On the Analysis and Improvement of the Operational Reliability of Semiconductor Amplifiers in MUS-D Electric Drives

V. A. Sebastyanov, A. P. Ineshin, A. I. Esin

(Recommended by the scientific seminar of the Department of Electric Drive, Tomsk Polytechnic Institute)

An intermediate amplifier (IA) is the most important link in a modern automated electric drive with a wide range of speed control. Almost all feedbacks of the automatic-control system (ACS) are closed through the IA; therefore, most emergency operating modes are associated with its malfunction.

In the present article, on the basis of experience in introducing semiconductor magnetic amplifiers into the electric drives of heavy machine tools at UZTS, an analysis is given and the specific features of the circuit solutions and operation of intermediate semiconductor amplifiers—PSAs—in the control circuit of a power self-saturating magnetic amplifier—MUS—are considered; recommendations are also given for improving the reliability of their operation. In electric-drive systems with a dc actuator motor, dc signals act at the input and output of the IA. Therefore, when designing intermediate amplifiers for such systems, preference was given to dc semiconductor amplifiers (PSAs).

A typical example in this respect is the transistor version of the PSA for drives of the PMU-9M series of ENIMS [1]. The circuit of such an amplifier is shown in Fig. 1 and is an adapted version of the UPP-3 type amplifier, developed earlier by ENIMS for EMU-D drives.

The circuit is a three-stage nonreversing dc PSA, whose semiconductor triodes are connected in a common-emitter circuit with direct coupling between stages. In this arrangement, sufficiently high voltage and current amplification factors are obtained with maximum simplicity of the amplifier circuit itself.

In order to reduce the time constant of the bias circuit, the supply voltage was chosen to be increased—60 V—and an additional ohmic resistance \(R_7\) is connected in series with the bias winding.

The supply voltage of the collector circuits of the DC amplifier was chosen with a margin and is 25 V, which, for the types of semiconductor triodes used, is less than the permissible collector voltage—40 V.

Fig. 1.

Fig. 2.

The adjustment of this amplifier is carried out by selecting the resistance \(R_4\) in the base-bias circuit of \(PT_1\) with the input circuit of the PPU disconnected, and is intended to achieve a minimum dead zone of the input—output characteristic of the amplifier—Fig. 2a (curve 2). In the actual drive circuit, when the PPU input circuit is connected, this characteristic has a somewhat enlarged dead zone, Fig. 2a (curve 1). When the input circuit of the PPU is short-circuited, the dead zone increases still more, Fig. 2a (curve 3, taken at a low resistance of the amplifier input circuit). Thus, the typical characteristic of the PPU has a dead zone that is significant in comparison with the width of the linear portion.

The slope of the operating linear portion of the amplifier characteristic depends on the nature of its load and is numerically determined by the dynamic gain:

\[ K_d=\frac{\Delta U_{\text{out}}}{\Delta U_{\text{in}}} =\frac{\Delta I_{\text{ou}}\cdot R_{\text{ou}}}{\Delta U_y} =500 \div 1000, \]

where the increments are taken on the linear part of the characteristic. The static gain, equal to

\[ K_{\text{st}}=\frac{U_{\text{out}\cdot n}}{U_{\text{in}\cdot n}} =\frac{I_{\text{ou}\cdot n}\cdot R_{\text{ou}}}{U_{y\cdot n}} =70 \div 150, \]

characterizes the amplifying properties of the PPU with allowance for the dead zone. Here the subscript “\(n\)” denotes “nominal.”

The questions of temperature stabilization of a semiconductor amplifier with direct interstage coupling have been considered rather fully in [2], where it is shown that, in such amplifiers, the temperature stability of the subsequent stages is determined by that of the preceding ones, while the temperature stability of the entire amplifier depends mainly on the temperature characteristics of the first stage.

The instability of a stage is customarily characterized by the instability coefficient \(S\); as it increases, the stability of the amplifier stage decreases, and conversely. When high temperature stability is required of the circuit, it is recommended to take \(S\) within the limits \(1.5 \div 4\) [2], whereas for the first stage of the PPU (Fig. 1) the instability coefficient is equal to

\[ S=(1+B)>(1+15),\quad \text{where } B=\frac{\alpha}{1-\alpha}>15 \quad (\text{for } PT3), \]

here \(\alpha\) is the current gain of the triode for the common-base connection, and \(B\), correspondingly, for the common-emitter connection. This leads to the fact that, as the temperature rises, the normally closed triodes \(PT_1\) and \(PT_3\) open slightly; the input—output characteristics of the PPU shift to the left relative to the \(I_{\text{ou}}\) axis, and the dead zone of the amplifier decreases, which is confirmed by the experimental characteristics, Fig. 3a.

As the temperature increases, the value of the zero current of the output triode \(PT_3\) increases. At \(t_1=20^\circ C\) it is equal to 4 mA, and at \(t_2=50^\circ C\), respectively, to 25 mA, which can lead to uncontrolled rotation (self-running) of the electric motor; therefore the question of temperature stabilization of the output triode \(PT_3\) deserves special attention. For circuits with direct connection there exists a condition of mutual compensation of the unstable currents \(I_{\text{co}}\) of adjacent stages. The increment of the voltage

from the temperature \(U_k(t)\) may be arbitrarily small with a small difference between the zero currents \(I_{ko.k}\) of the preceding stage and \(I_{kp\,(k+1)}\) of the following stage. The second stage of preliminary amplification of the PPU is assembled on a triode of type P13, and the output stage on the more powerful P4. Therefore the difference between the zero collector currents of PT\(_2\) and PT\(_3\) reaches a significant value.

Fig. 3.

The ambiguity (hysteresis) of the input–output characteristics of the PPU shown in Fig. 3a, when the input voltage \(U_y\) is increased and decreased, is explained by the insufficient temperature stability of the output triode PT\(_3\), owing to the adopted passive locking.

The greatest danger for the PPU circuit is posed by the action of the peak overvoltage that arises in the control windings of the MUS during the initial connection of its power circuits in the initial no-load operating mode of the drive. At the moment when the MUS is connected to the network, the current of the power circuit changes from zero to \(I_{xx}\). This induces a mutual-induction EMF in the control and bias windings of the MUS. The magnitude of the mutual-induction EMF pulse is determined by the rate of change of the current and reaches several hundred volts, Fig. 4.

Fig. 4.

In the bias winding, \(e_{\mathrm{sm}}\) acts in accordance with the supply voltage and causes only a short-time surge of current in its circuit. Therefore there are no faults in the bias circuit.

The mutual-induction EMF \(e_{\mathrm{oy}}\) in the control winding \(OU\), although it acts counter to the supply voltage \(E_k\) of the amplifier, is much larger than it in magnitude. Triode \(PT_3\), owing to the known reversibility of the collector-emitter terminals, is completely switched off when the polarity of the supply voltage is changed, and \(e_{\mathrm{oy}}\) is applied to the \(5V\) diodes in the nonconducting direction. Investigations have shown that bridge diodes selected with a large margin with respect to reverse voltage, for example \(D7Zh\) \((U_{\mathrm{rev}} = 400\ \mathrm{V})\), withstand 5--10 repeated switchings of the power circuit of the magnetic amplifier control system. Diodes with the nominally required \(U_{\mathrm{rev}}\), for example \(D7A\) \((U_{\mathrm{rev}} = 50\ \mathrm{V})\), fail at every switching.

From consideration of the parameters of the electrical operating mode of the semiconductor triodes of the amplifier it follows that the semiconductor triodes \(PT_1\) and \(PT_2\), because of the direct coupling between stages, operate with collector voltages and currents smaller than the rated values and, consequently, are in a lightly loaded mode. Since the output triode \(PT_3\) operates into an inductive load \((OU\) of the magnetic-amplifier control system) and has an unstable temperature characteristic, it is the weakest point of the semiconductor amplifier. In the event of electrical or thermal breakdown of its collector junction, \(I_{\mathrm{oy}}\) tends toward its maximum value and the drive begins to rotate uncontrollably. Zero output of the semiconductor amplifier and stopping of the electric drive in this case can be obtained by changing the operating mode of the output triode so that, in the absence of a control signal at the input of the semiconductor amplifier, the output \(PT_3\) is fully open. This method of constructing the semiconductor-amplifier circuit has also been adopted, for reliability reasons, in the known UPP-1 ENIMS circuit [3] for EMU-D systems, where, in order to compensate the initial ampere-turns of the control winding, differential connection of an additional bias circuit is provided.

Figure 4 shows the circuit of a modified version of the semiconductor amplifier, investigated in the branch laboratory of the Ulyanovsk Polytechnic Institute.

Fig. 5.

institute, for the MYS-D electric drive and making maximum use of the circuit of the commercially produced SPU for the PMU-9M electric drive of ENIMS. Implementation of this variant gives, in comparison with the EMU-D system with UPP-1, an additional circuit advantage: since the output PT is normally open, the MYS winding included in its collector circuit is normally traversed by the current \(I_{\mathrm{sm}}\) and combines the functions of control and bias. It has been experimentally established that in a MYS circuit with external contour feedback, when regulating the motor speed up to nominal, reversal of the ampere-turns is not required, i.e., it is sufficient to regulate the current \(I_{\mathrm{sm}}\) in the combined control-bias winding from maximum to zero. In the case of a MYS circuit with bridge contour feedback, when reversal of the control ampere-turns is required, the SPU circuit remains the same and a separate bias circuit creates additional positive ampere-turns.

The adjustment of the improved SPU variant is carried out in the same way as in the circuit [1], by selecting the resistance \(R_4\) in the base-bias circuit of the PT triode, with the input terminals of the SPU short-circuited, and has the purpose of attaining the minimum dead zone of the amplifier input-output characteristic with the output PT triode completely open, providing the initial bias of the MYS.

When the SPU operates with a real input circuit, its characteristic also acquires an enlarged dead zone, Fig. 2 (curve 1), which increases still more when the input circuit of the SPU is broken, bringing it to zero output and stopping the drive, Fig. 2b (curve 2).

In the circuit of Fig. 5, owing to the changed operating mode of the semiconductor triodes, an increase in temperature leads to a more favorable, from the standpoint of zero output, rightward shift of the output-input characteristic of the SPU relative to the \(I_{\mathrm{uy}}\) axis, to an increase in its dead zone, and to a decrease in the probability of the occurrence of self-running when the SPU input circuit is broken.

The increase of leakage current through the collector junction of the PT when it is in the closed state therefore manifests itself only in the form of a certain decrease in the value of the maximum output of the SPU, which may be partially compensated by changing the type of output PT (P201 instead of P4) or, if necessary, almost completely by installing a supporting bias (the \(D_3\)—\(R_9\) circuit, shown in Fig. 5 by a dashed line).

In order to limit the magnitude of the collector overvoltages of the output triode PT to the level of the supply voltage of the collector circuits of the SPU, its collector load (the series connection of the MYS control windings and the additional ohmic resistance \(R_7\)) is shunted by a reverse valve. In this case, when the MYS is first connected to the alternating-current network, the overvoltage pulse arising in the combined control-bias winding also closes in the aforementioned discharge circuit \(D_2\)—OU MYS and is limited to 100–150 V (Fig. 4).

Thus, by simple circuit solutions, while preserving the basic circuit of the commercially produced semiconductor amplifier for PMU-9M-2 drives, most of the possible faults of the original circuit are eliminated. In Table 1, for comparison, the most probable typical faults of both SPU circuits and their investigations as applied to the MYS-D electric drive with these types of SPU are given.

The questions of matching the intermediate SPU with the power MYS in order to achieve the optimum dynamic characteristic of this

converter type are complicated by the effect of interference on the PPU in actual drives and require additional investigation and consideration.

Table 1

REFERENCES

1. A. Ya. Petrovsky, Ya. B. Rozman. Regulated Electric Drive with Magnetic Amplifiers, Energiya, 1964.

2. I. F. Nikolaevsky. Operating Parameters and Features of the Use of Transistors, Svyazizdat, 1963.

3. Ya. S. Brovman, V. T. Kagan, F. D. Kochubievsky, V. A. Naidis. Electric Drives with Semiconductor Control. Systems with Electric-Machine Amplifiers, Energiya, 1964.