EXPERIMENTAL AND ANALYTICAL STUDY OF COMMUTATION IN A CROSS-FIELD EMA

B. I. Kostylev, A. I. Skorospeleshkin

(Recommended by the seminar of the departments of electrical machines and general electrical engineering)

This paper presents some results of a study of commutation in a cross-field EMA of the EMA-12A type. The purpose of these investigations is to identify the principal factors affecting the commutation conditions of an EMA and to verify the method for calculating commutation with the aid of an electronic digital computer, proposed in [1, 2] for commutator electric machines. In [1, 2], calculations of EMA commutation were carried out for operation on direct and alternating current for a brush overlap \(\beta = 1.5\), and these are well confirmed by experiment. In the present work, commutation studies were carried out only on direct current, but with different brush overlap and with different degrees of compensation of the armature reaction along the longitudinal axis of the EMA.

The calculation of commutation on direct current is simplified, since in the sections being commutated there is no transformer e.m.f.; moreover, it is unnecessary to take into account the time shift of the active currents and e.m.f.’s. A change in the brush overlap changes the system of differential equations, which makes it necessary to change the calculation program.

We carried out studies of commutation in the longitudinal and transverse circuits of the amplifier both in their separate and in their joint operation. The brush overlap was 1 and 1.5. The reduction of brush overlap was achieved by filing down the brush to the width of one commutator segment. Oscillographic recording of the current curves was performed on an electronic oscillograph by means of bifilar turns laid together with the main armature winding.

Figures 1 and 2 show the calculated and experimental current curves of commutation of three sections of a slot for the longitudinal and transverse circuits during their joint operation. These curves, first of all, show satisfactory agreement between the calculated and experimental results. All experimental curves are characterized by premature completion of current reversal, which is explained by instability of the contact at the trailing edge. Comparison of the curves for different sections shows that the sections that are last to complete commutation in the slot are under the most unfavorable conditions. This occurs because of significant inductive coupling between the sections, especially at \(\beta = 1.5\). Since, for the examples presented, the compensation was set to unity, the commutation of the longitudinal circuit is very close to rectilinear. In the transverse circuit, commutation is retarded because of the reactive e.m.f. and the rotational e.m.f. due to the armature field, which are not compensated by anything.

With a decrease in brush overlap, despite an increase in

Fig. 1. Curves of the commutation current of the transverse and longitudinal circuits of the EMA at \(\beta = 1.5\).

Fig. 2. Curves of the commutation current of the transverse and longitudinal circuits of the EMA at \(\beta = 1\).

the current density in the brush contact by a factor of 1.5, the commutation conditions practically did not change. Sparking under the transverse brushes did not exceed \(1^{1/2}\), and under the longitudinal brushes, \(1^{1/4}\) points on the GOST scale.

The current curves likewise do not show any significant changes in commutation with a reduction in brush overlap. Moreover, the current curves for different brush overlaps show that, with a smaller brush width, the contact is more stable both at the entering and at the

at the receding edges of the brush. This is probably due to the better distribution of current in the contact at an increased current density. This is especially characteristic of the transverse circuit.

With \(\beta = 1\), mutual-inductive coupling between the sections commutated under one brush is eliminated. Therefore, the experimental current curves for the various sections are very similar in character. Certain differences are explained by mutual-inductive coupling with sections commutated by the other brush. However, this coupling is insignificant because of the shortened winding pitch.

Studies of commutation with different compensation have shown that the degree of compensation largely determines the character of commutation in the longitudinal circuit. The role of the interpoles in this case is very small. This is as it should be, since in the EMM the number of turns of the interpole winding is very small in comparison with the compensating winding and the armature winding. Thus, in the machines investigated, the number of turns of the interpole winding is 5, while the number of turns of the armature winding per pole is 87. Therefore, an undercompensation of only 5% practically reduces the role of the interpole winding to nothing.

Figure 3 presents experimental curves of the commutation current of three sections of one slot in the longitudinal circuit at different degrees of adjustment of the compensation of the EMU-12A. The brush overlap is \(\beta = 1\); the current of the armature and of the interpole winding is \(5\ a\). The designations of the curves in Fig. 3 correspond to:

1 — 50% undercompensation,
2 — 2% undercompensation,
3 — unit compensation,
4 — 3% overcompensation,

Fig. 3. Curves of the commutation current of three slot sections at \(\beta = 1\) and different degrees of compensation.

5 — overcompensation 20%.

The curves obtained clearly show that the commutation conditions in the longitudinal circuit are determined mainly by the degree of compensation of the armature reaction, while the favorable influence of the auxiliary poles is felt only at a compensation that differs little from unity. Therefore, removing the auxiliary poles along the longitudinal axis of the amplifier, as was proposed in [3], will not lead to any substantial deterioration of the commutation conditions even at unity compensation.

At the same time, if the compensation differs significantly from unity, the absence of auxiliary poles will lead only to improved commutation. This will occur because the field from the uncompensated magnetomotive force along the longitudinal axis in the commutation zone will be considerably weakened owing to the presence of a large air gap.

In the presence of auxiliary poles, as was the case when taking the curves in Fig. 3, the uncompensated magnetomotive force of the longitudinal axis creates a considerable field in the commutation zone, which leads to sharp disturbances of the commutation conditions (curves 1, 5). In magnitude this magnetomotive force is smaller than the magnetomotive force along the transverse axis by a factor of 1.34 for curve 1 and by a factor of 1.4 for curve 5. Nevertheless, in the transverse circuit such a sharp disturbance of commutation is not observed (Fig. 2). This is explained by a considerable weakening of the field, since the commutation zone of the transverse circuit is located under the large stator slot. Consequently, the presence of auxiliary poles along the longitudinal axis, with compensation differing significantly from unity, leads not to an improvement but to a deterioration of the commutation conditions.

The experimental investigations carried out showed that deviations of the compensation by 2–3% from unity do not lead to a noticeable deterioration of commutation in the longitudinal circuit. Thus, when curves 2, 3, and 4 were taken, sparking under the longitudinal brushes was completely absent. With a further deviation of the compensation to 4–5%, sparking begins to develop, but it does not exceed 1½ points on the GOST scale. When curves 1 and 5 were taken, i.e., with compensation deviations of 50 and 20%, the sparking already becomes 2 points or worse, with burning of the trailing edge of the brushes, which is impermissible for an electric-machine amplifier.

The investigations carried out make it possible to draw the following conclusions:

1. The method for calculating commutation proposed in [1, 2] gives satisfactory accuracy for practice.

2. The commutation conditions in the longitudinal circuit of an electric-machine amplifier are determined mainly by the adjustment of the armature-reaction compensation.

For the machines investigated, the deviation of the compensation from unity should not exceed 4–5%. From the point of view of preserving normal commutation, the most expedient deviation is 2–3%.

1. For a transverse-field EMA, it is possible to remove the auxiliary poles along the longitudinal axis without substantially disturbing commutation. At the same time this will lead to an increase in the permissible deviation of compensation according to the commutation conditions.

REFERENCES

1. B. I. Kostylev. Investigation of a transverse-field electric-machine amplifier when operating with a variable control signal. Dissertation, Tomsk, 1966.

2. A. I. Skorospeshkin, B. I. Kostylev, G. G. Konstantinov. Analytical study of the commutation of commutator electrical machines with the use of an electronic digital computer. Izvestiya TPI, vol. 166, Tomsk, 1966.

3. A. I. Skorospeshkin. Some problems of commutation in transverse-field electric-machine amplifiers. Izvestiya TPI, vol. 117, Tomsk, 1963.