Coding schemes

Another aspect of digital transmission systems is the algorithm used for data transmission on the line. There are various ways in which the data can be represented as signals on the line. One of the important factors of a transmission system is the number of discrete levels a signal can have. For example, TTL has two levels to encode the binary data. One voltage level indicates a logic 1, and the other voltage level a logic 0. A single switching point, or decision level, is used to set the threshold. By comparing the coded signal to the threshold level, the binary information can be decoded. If the signal level is above the threshold, then a logic 1 is detected; conversely, if it is lower, a logic 0 is detected. Because the signal has two levels, it is known as a unipolar code. Unipolar coding is a very good system for TTL systems but is not suitable for transmission systems using copper cables over long distances.

The reason why lies in the construction and characteristics of such cables. A transmission cable is made of a group of wires, having two electrical properties. One is the d.c. resistance and the second is the self inductance value of the wire. When a group of wires is placed in a casing to form the cable, a third electrical property, capacitance between the wires, is brought to bear. The transmission cable now behaves like a low pass filter.

If a unipolar code is transmitted down the line, each high level signal will inject energy into the capacitor/inductor of the cable. Conversely, a low level signal will discharge the line. If the number of highs and lows is matched, then the net energy level on the line will be zero. But if the number is not matched (which is normally the case), the transmission line will have periods during which energy is stored. This stored energy will result in a d.c. offset being superimposed on the transmitted signal. The d.c. offset will interfere with the decoding of the received pulse by reducing the difference between the high and low signal levels and the threshold level. For instance, if the threshold level is 1.3 V and a low level signal is 0.8V, then a DC offset of greater than 0.5 V will stop the decoding of a low level signal and affect the overall performance of the transmission system.

To overcome this, transmission codes with three levels, positive high, zero, and negative low, have been developed. The highs and lows are used to represent the same logic level. The main difference between this system and the unipolar system is that the highs and lows are alternated for the same logic level. For example, in the system of coding known as alternate mark inversion (AMI), a high or a low level represents the logic level 1. A zero level represents the logic level 0. To transmit the sequence 10101 the following pattern would be output: high/zero/low/zero/high. This type of coding is known as pseudoternary coding. It is called "pseudo" because the zero voltage level is not really classified as a discrete level. The receiver needs two decision levels to decode the incom­ing data and also has to keep track of the level of the last logic 1 transmitted so that the correct level of the next logic 1 is sent as the code alternates. In the case in which the wrong level is detected, i.e. a high followed by a high, a code violation is recorded. The advant­age of this type of coding is that the d.c. level, or balance, is maintained by the transmission of alternate high and low pulses. The disadvantage is the extra circuitry required.

Many receivers extract clocking information from incoming data by detecting the occurrences of the incoming signal data crossing a decision threshold. Each time the signal crosses this point, a PLL can lock its output to it. To perform this task effectively there have to be sufficient crossings over a period of time. If the number of crossings is reduced, the PLL can drift from the frequency of the incoming signal. This will cause the incoming signal to be sampled incorrectly and result in data errors. With a ternary code, the logic levels encoded into alternate high and low output signals provide the clocking information. For example, in AMI coding, each time a logic 1 is transmitted the receiver can lock the PLL onto the incoming signal. Unfortunately, long series of zeros, which are impossible to prevent, may deprive the PLL of all synchronisation information. To avoid this, a group of several consecutive zeros is replaced by a group containing a factitious 1, signalled as such to the receiver by a polarity in violation of the law of alternation of the AMI mode. This principle is systematically applied to all primary and second order digital transmission systems.

The European digital system makes use of a pseudo-ternary mode called a high density bipolar 3-zero maximum code HDB3, which avoids the appearance of more than three consecutive zero symbols, as illustrated in Figure 21.8. It consists of replacing groups of four binary zeros by groups of four ternary symbols of which the the last is non-zero and transmitted with the same polarity as the last non-zero symbol, i.e. in violation of the alternation law of the AMI code. This allows the easy identification of such group at reception and its interpretation as four binary zeros. Furthermore, the first of the four ternary symbols is chosen to be positive, zero, or negative in such way as to maintain or reset the d.c component to a zero value.

The American Tl system generally uses a technique called B7 zero code suppression (B7ZS). To keep telephone company Tl line repeaters and channel service units (CSU) in synchronisation, the digital bit stream permits a maximum of 15 consecutive zeros. As already mentioned, this is known as the Telco's ones density re­quirement (also refer to Section 21.4.3). To obtain compliance with the ones density requirement, communications carriers use the B7ZS technique.

As an example of B7 zero code suppression, consider the DSO time slot illustrated in Figure 21.9. If all 8 the bits are zeros, B7 zero code suppression will substitute a 1-bit in position 7. The adapted time slot is also depicted in Figure 3.9.

Figure 3.10 shows the worst case scenario, where channel 24 is followed by a 0 frame bit, and all bits in channel 1 are zeros, resulting in a total of 16 consecutive zeros. The illustration shows that in this case B7 zero code suppression reduces the number of consecutive zeros to 14.

As already mentioned in Section 21.4.3, this coding technique followed from the assumption that, in order to guarantee synchroni­sation, at least one eighth, i.e. 12.5%, of the incoming bits were Is. However, if a data channel contains all 0s, the data will be corrupted due to B7 zero code suppression. As a result, a data channel is normally restricted to seven usable bits, with one bit set to a 1. This prevents the user data from being corrupted but limits the actual bandwidth to only 56kbit/s. When one bit is set to a 1 on a DSO channel, the channel is said to be a non-clear channel. The 56 kbit/s on a non-clear channel is also known as a DS-A channel.

Clear channel capability can be obtained by using bipolar trans­mission with bipolar 8 zero substitution (B8ZS) in the Tl bit stream (Figure 21.11). With B8ZS no more than eight logic 0s can be transmitted sequentially. Therefore, the data stream to be trans­mitted is examined to determine if a long sequence of logic 0s is about to be transmitted. If such a sequence is detected, each eight consecutive Os in a byte are replaced by a special pattern. If the pulse preceding the all zero byte is positive, the inserted B8ZS code is 000 +- 0 -+. If the pulse preceding the all zero byte is negative, the inserted B8ZS code is 000 - + 0 + -. Both examples result in bipolar violations occurring in the fourth and seventh bit positions. This special pattern is unique because of the embedded code violations. When detected at the receive side it is removed and not seen as a code violation.

If a string of nine logic 0s were to be transmitted, then the first eight logic 0s would be replaced by the special pattern while the ninth 0 would be transmitted normally. At the receive end, the special pattern, which is expected by the receiver, is decoded back to the eight zeros removed. The data output by the receiver is still the nine logic 0s that were to be transferred. Two things can be seen from the example. The addition of the extra, special pattern gave the receiver the clocking data it required to allow the PLL to maintain synchronisation. Secondly, to transmit the sequence of nine binary digits still requires only eight pulses on the line. Although the receiver will be less likely to lose synchronisation, more compli­cated transmitters and receivers have to be used.

1 Learn the words & word combinations:

Principal	Главный
Differentiate	Различать
Failure	Сбой
Cause	Причина, вызывать
Slip	Проскальзывание
Timing problems	Проблемы синхронизации
Jitter	Вибрация
Ominous	Зловещий
Intermittent	Прерывистый, механизм прерывистого действия
Inherent	Присущий
Wander	Смещение, дрейф

2 Find Russian equivalents:

ü Principal alarms; ü The transmitting terminal; ü Being a failure to synchronize…;

ü Impairment; ü Violations; ü An entire multiframe.

3 Find English equivalents:

ü Последовательно; ü Доступный; ü Несоответствия; ü Точность;

ü Частотные колебания; ü Добавление бита; ü Устранение бита; ü Цифровое оборудование.

4 Answer the questions:

1 What are the principal alarms?

2 What does PCM – 30 differentiate?

3 What is a bipolar violation?

4 What is a second possible T1/Е1 error condition?

5 What is another impairment?

6 What coding systems do you know?

Дата добавления: 2015-03-31