EP1557005A1

EP1557005A1 - Output unit, input unit, data transmission system for a motor vehicle and corresponding method

Info

Publication number: EP1557005A1
Application number: EP03748077A
Authority: EP
Inventors: Bernd Pfaffeneder
Original assignee: Siemens AG
Current assignee: Continental Automotive GmbH
Priority date: 2002-10-31
Filing date: 2003-09-23
Publication date: 2005-07-27
Also published as: DE10250920A1; US20050273210A1; DE10250920B4; JP4268939B2; WO2004040852A1; KR20050075371A; US7016770B2; JP2006516073A; KR100667389B1; AU2003267400A1

Abstract

The invention relates to the use of a first coding or decoding rule for a normal operation and of a second coding or decoding rule for a particular operation. Said invention provides the opportunity for speedily transmitting data between different satellite devices arranged in a motor vehicle and an evaluation unit arranged, for example in the central part thereof, even when the data bus (31, 32) of a transmission channel CAN (3) is externally short circuited, for example by a traffic accident, i.e. the data bus BUS L (32) and the data bus BUS H (31) are connected to earth (GND) or to a battery (Vbat). The invention can be used, in particular for protecting passengers.

Description

description

Output unit, receiving unit, arrangement for data transmission in a motor vehicle and method therefor

The invention relates to a unit for outputting a signal on a transmission channel in a motor vehicle, a unit for receiving a signal from a transmission channel in a motor vehicle, an arrangement for data transmission in a motor vehicle via a transmission channel and a method for data transmission or Data acceptance in a motor vehicle.

Motor vehicles often have distributed control or computing units. Such distributed control or computing units are usually understood to mean units which are arranged at different points in the motor vehicle. Due to their need to exchange data, these control and computing units are connected to one another via a transmission channel - contactless or wired. In this case, sensors arranged in the engine compartment, in the doors or in tires, for example, transmit data to central processing units which use the received data algorithmically and actuate corresponding actuators.

The wired networking of control or computing units with sensor units is usually implemented using a bus system. Such a bus system is, for example, the well-known CAN bus (CAN = Controller Area Network). Special transmission and reception devices or driver modules, in particular so-called transceivers, are provided as access to the bus transmission channel. Depending on the application, a distinction is made between a high-speed CAN with data rates of 500 kbit / s, for example for the drive train in a motor vehicle, and the so-called low-speed CAN with data rates of less than 125 kbit / s, for example 83 kbit / s for the area Carbody. In the context of a traffic accident, the CAN transmission channel can be damaged and short-circuited. This is also referred to as the external fault of the bus lines (Bus_L; Bus_H) of the transmission channel. Four cases can be distinguished: Bus_L to ground or ground (GND); Bus_L to battery (Vbat or BAT) with a voltage of e.g. Zt. 12V, soon 42V; Bus_H to GND and Bus_H to BAT. In contrast to the low-speed range, no fault-tolerant driver blocks are currently available in the high-speed range or tolerate at most two cases of external short-circuiting of one of the bus lines in the transmission channel: namely either Bus_H to BAT or a termination of the Bus-L line to GND. The disadvantage of this is that the transmission of sensor signals is not guaranteed when this is vital under certain circumstances, in particular when it comes to sensor signals from a passive safety system such as an airbag system, belt tensioner system or the like.

The data transmission between the aforementioned units is usually asynchronous. To correctly reconstruct the data in the receiver, the receiver must therefore know the clock information of the sending unit. Because of this, this clock information must be transmitted over the transmission path from the transmitter to the receiver. If clock information is transmitted in addition to the other information, the transmission bandwidth increases. The data transmission has an overhead.

The data generated in a unit - for example a sensor - is encoded in this unit for the purpose of data transmission to a remote location. The experts here also speak of channel coding, which brings the generated data into a form suitable for transmission. This takes place on the basis of a coding rule which transcodes the sensor signal into the signal to be transmitted. In the following, the term sensor signal is always used for the Transmitter present signal used, the information to be transmitted to the receiver.

In automotive engineering, such sensor signals are usually coded according to the NRZ code or the Manchester code and then transmitted. In this context, FIG. 5 shows these known coding methods for data transmission in a motor vehicle.

5a shows a sensor signal DATA over time t, the information of which is to be transmitted to a receiver via a transmission channel. The sensor signal DATA is binary coded, that is, it has a character set of two characters, namely a "0" and a "1". Individual signal units of the sensor signal DATA have a duration T.

Several such signal units lined up and each occupied with a character from the character set result in a data word that is physically present as a signal, characterized by its voltage or current states. 5b shows a work cycle TAKT of the sending output unit over time t.

FIGS. 5c and 5d show signals CHAN which are associated with the sensor signal DATA and are channel-coded according to certain coding regulations, FIG. 5c showing a signal to be transmitted

Signal CHAN shows that the sensor signal DATA was obtained after the NRZ coding. This coding is initially a 1: 1 mapping of the sensor signal. With the UART (Universal A-synchronous Receive Transmit) standard, the receiver is only synchronized by a start signal. The free-running oscillator provided in the receiver for clock generation may not leave a predetermined tolerance range until further synchronization with the transmitter by a further start signal. This requires either a high-precision oscillator in the receiving unit or a high-frequency synchronization, which is at the expense of the transmission bandwidth. 5d shows a channel-coded signal CHAN to be transmitted, which was obtained from the sensor signal DATA by Manchester coding. Manchester coding is characterized by the fact that, like NRZ coding, it uses a binary character set. Within a signal time unit T of the sensor signal, however, two characters / signal states are provided in the Manchester-coded signal. The change from one character of the sensor signal to its complementary character in the subsequent signal state is implemented by the Manchester code by a phase change. The Manchester Code thus offers the possibility of clock recovery in the receiver within a theoretical tolerance range of 50%. However, this option of clock recovery is purchased by doubling the bandwidth, since one signal unit (bit) of the sensor signal is represented by two signal states during the same time period T in the signal to be transmitted.

A channel coding based on current modulation is known from WO 98/52 792-A. The channel coding has a character set of three characters, HIGH, LOW and zero. The sensor signal provides a binary code. According to the coding specification, ones of the sensor signal are alternately encoded in HIGH and LOW pulses in the signal to be transmitted. ZEROs of the sensor signal remain ZERO levels in the signal to be transmitted.

In this known data transmission method, the time average of the signals to be transmitted is kept constant. However, no work cycle can be derived from the signal to be transmitted.

A data transmission method is known from EP 0 384 258 A2, in which a binary sensor signal by means of an AMI (Alternate Mark Inversion) code in connection with a

Pulse width modulation is channel coded. The sensor signal is first pulse-width modulated before it is formed in this way. The pulse width modulated signal is subjected to the alternate mark inversion.

A disadvantage of this data transmission method is the increased bandwidth in the signal to be transmitted compared to the sensor signal. In addition, the narrow pulses generated by pulse width modulation pose problems with regard to electromagnetic compatibility (EMC).

To reduce these problems, it is proposed in DE 101 32 048 to design a channel coding such that the code for the signal to be transmitted via the channel contains at least one character more in its character set than the character set from which the sensor signal is formed, the Information should ultimately be transferred. For example, a binary code is provided for the sensor signal, then the signal to be transmitted is formed at least from a ternary code, i.e. There are three different characters available, which are represented, for example, by three different signal states on the line, for forming a signal. In general, there are n characters available for the sensor signal, with n as an integer, and at least n + 1 characters for the signal to be transmitted.

The present invention is based on the object of specifying an arrangement for transmitting data in a motor vehicle, an associated output unit and an associated receiving unit and a data transmission or acceptance method, in which, when using a high-speed CAN, the external interference immunity of both bus lines (Bus_L, Bus_H) is guaranteed against both GND and BAT. In addition, the transmission bandwidth should be kept small and sufficient information about the work cycle should be transmitted to the receiving unit.

The part of the task relating to the output unit is solved by the features of patent claim 1. The part of the task relating to the receiving unit is achieved by the features of patent claim 10.

The part of the task relating to the arrangement is solved by the features of patent claim 18.

The part of the object of the invention relating to the method is solved by the features of patent claim 19 or 20.

Advantageous refinements and developments, which can be used individually or in combination with one another, are the subject of the respective dependent claims.

The output unit according to the invention, which operates according to various case-dependent coding regulations - laid down below - for outputting a signal CHAN to a transmission channel, formed from at least two bus lines, in a motor vehicle, comprises: a fault-tolerant coding unit for converting a sensor signal DATA into outgoing Transmit signals TxA or TxB; at least two high-seed driver modules connected downstream of the coding unit and connected antiparallel to one another for connecting the output unit to the transmission channel and converting the transmit signals TxA and TxB into the signal CHAN to be output; and a comparison unit which allows a voltage comparison of the outgoing transmission signals TxA and TxB with incoming reception signals RxA and RxB.

The receiving unit according to the invention, which operates according to various case-dependent decoding regulations - laid down below - for receiving a signal CHAN from a transmission channel, formed from at least two bus lines, in a motor vehicle, comprises: a decoding unit for converting incoming received signals RxA and RxB into a working signal DATA; at least two high- connected upstream of the decoding unit anti-parallel with each other Speed driver modules for connecting the receiving unit to the transmission channel and converting the signal CHAN to be received into incoming signals RxA and RxB; and a detection unit, which allows the detection of clock edges from the incoming signals RxA and RxB.

The arrangement for data transmission in a motor vehicle via a transmission channel, comprising at least two bus lines, according to the invention makes use of an output unit according to one of claims 1 to 9 and a receiving unit according to one of claims 10 to 17.

The method according to the invention for data transmission or data acceptance in a motor vehicle is characterized in particular by the use of a first coding regulation or decoding regulation for a normal operating mode and a second coding regulation or decoding regulation for a special operating mode.

The channel coding takes place according to the invention by means of a first coding regulation for the normal operating mode with the equality of the voltages of TxA and RxA or TxB and RxB detected by the comparison unit. A second coding rule is provided according to the invention if the inequalities mentioned are detected accordingly, that is if one of the bus lines BUS_L or BUS_H is connected to ground (GND) or battery (Vbat).

The channel decoding takes place according to the invention by means of a first decoding rule for normal operation of the decoding unit with synchronism of the clock edges detected by the detection unit at a defined signal time unit T. A second decoding rule is provided according to the invention with a correspondingly detected asynchrony of said clock edges for the defined duration of the signal time unit T. In both cases, the channel is decoded according to the invention in such a way that the code for the decoded work signal DATA provides a character set of only n characters if the character set for the signal CHAN to be accepted or the incoming receive signals RxA and RxB have at least n + 1 characters.

Likewise, the channel coding is carried out according to the invention in both cases in such a way that the code for the outgoing sensor designals TxA and TxB or the signal CHAN to be transmitted via the channel contains at least one more character in its character set than the character set from which the sensor signal DATA is formed, the information of which is ultimately to be transmitted. If a binary code is provided for the sensor signal, the signal to be transmitted is formed at least from a ternary code, i.e. at least three different characters, which are represented, for example, by three different signal states on the line, are available for forming a signal. More generally, n characters are available for the sensor signal, with n as an integer, and at least n + 1 characters for the signal to be transmitted.

It should also be noted that preferably a signal time unit of the sensor signal is preferably mapped one to one or as a corresponding divisible time unit thereof to a signal time unit of the signal to be transmitted. The signal time units of the sensor signal and the signal to be transmitted thus have the same time periods or time division ratios. In addition, the invention provides in a further embodiment that two successive signal time units always have different characters from the assigned character set in the signal to be transmitted. The implementation of this feature is achieved in that the character set of the channel code comprises at least one character more than the character set assigned to the sensor signal. In this way, a character can always be Change and thus a change of state take place, even if the sensor signal has the same sign and thus the same state over several signal time units. The same applies to decoding.

The continuous change of state, for example in the signal to be transmitted, in turn now contributes to the fact that the working cycle of the remotely located output unit can advantageously be recovered in the receiving unit. This is preferably done by means of a derivation unit. Since the time periods of the signal units of the sensor signal in the output unit and of the signal received by the receiving unit correspond and at least after each signal time unit there is a change of state, the receiving unit only needs to record the changes of state in the received signal in order to be able to derive the work cycle of the output unit. At the same time, however, the bandwidth is not increased, such as. B. in the Manchester coding introduced at the outset, since the time units for the individual bits (signal time units) can always be of the same duration or can be derived accordingly in the corresponding case described.

It is an advantage of the invention that, due to changing operating modes, a communication system can be provided both in the output unit and in the receiving unit, which can use high-speed driver modules to tolerate all types of external faults mentioned in a bus line, and thus one Security against third-party connections guaranteed, which has not been provided in the prior art. In addition, the invention has the advantage that no or only imprecise oscillators have to be used in the receiver. This allows a more economical overall arrangement. The inaccurate oscillators can be integrated on a chip. Standard bus high-speed drivers can also be used. The invention can always be used in a motor vehicle as soon as data are to be transmitted between two computing or control units. The invention is used in particular where sensor data with a high degree of security from sensors distributed over the vehicle are connected to control units arranged in the vehicle center and these control units are to be supplied with sensor data. In particular, the invention is used in occupant protection technology for high-speed transmission of sensor data from impact sensors arranged, for example, on the vehicle front or on the vehicle sides to an evaluation unit arranged in the vehicle center. The impact sensors can be acceleration sensors with downstream signal processing and a corresponding interface, or they can also be pressure sensors.

Exemplary embodiments of the invention and advantageous developments are explained in more detail below with reference to the drawings.

They show schematically:

1 shows the block diagram of an arrangement with two combined output and reception units according to the invention;

2 shows a table for driver activation of an output unit according to the invention and a reception unit according to the invention;

FIG. 3 shows an enlarged detail of the output unit according to the invention compared to FIG. 1;

4 shows the logic circuit of a high-speed driver module, for example the output unit according to FIG. 3 in detail; 5 shows waveforms belonging to known coding methods;

6 shows the signal curve of the coding method according to the invention for the normal operating mode;

FIG. 7 shows the course of a signal CHAN to be transmitted to a bus logically represented in FIG. 6d with regard to its physical course;

8 shows how the termination of a bus line according to GND or Vbat has an effect when coding only according to the first coding regulation, in the case of a ternary signal;

9 shows the signal curve of the coding method according to the invention for the special operating mode;

10 shows how the termination of a bus line according to GND or Vbat when coding according to the second coding regulation has an effect on a ternary or higher-value signal;

11 shows a status table belonging to the output unit according to FIG. 12;

12 components of an output unit;

FIG. 13 shows an enlarged detail of the receiving unit according to the invention compared to FIG. 1;

14 shows the process of clock recovery; and

15 is a table according to which the sampled signals RxA and RxB are assigned to the output value by logic, for example.

Identical elements or signals are given the same reference symbols in all figures. 1 shows the block diagram of an arrangement 4 with two combined output and reception units 4 according to the invention, which are connected via a transmission medium 3, which in turn has two bus lines 31 and 32. The first output and reception unit 4 contains a microcontroller 13 with an interface 131, an encoder 11, a decoder 21 and two high-speed driver modules 12. The high-speed drivers 12 are CAN drivers in the form of standard components that can use cables and connectors standardized according to DIN ISO 11898. The high-speed drivers 12 are connected in anti-parallel with one another and with the transmission medium 3. The CAN-HIGH output of the first high-speed driver 12 (component A) and the CAN-LOW output of the further high-speed driver 12 (component B) are connected to the first bus line 31. Likewise, the CAN-LOW output of the first high-speed driver 12 (component A) is connected to the CAN-HIGH output of the second high-speed driver 12 (component B) with the second bus line 32. As a result of this interconnection of the high-speed driver modules 12, three bus states (HIGH, LOW and NULL) can be generated between the bus lines 31 and 32.

2 shows a table for driver activation of the output unit 1 and receiver unit 2 according to the invention, from which, among other things, it can be seen how the inputs TxA and TxB of the driver 12 are to be assigned in order to obtain the bus states LOW, NULL and HIGH. For example, for a LOW bus state TxA should be controlled with "1" and at the same time TxB with "0". Such an interconnection of high-speed drivers 12 is only permitted if it is excluded that the two drivers 12 are not actively different Driving potential. According to table Fig. 2, a state must be prevented in which both driver inputs TxA and TxB are not used. Tx = 5 volts is the recessive state and Tx = 0 volts is the dominant state. FIG. 3 shows, in a detail enlarged compared to FIG. 1, an output unit 1 according to the invention, comprising the coding unit 11 for converting a sensor signal DATA into outgoing transmission signals TxA and TxB. The data information to be transmitted is supplied as a sensor signal DATA from the microcontroller 13 to the coding unit 11 via its SPI interface 131 (SPI = Serial Peripheral Interface). This SPI interface 131 makes it possible to read in and output data synchronously via a data and clock line. Downstream of the coding unit 11 are the two high-speed driver modules 12 which are connected in anti-parallel with one another and which serve to connect the output unit 1 to the transmission channel 3 and to convert the outgoing transmission signals TxA and TxB into a signal CHAN to be output. For this purpose, the input TxA is assigned to the first high-speed driver module 12, the input TxB to the second high-seed driver module 12. The driver modules 12 are controlled via the signals TxA and TxB by the coding unit 11, which implements coding regulations according to the invention , Thus, the coding unit 11 works according to a first coding rule for normal operation with a detected equality of the voltages of the outgoing transmission signal TxA with an incoming reception signal RxA. There is also a comparison of TxB with RxB. For this purpose, the output unit 1 has a comparison unit 111, which allows a voltage comparison of the outgoing transmit signals TxA and TxB with incoming receive signals RxA and RxB. Instead of according to the first coding regulation, the coding unit 11 operates according to a second coding regulation when the inequality of the voltages of TxA and RxA or TxB and RxB detected by the comparison unit 111, that is to say in particular when at least one of the bus lines 31 and 32 is externally connected to GND or BAT.

FIG. 4 shows the logic circuit of a high-speed driver module 12, for example the output unit 1 according to FIG. 3, in detail. 5 schematically shows associated signal profiles for the known coding methods already recognized in the introduction to the description.

6 shows the signal curve of the coding method according to the invention for the normal operating mode of a coding unit 11. Each character is represented by a discrete, electrical signal state. 6a shows an exemplary binary sensor signal DATA over time t with exemplary four signal time units (bits), each with a time duration T. The binary character set is exhausted by a "1" and a "0" sign. The "1" sign is identified in the output unit 1 by a 5 volt voltage state, the "0" sign by a 0 volt voltage state. The exemplary sensor signal DATA contains the following characters sequentially: "1", "0", "0", "1".

The outgoing transmission signals TxA and TxB associated with the sensor signal DATA according to FIG. 6a can be seen in FIGS. 6b and 6c and correspond to the tabular values for driver activation shown in FIG. 2. The "1" sign preferably corresponds to a plus 5 volt voltage pulse, the "0" sign preferably corresponds to an O volt voltage state.

The signal CHAN, which is associated with the outgoing transmission signals TxA and TxB and is converted by the driver modules 12, can be seen with regard to its logical course according to FIG. 6d. Basically, a character set with three characters - HIGH, LOW, NULL - is provided for the signal CHAN to be transmitted. Each signal time unit T of the signals DATA, TxA or TxB corresponds to a signal time unit T of the signal CHAN to be transmitted. The bit times are thus the same in all signals, so that there is advantageously no increase or decrease in bandwidth. The HIGH sign preferably corresponds to a plus 2 volt

Voltage pulse, the LOW sign preferably a minus 2 Volt voltage pulse, the ZERO sign preferably corresponds to an O volt voltage state.

The first coding rule for normal operation provides the following rules:

A "1" sign in the DATA sensor signal is always encoded in a HIGH sign in the outgoing transmission signal TxA or TxB. A "0" sign in the DATA sensor signal is generally encoded in a LOW sign in the outgoing transmission signal TxA or TxB. However, if another "0" character follows a "0" character in the input signal DATA, then this further "0" character in the transmit signal TxA or TxB is not coded in a further LOW character, but in a NULL character , The same applies to two successive "1" characters in the

Sensor signal DATA. Here too, a "1" character immediately following a "1" character is encoded by a NULL character in the outgoing transmission signal TxA or TxB.

However, if the preceding character is a ZERO character in the outgoing transmission signal TxA or TxB, coding is carried out according to the basic coding presented above, so that a further "0" character in the DATA sensor signal with a LOW character in the outgoing transmission signal TxA or TxB is encoded, or a following "1" character in the DATA sensor signal is encoded to a HIGH character in the outgoing transmission signal TxA or TxB.

Protection encompasses other encoding variants, of course. B. basically a "0" sign in the sensor signal can be converted into a HIGH sign in the outgoing transmission signal TxA or TxB.

With this type of coding, a change of state can always be generated on the transmission medium 3 between two signal time units. In any case, there is an edge between two bits. Any code protected by the protection It must therefore be ensured that a change of state takes place after each signal time unit.

The outgoing transmission signals TxA and TxB are converted into a signal CHAN to be output by the high-speed driver modules 12 connected in anti-parallel with one another.

FIG. 7 shows the course of the signal CHAN transmitted on a bus 3 logically represented in FIG. 6d with regard to its physical course, i.e. recorded after the

Course of the signal potentials on the individual bus lines 31 and 32, that is, with respect to a two-wire transmission medium. The differential voltage between these two bus lines 31 and 32 provides the signal level of the signal CHAN to be transmitted.

FIG. 8 shows how, in the case of a ternary signal, the termination of a bus line according to GND or Vbat has an effect when coding only according to the first coding regulation. 8d shows the respective bus voltages. So the end of BUS_L to GND or BUS-H to Vbat is tolerated. However, the driver structure does not allow a termination from BUS_H to GND or BUS__L to Vbat. In these two cases, a sufficient bus differential voltage is not generated, which is shown in FIG. 8c - possibly with fatal consequences - which leads to a breakdown in communication. 8b shows the characters sent by the transmitter and FIG. 8a shows the characters recognized by the receiver.

9 shows the signal curve of the coding method according to the invention for the special operating mode of a coding unit

11. FIGS. 9a and 9b show the respective signal curve of TxA and TxB in the event of an unequal tension between TxA and RxA or TxB and RxB. This state can be detected by the comparison unit 111. The comparison unit 111 thus recognizes that a dominant signal cannot be represented on the bus 3 and causes the outgoing signals TxA and TxB to be re-encoded using a time condition. The resulting signal curve is shown in FIG. 9c. For comparison, the dotted line shows the coding in normal operating mode, ie without an external fault on bus 3. A LIGH sign should follow a HIGH sign in the outgoing transmission signal. The Bus_L line 32 was short-circuited to ground (GND) due to, for example, damage to the transmission channel 3 due to an accident. As a result, the Pending LOW character is no longer transferable. This is recognized by the comparison unit 111, which preferably initiates a recoding after half a signal time unit T by changing the voltage of the character LOW which is about to be transmitted. Instead of a minus 2 volt voltage pulse, a pulse 2 volt voltage pulse is now generated, which in the present example begins with half a signal time unit and ends with a full signal time unit. Such a “high character ^λ thus has a time condition which allows a distinction from the characters of the previous character set (LOW, HIGH, NULL). Of course, high characters can also be generated with different expedient time conditions than the one mentioned above as an example. Depending on the application, it has been found to be advantageous to switch the character currently being transmitted (LOW, HIGH) in the range between 40% and 60% or between 30% and 70% of the signal time unit T to the other polarity. It is also advantageous to have the detection of the external fault independent of the time of the polarity change not only after half a signal time unit, but also earlier, for example after 40% or more clearly already after 30% or even 20% of the signal time unit T. Of course, safety routines are also conceivable which, for example, check predefined tolerance ranges in the case of the detected inequality of the voltages of TxA and TxB. Time condition rules designed in this way can thus advantageously take into account a wide variety of general conditions. The second coding rule for the special operating mode provides the following rules:

In the event of an external fault, Bus_L 32 to GND, a LOW character to be transmitted is transcoded into a high character with time condition in the transmit signal TxA or TxB; in the event of an external fault, Bus_L 32 on BAT, a HIGH sign pending for transmission is transcoded into a low sign with time condition in the transmit signal TxA or TxB; in the event of an external short circuit, Bus_H 31 to GND, a HIGH sign pending for transmission is transcoded into a low sign with time condition in the transmit signal TxA or TxB; in the event of an external short circuit Bus_H 31 to BAT, a LOW character pending for transmission in the transmission signal TxA or TxB is transcoded into a high character with time condition; whereby a recessive NULL character is transmitted as a NULL character in each of the above foreign-circuit cases.

As already described for the channel coding for the normal operating mode, the channel coding for the special operating mode also advantageously allows work cycles to be recovered in the receiving unit 2 due to the regular change of state in the signal CHAN to be transmitted without the aid of an additional oscillator.

The coding ratio also does not reduce the signal-to-noise ratio when using the ISO 11898 high-speed layer. At most, there are steeper edges when changing from a HIGH sign to a LOW sign.

FIG. 10 shows how the end of a bus line according to GND or Vbat has an effect when coding according to the second coding rule for a ternary or higher-value signal. With this approach, all four errors can be tolerated in an advantageous manner. The recessive NULL is transferred in all cases. The other two dominant characters

HIGH and LOW the transmitter compares the transmitted signal with the received signal. In the event of an error, these are different. The transmission logic recognizes this and in this case switches to the special operating mode and codes according to the second coding rule, in which only the recessive ZERO and one of the two dominant bus states low or high, but additionally transformed by a time condition, are used. 10a shows what the receiver recognizes; Fig. 10b shows what the transmitter previously sent. The transformation by means of a time condition thus advantageously allows the receiver to interpret the characters differently from what the bus differential voltage, which is shown in FIG. 10c, would normally prescribe. The respective bus voltages are shown in Fig. 10d.

In summary, a binary signal ("0", "1") is encoded into a ternary signal (LOW, HIGH, NULL) or a higher-order signal (LOW, HIGH, NULL, low, high) while maintaining the bit times or corresponding part-time units or recoded, that is, taking into account the change of state in the signal to be transmitted. Finally, the channel encodings according to the invention use at least three characters / states on a transmission bus 3 for the representation of two data characters / states. An overhead of log ₂ 3 = 1.58 = 36% is thus achieved in the value range. In contrast, four states (2 bits) are required for the representation of two data states in a Manchester code. Therefore a 0-head of log ₂ 4 = 2 = 50% is achieved in the time domain.

The coding described above can be implemented by software in a microcontroller or also in hardware, for example in a so-called state machine, which follows the status table according to FIG. 11. Tx is the input signal of the coding unit 11 and thus corresponds to the sensor signal DATA. The signals TxA and TxB correspond to the quantities Q2 and Ql in the table. The transmission signals TxA and TxB are constantly compared with the reception signals RxA and RxB. If a voltage difference is found here, generates a signal "Fehler_A" (FA) or "Fehler_B" (FB), for example at the inputs of a flip-flop.

The following initial equations can be obtained from the state table according to FIG. 11.

TxA = (NOT) Tx + (NOT) Ql * Q2 * (NOT) FA + Ql * (NOT) Q2 * FA; TxB = Tx + Ql * (NOT) Q2 * (NOT) FB + (NOT) Ql * Q2 * FB;

12 shows a conversion of these equations into a logic circuit of an output unit 1. If, for example, TxA is "0" and RxA is "1", it is recognized that the Bus_H line is GND or the Bus_L line is Vbat. The transceiver 12 (component A) cannot transmit a signal. In this case, the inverse state on bus 3 is set by a D flip-flop. This flip-flop is not triggered like known flip-flops with the rising edge, but preferably with the falling edge of a system clock signal (S-CLK) of the SPI interface 131. If the pulse duty factor is 50%, for example, this is preferably done after half the bit time. Likewise, in the case of TxB "0" and RxB "1", the signal is preferably inverted at the latest after half the bit time.

FIG. 13 shows a section of a receiving unit 2 according to the invention, enlarged compared to FIG. 1, comprising the decoding unit 21 for converting incoming received signals RxA and RxB into a working signal DATA. Upstream of the decoding unit 21 are two high-speed driver modules 22, which are connected in antiparallel with one another and shown in more detail in FIG. 4, for connecting the receiving unit 2 to the transmission channel 3 and converting the signal CHAN to be received into incoming reception signals RxA and RxB. This transformation process is based on the state table according to FIG. 2 with respect to the outputs RxA and RxB of the driver 12. The drivers 12 thus serve as a CAN bus transceiver. The signals RxA and RxB are fed to the coding unit 21. The received signals are decoded in the decoding unit 21 and fed as a working signal DATA to the microcontroller 23 for further processing via the interface 231.

The first decoding rule for normal operation provides the following rules: a LOW character in the received signal RxA or RxB is always converted into a "0" character or a "1" -

Characters decoded in the DATA working signal; a HIGH sign in the received signal RxA or RxB is basically decoded into a "1" sign or a "0" sign in the working signal DATA; so that the character in the work signal DATA, which is derived from a NULL character in the receive signal RxA or RxB, is identical to the preceding character "0" or "1" of the work signal DATA.

However, the character currently pending for decoding is interpreted under the condition of a foreign shot if the time between two clock edges occurring is less than 0.6 times to 0.9 times, in particular less than 0.75 times, or is greater than 1.1 times to 1.4 times, in particular greater than 1.25 times, a signal time unit (T).

This second decoding rule for the special operating mode provides the following rules: in the event of an external fault Bus_L 32 to GND, a recoded high character with time condition is decoded into a LOW character; Bus_L 32 on in the event of an external fault

BAT a decoded low character with time condition is decoded into a HIGH character; in the event of an external fault Bus_H 31 at GND, a recoded low character with time condition is decoded into a HIGH character; in the event of an external fault Bus_H 31 at BAT, a recoded high character with time condition is decoded into a LOW character; where a recessive null character in is decoded as a NULL character in each of the above interference cases.

Furthermore, a work cycle STROBE is derived from the received signals RxA or RxB by means of a deriving unit 211, which in turn is fed back to the coding unit 11.

Finally, the decoding unit 21 is operatively connected to a detection unit 212, which detects the

Clock edges permitted from the incoming receive signals RxA and RxB. Thus, the decoding unit 11 operates according to the first decoding rule for normal operation with the clock edges being detected by the detection unit 212 at a defined signal time unit T. According to the second decoding rule for the special operating modes, the decoding unit 21 operates with the asynchrony of the clock edges to the signal time unit T detected by the detection unit 212. This asynchrony corresponds to the time conditions already described above.

The further combined output and reception unit 4 is constructed symmetrically and in turn contains microcontroller 23 with interface 231, a coding unit 11, a decoding unit 21 and two high-speed drivers 22, all of whose functions have already been dealt with.

For example, the data transmission follows the following sequence: The microcontroller 13 sends a data sequence via the SPI interface 131. The coding unit 11 converts these into outgoing transmission signals TxA or TxB, which are sometimes also referred to as so-called tri-state signals (TxA, TxB). The CAN bus transceivers / drivers 12 then use these to generate the corresponding bus states. The bus transceiver 12 of the further combined output and reception unit 4 receives the signal CHAN and converts it accordingly into the signals RxA and RxB. The decoding unit 21 or the one with it deriving unit 211 of the receiver 2, which is operatively connected, generates the working signal DATA, which should otherwise be the same as the input signal DATA, and the working clock STROBE, which are supplied to the microcontroller 23 via the SPI interface 231.

The decoding unit 21 is clocked by the controller 23. The clock must be more than twice the data rate. The clock rate has no upper limit.

All components, in particular the coding 11 and decoding unit 21, can be implemented as hardware or as software in a microcontroller. Of course, the components in operative connection can also be integrated in a common ASIC. Due to the high-speed application, implementation in hardware is found to be particularly advantageous.

Depending on the number of bits to be transmitted in succession and the date, it may happen that the coding unit 11 does not end with a ZERO character but with a LOW character or a HIGH character. In the case of multiple access to the bus medium 3, however, the end bus state must be the idle state ZERO. There are several ways to ensure this: On the one hand, this end bus state can be achieved by a logical condition: If the number of bits of the same name last counted is odd, a pseudo bit of the same name is appended, which causes the coding unit to return to the ZERO state , This function can be carried out either in the microcontroller 13 and / or 23 or in the coding unit 11.

Alternatively, a further time condition can be introduced: the coding unit 11 sets the ZERO state if no state change has occurred after a certain time. 14 shows the process of clock recovery from the two input signals RxA and RxB in a derivation unit 211 that is operatively connected to the decoding unit 21. This is again done by detecting the edges. 14a and 14b correspond to FIGS. 9a and 9b, ie the former show the case of unequal voltages of TxA and RxA or TxB and RxB on bus 3; the latter show the bus differential voltages of characters coded according to the second coding rule. 14c shows clock edges detected with the aid of the detection unit 212. In Fig. 14d, the masking of the edges in the middle of the bit is shown by means of a window, which correspond to a coding without an external fault on the bus. 14e shows the remaining edges. According to FIG. 14f, the signals are delayed and added by one character duration in order to obtain the clock signal. A sampling signal can then be generated from the clock signal, with which the signals RxA and RxB can be sampled (FIG. 14g).

15 shows a table according to which the sampled signals RxA and RxB are assigned by logic, for example to the output value.

The subject of the present invention, which builds on the subject laid down in DE 101 32 048, the content of which is hereby expressly to be included in full, is particularly suitable for use in occupant protection technology for high-speed transmission of sensor data of various types in one Sensor satellite arranged in the motor vehicle and advantageously also ensures data transmission to an evaluation unit arranged, for example, in the vehicle center when the bus line 31, 32 in the CAN transmission channel 3 is subject to an external fault, for example due to an accident-related action the BUS_L- 32 or BUS_H line 31 is connected to GND or Vbat.

Claims

claims

1. Unit (1) for outputting a signal (CHAN) on a transmission channel (3), comprising at least two bus lines (31, 32), in a motor vehicle, with a fault-tolerant coding unit (11) for converting a sensor signal (DATA) in outgoing transmission signals (TxA, TxB); with at least two anti-parallel interconnections connected to the coding unit (11)

Speed driver modules (12) for connecting the output unit (1) to the transmission channel (3) and converting the transmission signals (TxA, TxB) into the signal to be output (CHAN); - With a comparison unit (111) which allows a voltage comparison of the outgoing transmission signals (TxA, TxB) with incoming reception signals (RxA, RxB); with a first coding rule for the normal operating mode of the coding unit (11) with the equality of the voltages of TxA and RxA or TxB and RxB detected by the comparison unit (111); and with a second coding rule for a special operating mode of the coding unit (11) in the event of an inequality in the voltages of TxA and RxA or TxB and RxB detected by the comparison unit (111), in particular if one of the bus lines (31, 32) is connected to GND or BAT; the coding regulations for the outgoing transmission signals (TxA, TxB) provide a character set of at least n + 1 characters (LOW, HIGH, NULL) if the character set for the sensor signal (DATA) has n characters ("0", "1") ,

2. Output unit (1) according to claim 1, wherein each character ("0", "1", LOW, HIGH, NULL, low, high) is represented by a discrete, electrical signal state, wherein in the detected case one External short-circuit in the transmission channel (3) LOW or HIGH characters currently pending for transmission in the transmission signal (TxA, TxB) can be changed with regard to their voltage.

3. Output unit (1) according to claim 1 or 2, wherein the character set for the sensor signal (DATA) has at least two different characters ("0", "1") and the character set for the send (TxA, TxB), receive - (RxA, RxB) and the signal to be output (CHAN) at least three (LOW, HIGH, NULL), preferably four, in particular five (LOW, HIGH, NULL, low, high), different characters.

4. Output unit (1) according to claim 1 to 3, in which the detection of an external fault in the transmission channel (3) takes place after half a signal time unit (T) at the latest, preferably after 40%, in particular after 30% of the signal time unit (T) at the latest ,

5. Output unit (1) according to one of the preceding claims, in which upon detection of an external fault, the change in the characters (LOW, HIGH) takes place in such a way that between 30% and 70% of the signal time unit (T), preferably between 40% and 60% , in particular at 50% of the signal time unit (T), the pending transmission

Character (LOW, HIGH) can be switched to another polarity.

6. Output unit (1) according to one of the preceding claims, in which the second coding rule is designed such that: in the event of an external fault Bus_L (32) to GND, a LOW character to be transmitted in the transmission signal (TxA, TxB) is high - character is transcoded with time condition; in the event of an external short circuit Bus_L (32) to BAT, a HIGH sign in the transmission signal (TxA, TxB) is transcoded into a low character with time condition; in the event of an external short-circuit Bus_H (31) to GND, a HIGH sign pending for transmission in the transmission signal (TxA, TxB) is converted into a low sign with time condition; in the event of an external short circuit Bus_H (31) to BAT, a LOW character pending for transmission in the transmission signal (TxA, TxB) is transcoded into a high character with time condition; and

- A recessive NULL character is transmitted as a NULL character in each of the above cases of interference.

7. Output unit (1) according to one of the preceding claims, in which at least the first coding specification for the signal time unit of the sensor signal (DATA) occupied by a character is a signal time unit with the same duration (T) in the transmission (TxA, TxB), reception ( RxA, RxB) and signal to be output (CHAN).

Output unit (1) according to one of the preceding claims, in which both coding regulations for two successive signal time units (T) provide different characters in the transmission signal (TxA, TxB).

9. Output unit (1) according to one of the preceding claims, in which the first coding rule is designed in such a way that a "0" character in the sensor signal (DATA) is basically a LOW character or a HIGH character in the

Transmission signal (TxA, TxB) is encoded; that a "1" character in the sensor signal (DATA) is basically encoded in a HIGH character or a LOW character in the transmission signal (TxA, TxB); - that a "0" character following a "0" character in the sensor signal (DATA) is encoded into a NULL character in the transmission signal (TxA, TxB), unless the previous the resulting character in the transmission signal (TxA, TxB) was a ZERO character; that a "1" character following a "1" character in the sensor signal (DATA) is encoded into a ZERO character in the transmission signal (TxA, TxB), unless the preceding character in the transmission signal (TxA, TxB) is already a ZERO - sign was; and that coding is carried out according to the basic coding if the preceding character in the transmission signal (TxA, TxB) was a ZERO character.

10. Unit (2) for receiving a signal (CHAN) from a transmission channel (3), comprising at least two bus lines (31, 32), in a motor vehicle, - with a decoding unit (21) for converting incoming reception signals (RxA, RxB) into a working signal (DATA); - with at least two of the decoding unit (21) upstream antiparallel interconnected with one another high-speed driver blocks (12) for connecting the receiving unit (2) to the transmission channel (3) and around ^¬ setting the assumed signal (CHAN) in incoming received signals (RxA, RxB); with a detection unit (212) which detects clock edges from the incoming signals

(RxA, RxB). with a first decoding rule for the normal operating mode of the decoding unit (21) with synchronism of the clock edges detected by the detection unit (212) at a defined signal time unit (T); with a second decoding rule for a special operating mode of the decoding unit (21) with asynchrony of the clock edges to the signal time unit (T) detected by the detection unit (212); - The decoding rules for the working signal (DATA) provide a character set of n characters ("0", "1") if the character set ^' for the incoming Receive signals (RxA, RxB) have at least n + 1 characters (LOW, HIGH, NULL).

11. receiving unit (2) according to claim 10, wherein each character ("0", "1", LOW, HIGH, NULL, low, high) is represented by a discrete, electrical signal state.

12. receiving unit (2) according to claim 10 or 11, wherein the character set for the working signal (DATA) has at least two different characters ("0", "1") and the character set for the receive (RxA, RxB) and that assumed signal (CHAN) at least three (LOW, HIGH, NULL), preferably four, in particular five (LOW, HIGH, NULL, low, high), different characters.

13. Receiver unit (2) according to one of claims 10 to 12, in which, if the time between two clock edges occurring is less than 0.6 times to 0.9 times, in particular less than 0.75 times , or greater than 1.1 times to 1.4 times, in particular greater than 1.25 times, a signal time unit (T) that is interpreted just pending characters under the condition of a foreign shot.

14. Receiving unit (2) according to one of claims 10 to 13, in which the second decoding rule is designed in such a way that, in the event of an external fault Bus_L (32) to GND, a recoded high character with time condition is decoded into a LOW character; in the event of an external fault Bus_L (32) at BAT, a recoded low character with time condition is decoded into a HIGH character; - In the event of an external fault, Bus_H (31) at GND, a recoded low character with time condition is decoded into a HIGH character; in the event of an external fault Bus_H (31) at BAT, a recoded high character with time condition is decoded into a LOW character; and a recessive NULL character is decoded as a NULL character in each of the above leakage cases.

15. receiving unit (2) according to one of claims 10 to 14, in which at least the first decoding rule for the signal time unit occupied by a character of the received signal (RxA, RxB) and the signal to be accepted (CHAN) is a signal time unit with the same duration (T ) in the working signal (DATA).

16. receiving unit (2) according to one of claims 10 to 15, wherein the first decoding rule is designed such that a LOW character in the received signal (RxA, RxB) in principle in a "0" character or a "1" character is decoded in the working signal (DATA), - that a HIGH character in the received signal (RxA, RxB) is generally decoded into a "1" character or a "0" character in the working signal (DATA), that the character in the working signal ( DATA), which is derived from a NULL character in the received signal (RxA, RxB), is identical to the previous one

Character ("0" or "1") of the working signal (DATA).

17. Em receiving unit (2) according to one of claims 10 to 16, with a unit (211) for deriving a clock signal (STROBE) from the incoming signals (RxA, RxB).

18. Arrangement (4) for data transmission in a motor vehicle via a transmission channel (3), comprising at least two bus lines (31, 32), - with an output unit (1) according to one of claims 1 to 9; and with a receiving unit (2) according to one of claims 10 to 17.

19. A method for data transmission in a motor vehicle, in which a sensor signal (DATA) is encoded by means of an output unit (1) according to one of claims 1 to 9 into a signal (CHAN) to be transmitted; in which signals (CHAN) thus formed are transmitted to a receiving unit (2).

20. A method for receiving data in a motor vehicle, in which a signal (CHAN), in particular formed according to claim 19, is decoded into a working signal (DATA) by means of a receiving unit (2) according to one of claims 10 to 17.