Proceedings of the Tomsk Polytechnic Institute named after S. M. Kirov,
Holder of the Order of the October Revolution and the Order of the Red Banner of Labor

Vol. 208　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　1974

On the Delay of Pulse-Width-Modulated Pulse Sequences

B. N. EPIFANTSEV, V. P. SOTNICHENKO

(Presented by the scientific seminar of the Department of Computer Engineering)

The problem of delaying pulse signals was posed long ago and, for most problems encountered in practice, has been solved. The basic idea underlying the solution of these problems consists in storing the positions of the fronts of pulses and shifting them along the time scale by means of delay elements. However, although implementation of this idea is simple for a periodic sequence with a large duty factor, when the delay is commensurate with the repetition period and information on the duration of incoming signals is preserved, there are no simple solutions (the use of electromagnetic delay lines, LZT, LZ, whose operating principle makes them unsuitable for a number of practical problems, is not considered here). Therefore in this case it is necessary to seek particular technical solutions, taking as the basis the criterion of simplicity and accuracy. Below, one such solution is considered, making it possible to realize an exact delay of a pulse-width-modulated pulse sequence by one repetition period. A similar problem arises in the construction of correlators that compute the correlation function of a signal obtained at the output of an “amplitude–time” converter.

Fig. 1.

In Fig. 1a a functional diagram of the device is given; Fig. 1b shows a timing diagram explaining the principle of its operation.

The device consists of a flip-flop \(T\), two coincidence circuits \(\mathrm{Sp}_1, \mathrm{Sp}_2\), a differentiating chain \(\mathrm{Dch}\), two monostable multivibrators \(\mathrm{Od}_1, \mathrm{Od}_2\), two anticoincidence circuits \(\mathrm{Asp}_1, \mathrm{Asp}_2\), and a collecting circuit \(\mathrm{Sb}\).

The pulse sequence, width-modulated as \(U_1(t)\), is differentiated as \(U_2(t)\) and switches the flip-flop \(T\), whose outputs are loaded by the circuits \(\mathrm{Sp}_{1,2}\) and \(\mathrm{Asp}_{1,2}\). Depending on which of the flip-flop outputs has the one level, the first \(\mathrm{Sp}_1\) or the second \(\mathrm{Sp}_2\) passes the negative overshoot formed after the differentiating chain. This overshoot triggers the monostable multivibrator \(\mathrm{Od}_1\) or \(\mathrm{Od}_2\), at whose output a signal of duration \(\tau\) appears (the repetition period of the input signals). Thus, a pulse of fixed duration from the monostable multivibrator and a pulse from one of the flip-flop outputs, of duration equal to the duration of the signal from the monostable multivibrator, are applied to the corresponding anticoincidence circuit. The indicated pulses are shifted in time by an interval equal in magnitude to the input pulse. As a result, the anticoincidence circuit will produce a signal equal in magnitude to the original one, but shifted relative to the latter by the time \(\tau\). At the output of the collecting circuit \(\mathrm{Sb}\), a pulse sequence appears shifted by the time \(\tau\) relative to the original one.

For operation of the circuit, the initial state of the flip-flop \(T\) is immaterial.

By connecting several delay elements of the type considered in series, one can obtain any time delay, which is determined by the formula

\[ t = n\tau,\quad (n = 1, 2, \ldots) \]

where \(n\) is the number of delay elements, and \(\tau\) is the delay time of one element.