CONTROL SCHEME FOR AN INDEPENDENT INVERTER

B. P. Soustin, S. I. Korolev

(Presented by the scientific-technical seminar of the Department of Electrical Equipment)

In frequency converters intended for controlling the speed of an asynchronous motor, the frequency and voltage must be varied according to the law of frequency control. In this case the voltage can be varied in two ways:

1. By using a regulated direct-current source;
2. By using pulse-width or pulse-width-modulation control of the inverter voltage with an unregulated direct-current source.

Inverter control schemes constructed according to the first method have been described in the literature and therefore are not considered in this article.

With pulse-width or pulse-width-modulation control of the inverter, the relative duration of the pulses of the output voltage is varied and, correspondingly, the rms value with an unregulated direct-current source. In this case it is necessary to provide a special device for turning off the inverter thyristors.

The inverter control system with pulse-width regulation of the output voltage (Fig. 1), in addition to the frequency regulator, includes a voltage regulator consisting of a pulse-width modulator (PWM), a ring recomputing circuit, and intermediate matching and amplifying elements.

A common element for both regulators is the variable-frequency master generator with rectangular output voltage. This generator is usually made either on magnetic elements or on the basis of a multivibrator. Fig. 2 shows a schematic diagram of a controlled transistor multivibrator as a precision converter of a continuously varying voltage into a pulse frequency, using separate functional elements.

The converter consists of the following elements: a controlled charging device on transistors $T_3$, $T'_3$; diodes $D_1$–$D_4$ and resistors $R_9$, $R'_9$, $R_{10}$, $R'_{10}$, $R_{14}$, $R_{15}$; two integrating capacitors $C_1$, $C'_1$; two comparing devices on transistors $T_2$, $T'_2$, $T_4$, $T'_4$; a reference-voltage source $U_{\mathrm{in}_2}$; a controlled flip-flop on transistors $T_5$, $T_6$; and two charging devices on transistors $T_1$, $T'_1$.

The output frequency of the master generator is determined from the expression

\[ f = \frac{I_{\mathrm{kz}}}{2 C_1 U_{\mathrm{in}_2}}, \]

where \(I_{kz}\) is the charging current of transistor \(T_3\);
\(C_1\) is the integrating capacitor;
\(U_{\mathrm{in}}\) is the comparison voltage, which, as is evident from the experimental data (Fig. 3), has a fairly linear dependence on the input voltage over a wide range of frequency variation.

Fig. 1

One of the complex control elements is the PWM, which regulates the output voltage when the frequency changes. In this case the following requirements are imposed on it:

Fig. 3

a) the dependence of the pulse duration on the control signal must be linear;
b) the output pulses must have a steep leading edge;

Figure 2

c) the possibility of obtaining several electrically unrelated output signals;

d) the convenience of summing signals at the modulator input.

Two modulator circuits were tested. In one of them a reactive flip-flop was used (Fig. 4), operating from a variable-frequency generator assembled on magnetic elements. The modulator characteristic is linear; however, in this case the frequency-variation range is small.

Fig. 4

The second circuit is made with a saturating transformer, the pulses to which are supplied from a variable-frequency generator. The magnetic core of the transformer has a high specific resistance and a high rectangularity coefficient of the loop. The maximum frequency of the generator is chosen from the condition

\[ \frac{T}{2} \geq \tau; \]

where \(\tau\) is the remagnetization time of the core and is a constant quantity.

When the frequency is varied, the ratio of voltage to frequency is constant, since

\[ U = \frac{\tau}{T} = \tau f,\quad \text{that is } \frac{U}{f} = \tau = \mathrm{const}, \]

and is determined by the remagnetization time of the saturating transformer.

The modulator with a saturating core, constructed on the basis of the variable-frequency generator described above, is shown in Fig. 2. The output pulses (in the circuit of Fig. 2) from transformer \(Tr_1\), with a core having a rectangular hysteresis loop, are fed to the starting input of the ring counting circuit of the voltage regulator through an intermediate amplifier, while the pulses from transformer \(Tr_2\), with a core made of ordinary steel, are fed to the starting input of the ring counting circuit of the frequency regulator.

The ring counting circuit is made using standard logic elements—ferrotransistor cells. The ferrotransistor

the cell is a combination of a ferrite ring with a rectangular hysteresis loop, used for recording, reading, and storing information, and an amplifying element—a transistor—used to amplify the signal during readout. The cell has two recording and two reading windings, and provision is made for compensation of interference on the trailing edge of the recording pulse and for secondary readout. Recording of a one into the counting circuit is performed automatically.

Setting the initial state of magnetization is carried out by applying a “destruction” signal to one of the readout windings of the ferrotransistor cell. The shift clock pulses to the two-stroke ring counting circuit in the frequency-control channel are supplied from windings \(\omega_{2-1}\) and \(\omega_{2-2}\) (Fig. 2) of the master oscillator through an amplifying element, and in the voltage-control channel—from windings \(\omega_{1-1}\), \(\omega_{1-2}\) to the two-stroke ring circuit of the voltage regulator.

The output pulses from the ferrotransistor cells are supplied to start blocking oscillators operating in a standby mode through matching transformers of the type MIT-3 (Fig. 5). The standby mode of the blocking oscillators is achieved by applying a negative bias from the divider \(R_1, R_2\) to the transistor emitter.

By varying the value of the blocking capacitance in the emitter circuit of the transistor, it is possible to regulate the width of the output pulse of the blocking oscillator supplied to the control electrode.

Fig. 5

By selecting the turns of the blocking-oscillator transformer and the value of the blocking capacitance, a pulse duration from 10 to 200 \(\mu\text{s}\) can be obtained.

With width control during the conducting part of the period \(T/3\), the main pair is switched on 2 times, and with pulse-width control—many times. In this case, the number of cells operating through one amplifier increases, and accordingly the frequency of the master oscillator increases.

REFERENCES

1. A. S. Sandler, R. S. Sarbatov. Frequency Converters for Controlling Asynchronous Motors. Library on Automation, “Energiya,” Moscow, 1966.

2. S. V. Kulikov. Controlled Multivibrators on Transistors. Library on Automation, “Energiya,” Moscow, 1966.

3. T. A. Glazenko. Pulsed Semiconductor Amplifiers in Electric Drives. Library on Automation, “Energiya,” Moscow, 1965.