Carrier Time Division is a Multiplexing method for digital modulation techniques where a number of carriers are time divided so that for each carrier at least two equispaced samples can be extracted with in its period to comply with Nyquist rate and Discrete Fourier Transform (DFT) process, the sampling rate Fs must be Fs ≥ 2*Fmax where Fs is the Sampling rate (Samples/sec) and Fmax is the largest frequency component in the signal that is going to be sampled.

Carrier time division process is to free time periods (TF) for each symbol in each carrier so that no any other signals values exist in these periods except for the previous signals as shown in Fig. 1.

Figure 1 is an example of a carrier time division multiplexed signals (3 QAM modulated signals), the first carrier frequency is 100 KHz, the second 125 KHz and the third 150 KHz, the time period from T0 to the end of the symbol time contains symbol 1 for signal 1, from T1 to T4 contains symbol 1 for signal 2 and from T2 to T3 contains symbol 1 for signal 3, the total symbol time (Tt) in this example is [3 / (4*100 KHz)] = 7.5 usec where only ¾ of the full sinusoidal cycle is used because of the fact that DFT requires equispaced samples and carrier time division requires free time period to be placed at the beginning and end of each symbol, this leads to the difficulty of using samples spaced by π and its multiples which if used will result in some constrains in the DFT process, as an example, a sine wave with zero phase and equispaced samples spaced by π or its multiples will have zero value for all its samples which leads to false results for the DFT process at the receiver, so it is not possible to use a full cycle sinusoidal symbol.

The processes of demodulating and demultiplexing of the first symbol for the received signal occurs by sampling the received signal at the free time periods of the first symbol for the first signal then processing a Fast Fourier Transform (FFT) or a direct low pass filtering of the samples to demodulate, for demultiplexing an Inverse Fast Fourier Transform (IFFT) and a subtraction process from the total signal or a direct subtraction process when direct low pass filtering is used to demultiplex, the same processes are repeated for the demodulation and demultiplexing of each symbol.

The minimum required bandwidth of a digital QAM equals to 1/TS where Ts is the Symbol time, Fig. 2 (a) shows the bandwidth allocation for a normal QAM signal and Fig. 2 (b) shows the bandwidth of a carrier time division multiplexed QAM, Fig. 2 (c) shows the single frequencies (Virtual Bandwidth) of the carrier time division multiplexing QAM in the demodulation and demultiplexing processes, as shown in Fig. 2 huge bandwidth saving can be attained by well allocating the carriers frequencies and the free time periods.

A block diagram for a simple carrier time division multiplexing transmitter is shown in Fig. 3, the serial data stream is split into N-1 sub-streams and each stream is modulated using a digital modulation scheme, Inverse Fast Fourier Transform (IFFT) is used to transform the results to time domain, symbols in time domain are shifted and centred to produce a carrier time division multiplexed signal, then the total sampled signal is converted to a continuous one by using a digital to analog converter (DAC), an optional RF modulator can be used if the operating frequency of the system is so high.

A block diagram for a simple carrier time division multiplexing receiver is shown in Fig. 4, the samples are extracted from the received signal at the free time periods of each symbol, a Fast Fourier Transform (FFT) is used to transform the time domain signal to its frequency domain so it can be demodulated, at the same time an Inverse Fourier Transform (IFFT) is used to convert the same signal to its time domain so it can be subtracted from the total received signal, the same processes processed for extracting the other symbols.