Features of PPU Operation in an Adjustable Electric Drive Based on the MUS-D System

A. P. Ineshin, V. A. Sevastyanov, V. A. Paramonov

(Recommended by the scientific seminar of the Electromechanical Faculty)

The development of automated electric drives for metal-cutting machine tools is proceeding both along the path of further expansion of the electric range of speed regulation and along the line of satisfying the technological requirement of increasing the response speed of the drive, especially under disturbing action from the load. The use of inertia-free intermediate amplifiers makes it possible, in this case, to realize practically any gain coefficient ensuring attainment of the necessary static accuracy, and also to improve significantly the dynamic characteristics of the drive.

The combination of an intermediate semiconductor amplifier and a power magnetic amplifier often gives the best solution to the problem of constructing a high-speed converter: PPU-MUS with sufficient output power. Thus, in [1] it is indicated that the introduction of a transistor amplifier into the control circuit of a magnetic amplifier improves the response speed of the latter at small signals. Indeed, the time constant of the control winding, in series with which the transistor is connected, is:

\[ T_{\mathrm{mu}}=\frac{L_{\mathrm{oy}}}{R_{\mathrm{oy}}+R_{\mathrm{tr}}}, \]

where \(L_{\mathrm{oy}}, R_{\mathrm{oy}}\) are, respectively, the inductance and resistance of the control winding;

\(R_{\mathrm{tr}}\) is the equivalent resistance of the transistor.

At small signals, when the triode is close to cutoff, \(R_{\mathrm{tr}}\) has a relatively large value and \(T_{\mathrm{mu}}\) is greatly reduced. This creates the impression that the use of a direct-current PPU, designed for a linear mode of operation, when the output triode has a large internal resistance, is the most acceptable variant of combining a PPU and an MUS in terms of dynamics. However, in [2] it is correctly shown that reduction of the duration of the transient process in an ordinary magnetic amplifier with self-saturation, when controlled by a semiconductor triode, occurs due to an increase in the potential output power of the transistor amplifier. Charac-

the nature of the transient process under consideration depends substantially on the parameter:

\[ E'_k=\frac{E'_k}{E_m}, \]

that is, on the ratio of the reduced EMF of the collector-circuit supply

\[ E'_k=E_k\frac{W_p}{W_y} \]

to the EMF of the supply of the working circuit of the magnetic amplifier \(E_m\).

Here \(W_y, W_p\) are, respectively, the numbers of turns of the control and working windings of the magnetic amplifier.

As applied to commercially produced electric drives of the PMU-P type [1] and to the transistor amplifier, the condition for attaining limiting speed of response is clearly not fulfilled, i.e.,

\[ E'_k=\frac{E_k}{E_m}\cdot\frac{W_p}{W_y}<1. \]

This leads to a disturbance of the forced-magnetization mode of the MUS, to the appearance of peaks of collector overvoltage when the output transistor is cut off, and to substantially different durations of the transient process when the control signal is increased and decreased. Shunting the MUS control winding with a capacitance in order to remove collector overvoltages at the output triode somewhat equalizes the values of the time constants of the transient processes. The noted circumstances make it possible to classify the DC converter PPU—MUS as an element of control systems whose inertia depends on the magnitude and sign of the mismatch.

The action of disturbances at the PPU input in a real adjustable-drive system practically rules out the possibility of operation of semiconductor triodes in the linear amplification mode. The principal disturbances arrive through the corrective-feedback circuit, and the disturbance spectrum is not random, but is the sum of two or three clearly expressed frequencies, among which the determining one is the disturbance of the fundamental harmonic of the output voltage of the power MUS.

The use of a structure with parallel correction in MUS-D drives is compulsory, owing to the substantial instability of the parameters of the PPU and MUS. Parallel corrective feedback, with an appropriate choice of its parameters, without changing the static characteristics of the system, ensures its stability and stabilizes the dynamic characteristics of the converter. However, in the MUS-D drive, the inclusion in the parallel-correction circuit of links with the largest amplification factors and with instability of the PPU and MUS, because of the known discrete nature of MUS operation and the presence of significant pulsations in its output voltage, leads to the transmission through the correction circuit to the PPU input of a very high level of disturbances. The unfavorable disturbance-to-useful-signal ratio in this case leads to a considerable change in the initial static characteristics of the system, and also affects its dynamic indices.

The input signal for the correction circuit, implemented in the form of an RC contour, is the armature voltage of the actuating motor [1], equal to

\[ U_y=C_e\cdot n+I_y R_y+L_y\frac{dI_y}{dt}. \]

When operating in the main, lower working portion of the speed-control range, the predominant components in the armature voltage \(U_i\) are pulsating components proportional to the armature current \(I_a\) and to its derivative. In Fig. 1a, characteristic 1 shows the experimentally obtained dependence of the disturbance \(U_i\) (the variable component of the armature voltage) on the armature current during operation of the actuator motor of the PMU-P drive at a speed \(n = 30\) rpm, \(D = 100:1\).

Fig. 1.

The magnitude of the disturbance voltage rises rapidly as the motor is loaded and reaches several tens of volts; here the form of the disturbance pulses, which follow at the ripple frequency of the rectified voltage of the power converter (300 Hz), changes as the saturation output angle of the MUS chokes is adjusted.

At the PPU input, in the steady operating mode of the drive, in addition to the error signal \(U_y\), equal to the difference between the reference voltage and the emf of the tachogenerator, a considerable alternating disturbance voltage transmitted through the RC correction circuit is present. In this case the PPU operates in a clear forced-switching mode with a frequency determined by the disturbance, similarly to known types of dc-voltage converters into a trapezoidal waveform.

The initial static input–output characteristic \(I_{oy}(U_y)\) (the dependence of the MUS control current \(I_{oy}\) on the error signal \(U_y\) of the reversible PPU), distinguished by the presence of a dead zone and a steep working linear section (Fig. 2, characteristic 1), now changes substantially under the influence of the disturbance (Fig. 2, characteristics 2–5, corresponding to different load currents). At small signals at the amplifier input, corresponding to low motor loads, the disturbance acts as an equivalent linearizing voltage, compensating the PPU dead zone and thereby somewhat increasing its initial static gain. At high loads of the actuator motor, the equivalent increase in the disturbance leads to a further shift of the PPU characteristics, forced opening of the amplifier at \(U_y = 0\) (disturbance detection), and substantial

…to a reduced gain coefficient \(K_{\text{PPU}}\). The latter, when the armature current changes from no-load to the rated value, falls from \(80 \div 100\) to \(5 \div 15\) (Fig. 1b, curve 1), worsening the static characteristics of the drive and the possible range of speed regulation.

Fig. 2.

Forced opening of the PPU when interference is detected leads to the phenomenon of self-running—uncontrolled rotation of the drive when the reference voltage is removed and the electric motor is initially loaded.

Since the interference voltage, associated with the magnitude of the armature current of the actuator motor, determines the switching mode of the PPU, also affecting the magnitude of its gain coefficient, the converter output and its input are additionally linked by nonlinear parasitic feedback through the interference. The experimentally measured rise time (\(t_{\text{н}} = 0.84\) s) and decay time (\(t_{\text{с}} = 0.26\) s) of the current in the power circuit characterize the transient process of opening—blocking of the PPU–MUS converter when operating into an ohmic load in an open-loop ACS and when an artificial interference is present at the PPU input, analogous in magnitude and frequency to that in the closed-loop drive system. In comparison with analogous data for the transient process in the absence of interference (\(t'_{\text{н}} = 0.52\) s and \(t'_{\text{с}} = 0.14\) s), the duration of the real transient process is substantially prolonged. In the closed-loop electric-drive system, because of the action of the interference, the transient process under a disturbing action proceeds most unfavorably, especially when the load is shed (oscillogram in Fig. 3), being characterized by large overshoots and oscillation.

The desire to eliminate the parasitic feedback through the interference and to remove the dependence of the interference-voltage parameters on the magnitude of the converter output voltage led the authors to the use of another type of correction, based on the operating features of a power magnetic amplifier with self-saturation [3]. In the applied MUS circuit with remote contour feedback, the voltages of the even harmonics in the control windings are suppressed. When a transient process arises in the MUS, a voltage appears on the free control windings

a signal proportional to the derivative of the “constant” component of the magnetic flux. The introduction of flexible feedback in this case is equivalent to the use of a differentiating transformer [3]. Experimental investigations of this type of electromagnetic correction have shown its high effectiveness and a significantly better signal-to-noise ratio for the voltage taken from the correction winding.

Fig. 3.

The DC PPU is also introduced into a forced switching mode by the initial disturbance, which is a suppressed 300-Hz pickup voltage induced in the correction winding from the power circuit of the MUS. Owing to the strong electromagnetic coupling between the control and correction windings of the MUS, during PPU switching an additional transformation takes place in the latter, in the form of an alternating component of the rectangular output voltage of the PPU. The total disturbance in the correction circuit therefore has an approximately constant amplitude (Fig. 1a, characteristic 2) and changes its form as the MUS and PPU open. Owing to the signal-to-noise ratio, which is better than with RC correction, and to the independence of the disturbance magnitude from the load current of the electric motor, it is possible, by installing the simplest filter in the electromagnetic correction circuit, to preserve the value of the static gain coefficient of the PPU under the action of disturbance at a level of 40, with its initial value 80–100, Fig. 1b (curve 2). This makes it possible to expand the speed-control range of the actuator motor to 200:1 with satisfactory dynamic indices and to avoid the phenomenon of “self-running.” However, the parasitic feedback between the input and output of the PPU through the disturbance (more precisely, through the disturbance waveform) along the correction circuit still remains, which continues to have an unfavorable effect on the dynamics of the electric drive.

The time constant of the self-saturating magnetic amplifier is unambiguously related to its power gain coefficient [3]

\[ T_{\mathrm{mu}} = \frac{K_p}{K_{\mathrm{dyn}}}, \]

where \(K_p\) is the MUS power gain coefficient;

\(K_{\mathrm{dyn}}\) is the dynamic coefficient of the MUS, given in catalogs.

A substantial reduction of \(T_{\mu}\) of the MUS, down to the elimination of self-magnetization \(T_{\mu} \to 0\), is achieved by increasing the output power of the PPU.

Increasing the output power of the PPU is equivalent to requiring a high efficiency of the output stage. Therefore, it is most expedient to construct PPU circuits as class-D pulse amplifiers with pulse-width modulation elements. It should be noted that PPU circuits with PWM have already been sufficiently developed and investigated as applied to electromechanical automation systems [4].

A pulse-width modulator consisting of a rheostat-capacitor circuit deserves attention; in order to obtain pulse-width conversion with a high gain coefficient, a circuit is used with successive discharge of capacitances into a controlled resistance, for which a semiconductor triode controlled by a DC input signal is employed.

The operation of a class-D PPU with such a PWM circuit, in comparison with a DC PPU in the presence of interference in the electromagnetic-correction circuit and with other conditions being equal (equal gain coefficients and output powers of the PPU), differs in the following two features: switching of the PPU is now performed not by the interference but by a special source of commutation voltage (in the simplest case, by the voltage of an industrial-frequency mains supply); the action of the DC input signal with superimposed interference on the pulse-width modulation process occurs during time intervals corresponding to the discharge of the capacitance, while during the remaining time the PWM is, as it were, self-disconnected from the source of the input signal and interference.

The latter circumstance makes it possible, conventionally, to regard the class-D PPU under consideration as a pulse element, to the input and from the output of which sequences of pulses are applied. Enclosing such a pulse element in a feedback loop reduces [5] to the transformation by the feedback element, according to a definite law, of the output sequence of pulses and its subsequent application to the input of the pulse element. In a particular case, the effect of such enclosure, with appropriate parameters of the feedback circuit, may also be zero. As applied to a class-D PPU, this means that the filter in the electromagnetic-correction circuit must be tuned not only for the purpose of suppressing the interference voltage, but also from the condition of ensuring a favorable phase relationship between the PWM commutation voltage and the interference. In this case, the interference pulses transformed from the PPU output into the correction circuit are suppressed and shifted in time by the filter of the feedback circuit. Thus, penetrating to the PPU input, they do not exert a substantial effect on the PWM. Since, in the process of regulating the speed of an electric motor under load, over the entire range of speed settings the relative on-time \(\gamma\) varies from zero to unity, it is not possible to obtain compensation of the parasitic feedback through interference over the entire range.

In practice it is important to ensure compensation of the interference in the range of low operating speeds, where \(\gamma\) is regulated from 0 to 0.3. On the basis of this condition, taking into account the frequency of the commutation source and the parameters of the electromagnetic-correction circuit, the filter is tuned. The oscillogram in Fig. 4, taken in a closed-loop drive system with a class-D PPU at a low rotational speed of the electric motor, shows an improvement in the dynamic characteristics of the system in comparison with the previously used—

[[unclear: beginning of word]] DC PPU, with approximately the same values of the gain coefficient and output power of the PPU. When the system operates at medium and upper speed settings, when $\gamma = 0.3 \div 1$, compensation of the effect of interference on the PPU is disrupted. However, owing to the greater effectiveness of the signal action through the main feedback circuit, this circumstance does not affect the course of the transient processes.

Fig. 4.

Moreover, the automatic reduction of the PPU gain coefficient under the influence of interference increases the stability margin of the system, whose overall open-loop gain coefficient is selected according to the known conditions for ensuring its static accuracy at the lower end of the range and, as a rule, proves excessive during operation in the middle and upper portions of the range.

Conclusions

1. Enclosing the PPU—MUS converter by an RC parallel-correction circuit leads to the appearance of parasitic feedbacks in the converter through the transmitted interference and to a deterioration of the static and dynamic characteristics of the electric-drive system.

2. Electromagnetic correction of the PPU—MUS converter is distinguished by a better signal-to-interference ratio; however, when a DC PPU is used in this case, parasitic feedback through interference is preserved.

3. The use of a class-D PPU with an internal commutation source in combination with electromagnetic correction makes it possible to improve the dynamic qualities of the drive by compensating feedbacks through interference and increasing the output power of the PPU.

References

1. A. Ya. Petrovsky, Ya. B. Rozman. Adjustable Electric Drive with Magnetic Amplifiers. “Energiya,” 1964.

2. R. A. Lipman, I. B. Negnevitsky. High-Speed Magnetic and Magnetic-Semiconductor Amplifiers. TEI, 1960.

3. M. A. Rozenblat. Magnetic Amplifiers. “Sovetskoe radio,” 1960.

4. T. A. Glazenko. Class D Semiconductor Amplifiers for DC Electrical Machine Automation Systems. Leningrad DNTP, 1963.

5. S. L. Popkov, Yu. S. Popkov. Continuous and Discrete Tracking Systems. “Energiya,” 1964.