DEVICE FOR INVESTIGATING MECHANICAL FACTORS IN COMMUTATOR ELECTRICAL MACHINES

A. I. Skorospeshkin, L. Ya. Zinner, A. I. Proshin

(Presented by the seminar of the Department of Electrical Machines and General Electrical Engineering)

Satisfactory commutation in commutator electrical machines is determined to a considerable extent by the quality of manufacture of the commutator-and-brush apparatus and, consequently, by the condition of the working surface.

One of the causes of sparking is instability of the commutator, which consists in the fact that, at high peripheral speeds, centrifugal forces cause a relative displacement of individual bars or groups of bars.

In addition, factors such as eccentricity, ovality of the commutator, and vibration of the bearing assembly have a great influence on the commutation process, although, being of lower frequency, they affect commutation to a lesser degree than radial displacement of the bars.

All these factors lead to a sharp deterioration of commutation and to a reduction in brush service life.

Until recently, in factory practice the condition of commutator surfaces was generally evaluated under static conditions by means of minimeter heads. However, as numerous tests of commutators on spin-test installations have shown, the surface relief of commutators under dynamic conditions differs significantly from the relief under static conditions. This makes it necessary to pay the most serious attention to ensuring the required rigidity of commutators and to monitoring their surface. If factory laboratories had the necessary monitoring and measuring equipment for evaluating commutator dynamics, the process of adjusting commutation would be greatly simplified. However, industry does not produce profilometers for electrical-machine building, and a rather large number of profilometers described in the literature have never advanced beyond the development stage.

The waveguide profilometers described in the literature [1, 2, 3] suffer from a number of substantial shortcomings. These are, above all, low resolving power, low sensitivity to small displacements, and the bulkiness of the instrument. The above-listed shortcomings of the waveguide profilometer have practically limited the scope of its application.

Recently there has been a tendency toward the creation of electronic profilometers using high-frequency capacitive and inductive sensors as the transducer. Works [4, 5] describe in detail the operating principle and the design of instruments in which a capacitive or inductive sensor is included in the tank circuit of an r.f. oscillator. As the distance from the commutator surface to the sensor changes, the oscillator frequency changes, i.e., the devia-

by which one can judge the condition of the commutator surface. However, these instruments are not without a number of substantial shortcomings. These are, above all, sensitivity to instability in the frequency of the measuring generator and to instability in the voltage of the power-supply sources. It should be noted, however, that by using measuring circuits based on beats with quartz stabilization of the reference-generator frequency, the measurement errors caused by instability in the frequency of the measuring generator can be reduced considerably.

As was already noted above, purely mechanical factors have a great influence on the commutation process in commutator electrical machines. In order to investigate them, the authors of the present article have developed and built a semiconductor profilometer which is largely free of the shortcomings inherent in the instruments described in the literature \[1, 2, 3, 4, 5\].

The operation of the instrument is based on the principle of eddy currents, which consists in the following: if a sensor that creates an alternating magnetic field is placed above a conducting surface, eddy currents arise in the surface of the conducting material. Their direction is such that the magnetic flux they produce is shifted by \(1/2\) period relative to the exciting field. The interaction of the exciting and induced magnetic fluxes leads to a weakening of the resultant magnetic flux. By monitoring, by one of the existing methods, the change in the resultant magnetic flux, one can thereby monitor the magnitude of the change in the gap between the sensor and the conducting surface.

In the instrument, an inductive small-displacement sensor is used as the transducer (Fig. 1); it contains two windings—the exciting and the measuring windings—and is a kind of high-frequency transformer. When the gap between the core and the surface of the commutator lamella is varied over a wide range, the e.m.f. induced in the measuring winding changes in proportion to the change in the gap (Fig. 2). In order to reduce lateral leakage fluxes and, consequently, to increase the resolving power of the sensor, the sensor windings are placed on a ferrite core and the entire sensor is placed in a copper shield.

Shielding of the sensor is very important, since it provides both an increase in resolving power and an increase in the sensitivity of the entire instrument as a whole. In addition, the shield gives the entire structure the necessary rigidity, shields the sensor from external electromagnetic fields, and is a good radiator providing complete heat removal.

Fig. 1. Calibration disk with the measuring sensor installed.

The resolving power of the instrument depends to a significant extent on the frequency of the supply current (Fig. 3). As the frequency increases, it increases; however, the operating frequency of the instrument is limited by the resonant frequency of the sensor itself, which is \(1.5\) MHz. When the frequency of the supply current is increased above \(1.5\) MHz, the sensitivity begins to decrease sharply.

Fig. 2. Calibration curve of the measuring transducer.

Fig. 3. Dependence of the pulse shape (resolving capability) on the frequency of the supply current.

Fig. 4. Schematic circuit diagram of the profilometer.

In order to increase the proportion of the useful signal in the overall signal at the sensor output and to facilitate adjustment to the optimum gap, an identical compensating sensor is connected in series with the measuring sensor, in such a way that the emf in its secondary winding is in antiphase with the emf of the measuring sensor.

The excitation windings \(L_3L_4\) of both sensors are connected in series and constitute the load of a separately excited generator (Fig. 4), assembled on a semiconductor triode P402-\(T_2\). Excitation with a frequency of \(1.5\) MHz and an amplitude of \(0.3\) V is applied to the input of \(T_2\) from the circuit of an LC generator.

As the size of the gap between the sensor and the commutator changes, the reactance of the excitation winding changes by approximately a factor of two, which entails a corresponding decrease in the exciting current at fixed \(U_{\mathrm{v}}\) on the circuit. In order to stabilize the excitation current and, consequently, to eliminate measurement error due to its instability, the separately excited generator \(T_2\) has a large output resistance, exceeding by 15–20 times the maximum possible resistance of the sensor winding. The differential emf, amplitude-modulated by the commutator plates, is fed from the output of the sensor \(L_5L_6\) to the emitter follower \(T_3\) (P403) and then to the input of the resonant amplifier \(T_4\) (P403). The signal, amplified 150 times, is taken from the coupling turns of the resonant-amplifier circuit, detected on D2E diodes, and applied to the input of the video amplifier \(T_5\), \(T_6\), \(T_7\) (P403). The last stage of the video amplifier, \(T_7\), is an emitter follower and serves to match the output of the video amplifier with the input of the EO-58M oscilloscope. To set the optimum working gap according to the minimum emf at the sensor output, the circuit contains a millivoltmeter, assembled on \(T_8\), \(T_9\) (P403), which is used simultaneously for calibrating the generator current.

The generator and the measuring circuit are powered from a semiconductor voltage stabilizer that supplies \(12 \div 21\) V to the load.

In order to eliminate measurement error due to instability of the amplitude of the signals taken from the stages of the measuring circuit, the stabilization factor of the power supply has been brought to 0.02%.

For adjustment and calibration of the instrument, a calibration setup was constructed. The measuring sensor is mounted on a bracket opposite the calibration disk and is rigidly fastened to the plate of the setup (Fig. 1). On the outer cylindrical surface of the disk two groups of teeth are cut, 10 in each group, separated by sections with a smooth surface. The width of the teeth in one group is \(2.5\) mm, and in the other \(4\) mm; their height is \(2.5\) mm.

Figure 5 shows oscillograms of the calibration disk for a full revolution and a quarter revolution. From the oscillograms it is evident that the sensitivity of the instrument to displacements of the commutator surface is sufficiently high and is approximately \(0.1\)–\(0.2\) V/µm. The resolving power with respect to the width of a plate is somewhat worse for plates \(2.5\) mm wide and considerably better for plates \(4\) mm wide.

In order to study the influence of plate width on the shape of the output signal, a \(5\) mm tooth was cut. In the oscillogram of Fig. 5a, a broad pulse with a flat section corresponding to this tooth is clearly distinguished. The rounding of the leading and trailing edges of the pulses from the plates is explained by the nature of the distribution of eddy currents in the metal surface. This rounding can be eliminated by increasing the frequency of the supply current or by reducing the cross-section of the sensor core. In that case, however, the sensitivity of the profilometer decreases.

The sweep of the oscilloscope was synchronized by means of a \(5\) mm tooth, onto which foil \(12\) µm thick was glued. Thus, the pulse

from a tooth, being the synchronizing one, at the same time serves as a standard for calibrating the instrument.

The working gap between the surface of the calibration disk and the measuring transducer is 100 μm. When this gap is reduced to 75–50 μm, the sensitivity of the instrument increases sharply.

For setting the initial gap, the measuring and compensating transducers are provided with micrometer screws.

As can be seen from the oscillogram, the instrument makes it possible to assess both the eccentricity of the commutator and its microrelief. The blurring of the tops of the pulses from the teeth and of the oscilloscope sweep is explained by quasi-random processes occurring in the bearings when the balls move, owing to the nonuniformity of their dimensions.

If it is necessary to assess oscillations of the commutator surface without superposed oscillations due to bearing mobility and inaccurate balancing of the machine armature, it is convenient to have two identical transducers, one of which would be placed at the surface of the commutator, and the other at the surface of the shaft. With opposing connection of the signal windings of the transducers, the differential e.m.f. exactly corresponds to the shape of the commutator profile. Thanks to the presence of two transducers, the device created makes it possible to assess the microrelief of commutator surfaces in pure form, without superposed osci-

Fig. 5. Profilogram of the calibration disk.
a — first group of plates,
b — second group of plates,
c — full revolution of the disk.

vibrations due to shaft runout, inaccuracies in armature balancing, etc.

Thus, the profilometer developed is a sufficiently accurate instrument, making possible a comprehensive investigation of the mechanical factors affecting commutation in commutator electrical machines. The instrument makes it possible to record displacements of the surfaces of the commutators under investigation on the order of several thousandths of a millimeter. By reducing the installation clearance and increasing the gain of the measuring circuit, this value can be considerably reduced.

REFERENCES

1. K. K. Namitokov, V. F. Chepura. An instrument for investigating the surface form of commutators of electrical machines. Elektrichestvo, No. 5, 1958.

2. I. A. Baranov. Measurement of deformation of commutators of electrical machines in dynamic operation. Vestnik elektropromyshlennosti, No. 5, 1958.

3. K. K. Namitokov, V. F. Chepura. Measurements of vibrations and linear displacements by means of microwave radio waves. Izmeritel’naya tekhnika, No. 5, 1960.

4. B. Z. Mikhlin. High-frequency capacitive and inductive transducers. GEI, 1962.

5. K. K. Namitokov, V. F. Chepura, V. G. Brezinsky. An instrument for investigating dynamic changes in the surface form of commutators of electrical machines. Elektromekhanika, No. 1, 1966.