DE10058968A1 - Coding and decoding data involves specifying information by time duration of signal in first or second signal states within time period - Google Patents

Abstract

The method involves converting the data into signals that can adopt a first and a second state. Several data items representing an information item are converted into a signal, whereby the information is specified by the time duration of the signal in the first or second state within a time period. The data are contained in a data stream with rising and falling edges synchronized to a clock signal.

Description

The invention relates to a method for coding data or for decoding signals.

Due to the growing data transfer rates between logic and / or memory modules are used for the secure detection time window always available smaller. With the 1 GHz transmission already achieved today frequencies change data signals at a time interval of 1 ns. Thus stand for the safe detection of the signal level usually less than 0.5 ns available.

When the transmission rate is doubled by using Double data rate (double data rate) data with stei leading and falling edges of a clock signal. Therefore, the transmitted data signal must be compared to the clock signal with an accuracy of 250 ps are available. The accuracy requirement is met in particular by Differences in transit times of the signals are more difficult. This will make the limited maximum transmission frequency, which leads to the occurrence of Transmission errors.

It is therefore an object of this invention to provide an improved Ver drive to encode and decode data signals see with which higher transmission frequencies are achieved can and the probability of the occurrence of Feh learning is reduced.

This task is accomplished by the method of coding according to An saying 1 and the method for decoding according to claim 4 ge solves. Further advantageous embodiments of this invention are specified in the dependent claims.  

According to an inventive method for coding Da It is envisaged that the data will be converted into signals be a first state and a second state can accept. Multiple data that present information animals are converted into a signal. The In formation by the duration of the signal in the first or in the second state within a clock period.

This method has the advantage of being a coding method is provided with which two or more Data e.g. B. transmitted within a clock period have it encoded into a signal. In contrast to uncoding With the coded signals it is not necessary to that flanks that change when the state changes first and second states occur at a clock edge are synchronized. It should only be provided that the encoded signal during the clock period for the predetermined one duration in the first or second state remains there with the associated information is clearly determined. This Coding method has the advantage of being a coded signal is generated that on the recipient side only by a De coding process converted back into the output information can be. This represents an improvement over the un encoded data, since the several data must be detected individually. I.e. at egg ner data transmission with double data rate that on each Edge of the clock signal a date must be detected. in the In contrast, the encoded signal can be replaced by a single one Decoding process recognized and recovered in the output data be delt.

According to a preferred embodiment, the Co dierverfahren coded two data each. in particular especially in the case of transmission methods with two data each during a clock period with twice the data rate the method according to the invention is special suitable. These data are according to the Ver  drive coded in signals. These signals are designed that they can represent four different pieces of information NEN. The signals are preferably designed such that Representation of information the signal for a temporal Duration of 0/3, 1/3, 2/3 or 3/3 of the clock period in the first or remains in the second state. Let it this way the data in a signal with four different informa encode content. This advantageously corresponds to Transmission time of the signal of the transmission time of the two Data, i.e. H. the signal has a duration of one clock period, thus coding the data stream with the output data can run without intermediate data storage.

According to a further embodiment of the invention, a Method for decoding signals provided, the Signals a first state and a second state can take. The duration of the first or two Information is assigned to the th state, the In formation is converted into multiple data.

In this way it is advantageously achieved that the coded Signals are converted into the corresponding multiple data that can. In this way, the data for one after following processing made accessible. By this ver driving is still achieved on the receiving end a lower requirement for the accuracy of the signal necessary is. It is also advantageous that this De coding method is scalable, d. H. it can higher data transfer rates can be achieved if the on solution of the coded signal and the accuracy of the transmitter the decoding of more than four different pulse widths allowed within one clock period.

According to a further preferred embodiment, it is provided see that the assignment of information to temporal Duration of the first or second state by a temporal Integration of the signal carried out over a clock period  the integration result is supplied with information is arranged. The coding of the signal is that several states can be assumed within the signal that are pending for a certain time. These states are usually shown with voltage levels. To a To convert such a signal back, according to the invention, an In tegrator be provided depending on the voltage level and duration of the applied voltage level an int egration over a clock period. The result of Integration represents the transmitted information. there it is irrelevant at what point in time the particular To stood during the clock period is assumed. It is single Lich necessary that the certain state for a certain Duration exists within the clock period. For this Basically, it is advantageous to use an integrator for decoding to use the coded signal. To such an integra the resetting of the integrator capacity and the decoration To enable the result to be determined, it makes sense to provide several integrators, one after the other the following encode and decode signals.

The advantageous embodiments are based on the drawing nations explained in more detail. Show it

Figure 1 shows an encoding or decoding system which operates according to the method according to the invention.

Fig. 2 shows a possible data transmission protocol double data rate; and

Fig. 3 coded signals, which have been encoded according to the inventive method.

Fig. 1 shows a block diagram of an apparatus, encoded in the signals by the inventive method and is decoded by the inventive method. It is provided that data bits in an input data stream DATA are present on a data input line 2 in synchronism with a clock signal CLK on a clock signal line 1 . The input data on the data input line 2 are read into a latch 3 with each rising edge of the clock signal and into another latch 4 with each falling edge. The outputs of the latch 3 , 4 are each connected to the data inputs D1, D2 of a coding device 5 .

The coding device 5 is also connected to the clock signal line 1 . As soon as the input data are present at the data inputs D1, D2 of the coding device 5 , the data in the latches 3 , 4 are taken over into the coding device 5 with the next rising edge of the clock signal CLK and coding is carried out by changing the four possible states to the A pulse length is assigned to data inputs D1, D2. So z. B. provided that the states "00" at the data inputs D1, D2 by a sustained low signal, "01" with a high signal whose length corresponds to a third of the clock length, "10" with a high signal, the Length corresponds to two thirds of the clock length, and "11" are encoded with a high signal, the length of which corresponds to the total clock length. The encoded signal is then output on a transmission line 6 , it not being necessary for the encoded signal to be synchronous with the clock signal CLK.

Before a coded signal can be encoded on the transmission line 6 , the clock signal on the clock signal line 1 must first be synchronized in an initialization step by z. B. the rising edges of a coded calibration signal with the rising edges of the clock signal CLK are brought into temporal agreement. This takes place in a phase shifter 7 , which generates a signal CLK 'adapted to the transmitted calibration signal on the transmission line 6 on a further clock signal line 8 . The synchronization can also take place with respect to one of their regular characteristics of the calibration signals. The clock signal line 8 and the transmission line 6 are connected to a decoding device 9 .

When decoding, the decoding device 9 decodes the coded signal on the transmission line 6 by, for example, converting the applied signal by time integration over a clock period corresponding to the pulse length into a voltage signal. The size of this voltage signal corresponds essentially to the pulse duration on the transmission line 6th The amount of voltage means that information can be clearly assigned to a coded signal. The information is converted back into two data bits in accordance with the scheme described above and output via the data outputs B1, B2 of the decoding device 9 . By means of a suitable demultiplexer device (not shown), the data on the data outputs B1, B2 can be combined again to form a data stream.

The decoding device 9 comprises a plurality of integration devices, with successive coded signals being assigned to different integration devices. In this way, each integration device can be reset and evaluated after an integration in the subsequent clock period.

In FIG. 2, a data transmission double data rate is shown in a state diagram. During a clock cycle ses T, data is transmitted with every edge of the clock signal CLK. The data DATA is preferably present for a certain time before and after the arrival of an edge of the clock signal, so that the respective date is safe, e.g. B. can be adopted in a subsequent latch. If there are shifts in the data signal and clock signal due to interference or delay differences, transmission errors can occur if a shift of approximately a quarter of the clock cycle length or greater occurs between an edge of a clock signal and the time when data of the data signal is present. With this type of data transmission, care must be taken that the data signal is synchronized with the edges of a clock signal at all times.

In contrast, in Fig. 3, the coded signals are Darge, which have been generated with the inventive method. It can be seen that, depending on the information to be displayed, the coded signal CODE is during a clock cycle of the clock signal CLK a certain duration on high and a certain duration on low. According to the coding listed above, the coded signal corresponds to the 2-bit information "00" if it is high during 0/3 of the clock length, "01" if it is high and 2/3 during 1/3 of the clock length the clock length is low, "10" if it is 2/3 high and 1/3 low and "11" if it is high for most of the clock length. It can also be seen that the length of time that the coded signal CODE is high during a clock cycle does not start at the beginning of the clock period, but can also be present at any time within the clock period. It is only essential that the total share within the relevant clock period, during which a state of the coded signal is assumed, corresponds to the assigned share according to the coding of the signal.

When receiving the coded signal, only the pulse has to length within the period is accurate to +/- 1/6 period be true, and it is done unlike the uncoded Data is only decoded within one clock period. This ensures that the permitted inaccuracy of the Pulse length may be 1/3 of the clock period, while at the transmission with double data rate described above the data signal with an accuracy of ¼ to the position of the Clock edge would have to be transmitted.

The in the preceding description, the claims and the Drawings disclosed features of the invention can both individually as well as in any combination for the forfeiture  Lichung the invention in its various configurations be essential.  

LIST OF REFERENCE NUMBERS

1

Clock signal line

2

Input data line

3

latch

4

latch
D1, D2 data inputs

5

coding

6

transmission line

7

phase shifter

8th

Clock signal line

9

Decoding device
DATA input data
CLK clock signal

Claims (7)

1. A method for encoding data, the data being converted into signals which can assume a first state and a second state, characterized in that a plurality of data representing information is converted into a signal by the information being transmitted by the duration of the signal in the first or in the second state is fixed within a clock period. 2. The method according to claim 1, characterized in that the number of data coded together is two. 3. The method according to any one of the preceding claims, since characterized by that the data before in a data stream lie, with the rising and falling edges of the data are synchronized to a clock signal in the data stream. 4. Method of decoding signals, the signals can assume a first state and a second state NEN, characterized in that the duration of the first or second state an information is assigned, the information in multiple data is converted. 5. The method according to claim 4, characterized in that the multiple data are synchronized to a clock signal, so the multiple data during a period of the clock signals are made available. 6. The method according to claim 4 or 5, characterized in net that the assignment of information to the duration of the first or second state by a temporal In tegration of the signal over a clock period  the integration result is supplied with information is arranged. 7. The method according to any one of the preceding claims, since characterized by that the data corresponds to data bits.