DE2360212A1 - METHOD OF CONTROLLING AN COMBUSTION MACHINE - Google Patents

Description

DB. EBICH NEW BUILD "

ok'A / Srt ^ -i '*, ⁸ MUNICH

PATENT ADVERTISER 2360212 TEt-BFON (OS

8 MUNICH 2β - POST BOX 31 TEI.EGBAMSLÄ.BBES8E:

ZWEIBBÜCKENSTBASSE 10 % ££ IAPATENT MUNICH

(KEBEJT SEX DETJT S CHHJi FATEKTAXT)

3rd December 1973

1 Ä - 3255

SOCIETE DES PROCEDES MODERNES D'INJECTION SOPROMI

CLICHY / FRANCE

Method for controlling an internal combustion engine

The invention relates to a method for controlling an internal combustion engine and in particular a motor vehicle engine .

It is well known that one can get the handiness of steering that Safety, specifications and the elimination of vehicle pollution from use of electronic circuit arrangements can improve. These circuit arrangements are intended to control be it to control the function of certain elements or to inform the driver of the current one State of the function of this or that element. It is clear that such circuit arrangements in general in the field of information technology or by Use techniques known to industrial process computers.

In vehicles equipped with an electronic control system, the "computer" receives information in

409824/0329

various forms of transducers that constantly use the Determine functional conditions of the vehicle or this or that part of the vehicle. If you can with this information wants to create a complex function that depends on these variables to control elements or just size To make the function audible or visible to the driver, it is often advantageous to use numerical techniques, which are now sufficiently easily accessible due to the presence of numerous integrated circuits.

The invention is based on the object of a simple system for generating information in numerical form, i.e. in coded form supplied by the transducers for conversion and use of the coded ones To create information.

This object is achieved according to the invention in that the information that are supplied with each sampling of transducers for the measurement of physical quantities that are related to the function of the motor are related, each in the form of a number or a word segment comprising a limited number of digits, such are coded that the mean value over several scans of the word segment is characteristic of the mean value of the corresponding physical quantity is that the word segments are grouped to form a first complete word that the first word is converted from a programmed memory into a second word that the second word is converted by a conversion circuit with a function whose mean value is characteristic of the Is the mean value of the regulatory function sought, and that this regulating function is converted into an engine control variable.

The invention is explained below with reference to FIGS. 1 to 19, for example. It shows:

409824/0329

Figure 1 is a circuit diagram of a circuit for generating the sampling frequency.

Figure 2 is a circuit diagram of a circle for coding a frequency measurement with interpolation and zero Shift, where the sampling frequency is significantly lower than the measured frequency,

FIG. 3 shows a circuit diagram of a circuit for coding a frequency measurement with elimination of the stroboscopic Effect, the sampling frequency being of the same order of magnitude as the measured frequency is,

FIG. 4 shows a circuit diagram of a circle for measuring the frequency with the restraint of the first numbers,

FIG. 5 is a circuit diagram of a circuit for generating a sawtooth signal and its use on one Comparator,

Figure 6 is a circuit diagram of a circuit for coding a voltage,

FIG. 7 is a circuit diagram of a circuit for generating a linear function,

FIG. 8 shows a group of circles for generating a linear one Function from coded information,

FIG. 9 shows a circle for generating the mean value of a linear one Function in numerical form, '

FIG. 10 shows a circle for converting a voltage into a. Duration,

409824/0329

FIG. 11 shows the design of a speed measuring transducer (lla), the arrangement of the measuring transducer (Hb) and the circuit diagram of a circuit with this transducer for generating a signal with double frequency,

FIG. 12 is a circuit diagram of a circuit for generating two signals in the special case of controlling the Ignition of an engine,

FIG. 13 is a circuit diagram of a complete circuit for carrying out the method according to the invention,

FIG. 14 is a circuit diagram of a numerical counter which is used for measuring a physical quantity Scanning is used.

FIG. 15 is a diagram of trigger signals and output signals a counter with preselection,

Figure 16 and 17 curves representing the changes in the cyclic Relationship are characteristic,

FIG. 18 a simplified circuit diagram of the electronic circuit for generating the function of FIG. 3,

and

FIG. 19 shows a diagram of the output signals of the circle in FIG. 18 for various values of the cyclic Relationship.

409824/03 2.9

In the invention, the signal supplied by a transducer is converted into numerical form. In a motor vehicle A variety of transducers are used, which are very different types. They convert their measurement into that Change of an electrical quantity by. This information can be available directly in numerical form, e.g. in Shape of the location of an element that is provided with an encoder, which obviously does not pose any particular problems prepares.

The information can also be represented by pulses of variable frequency, as is the case with a transducer, for example is the case in the manner of a distance detector which is arranged opposite a toothed wheel. In this case, as will be explained later, for example, the frequency of pulses, i.e. the repetition frequency, is measured. In general, one chooses between these two possibilities depending on the problem at hand and above all the following versions:

If, as is often the case, you take the mean of a electrical signal for · representation of the mean value of a physical quantity, which is a given If the position changes, it is advisable that the response of the transducer is linear so that the mode of operation is characteristic is.

The signal from the transducer can also be in the form of a voltage are present. In this case one can use known voltage-frequency converters use, i.e. periodically compare the voltage with a variable voltage (sawtooth) and measure the number of pulses delivered by a clock, while the voltage to be measured is smaller (or greater) than the reference voltage. It is here a technique known for making electronic voltmeters.

409824/0329

Finally, the transmitter can transmit its information in the form of changes in the electrical resistance, the electrical Supply inductance or electrical capacitance. Such changes can easily result in voltage or frequency changes being transformed.

The part of the computer that is intended to receive data must be periodic at a frequency known as the sampling frequency provide coded information in the form of a number (or a word). Every word represents a limited number of digits (or "digits"), but in such a way that the mean value after a sufficient Number of samples for the mean value of the physical input variable is characteristic.

As shown in FIG. 1, the size or sizes are measured during a time T. The time T is determined e.g. by frequency division the pulses of a clock H received.

The clock generator H sends pulses of the frequency 1 / Tl to a divider 1, the intermediate _{outputs of} which are labeled A, A- ... A -1. At the output A. the frequency is equal to the clock frequency divided by 2.

Using AND gates 2, 3 and 5 and using inverter 4, the quantities are formed:

S ₃ = SA ₁ -H

The sampling period is e.g. T = T, χ 2, which can be found for S, S_ or S- is found, as indicated in the diagram attached to FIG. 1.

Measurement of a frequency

The measurement of a frequency is very simple in principle. One counts, for example, how many impulses from the through a frequency

409824/0329

expressed quantity can be generated during the time T.

If one takes into account the requirements that were set out earlier, the problem is a little more difficult.

For an easier explanation one assumes, for example, that the frequency the clock pulse is 1 MHz (T. = 10 ~ seconds). In addition, one assumes that a division by 2 (n = 10) performs, i.e. T = 512 microseconds (the division by a power of 2 is not absolutely necessary). Finally, it is assumed that the frequency to be measured is between 0.4 and 0.8 MHz changes.

Using the aforementioned procedure, the ■ calculated number of pulses is between:

6 -4
O, 4.10 χ 5.12.10 * = 204.6 pulses, and

0.8.ΙΟ ⁶ χ 5.12.10 ~ ⁴ = 409.6 pulses.

You can now see that you have to increase the number of received To measure impulses down to one impulse, a counter

with nine digits (2 = 512), but always at least 204 indicates "which is completely inexpedient."

in order to eliminate these disadvantages, using the device of Figure "2", the device 2-1 is a NAND gate (eintiND-NI (SIT gate) ₉ as well as the apparatus 2-5 (or an AND gate) and the device 2 = 4.

2 = 2 β 2-3 and 2 = 6 are binary counters »2 = 7 is a transmission device (latches)"

The working method is as follows %

The NAND gate 2-1 lets the pulses through _? their frequency

n — 1

one decreased by the time during the pages T - 2 T ₁

"I want to measure Ap to the power of ¹⁸ _{* d o} h _o in the previous example 512 = 2x1 = 510 microseconds.» It follows from this that M pulses pass through a sampling time ^ where H is between

see 0.4 χ 1.5.ΙΟ ² = 204 and 0.8 χ 5.1.ΙΟ ² = 408 pulses.

The counter 2.2 is connected as a divider dividing by D, so that the number of exiting from 2.2 and to 2.3 and

N
2.5 impulses arriving — g “. These pulses are given on 2.3, which is also a counter. The output signals of certain stages of 2.3 are applied to the inputs of an AND gate 2.4, so that the output signal of 2.4 is only high when the signals applied to the input of 2.4 are all high. This output signal is given to a special input of 2.3, which blocks the counter, and on the other hand to 2.5. The impulses that emerge from 2.2 therefore only pass through 2.5 if 2.3 has reached the fixed value through the selection of its output signals given to 2.4. 2.6 is also a counter.

At the end of period T, 2.6 shows a number N:

N
N- = —ττ— - η, if η is the number of the

Coding of the output signal from 2.3, which is given to 2.4, is certain number.

Assume, for example, that D = 100 and that f = 0.4 MHz, then

This means that two pulses emerge from 2.2 during T and that at the end of T 2.2 indicates 4.

It is assumed that η = 2. This means that at the end of T, when all counters were initially at zero, the following situation prevails:

409824/0329

Number of pulses :. - .204

Counter 2.2 at the end of T:

Counter 2.3 at the end of T:

Counter 2.6 at the end of T:

It is assumed that the cycle begins again after _{the counters 2, 6 and 2.3 have been cleared by S 2} (see FIG. 2), which is applied to suitable inputs, during the time S ₁ , but without resetting 2.2 to zero .

At the end of the second scan you then have:

2.2 > 8

2.3 -> 2

2.6> O

At the end of the third scan you have:

2.2 =? 12th

2.3> 2

2.6 -> O

etc.; at the end of the 25th scan one has:

2.2 > O

2.3 - > 2

2.6> 1

This means that the device performs two operations:

a) it interpolates the values

b) it eliminates the unusable numbers, if you want, it shifts the zero.

In other words, the device indicates 0 for N «= i 200, for N = 300, 4 for N = 700.

409824/0329

If a value of N is between 200 and 300, e.g. 204, the output counter for (x + y) samples shows χ times O and y times 1, so that:

= 0.04

χ + y

that is, the interpolation is simply done by the instantaneous frequency of two numbers that represent the value of N. surrounded so that the numerical value is equal to the encoded value that would be assigned to N.

The device 2.7 receives S_ and transmits under it Command sends the value received at its inputs to the outputs for further use in each cycle.

The device thus makes it possible to eliminate the unusable numbers (savings in digits) and also between Interpolate values on condition that their frequency is high with respect to the sampling frequency. If this is not the case, the device of FIG. 3 is used,

This device makes it possible to carry out operations analogous to those of the circuit of FIG. 2, however under the hypothesis that the frequency to be measured is not very large with respect to the sampling frequency. The problem is therefore the following: In order to get to know the size of the frequency f, it is sufficient in principle to measure the pulses during a count for sufficient time to achieve the desired accuracy. In the case of a computer for a vehicle one wants to know approximate sizes of this frequency to be measured with a relatively high frequency in order to get it through express a number comprising a few digits or digits, but in such a way that the measurements are carried out in this way ensure that the mean value of the displayed quantities is equal to the measured value.

If these requirements are taken into account, then, as in the case of the circuit arrangement of FIG. 2, one period

40 9 8 24/0329

to reset to zero, during which no counting is carried out * There is a risk of getting a wrong result due to the "stroboscopic" effect. For example, if the frequency to be measured is four times the sampling frequency, the counter shows 3 or 4 accordingly
the relative phase of the sampling stage and the pulses
from f. One solution is to keep the pulses f constantly

X X, "

to count.

The operation of the device of Fig. 3 is as follows:

The pulse of the beginning of the scan (S) is applied to a
of the inputs of the RS flip-flop 3.1-given and brings the output of 3.1 to a high level, which is connected to one of the inputs of the AND gate 3.6. It also follows from this that the complementary output of 3 »1, which is connected to an input of 3.5, is brought to a low level at the same moment. The impulses f reach the other inputs of 3.6 and 3.5. From the switch from 3.1 to the "working position" at the beginning of the occurrence of S it follows that these pulses f in this position 3.6 through

X X

can run, but they are locked at 3.5. The occurrence of a pulse f is thus on 3.3 over 3.6
transmitted and ^ causes the transition of the RS flip-flop 3.3 in the working state, and the output of 3.3, the
is connected to one of the inputs of the element 3.4, is

At the other input of 3.4, S (S ₁ , is inverted
by the inverter 3.2) so that the output of
3.4 moves to a high level if:

a) its input connected to 3.3 is at a high level, i.e., when the pulse f has been counted, and

b) its input connected to S is at a high level,
ie that the dead time S has expired.

409824/0329

The thus the inputs of R.l and R.3 to reset Output connected to one causes R.3 and R.l to tilt back and link 3.5 to open.

Elements 3.7, 3.9, 3.8, 3.10, 3.11 correspond to elements 2.3, 2.4, 2.5, 2.6 and 2.7 of FIG. 2 with the same purpose (removal of the unusable numbers), counting processes> reset to zero by S ₂ and transmission by S. perform. With a suitable arrangement of the inputs and polarities, the circuit can be designed in such a way that the toggle of 3.3 is counted or not. If it is not counted, you get a direct elimination of the first number.

Finally, the counting is permanent and in general the counting element is 3.10 (or 3.9), but during the dead time S ₁ and until the occurrence of the first pulse f following the beginning of S. the counting element (of 1) consists of 3.3 (and 3.1) . This of course presupposes that the duration of S _{1 is} highly lower than the minimum period of f, which is practically always the case (low frequency of f); in the opposite case one can obviously replace 3.3 with a device 3.9 and count several pulses before opening the main counter.

Fig. 4 shows a device 'which is somewhat simpler than that of Fig. 3 and which serves a corresponding purpose.

An RS flip-flop 4.1 receives at its input for putting in the working state S, which is the output with the AND gate 4.2 is connected to a high level, and the output, which is connected to the AND gate 4.7, to a low level Brings level. It follows from this that the impulses f can pass 4.2 and cannot pass 4.7. 4.3 is a 2.3 Corresponding counter that "holds back" the first numbers.

409824/0329

If after the start of the scan a number of pulses equal to that of 4.3, 4.4 coded (as for 2.3, 2.4 in 2) has passed through 4.2, the output goes from 4.4 to a high level, so that 4.1 in the Rested position and thus blocked and 4.2

4.7 is opened. The counting is now continued at 4.5, which was _{reset to zero during S by S 2} after transmission to the outputs of 4.6 (latches) by S_.

Measure a voltage

A voltage is measured by measuring the time while the voltage to be measured is e.g. less than a linear voltage (sawtooth). The circuit arrangement is shown in FIG.

The signal S., which is given to the base of the transistor 5.6 is discharged the capacitor 5.2, which is then over 5 * 5 recharges and is connected to the inverting input connections and the output of a function amplifier 5.1 is connected, the non-inverting input of which is biased by the bridge circuit 5.3 / 5.4. This circuit arrangement is known and enables a very linear sawtooth signal to be obtained at output 5.1.

This voltage is compared by the comparator 5.7 with the voltage to be measured. The output of 5.7 is on low level as long as the sawtooth signal is greater than the voltage to be measured, and in the opposite Fall at a low level.

If the sawtooth voltage is denoted by V, then V, from the end of S ₁ onwards can be expressed by:

^v d = ^v do - ^ct

^V dO ^{un (} ^ ** ⁿ ^ ⁿ 9 ^{en from} 5.3, 5.4, 5.2 and 5.5 from «

409824/0329

The tilt period of 5.7 is such that

V _x = v _d0

_T _ ^v ao - ^v x

e k

The measuring time is e.g. the sampling time reduced by S and T-, i.e. in the above-mentioned case:

2 T ₁ - 2 ^X - T = A + BV

JL ti Jt

A and B are constants.

The circuit arrangement of FIG. 6, which hardly differs from that of FIG. 2, shows the measuring arrangement. If the output of 5.7 is at a high level (after tilting the comparator) and S is at a low level (ST is given to 6.1), the element 6.1 lets the pulses of the clock H through. These pulses are counted from 6.2 which reduces the number (interpolation) and programmed them to 6.3 (from 6.5 as explained above to eliminate the unusable digits). This makes it possible, as soon as this "remainder" has been eliminated, to count the pulses transmitted by 6.2 in 6.7 when 6.4 has become transparent. At the end of the cycle (S high) the counting stops, S ₂ resets 6.3 and 6.7 to zero, S secures the transmission on the transmission device 6.1 and, as previously mentioned, 6.2 is not reset to zero for the reasons stated became.

Decoding - application

By means of the means indicated, the signal supplied by one or more transducers of a physical quantity can be converted into one or more numbers which changes with each scan so that the mean value of the measured quantity is represented.

409824/0329

- vc-

It follows that each size at the end of each scan is expressed by a binary number, which according to the requirements a certain number of digits or places includes. ·

For explanation it is assumed that one wants to form a control function dependent on four variables LMNO.

It is assumed that the size L is divided by three digits (8 fixed points), size M by four places (16 fixed points), N by two places (4 points) and 0 by one Body expresses. That is, the scanning result is written in the form of a binary number with (3 + 4 + 2 + 1) = 10 digits can be.

^L l ^L 2 ^L 3 ^M l ^M 2 ^M 3 ^M 4 ^N l ^N 2 °

the values L., M., N. 0 are 0 or 1. The first word segment consists of LL ₃ L ₃ , the second word segment of M ₁ M ₃ MM ₄ , the third of NN and the fourth of 0.

011100110 1 expresses that the following applies to this word:

the changeable L has the value 3, the changeable M has the value 9 the changeable N has the value 2 the changeable 0 has the value 1

The total number of possible words is 8 x 16 χ 4 χ 2 = 1,024 = (2 ¹⁰ )

It is known from program memory technology that a number can be assigned to these 1,024 words, which represents the value is characteristic of the function of L M N 0 wants to give the point under consideration (the word "point" is used here in a broad sense, since it is obvious is not a geometric point since it is five. " Dimensions).

409824/0329

In the course of the scan, the word changes if each quantity does not exactly correspond to a "fixed value" of LMNO. This corresponds to the previously explained interpolation function. "It follows that the assigned value also changes, its mean being equal to the mean in linear Fann as a function of L M N O passing through fixed points.

For explanation it is assumed that L between L. and L- ₊₁ / I between M. unc
O and O ₂ lie.

M between M. and M.., N between N, and N, and 0 between

When you look at £ L the amount

MM.
M = τ. ETC.

is the weight of the amount F.j that L., M.,

• L · JCX J- j

Ν ,, O _{2 is} assigned, (1 - ε L). (1 -ίΜ). (1 - jN).

(1 -tO). That of F _{1 + 1} , ^ _-1 , _{fc + 1} , _{1 + χ} : (£ L. PM.

EN. € 0) .

That of F _{1 + 1} ,., _{K + 1} , _χ : £ .L (1- tn) £ .k (l - £ l) etc.

The center of the function F, the assigned function, is the center of gravity of the assigned points of the specified Weights. '

These numbers F are expressed by a binary code called includes certain places by the precision one wants to obtain, i.e .:

X ₁ X ₉ ... X _f , if f is the selected number of digits. X is the digit of the highest weight, X _f that of the lowest.

Decoding

Formation of a linear tension

Assume that f is less than (n-1), the number of digits of the circle of Fig. 1 (if this is not the case, one or more elements of this device

4098 2 4/0329

add to Fig. 1 and use only (n-1) of these to generate the to form previously specified sampling period.

The set is formed:

O ~ * ^^ 1 t ¹ ^ T "^ OO * ¹ ^ T ^^ O ***> O · · β ¹ ^ l ¹ ^ O · · · **. C · £

a _f X _f

In the course of a scan, the cyclic ratio high / (low + high) of this function is characteristic of F and independent of the frequency.

For ease of explanation it is assumed that f = 8 and n-1 = 9.

Over the course of a sampling period, the amount is: a = AT is 256 times high and 256 times low;

a ₂ = A AT is 128 times high, but this is while a is low;

a_ = A ₁ A ₂ AT is 64 times high, but during a and a ₂

are low;

a. = AA ₂ A_ AT is 32 times high, but while a, a ₂ , * a _{3 are} low;

a, - = A, A ₂ A ₃ A. AT is 16 times high, but while a, a ₂

a ₃ , a. are low;

a _ß = A ₁ A _? A_ A ₄ A ₅ AT is 8 times. high ... etc. a_ = A ₁ A ₂ A ₃ A. A ₅ Ag ÄT is 4 times high ... etc. a _o = A _T A "A- A- Ä _c Α ,, A_ ÄT is 2 times high ... etc.

O 12J4DD / O

It follows that the ratio of the high signal is S over the entire sampling time

256 X .. + 128 X _o + 64..X -, + 32 X. + 16 X _c +8 X _c +4 X_ + 2 X ₀

^R s'- : sir "

409824/0329

dVi R - - YI .- ■ γ ι γ j. · * · Γ J. ■ * ■ ν j. * ν ι ^ - γ ■ - * ■ ν

S 2 1 4 2 8 3 16 4 32 5 64 6 128 7 256 8 with X, ... X ₀ = O or I.

X ο

2 F This can be written: R_ =

This means that the sampled value of F is expressed twice by the sampling time T, if, on the other hand, F = (n) had a single meaning correspond to each sample.

Since it is a cyclical relationship, this is it Value independent of the frequency of the clock generator and the sampling frequency.

Fig. 7 shows the scheme.

The scheme uses exclusively NAND gates as an example

(AND-NOT elements).

The first link is the set:

The second link is:

A ₂ X ₂ = a ₂

etc.

The eighth link is the set:

A ₂ ... A ₇ A ₈ X ₈ = a ₈

The last link is:

^a l ^a 2 * ·· ^a 8 ^{= a} l ^{+ a} 2 " ⁺ * ^a 8 ~ ^S

409824/0329

In order to obtain an analog voltage characteristic of F, it is therefore sufficient to have an RC filter at the output of 5 (passive or active) to be arranged.

Another method for generating the voltage (FIG. 8) A further example of the circle for coding or for generating a voltage on the basis of a code written in a memory is shown in FIG.

For ease of explanation it is assumed that there is also the problem, the mean value of a regulating hypersurface depending on four parameters LMNO, the is transmitted at any moment by a code with multiple digits X., to be expressed in the form of a voltage.

For one or more of the previously explained methods of sampling information, the is directed at each moment Numbers L M N O to the inputs of a programmed memory 8.1.

You can also use the system without a "latches" circle between form a number characteristic of the measurement state of one of the variables and the memory.

To show this type of construction of the circle, Fig. 8. b shows this arrangement, assuming, for example, that the transferred quantity L is represented by a circle such as that of FIG. 6 is measured, the size M of a circle like that of FIG. 4, the size N of one Circle of the type as in Fig. 3 and the size 0 of a circle of the type as in Fig. 2.

409824/0329

Obviously, this choice is just an example
serves.

Of course, it is also permitted that the counting capacity of the circuits is reduced to what is necessary by selecting the counters was restricted.

At 6.7, 4.5, 3.10, 2.6 the elements of the circuits of FIGS. 6, 4, 3 and 2 are shown, from which connections to the programmed memory 8.1 are made. This means that these elements are linked directly to 8.1 and omitted the "latches" 6.8, 4.6, 3.11 and 2.7, respectively.

From this arrangement it follows that at the inputs 8.1.10 to 8.1.19 of the programmed memory 8.1 the places L to 0 change continuously during each scan, which obviously also leads to changes at the outputs X. of the "coding element" - Memory 8.1 leads. This has in the
Circuit arrangement is irrelevant. In fact, the outputs "X." from 8.1 connected to the inputs of a counter 6.2 with preselection. The outputs of this counter are
connected with an AND-NOT element 8.3. The counter also receives 8.9 at its input, which is usually called
"Load input" is referred to as a transmission signal
like S_, previously described, at each sampling end pulse. When all of the 8.3 outputs are high, the 8.7 output is low. This allows
the pulses of the clock H, which are given to the input 8.5 of the ÜND-NOT element 8.4, do not reach the input of the counter 8.8 of 8.2, since the second input of 8.4 in this case is at a low level and therefore the output of 8.4 is always at a high level. If, on the other hand, at least one of the outputs 8.2 is low
Level is, the output of 8.7 is at a high level and
therefore 8.4 lets the pulses of the clock to 8.8 through and the counter 8.2 can count.

409824/0329

The working method is therefore as follows:

If the short transmission pulse S is given on 8.9, the counter 8.2 is loaded to a number η, namely:

η = 2 ⁷ . X ₁ + + 2 °. X.8

which just represents the value of the sample at the moment of the measurement.

It is assumed that the counter has eight digits and therefore its counting capacity ranges from 0 to 255.

From this it follows that 8.7 goes to a high level, except when η = 255. In general, therefore, the pulses H via 8 "4 (unlocked) reach 8.8- As soon as S _n ends, the counter 8.2 counts starting from ns (η + 1), (η + 2} ... When 255 is reached, all outputs go to high level, 8.4 blocks, 8.7 goes to low level and the counter remains in state 255 until the next pulse S_. It is assumed that that the repetition period of S_ η is H, that is, equal to η times the period of the clock (the period obtained simply by dividing the frequency of the clock by η by means of a known device), and it is assumed that S _{3 is} equal to a duration η "Η. This means n _T times the period of the clock generator: the output 8.7 is therefore at a high level during a time t".

t _H = n _T .H + (255-n) .H, or more generally t _R = η _τ · Η + (Nn) .H. If N is the capacity of 8.2, the ratio t is "

divided by the repetition period of the physical Size therefore:

η .H + (N-n) .H

7 = ~ ■

η + Nn
--S_ - -A-Bn

409824/0329

It follows from this that the simple filtering of the output 8.7 of 8.2 results in a linear voltage function of η, and consequently that its mean value is equal to the mean value of the regulation sought.

Formation of the numerical mean

In order to obtain the mean value of the function F in numerical form, proceed as follows: Count the impulses of the Clocks, which are present while S is high, for a certain number of samples, e.g. for 2, and one is used to represent the value of F only the number of pulses divided by 2, i.e. one uses the (b + 1) first digits of the counter not ..

The pulses S of the counter of FIG. 1 become a divider, which divides by 2 (Fig. 9), which, when it has received 2 pulses (or another code), the output signal of the coding member 9.2 brings to a high level, which is inverted by.

This inverted output signal is given to one of the inputs of elements 9.4, 9.5, 9.6. At element 9.4, the signal S _{1 is} applied to a second input, so that a low pulse of duration S ₁ appears at the output every 2 samples. This pulse is given to one of the inputs of the member 9.7, which also receives the pulses H and the signal S of the figure.

It follows that H is present at the output of 9.7 when S is high (decoding) and when the output of 9.4 is high, a lock necessarily occurs every 2 samples for a duration S. The pulses passing through are counted by the counter 9.8, which divides their number by 2 and therefore transmits a signal to 9.10 every 2 pulses.

This means that, starting from a zero at the beginning of a cycle, after 2 scans, the counter 9.10 displays the number N: 409824/03 29

2 ^b

N = R x -R = F., _. ■,.
s J ° ^s medium

The members 9.5 and 9.6 enable the transfer to in accordance with an arrangement already described. the carryover device 9.9 (through 9.6) and the effectiveness of
9.8 and 9.10 for each mean value cycle through S and 9.5.

Converting the voltage (or code) to a duration
It can happen that one wants to convert the signal S, which is coded in a cyclic ratio modulation, into a duration. This is the case with injection, for example.

The circuit arrangement is that of FIG.

A positive triggering signal is applied to the base of the transistor 10.1 (which is related to the control mechanism, for example, the injection triggering), the 10.1 being conductive
makes and thus 10.5 via 10.4 (current limitation) down to the potential of the emitter of 10.1, which is emptied by the
Bridge 10.15, 10.14 fixed and because of the capacitor
10.16 is only slightly changeable.

It follows that the output of 10.6 goes to a high level. This signal is sent to one of the inputs of the ÜND-NOT element 10.9, which receives the output signal inverted by 1Ο.17 at its other input, so that even when the output of 10.6 is high, the output of 10.9 is high by the end of the trip signal, so this is not an accurate one Must have duration.

At the end of the signal, the output will go low from 10.9 to
the output of 10.6 goes low.

This happens when the voltage of 10.5 has become greater than that applied to the non-inverted input of
10.6 is created.

409824/0329

At this input the voltage is that obtained by the RC filter 10.12, 10.11 that is sent to the collector from 10.2 is connected, which sends the signal S which Fig. 7 receives.

The characteristic is exponential, i.e. for a certain change in V the change in absolute value corresponds to its absolute value the time is different. This is often of interest for an injection circuit, for example, but it is obvious that one can linearize the characteristic if one replaces 10.3 with a constant current generator.

Angle converter

It can happen that you want to convert the signal S into an angle. You can do this very easily with an appropriate Achieve circuit arrangement, with the exception that the resistor 103 is replaced by a generator that delivers a certain amount of electricity each time a Angle marking is counted.

For the sake of explanation, it is given in full how such signals are and will be generated in a particular case sufficiently completely simpler cases, which are obviously derived from this, are given.

The problem chosen is the following: it is a matter of generating a signal twice per revolution e.g. of a motor. These signals are essentially diametrically opposed to each other, but with an angular position which is a multidimensional function of several parameters. For example, the signals must be set to two different Ways to be given away. This is the case, for example, with an ignition system for engines with four cylinders without a distributor, where every signal triggers the ignition of two cylinders that are simultaneously controlled by the same coil.

For example, one can determine the angle by counting the number of teeth of the ring gear of the flywheel of the engine.

409824/0329

For this purpose, as indicated in FIG. 11a, a transducer is arranged opposite the teeth. This transducer can be of different types, e.g. optical, with a blocking transducer, electromagnetic etc. The flywheel 11.1 carries a starter ring gear 11.2, opposite to it the distance measuring transducer 11.3 is arranged, which emits signals at the output 11.4. It will also be a pion or a Slot 11.5 is provided for a synchronization transducer 11.6 is "visible" and emits a signal once per revolution. Of course this will be Slot or this pion arranged in front of the mechanical balancing of the flywheel.

In the case of the problem posed, one operation is carried out for every half revolution of the motor. But now it is known that the number of teeth on a starter device is generally prime and therefore not divisible by two is.

In this case it is advisable to first get a double frequency, i.e. a signal every time a Angle equal to half the angle of two consecutive teeth is traversed. There is an obvious, Solution, not shown, is to place two transducers like 11.3 against each other by (η + ^). o (to be staggered if o ( is the angle separating the two consecutive teeth and jn is any integer.

The outputs of the two transducers are after possible Impedance matching is connected to an OR gate and the signal is obviously obtained at the output of this OR gate with the frequency you are looking for. By the same method, one can generate a signal with a frequency k times greater than that is obtained, which is generated only by the teeth by arranging k transducers such as the transducer 11.3, the among each other by angles (n. + - ^ 4—-) ο o (with i = 1,2,3 ...

409824/0329

(k-l) are separated. The first transducer is zero numbered and taken as the origin of the dislocations. This of course enables angular resolution to obtain that Jc is times more accurate than the natural frequency when the teeth pass by.

In the particular case of the example chosen, in which k = 2, one can find a satisfactory solution with a single Transducers like 11.3 obtained by using a magnetic type transducer with variable reluctance chooses. This transducer, which is shown in Fig. 11b, has a magnetic armature 11.8, which has two teeth with a distance equal to that of two consecutive teeth of the flywheel and a coil 11.9, one end of which is e.g. connected to the ground of the system BfcmJdeien other end 11.4 is free.

The transducer is now in a shown in Fig. Lic electronic circuit used.

The coil 11.9 is fed via the resistor 11.10, which is connected to the positive pole of the system.

The passage of teeth 11.2 causes changes in the voltage at the connections of 11.9, the voltage peaks being obtained essentially at the minimum and maximum of the magnetic resistance, ie essentially when the teeth are opposite and when the angular phase shift is maximum. This selected voltage is transmitted directly via 11.11 to one of the inputs of a differential amplifier 11.15, to the other input of which the mean value of this voltage is applied via 11.14, which is obtained by the RC filter, which consists of the resistor 11.12 and the capacitor 11.13. The amplifier 11.15 now passes, for example, from one high state to the other, lower one, when the teeth are opposite, and from the low one

409824/0329

in the high, when the teeth are maximally out of phase.

The transition from low to high, which is derived by the RC circuit 11.20, 11.26, is transmitted via the diode 11.24 to the base of the transistor 11.28 , which is normally blocked by the resistor 11.27. Resistor 11.21 allows 11.20 to discharge while 11.16 output of 11.15 is low. It follows that a brief low signal is obtained at collector 11.28 with each transition from low to high of 11.16. The signal from 11.16, which is inverted by the transistor 11.19, which receives the signal from 11.16 at its base via 11.1.7, generates a positive signal at the collector of 11.19, which is fed via 11.18, every time 11.16 goes from high to low passes. This signal is used as the direct signal from 11.16, ie derived via the RC circuit 11.25, 11.29 and the diode 11.23 and also given on the basis of 11.28. Resistor 11.22 is used to discharge 11.29 while 11.16 is high. It follows that a brief low signal is also obtained when 11.16 transitions from high to low. Finally, at this collector, one receives the signal with twice the frequency of the natural frequency, which is the desired goal. '■

Fig. 12 shows how to use the foregoing information to solve the problem posed.

The double frequency pulses that are generated as has just been explained, are given to input 12.1 of the circuit shown, while the pulses that are accepted are negative and have an angular width less than o (/ 2, can be given to input 12.6 of the same circle.

The device 12.2 is a transducer with ρ digits a counting capacity equal to (2 ^P -1), so that (2 ^P -1) ■> · 2 ^ J, if one

409824/0329

with J ^ denotes the number of teeth.

The device 12.3 is a device which emits a (low) negative signal at its output 12.41 when the Outputs of counter 12.2 pass through a group of values, one of which is preset by the internal connections of 12.3 Represent number. The device 12.4 is analogous.

The device 12.3 is equal to 2 ^ by a number coded. If A is the signal leaving 12.3 at 12.41 and B is the signal leaving at 12.8 by the synchronization transducer is supplied, then a signal C is obtained at the output of 12.5

C = aTb = A + B.

As a result, this signal goes high when A or B or both go low. This high signal that is on the corresponding input of 12.2 is given, secures their resetting to zero.

Since the frequency of occurrence of B (after one motor revolution, i.e. 2 ^ pulses at 12.1) is the same as that of A (12.3 switched for 2 Z ^) is the reset to zero of 12.2 after a maximum of one motor revolution in such a way that at every moment the number appearing at the output of 12.2 is equal to twice the number of teeth that since the passage through the location of the slot or the pion 11.5 of FIG. 11a in front of the transducer 11.6 whose output 11.7 is connected to 12.6, as mentioned.

The signal A, which is negative and very short, acts as it causes the reset from 12.2 to zero, via the device 12.3 that it disappears.

409824/0329

This negative signal therefore triggers an RS flip-flop, the consists of the AND-NOT elements 12.8 and 12.9. The tilt brings the output of 12.8 to a high level. The first Pulse / following the reset from 12.2 to zero, the output that is currently the lowest (2) goes from low after high. This signal is sent to the inverter 12.7, which is connected to the second input of 12.9 and causes the RS flip-flop to tilt back. It follows, that at every revolution of the motor starting from the angular position of the reset to zero, which is taken as the origin, the Output of 12.8 becomes high during an angle equal to o (/ 2.

The device 12.4, which is analogous to 12.3, is with a value equal - ^ coded. It follows from this that ^) pulses after resetting to zero, i.e. at an angular position equal to 180 °, its output becomes low and it remains until 12.2 (J ^ + 1) indicates i.e. during an angular sequence equal to o (/ 2. This signal is inverted by the inverter 12.12.

This inverted signal and the 12.8 exiting signal generated as previously mentioned are passed through the diodes 12.13 and 12.14 given to the base of the transistor 12.15, so that this from the position zero to the position o (/ 2 and is made conductive from the position 180 ° to the position (180 ° + o (/ 2). This has the discharge of the capacitor 12.28 apart from essentially the potential of the emitter 12.15 to be effective. This potential is from the voltage divider 12.21-12.20 generated and stabilized by the bias capacitor 12.19.

In addition, to the base of the transistor 12.17 via the Resistor 12.18 at 12.16 applied the square-wave signals, the cyclical ratio of which is characteristic of the regulation e.g. those signals coming from output 8.7 of 8.3 of Fig. 8a.

This transistor, which is fed at its collector via 12.22, which is connected to the positive voltage source pole

409824/0329

is bound, caused by the RC filter 12.23, 12.24 over 12.25 at the non-inverted input of the functional amplifier 12.35 the application of a voltage, the size of which, the between the supply voltage and the voltage applied to the emitter 12.17, which corresponds to that of 12.15 is connected, is characteristic of the cyclical mean value of the signals applied at 12.16. The discharge of 12.28, as previously mentioned, causes the inverted input of 12.35 to be low and consequently Output goes high. This state obviously remains as long as 12.15 conducts, i.e. during an angle o (/ 2. Then the capacitor 12.28 is suddenly charged again in the following way: With every pulse arriving at 12.1 the monostable multivibrator (which is shown here in a known manner as an integrated circuit) 12.36 is triggered. The negative pulses of a fixed duration which are supplied by this -are transmitted via 12.37 to the base of the transistor 12.34 given. During these pulses, 12.34 is therefore blocked and enables the high voltage of the output of to be applied 12.35 via the voltage divider 12.32 / 12.33 to the base of 12.30. It follows that the 12.35 output is high and only during this time negative pulses with a duration equal to that of the pulses of the monostable multivibrator 12.36- is to be applied via 12.31 to the base of 12.29 and open it by impulses and also 12.28 load. From this it follows that from the moment when 12.15 is blocked, the via 12.37 to the inverted input of 12.35 applied voltage suddenly increases. When this voltage becomes equal to that on the non-inverted Input is given, 12.35 flips back again.

This tilting consequently produces a number η of pulses,

XV

which reached 12.1 after the end of the conductive state of 12.15. This number n ^ depends on the size of the mean Voltage caused by the signals applied at 12.16, with all other factors (the

409824/0329

- 3β -

Size of 12:37, the duration of 12:36) can be assumed to be fixed.

The introduced angular delay is:.

“R

since the impulses are separated by the angles ——- and there

of the angle of the conductive state of 12.15 - ~ - is how previously explained.

Finally, a signal (transition from high to low of the output of 12.35) occurs at angles (O + Δο () and (180 ° + Λ o (), where / \ o (is a function of the program of the memory.

When tilting back from 12.35, the reaction resistance decreases 12.50 the potential of the non-inverted input in order to avoid oscillation of the system. It is with one other embodiment ^ shown in dashed lines is possible proceed in a different way, e.g. the inverted input to bring to a high potential by means of a transistor 12.52, which is blocked when the output of 12.35 is low and that above 12.v53 and the diode 12.55 ensures the high potential of 12.28. The diode 12.55 avoids any influence during the sudden charge.

The evaluation of the signal when tilting back 12.35, which is used here as an example, assumes that you have a monostable Circuit 12.38 triggers, which controls when 12o35 tilts back at its input connected to 12.35 and now emits a positive signal at its output. The elements 12.39 and 12.40 enable this signal; to guide in two ways.

409824/0329

In fact, in position O, the negative pulse that leaves at 12.41, 12.3 causes the RS flip-flop, which consists of the AND-NOT gates 12.10 and 12.11, to flip. The toggle causes the 12.11 output to go high. From this it follows that the signal which immediately afterwards (/ lo (after) leaves 12.38) runs through 12.39 (AND-NOT element) and 180 ° further the negative pulse, which leaves 12.4 at 12.42, causes the RS- Flip-flops and that the output of 12. IO is high. In the position ( / l o (+ 180 °) this leads to the appearance of the signal of 12.38 (inverted) at the output of 12.40.

In the symbolic diagram of Fig. 13, for example, there is the circuit for controlling the duration of the injection of a motor vehicle engine shown.

The regulation of the duration of the injection takes place on the basis of information received from a speed sensor 13.1 and a pressure transducer 13.2. The first provides a signal in the form of pulses changeable Frequency f, the second a characteristic analog voltage signal V.

A clock generator 13.3 according to the scheme of FIG. 1 supplies signals S ₁ , S ₂ , S-, which are given to the two counting and coding devices 13.4 and 13.5.

The device 13.4 is one of the measuring circuits of FIGS. 2, 3 or 4. It generates binary signals at its four outputs corresponding to a word segment, so that the mean value of this word segment over a certain number of samples for the mean value of the speed measured by the transducer 13.1 is characteristic.

The device 13.5 also includes the circles 5 and 6. You emits binary signals corresponding to a word segment at its four outputs, so that the mean value of this word

409824/0329

segments over a certain number of samples for the Mean value of the pressure measured by the transducer 13.2 is characteristic.

These two word segments are placed next to each other to form a first word that will respond to the input of the programmed one Memory 13.6 is given. This programmed memory outputs a second word, which is also made up of word segments can be composed. For the sake of simplicity, assume that the programmed memory is only one word to regulate the duration of the injection.

The second word is given to the conversion circuit 13.7, which is that of FIG. This circle transforms the second Word into a function S, which is coded in the cyclic ratio modulation and its mean value for the mean value is characteristic of the regulatory function sought.

This function S in turn is converted by circuit 13.8, which corresponds to that of FIG. 10, into a signal R of the injection duration converted.

Thus, in the invention with a certain frequency, the sampling frequency, the information that is sent by the transducers for each sampling to measure the physical Quantities are supplied which are related to the function of the engine, in each case in the form of a number or a word segment encoded, which comprises a limited number of digits, so that the mean value over several samples of the Word segment is characteristic of the mean value of the corresponding physical quantity, the word segments can be grouped to form a first complete word. This first word is programmed by a Memory converted to a second word. The second word is converted into a function by a conversion circle, whose mean value is characteristic of the mean value of the regulating function sought, and this regulating function is converted into an engine control variable.

409824/0329

In addition, the sampling frequency is determined by dividing the clock frequency and the code used is a binary code. The word segments fed into programmed memory represent the coding of the digits with high weight of the counted numbers. The digits with lower weight are retained and added to the measurement result at the next scan.

In order to ensure effective control of the motor at the same time with several measured physical quantities, the invention provides a method for controlling an internal combustion engine, in which with a certain frequency, the sampling frequency, the information that is provided by transducers for measuring physical quantities with each sampling which are related to the function of the engine, each in the form of a number or a word segment are encoded so that the mean value of the word segment over several samples for the mean value of the corresponding physical quantity is characteristic, with each word segment encoding digits with high weight are characteristic, which are determined by counters with preselection, and the digits with lower weight withheld and fed to the measurement result of the next scan. This method is characterized by the fact that the preselected Values of the counters, which determine the digits with high weight of each measured physical quantity, one below the other Are prime numbers.

In Fig. 14, the member 14.1 receives at the input 14.2 a square wave signal with a duration that is proportional to the to measuring size is χ. The other input 14.3 receives clock pulses. The signals that leave the element 14.1 are sent to the counter input 14.6 of a counter 14.4 with a preselection given. The outputs of the counter 14.4 are connected to a known loop counter 14.5. If the number counted at 14.4 reaches a predetermined value ρ,

409824/0329

the device 14.5 emits a signal at its output 14.8, which is given to the counter 14.4 at 14.7 in order to control it reset to zero, and this is also given to the counting input of a counter 14.9. The outputs 11-14 of this Counters 14.9 are connected to the rest of the calculator. This counter is reset to zero by a signal, which is given to its input 14.10 between two scans. The counter 14.4 is used between samples not reset to zero.

Let B be the value present at the outputs of the counter 14.4 at the beginning of a sampling comprising (A. Ρ + B) pulses. At the end of the scan, the counter 11 shows the conversion value A or (A + 1) , depending on whether (B - + B)
is less than or greater than ρ. The counter 4 shows the
transfer value (B + B) or (B "+ B - p).

OjC 2C Oa X

In the case of two quantities to be measured, they can be expressed by:

η = A ·. P + B - xxx

n _y = A _y . g + B _y

A point z_ is assigned to a point η, η in such a way that χ y Aj £, α

that the value of ζ is the one given by the memory to the

Pair (A, A) is assigned. This may bring a determination y.

ming problem. It is assumed for explanation that B ^ = p / 2 and B = q / 2. It follows that the conversion m is +1 once for two A's and once for two A's. Likewise takes place -the measurement of y by alternately converting A and

(A + 1). The mean of ζ is therefore:

that of the mean values of z- _s and z, ~ .,. , _

x ' ^Ä y ^iA x ^1} ' ^
if the changes are made at the same time,

.. and s . ...

the mean of ζ, .. and s . .. ' . " If the

Changes are out of phase.

40982 4/0329

- 95 -

In general, when a value of η corresponds to a measurement between A and (A + 1), and a value of η corresponds to a measurement between A and (A + 1) corresponds to the mean of 2, which is assigned after a sufficient number of samples, is indeterminate, i.e., it is within of the tetrahedron, which is determined by the points:

VV ² A _x , A _y ; V ^(A y ^{+ 1}} ' ² A _x , (A _y + 1)

^(A x ⁺¹⁾ 'V ^Z (A +1), A _v ; ^(A x ⁺¹⁾ ' ^(A y ⁺¹⁾ '"(A +1), (A + l) χ y χ y

In other words, for every pair of values η and η, like the conversion mean values (A +1),

ji y ^ - **

A and A, (A + 1), the measurements m and m, the η and η x y y χ y χ y

yield correctly, points ζ of between two planes of the tetrahedron maps to enclosed values. In general, the mean of ζ is that divided by the mean of Values of all possible pairs A, A are obtained, except in the aforementioned case, in which one according to the Phase of the measurements always receives one or the other of the possible extreme values.

In order to avoid such a phenomenon, it suffices to choose the values of ρ and q appropriately.

In fact, a measurement slides m. (or m) at each scan with respect to the value of the conversion of a quantity B / p or B / q and it follows that from the measurements of the both quantities χ and y, one with respect to the other slides by an amount (B / p-B / q) with each scan.

If the two measurements do not slide relative to each other, the above phenomenon can be obtained, i.e. that the Interpolation pairs do not change in time, but they can be different according to the phase, i.e. dependent from the initial state.

409824/032 9

- ze -

The speed of the sliding quantity (B / p-B / q) is obvious-

Xy

borrowed zero when B = B = O. This means that the values

χ y

correspond exactly to the conversion values. So there is no uncertainty. If, on the other hand, B and B are not zero at the same time sliding occurs except when B. q = B. p, i.e. also if q / p = B / B. This inevitably means sliding

y *

occurs, it is sufficient that this equation is not satisfied.

p, q, B and B are integers and B <ρ, Β / q.

y χ χ y.

For example, when q = 27 and ρ = 9, q / p = 3, an appearance occurs which is e.g. analogous to the one mentioned above, if B = 3 B,

y 2c

i.e. for a range of values

^B x - 1 ^B x = 2 ^B x = 3 ^B x ■ 4th B.
X 5

^B y = 3 ^B y = 6 ^B y s 9 ^B y = 12 B.
y = 15, etc.

The phenomenon can also occur when q / p is a rational fraction, e.g. _c for q = 32, ρ = 5, q / p = 32/6 = 16/3, the phenomenon occurs for B = 3 and B = 16 .

In order that the phenomenon is not possible, it is sufficient that q / p is not in the form of an integer or a rational fraction can be expressed, i.e. that ρ and q are mutually prime numbers.

In the case of a system with more than two quantities, it is sufficient that the division factors p, q, r, s ..., which have the meaning of ρ and q in the two previous two-variable examples are prime numbers among themselves, so that each Any fraction formed by any of these numbers taken as twos is irrational.

409824/0329

If several device lengths of this type are used at the same time for measurements η, η etc., the counters of type 14.4, χ y

which are available for each measurement, with predetermined values p, q, r ..., which are prime numbers to one another, to one Loop switched.

Corresponding to the interpolation method used, which is based on the principle of sampling, with a physical one Size is represented by the duration of a voltage square pulse, one measures the number of during this Square pulse emitted clock pulse. Thus, the value χ of the quantity is assigned a number η of clock generators

pulse to.

To determine the number of digits belonging to the measuring system To limit the memory, only the heavily weighted digits are transmitted, the rest is retained. E.g. for a Number of pulses η = A. ρ + B, where ρ is a whole Number is transferred for the first measurement starting from zero, i.e. m., A and retains B. At the fol-

XX X X

The next measurement is the reading

^m 2x = ⁿ x ^{+ B} x - ^A x * ^{P + 2 B} x
If 2 B is lower than ρ, transfer A and keep

X X

2 B back. At the i-th measurement, one has

^m lx - ⁿ x ^{+ (i} - ^{1} B} x - ^A x * P ^{+ l} ' ^B x and if IB is higher than ρ, one has

^m ix ^{= (A} x ^{+ 1} p + (i} ' ^B x " ^p)
Hence one transfers (A + 1) and keeps (i. B - p) to-

X X

From this procedure it follows that the transfer soon A, soon

Ji

(A +1) is. The mean after a sufficient number

of samples is equal to η _χ .

40 9 8 24/0 329

In general, the one corresponding to the regulatory function sought Signal converted into a signal with the duration of a square pulse or a signal that corresponds to an angular rotation of the engine.

Finally, it should be noted that the second word output from the programmed memory is one of several Word segments may consist of each of which has a function to regulate the operation of the engine or corresponds to a function for controlling an auxiliary device of the engine or the vehicle equipped with the engine is. These different word segments are each converted by an independent electronic chain.

In the previous examples, the analog voltage, which is characteristic of the linear function, was determined using received an RC filter »

The invention can also be used to generate a non-linear function from this linear function which is used for the the regulatory function sought is characteristic. By the invention is a method for controlling internal combustion engines created, in which the physical quantities, which are related to the function of the engine, in the form of word segments that are grouped to form a first word that is programmed by a Memory is converted into a second word, which itself is converted into a linear one by a conversion circuit Function converted with cyclic behavior modulation coding will. This method is characterized in that the linear function is converted into a non-linear function corresponding to the regulation function sought will.

For this non-linear function ' characteristic curve made of straight sections.

4098 24/03 29

On the first line, FIG. 15 shows the signal A for resetting the counters to zero, which occurs during the short zero reset period is high and low the rest of the time. The second line shows the signal Tr for transmission or loading the counter with preselection and the third line is the output signal of the counter with preselection, usually this output signal is high as long as the counter is not full and low when full.

The frequency of the last signal is that of the sample and is obtained by numerically dividing the clock frequency. The division takes place in such a way that the sampling period, i.e. the repetition of the signal, expressed in number of clock pulses at least equal and generally greater than the maximum capacity of the counter or the counter with preselection. The length expressed by the number of Clock pulses, the high part of the signal of the third straight line depends on the size of the charge on the counters, i.e. of the one present at the output of the matrix at the moment of transmission Word. .

If you enter this length with N (N is the number of clock pulses, that you measure), and if you place an RC filter at the output of the signal, you get at the output of the Filters a voltage V which is proportional to the supply voltage and N (Fig. 16).

According to the present improvement, it is proposed to form a curve of V as a function of N, such as it is shown in Fig. 17, i.e. that it is proposed, instead of a straight line passing through the origin of the plane (N, V) a value V = El for N = O (N obviously cannot assume negative values), V = E2 for N = 1, V linear increasing from E2 to E3 for N between N = I to N = Nr, then linear with a different slope from E3 to E4 for N between Nr and Nm.

409824/0329

to do this, use the arrangement of Fig. 18, A bistable RSH? Lip flop receives at its input R das Signal A after derivation by the element d / dt, so that the output signal of this flip-flop at the moment of occurrence of signal A (positive edge on the first line in Fig. 15) goes high. The same flip-flop gets on its other Input S the signal N, i.e. the complementary signal of the signal N (third line in Fig. 15).

The circuit consists of three transistors T1, T2, T3, whose emitters are connected to ground. There are two resistors between the supply source E and the collector of Tl Rl and R2 connected in series. The output signal S is picked up at their common point

The output of the flip-flop FF is connected to the base of Tl.

An analog circuit is connected in parallel to R2 and Tl, which consists of a resistor R3 and a transistor T2, the base of which receives the signal N. Parallel to T2 is a Connected circuit, which consists of a resistor R4 and a transistor T3, the base of which receives the signal No. Finally, a resistor R5 is connected in parallel with T3.

In addition, the sampling period is obtained by numerically dividing the clock frequency. At the appropriate divider an output signal is picked up which, for example, goes high when No clock pulses have passed since the signal Tr (this circle is not shown as it is normally structured is) .

For different values of N the output signal in Fig. 19 shown.

The way it works is as follows:

409824/0329

Output FF high ........ Tl open

Output FF low Tl blocked

N high (N low) T2 locked

N low (N high) T2 open

Number of clock pulses since Tr lower than No: T3
open minded.

Number of clock pulses since Tr greater than No: T3 blocked.

At the moment of the transfer, T1 is therefore opened (tilting of FF) and remains open if N = O, or is blocked again immediately if N is greater than or equal to 1 (immediate tilting back from FF).

The regulation of the device is carried out as follows:

a) The transistor T1 is kept blocked and the transistors T2 and T3 are kept open by means of a positive voltage applied to their base. The voltage Vs.
is therefore the same:

R3

E. when E denotes the supply voltage.

R3 is regulated, whereby R1 is assumed to be non-adjustable becomes such that one obtains

Vs = V2 (R4 and R5 short-circuited by T2 are without
Effect).

b) T3 is left open and T2 is blocked. You now have (Tl is always blocked):

^{VS = E} * R1 + R3 + R4

Adjust R4 so that this value is equal to the desired value V3.

c) Tl and T2 are left blocked and T3 is blocked. One has

well:

409824/0329

-A-

VS = E. R 3 + R4 + R5

R1 + R3 + R4 + R5
Set R5 so that this value is equal to V4.

d) You open Tl, T2, T3 and you now have:

Vs = E.

R2.R3 R2fR3

R 1 + R2.R3

R2 + R3
Adjust R2 so that this value is equal to Vl.

The curve in FIG. 17 is obtained after filtering through an RC circuit of the signals in FIG. 19.

According to the invention, one can use a non-linear function get more than two straight lines with different slopes.

The non-linear function is achieved by modulating the amplitude of the signal with a variable cyclic ratio depending on the code of the second word.

409824/0329

Claims (26)

- 4T - Expectations | 1./ Method for controlling an internal combustion engine, characterized ^ - ^ characterized / that with a certain sampling frequency the information that is supplied with each sampling of transducers for measuring physical quantities that are related to the function of the engine, in each case in the form a number or a word segment that includes a limited number of digits are encoded in such a way that the mean value over several scans of the word segment is characteristic of the mean value of the corresponding physical quantity, that the word segments are grouped to form a first complete word that the first word is converted from a programmed memory into a second word, that the second word is converted by a conversion circuit with a function whose mean value is characteristic of the mean value of the regulating function sought, and that this regulating function is converted into an engine control variable. 2. The method according to claim 1, characterized in that the sampling frequency by dividing the frequency of a clock generator is determined. 3. The method according to claim 1, characterized in that the code used is a binary code. 4. The method according to claim 1, characterized in that the coding of each word segment for each scan Digits with high weight represents, and that the digits with lower weight withheld and the result added to the measurement in the following scan. 5. The method according to claim 4, characterized in that in the case of measuring a frequency which is substantially greater than the sampling frequency, the process of coding 409824/0329 - 4-3 - by transferring the withheld digits and by shifting the zero he follows. 6. The method according to claim 4, characterized in that im Case of measuring a frequency of the same order of magnitude as the sampling frequency used by the transducers during the duration of the execution of the scan information supplied and the result of the measurement in the following scan can be added to the strobe effect to eliminate. 7. The method according to claim 1, characterized in that in the case of measuring a voltage, the process of coding by transferring the withheld digits and by zero shift he follows. 8. The method according to claim 1, characterized in that the Average value of the function generated by the conversion circuit independent of the frequency of the clock generator and the sampling frequency is. 9. The method according to claim 8, characterized in that the mean value of the function by filtering the signal of the Output of a counter with preselection is received. 10. The method according to claim 8, characterized in that the Mean value of the function in numerical form after dividing the number of pulses from the clock by the number of Samples is obtained. 11. The method according to claim 1, characterized in that the signal corresponding to the regulating function in width a square wave signal is converted. 12. The method according to claim 1, characterized in that the . signal corresponding to the regulating function into a signal is converted, which corresponds to an angle of rotation of the motor, are prime numbers among each other. 409824/0329. - 44 - 13. The method according to claim 1, characterized in that the . second word consists of several word segments, each of which has a function to regulate the operation of the Engine. 14. The method according to claim 13, characterized in that
at least one of the word segments corresponds to a function for controlling an additional device of the engine. 15. A method for controlling an internal combustion engine, in which with a certain sampling frequency, the information that is supplied with each sampling of transducers for measuring physical quantities that are related to the function of the engine, each coded in the form of a number or a word segment that the mean value of the word segment over several samples for the
Mean value of the corresponding physical quantity is characteristic, with each word segment encoding the
Digits with heavy weight represents those of counters with
Preselection are determined, and wherein the digits with lower weight are retained and added to the result of the measurement in the following scan, characterized in that the preselected values of the counters, which determine the digits with high weight of each physical quantity measured, are among themselves prime numbers . 16. Device for performing the method according to claim 15, characterized by a counter with a preselection, the outputs of which are connected to a loop coding device
that feeds a zero reset signal to a counter,
if this has counted a number that is equal to the preset value. 17. The device according to claim 16, characterized by a plurality of counters for measuring a plurality of physical quantities which values are selected that correspond to numbers that 409824/0329 -4S- 18. Method for controlling internal combustion engines, in which the physical Variables that are related to the function of the engine are coded in the form of word segments, which are grouped to form a first word which is converted into a second word by programmed memory which is itself converted into a linear function by a conversion circuit, which in a cyclic Ratio modulation is coded, characterized in that the linear function is converted into a non-linear function that corresponds to the regulatory function sought. . . · The method according to claim 18 for control by sampling at a frequency obtained by dividing the frequency of a clock generator is determined by internal combustion engines in which the Umwahdlungskreis consists of a counter with preselection, thereby characterized in that the non-linear function consists of a constant level step with a first even part and a second straight part with a different slope consists. 20. The method according to claim 19, characterized in that the Level constant level corresponds to a code of the second word equal to zero, that the first even part corresponds to that from one to codes reaching to a predetermined fixed code, and that the second even part corresponds to the remaining codes is equivalent to. ' 21. The method according to claim 20, characterized in that the non-linear function by modulating the amplitude of the Signal is formed with a variable cyclic ratio as a function of the code of the second word. 22. Electronic circuit arrangement for performing the method according to claim 19, characterized by a - ' first transistor connected in series with two resistors between a supply source and ground, whose common point the connection of the output of the circuit 409824/032 9- arrangement forms, a second transistor with a third resistor in series between the terminal of the output and connected to ground, a third transistor connected in series with a fourth resistor between ground and the common point of the second transistor and the third resistor is connected, and a fifth Resistor connected in parallel with the third transistor. 23. Circuit arrangement according to claim 22, characterized in that the base of the first transistor with the output a bistable flip-flop is connected, whose zero reset input a signal with the frequency of the changeable cyclic ratio before reloading the counter with preselection, which is used to create the changeable cyclic ratio is used, and the other input is the complementary signal of the changeable cyclic ratio, which is supplied by the counter with preselection, which signal is only available when the code of the second word is non-zero. 24. Circuit arrangement according to claim 22, characterized in that the base of the second transistor is the complementary Receives the signal of the output of the counter with preselection. 25. Circuit arrangement according to claim 22, characterized in that the base of the third transistor is the complementary Receives a signal of a signal of the predetermined fixed cyclic ratio corresponding to the predetermined fixed Code corresponds. 26. The method according to claim 19, characterized in that the linear function has more than two straight parts with different Has gradients. 409824/0329 Blank page