In the waveform coding of speech, due to the large number of pauses in the conversational language, most of the possible speech amplitudes are the lowest amplitudes. In addition, the use of waveform coding means that frequent zero crossings of any waveform will produce many zero amplitude codes. As a result, there may be quite a lot of zero strings in the sending stream. Because the receiver must derive its clock from the received data stream, it must receive enough ones on a regular basis to allow its sequential circuits to function as required. The method of minimizing the long string of zeros caused by the pause of the voice is called zero-code suppression. Scramblers can also be used to prevent long strings of 0s from appearing.

Another reason to use a scrambler is a data stream. For example, the data stream generated by signaling and management functions may have a repetitive pattern with highly discrete frequency components. If such a mode is sent, due to the line spectrum it generates, it may constitute a serious source of interference to other DS1s that use the same equipment. By adding a pseudo-random number sequence to the transmitted data, and then subtracting the same sequence from the received stream, the scrambling is completed. This can be done by using a pseudo-random number generator. The pseudo-random number generator is uniformly constructed, synchronization starts, and uses the same rate as the clock. Of course, the clock rate is directly derived from the received data stream itself. Figure 1 illustrates the scrambling and subsequent descrambling of a digital stream.