A time division multiplex system conveys digital data, and speech must first be converted to data. TDM allocates short-duration time slots within a wider time frame to each information channel. For example, a continuous stream of data sent over a link at a rate of 2400 bit/s could convey the information contained in four 600 bit/s channels in short sequential bursts.
If the duration of one input bit is 1/600 s or 1.666 ms, a sevenbit character occupies 11.66 ms and 85 such characters can be sent per second. If the transmitted rate can be speeded up each bit sent at 2400 bit/s has a duration of 416µs and 343 seven-bit characters can be sent per second. Such a system is shown in Figure 13.2. The data is stored in the buffers at the transmitter and the clock pulses are applied to each store/gate sequentially allowing one character from each data channel to be transmitted at a rate of 2400 bit/s. Perfect synchronization must be maintained between all channels and the transmitter and receiver to avoid data errors.
Store andLink 3 2400 bit /s
Link 3 600 bit /s
The principle is illustrated in Figure 13.3 where a 1 s time frame at 2400 bit/s contains 343 time slots, each of 2.92 ms and containing a character from a specific information channel. Every second, therefore:
channel A would occupy slots 1, 5, 9, ..., 337 channel B would occupy slots 2, 6, 10, ..., 338 channel C would occupy slots 3, 7, 11, ..., 339 channel D would occupy slots 4, 8, 12, ..., 340
This leaves three blank slots; in practice slots are also allocated for preamble, address and synchronization purposes.1 s frame
Time slots of 2.92 ms each containing one seven-bit character at 2400 bps 12345 6 7 8 9 339340341342343 AA A CD
Information channelsThe digital base band signal adopted in Europe operates at 64 kbit/s and the multiplexed signal at 2048 kbit/s. An eight-bit word or sample of a PCM voice channel (see Chapter 14) occupies 3.9µs, and the interval between successive samples of a channel is 125µs, the time frame duration. Therefore, the number of channels (time slots) that can be accommodated in one frame is 125/3.9= 32. Thirty of the slots are used for information channels and two for control purposes (Figure 13.4).
125 ms3.9 ms Each voice channel time slot contains 8 bits 0.488 ms wide Figure 13.4 TS0, TS16 are used for signalling; TS1–TS15 and TS17–TS31 are used for voice channels
In the European E1 system, timeslots are numbered 0 to 32. Timeslot zero (TS0) is used for synchronization and TS16 is used for signalling. The signalling channel is therefore operating at 64 kbit/s. Digitized voice is carried in the remaining 30 timeslots. Voice channels 1 to 15 are carried over TS1–15. Voice channels 16–30 are carried over TS17–31.
The American T1 system is a 1.544 Mbit/s transmission system. The data format should be referred to as DS-1, but T1 is sometimes used instead. Timeslots (TS), or channels, in this system are numbered 1 to 24; there are 23 available to carry traffic and TS24 is used for synchronization. Each timeslot carries 8 bits, sampled at 8 kHz. An additional bit is sent at the end of each frame (known as the framebit or F-bit) so there are 24, 8-bit timeslots, plus 1 bit= 193 bits per frame.
The synchronization word sent in TS24 is coded as 1 0 1 1 1 Y R 0. The Y and R are dependent upon the data multiplexer and are for the manufacturer’s use.
Signalling information is transmitted using ‘stolen’ bits from each channel, rather than using a separate timeslot as in the European E1 system. Bits are not stolen from timeslots in every frame, because this would degrade the quality of a voice channel. Instead, the least significant bit of every sixth frame is used for signalling. Thus the effective word length of the encoded voice is 7.833 bits, rather than 8 bits of the E1 system.
There are two virtual signalling channels (A and B) created by the use of stolen bits. Signalling Channel A is derived from bits stolen from timeslots during the sixth frame. Signalling Channel B is derived from bits stolen from timeslots during the twelfth frame. This is repeated for every superframe. Each channel has a data rate of 24 bits per 12 frames (1.5 ms), or 16 kbit/s, giving 32 kbit/s signalling in total. Although this data rate is half that of the E1 system, there are only 23 voice channels instead of the 30 used in the E1 system.
The F-bit is used to identify the frame, which is important when extracting the signalling information. Because only one in six frames are used for signalling, it is vital to know which ones carry the information and which ones carry voice traffic. Also there are two signalling channels and these must be separated at the receiver end of the system. A six bit data pattern of 1 0 0 0 1 1 is transmitted, followed by the inverse: 0 1 1 1 0 0. The complete 12-frame sequence is known as a superframe.