CLKO through the first and second delay elements (DEL) 42-1 and 42-2, and is applied to the reset input of the binary counter 43. The binary counter 43 is made up of n+m D's connected in series, and each D is constituted by a flip-flop circuit or the like.

Therefore, D1 is the clock pulse C from the reference oscillation circuit 44.
When the frequency of LK is f, a pulse train Bll with a frequency of f/21 is output as 0101010... The next stage D2 receives a pulse train with a frequency of f/21 and converts the pulse train Bl2 with a frequency of f/7 to 000.
Output as 110011... and do the same below as Dn+
．． -1 is f/2n+m? Pulse train Bn+ with a frequency of 1
．． -1 is sent. Therefore, the clock pulse CLK from the oscillation circuit 44 must be highly accurate, and a crystal oscillation circuit is used in the embodiment. The frequency f may be determined arbitrarily, but it needs to be determined appropriately depending on how many fractions of a revolution the reference detection pulse P is obtained and how high the resolution of the crank angle is selected. In the embodiment, the reference detection pulse was obtained every 1/2 revolution, and the resolution of the crank angle was set to 0.7, so the frequency f was set to WM. In addition, in the case of WM field, the crankshaft is, for example, 6000r.
''.P.m., the number of clock pulses CLK sent out during 1/2 rotation is 5000@, and subsequent digital processing can be performed appropriately.On the other hand, The number of Dl, D2,...Dn+ is n
+m was selected to be 24. This is 10IS-
The number is sufficient to prevent the binary count value from overflowing from the binary counter 43 when the 4 Hz clock pulse CLK is applied to the binary counter 43. Normally, the minimum rotational speed of the crankshaft is about 30r''.Pm (during cranking), so even at such a low speed, the cumulative count value will not overflow the 24-stage binary counter 43. When n+m=24, the minimum allowable rotation speed is (74-1) mowing 00r1S-.1.67S/half rotation, or about 18r'.P.m.As already mentioned, the first criterion is Detection pulse P1 is set as CLKO to all stages Dl, D2...Dn of binary counter 43
+. Since it has been reset, its contents are all 0.

After this, the binary counter 43 starts counting the clock pulses CLK. Subsequently, the second reference detection pulse P2 is applied to the input terminal 411n after a period T, -1,
When passing through the first delay element 42-1, the clock pulse C
It is applied as LKI to the second latch circuit 45-■. The second latch circuit 45-■ includes m D-type flip-flops FFll, FFl2, . . . FFl. ．． When the above-mentioned clock pulse CLKI is applied to each clock input CL, the "゜1゛゜゜0゛ information given to each S input is taken in and sent from the O output and stored. These S inputs The upper digit outputs of the binary counter 43 are applied to the binary counter 43 with the bits aligned.It should be noted here that the information taken into the second latch circuit 45-■ is applied to the binary counter 43's count. The numerical value (in this case, the number of clock pulses CLK counted at period T, -1 during (1-1)/2 rotations) has already been set to 27 (=128).
The value is divided by . Therefore, in this example, D
type flip-flop FFll is D of the binary counter 43.
The output Bl8 of the 8th stage is input, and the outputs of FF12 and below are the outputs of the D9 stage and below, respectively. In other words, n in Figure 6 is 8
Since n+m=24, m is 16. After the second latch circuit 45-2 takes in and stores the information T, -1 by the second reference detection pulse P2, the detection pulse P2 resets the binary counter 43 as CLKO. Similarly, the third reference detection pulse P″1 is C
As LK■, the lower D-type flip-flops F2l, FF22, . . .FF2(.
n-1), the contents of 45-■ of the second latch circuit T, -1
Let's incorporate. Thereafter, P''1 (ie, CLKI) causes the content Ti of the binary counter 43 to be newly taken into the second latch circuit 45-2 through the delay circuit 42-1.
Here, in each stage of the first and second latch circuits, there is a binary counter string (B3l, B32, . . . ) obtained by dividing the period T, -1 at the time of (1-1)/2 rotation by 27. B3(m-
1) and a binary bit string (B2O, B2l, . . . B2-1) obtained by dividing the period Ti during i/2 rotation by 27 is stored. These bit strings B are T,-1/27
3l, B32,...B3 (.n-0 and Ti/27
The bit string B2O, B2l,...B,,, -1)
are the adders Al, A2, and
...A. ．． is applied to, and here the future (1+summer−/
The period T, +1 that should be present during two rotations is also rendered high.
Note that the expression A means the bit inversion of A; for example, if A is "°10r", x becomes "010". any 1
The two adders Ak have a configuration as shown in FIG. 7, and have two inputs as B2(k-1) and B3k, and Ck-1 and Ck for carry from the lower order and carry to the upper order, respectively.
Assuming that the calculation result of k digits is Rk, it exhibits eight functions as shown in the table below. However, in this example, it is Izumi! Therefore, the bit string representing the B2(k−,) input, i.e., T·/2″, must be inverted, while the bit string representing the B3k input, i.e., T,−1/7, must be given the inverse of the inversion. Therefore, as shown in the figure, for the input 8 (k-o (k=1, 2,..., m), these are while input B3k
For (k=1, 2, (m-1)), D-type flip-flops FF2l, FF22, . . . F2, are used. -o outputs O from adders Al, A2, . . . A. ．． Further carry C to the higher order of the least significant digit A1 so that it is applied to
. We decided to always give ゜゜0゛ to . In this way, the adders Al, A2, . . . A. The calculation results Rl, R2, . . . R. ．． has a bit pattern corresponding to Ti+1/23. As already explained, the calculation results Rl, R2, . . . Rn. T appeared in
The bit pattern that is the same as the bit pattern corresponding to i+1/2B is the counter R4. appears 7 times evenly in m digits. Therefore, the comparison circuit 47 (FIG. 4) compares the calculation result Rl
, R2, ... Rm is the first input, and CLK is the second input through the route R, an m-ary re-inputable counter (
R4c) and m from the counter R4C.
It was decided that the outputs of 1000 and 1000 would be input to an m-bit NAND. Rl, R2,...R. From ゜゜1゛、゜゜0
``The output corresponds to r+1/8, so when the clock from route R4 is Ti+, and the pulse corresponding to /8 is input, R4. The outputs Tl, T2, . . . T. ．． Since all output ゜゜1゛ (A + 82゜゜11...1゛),
The m-bit NAND N, AND outputs "0" and inputs Rl, R2, . . . R.n through L again to R4.
At this time, Tl, T2,...T. at least one of
Since it outputs “゜0゛”, m-bit NAND
outputs ゜゜1゛ again. In this way, terminal 410
This pulse '601 sent from ut becomes the seven crank angle pulses Pl, P2, . . . P256 to be stored in the desired future period Ti+1. As explained above, according to the present invention, at most several (one
By simply using the standard detection pulses as an information source, it is possible to generate a specified number of crank angle pulses with ultra-high resolution that is practically impossible with mechanical means. When obtaining the pulse period of the crank angle pulse, which should be changed according to the fluctuation of the rotational speed of the shaft, it is necessary to The period Ti+1 that will be required for the future (1+1)/N rotation is predicted based on the period T, 1, and Ti required for the /N rotation, and the pulse period of the crank angle pulse is calculated according to this prediction result. Therefore, even if the rotation of the crankshaft randomly shifts to a constant speed, acceleration, or deceleration state, a crank angle pulse corresponding to a normal crank angle can always be obtained.

In addition, in the present invention, the calculation for the prediction is based on the future period T.
, +1, but at the same time as the start of this period T,+1, for a short time (about 10011S), so it is possible to use the data just before the period T,+1 for this prediction. The reliability of the prediction is extremely high. Furthermore, in the above explanation, it is assumed that the pulse period of the crank angle pulse generated during each 1/N rotation is a constant value during that i/N rotation, but this does not necessarily have to be a constant value, For example, the oscillation circuit 44 is constituted by a voltage controlled oscillator (VCO), and adders Al, A2, . . . A. ．． The calculation result Rl
,R2,...R. ．． In order to obtain an analog output voltage by D/A converting the voltage, and to use the voltage passed through a CR circuit with a certain time constant, the pulse period gradually changes even during the i/N rotation, and the total pulse It is also possible to obtain a constant number of crank angle pulses.

[Brief Description of the Drawings] Fig. 1 is a diagram showing an example of the reference detection pulse detection means used in the present invention, Fig. 2 (1), Fig. 2 (■), and Fig. 2 (■) are the reference detection pulses. Time chart showing the relationship between detection pulse and crank angle pulse, Figure 3 (1), Figure 3 (■
) and Fig. 3 (■) are graphs for explaining the basic principle of the present invention, Fig. 4 is a block diagram showing the basic configuration of the present invention, and Fig. 5 is a timing chart for explaining the basic operation of the present invention. 6 is a block diagram showing an embodiment based on the present invention, and FIG. 7 is a diagram showing the configuration of an adder used in an embodiment of the present invention.

Claims (1)

[Claims] 1. A crank angle pulse representing the rotation angle of a crankshaft rotationally driven in an internal combustion engine with respect to a reference point, and which is a crank angle pulse per revolution of 1/N (N is a predetermined constant positive integer). The number of occurrences is always 2 regardless of rotational fluctuations of the crankshaft.
A circuit that generates ^n (n is a predetermined positive integer) rank angle pulses, and calculates the next crank angle by sequentially calculating the acceleration of rotation of the crankshaft. A crank angle pulse generation circuit that predicts a generation period of an angular pulse and sequentially sends out the crank angle pulse having the predicted generation period,
a reference oscillator that outputs a clock pulse that serves as a time reference for the crank angle pulse to be generated, and a reference detection pulse (P_i_-_1 , P_i・P_i
__+_1, P_i_+_2...
a second delay element that provides a further fixed delay to the reference detection pulses successively output from the delay element; and an output from the reference oscillator each time the reference detection pulse is sequentially reset by the output from the second delay element. a binary counter that starts counting the clock pulses to be detected; a latch circuit that captures the count contents of the secondary counter at the time when the reference detection pulse is output from the first delay element; Another latch circuit that captures and temporarily holds the temporarily held count contents at the time when the reference detection pulse is input to the first delay element, wherein the reference detection pulse P_i_-_1 is Period T from when the reference detection pulse P_i is output until the reference detection pulse P_i is output
A first latch circuit that temporarily holds the count contents corresponding to _i_−_1 and a first latch circuit that is cascaded thereto and corresponds to a period T_i from outputting the reference detection pulse P_i to outputting the reference detection pulse P_i_+_1. a second latch circuit that temporarily holds the count contents; and the period T_i held in the first latch circuit.
____1 and the period T held in the second latch circuit
The calculation 2T_i-T_i_-_1 is performed with each input as _i to calculate the period T_i_+_1 from the output of the reference detection pulse P_i_+_1 to the output of the reference detection pulse P_i_+_2, and represent the value of the period T_i_+_1. an arithmetic circuit that outputs a bit pattern obtained by dividing a binary bit string by the 2^n; one side inputting the bit pattern from the arithmetic circuit, and the other inputting the clock pulse from the reference oscillator; A crank angle pulse generation circuit comprising: a comparison circuit that sequentially generates crank angle pulses having a period such that the 2^n pulse train falls within a period T_i_+_1.