STUDY OF COMMUTATION IN SERIES “P” MACHINES

A. I. Skorospelshkin, Yu. P. Galishnikov

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

This paper presents some results of a study of the commutation of direct-current machines of the “P” series, namely motors of the 3rd frame size. The study was undertaken in order to assess the commutational stability and to examine the electromagnetic conditions of commutation of the indicated machines.

Ensuring satisfactory commutation in highly utilized machines, such as the motors under study, is generally difficult; in the present case the matter is further complicated by the fact that the motor supply voltage varies smoothly within the range from 320 to 175 V, with a corresponding reduction in rotational speed from 3900 to 2400 rpm at a relatively constant torque. Obviously, the most severe condition with respect to commutation is operation of the motor at a supply voltage of 320 V, when saturation of the machine magnetic circuit reaches its greatest degree, and the influence of mechanical factors (eccentricity and irregularities of the commutator surface, etc.) is also large owing to the considerable rotational speed.

Let us briefly consider the basic parameters of the motor under study. The number of pole pairs is \(p = 1\). The motor is provided with two additional poles. The winding is a simple lap winding with the number of sections per slot equal to 4. The winding pitch is diametral. The brush covers 3 commutator bars. The number of brushes of one brush bolt is 2.

A preliminary examination of the motors subject to study, which included taking profilograms of the commutators, measuring the air gaps under the main and additional poles, and also the distances between the edges of the pole shoes of adjacent poles, showed that mechanically the given specimens are satisfactory.

In the study, as comprehensive an evaluation as possible was made of the commutation conditions of the machines under investigation. The research program included the following:

1. Taking adjustment curves.
2. Oscillographing the current curves and the e.m.f. of the commutated sections.
3. Oscillographing the pulses of the transient voltage drop under the trailing edge of the brush.

In order to determine the commutation period of the armature sections, a device was used that makes it possible to obtain pulses, co-

corresponding to the running of each commutator bar under the brush. Sparking was monitored visually. A comparative analysis of the no-spark-zone curves obtained by the no-spark-zone method for voltages of 320, 220, and 175 V, respectively, shows that a considerable narrowing of the machine’s no-spark zone takes place as the supply voltage is increased. Indeed, if at a voltage of 175 V the width of the no-spark zone at rated load is 13.5%, then at 220 V it is 8%, and at 320 V only 1.5%. At the same time the recommended value \(\pm J_{\mathrm{p}}/J_{\mathrm{n}}\cdot 100\%\) for small machines is \(10 \div 25\%\) [1].

The reason for such low commutating stability of the machines under study, in our opinion, is twofold.

First, the presence of a certain uncompensated e.m.f. is of considerable importance; its magnitude increases with increasing supply voltage and, consequently, with increasing rotational speed of the machines. Oscillograms confirm the existence of such an e.m.f., which has the form of pulses corresponding to the beginning and completion (in the case of delayed commutation) of commutation of the sections.

Second, the field of the main poles has a significant effect on the narrowing of the no-spark zone as a result of the increased degree of saturation of the machine’s magnetic circuit when the supply voltage is raised. As the calculations performed show, the magnitude of the e.m.f. of the commutated section due to the field of the main poles at the beginning and at the end of the commutation zone increases by more than a factor of 2.5 when the supply voltage is raised, and at a voltage of 320 V is approximately equal to one half of the reactive e.m.f. \(e_{\mathrm{r}}\). Being directed in accordance with \(e_{\mathrm{r}}\), this e.m.f. acts on the commutation of the section in a retarding manner.

The consequences of insufficient commutating stability of the machines are obvious. An unfavorable combination of mechanical factors (commutator irregularities, increased vibrations of the brush–commutator assembly, asymmetry of the magnetic circuit, etc.) can easily lead to sparking that cannot be eliminated by adjusting the interpoles; moreover, this sparking may arise not immediately, but after more or less prolonged operation of the machine under load.

To improve the commutating stability of the machines under study it is advisable to weaken the penetration of the field of the main poles into the commutation zone by reducing the degree of saturation of the magnetic circuit, or by decreasing the width of the pole shoe of the main pole. In both cases this will lead to some increase in the axial dimensions of the active parts of the machine.

A more detailed study of the commutation conditions was carried out by oscillographing the current curves of the commutated sections. As is known, such an investigation is effective only when, for study, some group of periodically repeating armature-winding sections characteristic with respect to the commutation conditions is singled out, on the assumption that for all other similar parts of the winding the commutation conditions are preserved to one degree or another. In the present case, such a characteristic group consists of 4 sections lying in one armature slot.

Taking into account that the condition of the brush contact and the polish on the commutator are of extremely great importance, the brushes were carefully run in before oscillographing the current curves.

Oscillograms of the currents of the commutated sections, obtained at a supply voltage of 320 V and rated load of the machine, are shown in Fig. 1.

In studying these curves, what is of primary interest is the fact that the actual commutation period is considerably shorter than the time during which sections are short-circuited by the brush. Commutation begins somewhat later and ends considerably earlier than the corresponding moments at which the commutator bars enter and leave the brush. This applies to a greater extent to sections undergoing accelerated commutation, for which the actual commutation period amounts to only about 70% of the calculated one. However, premature completion of the commutation cycle occurs even in the case of clearly delayed commutation.

Fig. 1. Current curves of commutated sections of one armature slot.
\(t_k\) — commutator-bar period, \(T_k\) — calculated commutation period.

This fact can apparently be explained by the discrete nature of the sliding contact. The points of direct contact, the number of which for one brush according to [2] is 8–10, are distributed randomly over the face surface of the brush, so that in most cases the edges of the brush prove to be inoperative. At the same time, the active area of the brush surface itself is, to a considerable degree, a random variable.

Further, in analyzing the current oscillograms, particular attention is drawn to the pronounced nonidentity in the course of the commutation-current curves of sections lying in one armature slot. Thus, if the first three sections undergo accelerated commutation, then the last, fourth section has markedly delayed commutation. This delay in commutation of every fourth section, causing a certain breaking current, leads to sparking of the machine. The fact that, despite the presence of this delay under all operating conditions, sparking occurs only when the load is increased can be explained by the dependence of sparking on the magnitude of the current interrupted by the brush at the end of the commutation cycle, and by the ability of the brush to carry out this interruption without sparking when the magnitude of the breaking current is below a certain critical value.

The presence of nonidentity in the commutation of the slot sections is especially clearly confirmed by an oscillogram of the pulses of the transient voltage drop at the leaving edge of the brush (Fig. 2). At the instants corresponding to the departure of every fourth commutator plate, the transient voltage drops increase substantially. Moreover, even when the machine is operating in the middle of the sparkless zone, this increase in transient voltage is quite perceptible.

Fig. 2. Oscillogram of pulses of the transient voltage drop under the leaving edge of the brush.

The regular character of sparking in the machine under investigation is also confirmed by the following. Sufficiently prolonged operation of the ma-

brushes with sparking of \(1 \tfrac{1}{2}\) points or more lead to slight scorching of the trailing edge of every fourth commutator bar.

Such a substantial violation of the identity of the commutation conditions within a single slot is apparently associated with the interaction of sections, consisting in the fact that the sections which begin and finish commutating earlier transfer part of the energy stored in them to the subsequent sections. In the worst position here is the section that is the last in the slot to complete commutation. In our opinion, the mutual influence of the sections can be weakened by reducing their mutual inductance.

The feature of the commutation process in the machines under study considered above imposes very stringent requirements on the brush: on its ability not only to ensure sparkless commutation, but also to help smooth the existing nonidentity by introducing into the commutated circuit the corresponding voltage, determined by the current densities flowing under the leading and trailing edges of the brush. As is known, the voltage drop between the leading and trailing edges of the brush acts in such a way that, in retarded commutation, it strengthens the action of the interpole, while in accelerated commutation it weakens it. Taking this effect of the brush on the course of the commutation process into account is, in our opinion, absolutely necessary. However, this accounting will be most effective only provided that the features of operation of the brush contact are known precisely.

At present it may be considered established that the actual electrical characteristics of a sliding contact are not something completely stable for each grade of brush. As studies [2] show, the conductivity of the brush contact is determined to a significant extent by the state of the commutator film, which, in turn, depends not only on the grade of brushes sliding over the commutator surface, the atmosphere in which the sliding contact operates, and a number of other external factors, but also on the nature of the change in current in the sliding contact during the period of film formation.

In an actual machine, however, when there is a stable patina on the commutator, the electrical conductivity of the brush contact is, in essence, determined entirely by the magnitude and nature of the change in the current flowing through the brush.

Thus, the electrical characteristics of the brush contact are stable only to the extent that the conditions for formation of the commutator film and the conditions of current flow are stable.

Conclusions

1. The investigated Series “P” machines of the third frame size have low commutation stability owing to the strong penetration of the main-pole field into the commutation zone.

2. The immediate cause of sparking is the presence of a breaking current at the end of commutation of every fourth section of the armature winding, caused by the nonidentity of the electromagnetic commutation conditions as a result of the mutual influence of simultaneously commutating sections.

3. The electrodynamic characteristics of the sliding contact are determined by the magnitude and nature of the change in the current flowing through the brush. It is therefore most advisable to study these characteristics under conditions as close as possible to actual ones.

References

1. M. P. Kostenko, L. M. Piotrovskii. Electrical Machines, Part 1. Energiya, 1964.

2. R. Holm. Electrical Contacts. Foreign Literature Publishing House, 1961.