Proceedings of the Tomsk Order of the Red Banner of Labor Polytechnic Institute named after S. M. Kirov

Volume 149
1966

PARAMETRIC CONVERTER OF PULSE SIGNALS

B. A. SOLOMIN

(Presented by the scientific seminar of the Research Institute of High Voltages)

At the present time, in the theory and practice of radio communications, great attention is being paid to electrical systems in which, under the action of an external exciting signal, one or several reactive parameters (inductance, capacitance) change in a definite manner. Such systems may be constructed both with lumped and with distributed parameters, using nonlinear elements (ferrites, semiconductor diodes, ferroelectrics, electron beams). They are used for converting radio-frequency signals and are known as parametric amplifiers, parametric generators, and parametric frequency multipliers. The theory of the operation of these devices may be found, for example, in [1, 2, 3, 4, 5].

The technical devices known at present that operate on parametric principles as converters of electrical signals convert radio-frequency HF and UHF signals and, for effective operation, require the fulfillment of definite frequency relationships between the frequency of the controlled signal and that of the exciting signal.

Fig. 1.

Of considerable interest for pulse technology is the creation of nonresonant parametric converters for converting the energy and frequency spectrum of pulse signals. The theoretical possibility of creating such devices is shown, for example, in [6, 7, 8].

Figure 1 shows the block diagram of a pulse parametric converter. It consists of the following main elements:

1 — a parametrically coupled system of two nonlinear artificial lines;

2 — a synchronized generator of exciting pulses;

3 — a delay unit for the exciting signal;

4 — a load resistance $R_{\mathrm{n}}$, across which the converted pulse signal $U_{\mathrm{out}}(t)$ is taken;

5 — a matching resistance $R$, which serves to eliminate parasitic signals in the converter.

The input pulse signal \(U_{\text{in}}(t)\) is applied simultaneously to input \(a\) of the parametrically coupled lines 1 and to the trigger input of the exciting-pulse generator 2.

From the output of the generator, the exciting pulses \(U_{\text{exc}}\), with the necessary time shift determined by the delay unit 3, are applied to input \(b\) of the parametrically coupled lines 1.

The controlled parameter in the converter may be either the per-unit-length capacitance or the per-unit-length inductance of the line, or, in the most general case, the propagation velocity \(v\) of the principal type of electromagnetic wave in the line, determined by the expression

\[ v=\frac{1}{\sqrt{LC}}, \]

where \(L\) and \(C\) are, respectively, the per-unit-length inductance and the per-unit-length capacitance of the line. Further, for effective operation of the pulse converter, it is necessary that the exciting pulse signal, propagating along one of the two parametrically coupled lines (we shall call the line the controlling line), move with a greater velocity \(v_{\text{e}}\) than the pulse signal to be converted, which propagates along the second of the two parametrically coupled lines, which we shall call the controlled line.

For effective operation of the converter, the exciting pulse signal must have a power exceeding the power of the pulse signal being converted, and a flat top of such duration that the processes may be regarded as steady-state until the converted input pulse signal is separated at the load.

Further, for normal operation of the converter it is necessary that the exciting pulse be applied to the controlling line only when the pulse signal to be converted is already completely inside the controlled line; i.e., the delay unit of the exciting signal must provide a delay time \(T_{\text{d}}\) greater than or equal to \(t_{\text{i}}\):

\[ T_{\text{d}}\geq t_{\text{i}}, \]

where \(t_{\text{i}}\) is the duration of the pulse signal being converted at the input of the controlled line.

Let us suppose that the exciting pulse signal causes a change in the velocity of the electromagnetic wave from the initial value \(v_0\), equal to

\[ v_0=\frac{1}{\sqrt{L_0C_0}}, \]

to the final value \(v_k\), equal to

\[ v_k=\frac{1}{\sqrt{L_kC_k}}, \]

and let \(v_k>v_0\).

The exciting pulse signal, passing along the controlling transmission line and catching up with the signal being converted, will in this case accelerate it from \(v_0\) to \(v_k\). The electrical length of the converted pulse between its beginning and its end will also decrease, owing to the fact that the rear portions of the pulse will move with a greater velocity than the front portions, until the parameter front has traversed the entire path of the signal being converted.

This leads to the result that the signal being converted is shortened in duration; moreover, as calculations show [7], in the case of weak input signals the conversion occurs linearly, i.e., without distortion of the pulse shape. The energy of the converted pulse signal is proportional to the voltage and current of the signal in the line. As the velocity of motion of the signal being converted in the controlled line increases, its voltage and current, and consequently its energy, will increase.

Similarly, it may be noted that if the exciting signal decreases the propagation velocity of the signal in the controlled line, then the converted signal increases in duration, while its energy correspondingly decreases and is transferred to the circuit of the controlling signal.

From the foregoing it is clear that, when an input pulse signal propagates in parametrically coupled lines, i.e., in lines between which there is no direct electromagnetic interaction but there is interaction through a change in one of the basic parameters of the lines—the velocity of propagation of electromagnetic energy in them—and in the presence of a corresponding exciting pulse signal, the input pulse signal at the output of the system will be converted both in duration and in energy.

The converted pulse signal \(U_{\text{out}}(t)\) is taken off at the load \(R_{\text{н}}\), while the exciting pulse signal, in order to avoid reflections, is absorbed in the resistance \(R\).

Fig. 2.

The schematic diagram of parametrically coupled lines 1 is shown in Fig. 2. They consist of two artificial lines \(\mathcal{L}_1\) and \(\mathcal{L}_2\). Each cell of line \(\mathcal{L}_2\) consists of a capacitance \(C_1\) and two equal inductances \(L_1\) and \(L_2\). These inductances are made on two ferrite cores, each of which has two windings \(L_1\) and \(L_2\), where \(L_2\) is one of the two equal inductances of a cell of line \(\mathcal{L}_2\), so that \(L_1\) and \(L_2\) together form a transformer with a ferrite core. The cells of line \(\mathcal{L}_2\) also include capacitances \(C_2\). To eliminate electromagnetic interaction between \(\mathcal{L}_1\) and \(\mathcal{L}_2\), both transformers in each cell of the coupled lines are connected in antiphase in such a way that the signals induced in each of the coupled lines \(\mathcal{L}_1\) and \(\mathcal{L}_2\) from the adjacent line are compensated, and the electromagnetic coupling between the lines is substantially weakened.

The exciting pulse applied to line \(\mathcal{L}_1\) must move at a higher velocity than the input pulse applied to line \(\mathcal{L}_2\), i.e., the condition must be satisfied

\[ \frac{1}{\sqrt{L_1 C_1}} > \frac{1}{\sqrt{L_2 C_2}} . \]

The exciting pulse, passing through the inductances \(L_1\), creates a magnetic field that changes the magnetic permeability of the ferriteխ

cores, and consequently the inductance \(L_2\). Thus, in the line \(Л_2\) there is a traveling parameter wave, which causes the conversion of the input pulse signal.

Tests were carried out on the parametric converter of pulse signals as a compressor of microsecond pulses. Operation of the converter confirmed the possibility of increasing the energy and shortening the duration of the input pulse signals. The generator of exciting pulses (a synchronized blocking oscillator) produced pulses with an amplitude of 300 V and a duration of 7 μsec. Pulse signals with a duration of 10 μsec and an amplitude of 25 V were applied to the input of the converter. At the output of the converter, the input pulse signals were transformed into pulses with a duration of 1.4 μsec and an amplitude of 150 V.

Figure 3 shows oscillograms of the input and output pulse signals*.

Fig. 3.

Thus, the possibility of constructing broadband parametric converters of pulse signals has been practically demonstrated. Converters based on the principles set forth above can be used as amplifiers or generators of pulse signals for shortening pulse signals in time-compressor circuits and for other purposes.

REFERENCES

1. P. Tien, H. Suhe, Proc. IRE, 46, 700, 1958.
2. L. I. Mandelstam. Collected Works, vol. 2, Publishing House of the Academy of Sciences of the USSR, 1947.
3. S. M. Rytov. Proceedings of the Physical Institute, 2, 1, 1940.
4. S. M. Rytov. ZhTF, 17, 930, 1947.
5. A. A. Andronov. Collected Works, Publishing House of the Academy of Sciences of the USSR, 19, 1946.
6. S. I. Averkov, L. A. Ostrovskii. Izvestiya Vuzov, Radiofizika, 1, 4, 46, 1958.
7. S. I. Averkov, N. S. Stepanov. Izvestiya Vuzov, Radiofizika, vol. II, No. 2, 1959.
8. L. A. Ostrovskii. Izvestiya Vuzov, Radiofizika, vol. VI, issue 4, 1963.
9. M. Otuka. Abstracts of a report at the 13th URSI Symposium. 1965, Holland.

* Similar results were independently obtained in work [9].