Method for synchronising bi-directional data transmission

[0001] The invention relates to a method of time synchronization between a start pulse and data transmission between a transmitter and a receiver in accordance with the characterizing part of claim 1.

[0002] Two-way data transmission methods are known using which data is transmitted between a transmitter and a receiver. They are based on the transmission of packets with a predetermined bit length, with each packet having a first bit length for transmitting data from transmitter to receiver and a specified second bit length for transmitting data from receiver to transmitter. For example, in a first bit length, the transmitter may request a data receiver, which receiver then transmits to the transmitter in a second bit length. In parallel and regardless of the two-way data transmission method, the start pulse is transmitted from the transmitter to the receiver, with the help of which the reception of the data is triggered in the receiver, e.g. The start pulse is completely independent of the data transmission between the transmitter and receiver, and in particular its length is not exactly matched to the length of the packet. In known methods, the start pulse is digitized and then transmitted, with bit-length temporal inaccuracies, for example about 1 μs, as a result of digitization. However, this temporary inaccuracy is unacceptable in many applications, e.g. when measuring high-accuracy positions, for example with a coding device.

[0003] The prior art is patent applications EP 1 434 382 A1, US 5,428,603 and EP 0 100 386 A1.

[0004] The object of the present invention is to provide a method that allows as close synchronization as possible between the start pulse and data transmission between the transmitter and receiver.

[0005] The object of the present invention has been met by the method of implementation defined by the features of claim 1.

EP 1 855 413 [0006] Preferred embodiments and methods for carrying out the present invention are set out in the dependent claims.

[0007] The method of synchronizing bidirectional data transmission between a transmitter and a receiver according to the invention based on the transmission of packets of a certain bit length is characterized in that the packet has a time interval in which data is not transmitted from either the transmitter to the receiver or from the receiver to the transmitter, and in that in this time interval the start pulse is transmitted from the transmitter to the receiver. Therefore, digitization of the start pulse is unnecessary because the start pulse can be directly transmitted to the receiver within the time interval. Thanks to this, it is possible to avoid temporary inaccuracy resulting from digitization and ensure accurate synchronization between the start pulse and data reception, because exactly at the time at which the edge of the start pulse reaches the receiver, the data is received without time delays, which are caused by digitization of the start pulse and its transmission carried out independently of the data transmission.

[0008] In the present invention, the time interval is placed inside the packet in the first bit length. This ensures that no data will be transmitted from the receiver to the transmitter, and thus no additional time delay bits will be needed.

[0009] The length of the time interval is preferably from 3 to 8 bits. Thanks to this, a sufficiently long time window is available in which the start pulse can be transmitted, and in particular the start pulse edge that starts receiving data in the receiver.

[0010] The first bit length and the second bit length are preferably placed in the packet with an interval of several time delay bits, preferably 4 to 7 time delay bits. This avoids that the receiver tries to transmit data to the transmitter while the data is still being transmitted from the transmitter to the receiver.

[0011] The start pulse length usually does not exactly match the packet length. In this case, during the first transmission of the start pulse, the transmission would occur within the intended time interval in the packet, but during a larger number of transmissions the pulse would be shifted

EP 1 855 413 within the time slot. To guarantee the transmission of the start pulse always within the prescribed time interval, also in the case of differences between the length of the start pulse and the length of the packet, in a particularly preferred embodiment of the invention, one or more filler bits are inserted, depending on the length of the start pulse. The number of padding bits needed for insertion is determined by the offset of the start pulse position within the time interval between previous consecutive packets.

[0012] The inserted padding bits are preferably evenly distributed over a larger number of packets, in particular when a large number of padding bits are needed so that, despite the insertion of additional bits, substantially uniform data transmission is guaranteed.

[0013] When the number of padding bits is greater than or equal to N, the padding bits are preferably replaced by a synchronization packet. These types of synchronization packets contain additional information regarding synchronization between the transmitter and the receiver, e.g. a predefined bit order, which can be recognized by the receiver and used for further synchronization.

[0014] N is preferably 5 or 6, because already a bit order of 5 or 6 bits is sufficient to exchange further information regarding synchronization between transmitter and receiver.

[0015] The standard length of the first bit length is at least 14 bits and the length of the second bit length is 30 bits. The packet length is at least 60 bits and at most 120 bits depending on the number of padding bits inserted.

[0016] The data are preferably exchanged between the transmitter controller and the receiver measuring assemblies, the measuring assemblies being particularly preferably configured as position measuring assemblies, because when determining position, in particular in rotary decoders, high precision and accurate synchronization between the start pulse are required and data transmission between transmitter and receiver. The system according to the invention consists of a transmitter and a receiver, between which data is transmitted bidirectionally, with the data transmission being carried out according to the method in accordance with one of the attached

EP 1 855 413 patent claims. The data are preferably exchanged between the transmitter controller and the receiver measuring units, the measuring units being constructed in particular as position measuring units, mainly as rotary decoders.

[0017] The invention has been explained based on the attached drawings, in which

Fig. 1 shows a schematic representation of a packet according to the invention.

[0018] Fig. 1 shows a packet 10 with a regular, and thus a minimum, bit length of 60 bits. Inside the packet 10, at its beginning, the first bit length 20 is placed, which length is usually at least 14 bits, and which is intended for data transmission from the transmitter to the receiver. In this case, for example, there may be a data request.

[0019] Inside the packet 10, the data is transmitted bidirectionally. Therefore, the first bit length 20 is connected to the second bit length 30, which length is usually 30 bits, and which is provided for data transmission from the receiver to the transmitter. In the second bit length, for example, the data requested by the transmitter is transmitted from the receiver to the transmitter. The second bit length 30 is graphically represented as vertically displaced from the first bit length 20 to show that in this case the data is transmitted over the data transfer link in the reverse direction.

[0020] The first bit length 20 and the second bit length 30 are preferably separated from each other by several bits of time delay 60, preferably by 4 to a maximum of 7 bits of time delay 60, which ensure that data transmission from the receiver to the transmitter does not overlap for data transmission from transmitter to receiver. With this 7 bits of time delay 60 corresponds to a time delay of 1.12 μs.

[0021] Also, the second bit length 30 has several time delay bits 65, preferably from 4 to a maximum of 7 time delay bits 65 attached, to exclude the possibility of overlapping data transmission in this case as well.

[0022] Inside the first bit length 20 there is a time interval 50 within which data is not transmitted from the transmitter to the receiver or

Inversely, and in which time interval the start pulse, in particular the slope of the start pulse, is transmitted directly from the transmitter controller to the receiver. This makes digitization of the start pulse unnecessary, which could cause a time shift. Accurate synchronization between the start pulse and data reception initiated by the start pulse is therefore guaranteed. The length of the time interval 50 is from about 3 to 8 bits, so that also with possible shifts of the start pulse 40 within the time interval 50, the transmission of the start pulse 40 within the time interval 50 is guaranteed. The length of the first bit length 20 increases as a result of the standard 14 bits by the corresponding the number of bits for the time interval 50 and therefore the length of the first bit length 20 is defined as with "at least 14 bits".

[0023] If the length of the start pulse 40 and the length of the packet 10 exactly match, the slope of the start pulse 40 within each packet 10 is transmitted relative to the packet 10 at the same position within the time interval 50. The above case, however, usually does not occur because the pulse start 40 is produced irrespective of the data transmission between the transmitter and receiver. Therefore, despite the fact that during the first transmission of the start pulse 40 there would be a transmission within the prescribed time interval 50 within the packet 10, after the transmission of several packets 10 the start pulse 40 would shift within the time interval 50. Namely, the transmission of the start pulse would have occurred earlier compared to the transmission in the previous packet 10, in the event that the length of the start pulses 40 was shorter than the packets 10, and would have followed later than the transmission in the previous packet 10, if the length of the start pulses 40 was greater than package length 10. By comparing the positions of the start pulse 40, each time in two consecutive packets 10, you can determine the differences between the length of the start pulse 40 and the length of the packet 10. To guarantee transmission of the start pulse 40 always within the prescribed time interval 50, also in the event of differences between the length of the start pulses 40 and the length of the packet 10, one or more filler bits 70 are inserted immediately before the first bit length 20, depending on the length of the start pulses 40 The number of fill bits 70 that should be inserted is determined

EP 1 855 413 based on the offset of the start pulse position 40 within the time interval 50 between successive previous successive packets 10, since the location of the start pulse pulse 40 relative to the packet 10 within the time interval 40 is known from previous packets 10. The value by which the start pulse edge 40 was shifted. However, the maximum number of fill bits 70 is included in the packet 10, namely 60. Such a large number of fill bits 70 is necessary when the position of the start pulse edge 40 within the time interval 50, compared to the previous packet, is slightly earlier, than expected. If multiple filler bits 70 of this type are inserted, they will preferably be spread over more packets 10. If 50 filler bits 70 are inserted, they will preferably be added not as a block of 50 filler bits 70 before the following first bit length 20, but for example in 5 successive packets of 10 in blocks of 10 filler bits 70 each, or in 10 consecutive packets of 10 in blocks with 5 filler bits of 70 each time. This guarantees substantially uniform data transmission when inserting more filler bits 70.

[0024] To use filler bits 70 also for the transmission of useful information, N filler bits 70 are replaced by the synchronization packet whenever the number of filler bits becomes greater than or equal to N. Such synchronization packets contain additional information relating to synchronization between the transmitter and the receiver, for example a predefined bit string that can be recognized by the receiver and used for further synchronization. This predefined bit string should at the same time show as large changes in bit values ​​as possible and should not contain any fixed part so as to guarantee the most reliable transmission and recognition of this bit string coming from the receiver, as well as their accurate synchronization. Preferably, N is 5 or 6, because already a bit string of 5 or 6 bits is sufficient to exchange further information regarding synchronization between the transmitter and the receiver.

[0025] The above-described embodiment is preferably used for data transmission between the transmitter controller and the receiver measuring units. The measuring units are configured in particular as measuring units

EP 1 855 413 positions. It is precisely in the case of position measuring units, for example in rotary decoders, that the time requirements are particularly high, so that an accurate position determination can be guaranteed so that there is, particularly advantageously, accurate synchronization between the start pulse and data transmission between the transmitter and the receiver.

Proxy:

Ewa Grenda, patent attorney

EP 1 855 413