Line code

The line code line code and sets in the digital telecommunications, as used to transmit information symbols are transmitted on the physical plane. Certain consequences level, as light intensities on glass fibers or voltages or currents to the electric lines, the binary bit sequences are assigned to the data stream.

General

The object of the line coding (or line coding ), the signal to be transmitted to form spectrally so as optimally as possible to adapt it to the characteristics of a transmission medium. For example, to suppress the DC voltage component. In addition, clock recovery is possible. In some line codes the required link bandwidth is reduced. Transmission lines can be better utilized so because the transmission range depends on the cause, together with the insertion loss. To a first approximation for a metal cable that the damping constant is proportional to the square root of the frequency: higher frequencies are attenuated more than low frequencies. Therefore, the code should be at the lowest possible frequency of the maximum of the power density and spectral curve having a small bandwidth.

Not to be confused is the line coding with the channel coding or the source coding: Channel coding has the task to detect the addition of extra redundancy carry or memory error during the data transfer and data storage and correct can, while the source code unnecessary (redundant ) information of a reduces data source and referred to as data compression.

Some line codes are DC component, i.e., the time-averaged value is 0, which is important if the electric transmission of DC voltage across the channel is not possible in a particular application. The need for equal share of freedom can be specified for example by pulse transformers for galvanic isolation in the transmission, which can not pass DC.

Binary line codes

Also in the circuitry is the simplest case the logic states 0 and 1 are assigned to a logic level on the physical line, and is also referred to as a non return to zero (NRZ ), i.e., there is no condition on the line which carries no information.

For example corresponds to the EIA-232 serial interface ( RS -232) from the 1960s about a negative voltage to a logic one, a positive voltage to a logic zero, this is also referred to as antipodal coding. Now be all ones transmitted on the line does nothing. This means that only a few bits for asynchronous clock can be transmitted in a block that are marked with a start sequence, or is it an additional clock line for synchronization necessary. EIA-232 writes a 0 bit ( the start bit ) to the start of each sequence of 5 to 8 bits of user data at the moment ( plus any parity bit ) and a one or two bit ( stop bit ) to its end. The stop bits can be extended as a constant 1 therefore means "no data transmission ". This encoding is not DC-free ( depending on the data content can 1- bits or 0 bits clearly predominate), which long cables are impossible. EIA- 232 compensates for this disadvantage partly by the use of rather high voltages (typically ± 12 V), but this brings its own problems and has therefore not enforced in newer standards.

Serial transmission (EIA -232):

Start Stop    Data bits: 1 1 0 1 0 0 0 1    Line bits: .... 0 1 1 0 1 0 0 0 1 1 1 ....    Voltage level: ---- - - ------- Manchester encoding

When Manchester coding corresponds a zero-one result of a logical zero (rising edge ), a one-zero sequence ( falling edge) to a logical one:

Hereby is achieved that

It doubles the required symbol rate on the transmission channel.

In differential Manchester code polarity change is at the beginning of a measure for a logical zero ( two edge change per bit ) at a logic one no polarity change takes place at the beginning of a measure ( an edge change per bit ).

Clock recovery

By the continuous change of level, a clock recovery is possible. A separate line for the clock is therefore unnecessary. In the case of Manchester coding is a start sequence with a sequence of many one-zero data sequences ( 10-30 ) - 1010101010 ... on the line - which effectively corresponds to a 11111 - sequence at the input of the encoder - a phase-locked loop ( PLL) synchronized in the receiver. A one-zero sequence (1001 on line ) starts the actual data transmission. Followed by a packet of several thousand bits.

Clock recovery PLL by Manchester coding:

PLL Synchronize | PLL will remain synchronized                                            start Data bits: 0 1 1 0 0 0 0 0 0 1 ... Sync / Data: 1 1 1 1 1 ... 1 1 1 1 0 1 0 0 1 1 0 0 0 0 0 0 1 ... Line bits: 1010101010 1010101001 1001 ... 01101001010101010110 ... Level: - - - - - ... - - - - - - - - - - - - ... Encodings with more bits long blocks

The line code has in the Manchester encoding twice as many bits as the data stream, it is unfavorable in terms of the required data rate, bandwidth depends only on the edge of a bit from the steeper this slope is, the more bandwidth is required. Binary transfers are thus not suitable for channels with limited bandwidth ( eg, air); a channel encoder must be followed. Other encodings, about 4 to 5 bits coded bits ( 4B5B ), here are better error detection is next to it on several bits possible:

Instead of two ( / - binary) can in electrical transmission and three ( / = / -, ternary) are used or more voltage levels.

Block- code (4 bits 3 ternary level 4B3T ):

Data bits: 0 1 1 0 0 0 1 0   Data blocks: | 0110 | 0010 |   Ternary: | == | - = |   Voltage level: 0 0 - 0 Block Codes

So-called block codes are usually referred to by pBqX. In this case, "p" bits of a binary word, and are combined to form a block of q coded length in an X- display. The advantage of this approach is that the modulation rate is reduced by a factor of this code "p", thereby, the damping constant is reduced, and a greater transmission range is possible. However, the signal is also susceptible to interference with increasing number of stages.

• 4B5B is a code that maps each having 4 bits to 5 bits, or 16 (24) block code to 32 (25) line codes. The code is not dc-free. For Fast Ethernet 100BASE -TX is achieved by using MLT -3 with three voltage levels and the use of a scrambler under favorable circumstances dc freedom. In MLT -3, the required bandwidth of the transmission channel is reduced in comparison with simpler methods such as Manchester coding. A 4B5B coding plus subsequent MLT-3 coding, without further measures can not achieve a DC free. This can be checked through trial and error of the 32 possible 5 -bit code words.

• 4B3T forms corresponding to 4 bits each to a group of 3 with three voltage levels from ( Ternary signal), and thus 16 ( 24) block codes on 27 (33 ) line codes. Some block codes are mapped to multiple line codes so that "better " by changing the code may be the current DC component in each case reduced. To the sign of the fraction is accumulated in the transmission, the sum of influences the choice of the line code. Features are 4B3T clock recovery, low direct-current component and bandwidth reduction. This coding is used at 34 Mbit / s ( E3 ), and 140 Mbit / s (E4) and for the Koaxialübertragungsstrecken Uk0 interface of the ISDN basic access.

• The Manchester code is called 1B2B (1 bit to 2 bits ), namely, 2 (21) block code to 4 ( 22) line codes are used by the four although only two of line codes, 01 and 10, since the two other, no voltage change included. The coding is simple, it is DC-free and allows the clock recovery. The required link bandwidth is larger than any code.

• 2B1Q respectively forms two (22 ) bits to a voltage level of 4 (41) from steps ( Quaternary code). The code enabling bandwidth reduction, but no clock recovery and no DC free.

Conversion to voltage or intensity level

The encoding of a line code as 4B5B or 8B10B converts a binary sequence into another binary sequence. The encoded bit string must be converted to intensities then. With NRZ ( non return to zero english ) the simple assignment of logic values ​​(0 and 1) is referred to two levels. In NRZ -I also only two levels are used, the logic 1 will, however, associated with a level change, while a logic 0 level keeps constant. FDDI uses in transmission over optical fibers, for example, 4B5B and only then NRZ -I. In the electrical transmission of Ethernet with 100 Mbit / s over twisted -pair cable (100 - Base-TX ) 4B5B is also used initially. It is to reduce the DC component of the three - level Code MLT- 3 is added ( Pseudoternärsignal ). But the little used Ethernet over fiber (100 - Base -FX) used 4B5B/NRZ-I.

Known line codes (selection)

• NRZ code
• RZ code ( development of the NRZ code )
• MLT -3 code ( DC poor implementation of binary sequences on three voltage levels )
• Manchester - code (10 Mb / s Ethernet) and differential Manchester - code ( token ring)
• Biphase mark code
• Run Length Limited codes (RLL ), such as: 4B5B code (FDDI, 100 Mb / s Ethernet)
• 8B10B code (Gigabit Ethernet)