EXPERIMENTAL VERIFICATION OF A DC MACHINE WITH VALVE-MECHANICAL COMMUTATION

O. A. Bratkovsky, V. V. Ivashin, I. A. Miloradov

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

Recently, considerable attention has been devoted to the use of semiconductor devices for improving commutation in electrical machines. One example of such use is provided by valve-mechanical commutation (VMC) circuits, which significantly increase the commutation stability of DC machines. A fairly large number of VMC circuits is known (see the bibliography [1]); however, the number of works on their experimental investigation is very limited. This hinders the practical application of VMC, despite its positive qualities.

The present article gives the results of an experimental verification of one of the VMC circuits—with the inclusion of valves in the sections being commutated for the entire period of their short-circuiting by a subdivided brush, and with the valves loaded by the full current of the brush set [1]—on a low-power DC machine.

Experimental setup

The experimental setup consists of two identical DC electrical machines on a common bedplate, connected to each other by an elastic coupling. One of these machines \((M_1)\) is made with the usual commutation circuit, and the other \((M_2)\) with the VMC circuit (Fig. 1). The two-machine arrangement increases the capabilities of the setup, since each machine can be tested in motor and generator modes; in this case one of them serves as the load or motor for the other, and it is possible on a single setup to qualitatively compare VMC with ordinary commutation. The main data of the machines when operating in motor mode are:

armature supply voltage — up to 110 V,
supply voltage of the field windings — 27 V,
armature current — up to 10 A,
field current — up to 2.5 A,
rotational speed — up to 8000 rpm,
number of poles — 2,
armature diameter — 6 cm,
armature length — 7.5 cm.

The stator design of both machines is identical. The rotors of machines $M_1$ and $M_2$ are made from identical parts: shafts, armature-iron stacks, and commutators. The armature windings are made with the same wire, with the same total number of conductors. The pitch of both windings is diametral; there is no skew of the armature slots or pole shoes. The numbers of armature slots are also identical (16 each), as are the brush dimensions $(8 \times 12.5 \times 16\ \mathrm{mm})$.

The difference between $M_1$ and $M_2$ is as follows. Machine $M_1$ has one commutator with 32 commutator segments, two brushes (EG-71) spanning two commutator segments each, 8 turns in an armature-winding section, and 90 turns in the auxiliary-pole winding.

Machine $M_2$ has a double commutator, made for simplicity from two commutators of machine $M_1$, arranged on both sides of the armature stack. The number of commutator segments on one commutator is 16 (each consists of two adjacent segments connected electrically). The number of brushes (MGS-7I) is 4; the number of turns in a section is 16, and in the auxiliary-pole winding 130.

To create regimes of positive and negative boosting when determining the sparkless zones, a second (boosting) winding, supplied independently, was wound on the auxiliary poles of both machines in addition to the main one.

The schematic diagram of the brush–commutator unit of machine $M_2$ and the method of connecting the winding to the commutator segments are shown in Fig. 1.

Principle of operation of the VMK circuit

In the VMK circuit (Fig. 1), each brush is divided into two parts whose width is equal to one commutator division. In this case the commutated sections $(S_1—S_{16})$ are short-circuited not by the body of the brush (as in ordinary commutation), but by chains (1; 2 and 3; 4) made up of series-connected brush parts and diodes.

Fig. 1. Schematic diagram of the VMK of the experimental machine

When the rotor rotates, diodes 1; 2; 3; and 4 are switched in turn into the circuits of the commutated sections for the entire period of their short-circuiting by the divided brush. The change in current in the commutated sections occurs mainly under the action of the emf $e_k$ from the auxiliary poles. At the same time, the currents in the brushes and in diodes 1; 2 and 3; 4 change.

For sparkless operation of the trailing edge of the brushes in the nominal

in the mode, the EMF \(e_k\) must be of such a magnitude that the current in the sections being commutated reaches the value of the current of the corresponding parallel branch \((i_a)\) at the instant preceding the opening of the commutated section by the brush. At this same instant the corresponding brushes and diodes are de-energized. The increase of current in the commutated sections (under the action of the EMF \(e_k\)) above the value \(i_a\), from this instant until the section is opened, is prevented by the diodes, to which a reverse voltage approximately equal to \(e_k\) is applied. Thus, the brushes open sections that are already de-energized, and sparking is eliminated. It is evident that with a further increase of the EMF \(e_k\), the sparkless operation of the trailing edge of the brushes will not be disturbed.

When the EMF \(e_k\) is decreased to a value insufficient for the current of the commutated section to reach the value \(i_a\) by the instant of section opening, the brushes will interrupt the residual current and sparking will appear. A feature of the present VMK circuit is also that, with a further decrease and even a change in the sign of the EMF \(e_k\), the residual current interrupted by the brushes does not exceed \(2i_a\) (because of the presence of the diodes).

It is evident that the sparkless zone of a machine with the VMK circuit (Fig. 1) is bounded by only one lower limit.

Results of the Experimental Verification

The sparkless zones of the experimental machines were obtained by the method of V. T. Kas’yanov and plotted in the axes:

\[ I_p=f(I_a)\quad \text{at } n=\text{const}, \]

\[ I_p=f(n)\quad \text{at } I_a=\text{const}. \]

The feed current was determined by the formula

\[ I_p=i_p\frac{W_p}{W_d}. \]

Here \(I_a\) is the armature current,
\(n\) is the rotational speed,
\(I_p\) is the feed current,
\(i_p\) is the current in the feed winding,
\(W_p\) is the number of turns in the feed winding,
\(W_d\) is the number of turns in the interpole winding.

To eliminate the possible influence of changes in the main field on the shape of the sparkless zones, the latter, for both machines, were taken at a constant excitation current \((I_{\mathrm{e}})\). When machines \(M_1\) and \(M_2\) operated in motor and generator modes, the polarity of the brushes was not changed. The experiments showed that, with this procedure, the sparkless zones of the machines obtained in motor and generator modes practically coincide.

The boundaries of the zone of machine \(M_1\) were determined by the appearance of sparking (up to grade \(1\frac{1}{4}\)). The lower boundary of the zone of machine \(M_2\) was determined in two ways:

a) by the appearance of sparking of grade \(1\frac{1}{4}\),
b) by the disappearance of reverse-voltage pulses on the diodes.

The second method determines the boundary of the zone in the absence of commutating properties of the brushes and under conditions in which the reverse current of the diodes is equal to zero, while an infinitely small residual current causes sparking.

Figure 2 shows the sparkless zone for the trailing edge of the brushes of both experimental machines at \(n=4000\ \text{rpm}\). As is evident from the figure, the nature of the sparkless zones with ordinary commutation and with VMK is fundamentally different. If in machine \(M_1\) the sparkless zone is absent at \(I_a\) greater than \(6\ a\), then in \(M_2\) the width of the zone (more precisely, the part of it at

Fig. 2. Limits of the sparkless zones of the experimental machines at variable armature current ($I_{\mathrm{a}}$) and constant rotational speed ($n$): 1—under ordinary commutation, 2—with VMC without taking into account the commutating properties of the brushes, 3—with VMC taking into account the commutating properties of the brushes

with negative boosting) increases approximately in proportion to the armature current.

At an armature current greater than $1\,a$, the area bounded by curves 2 and 3 (Fig. 2) shows the influence of the commutating properties of the brushes, which expand the sparkless zone. At an armature current less than $1\,a$, the boundary of the zone of machine $M_2$ (curve 3, Fig. 2) drops sharply downward, and sparking does not appear at any currents $I_{\mathrm{p}}$. In machine $M_1$, however, there are two distinct boundaries (curve 1, Fig. 2). This is explained by the fact that the residual current interrupted by the brushes in machine $M_2$ does not exceed $I_{\mathrm{a}} = 2 i_{\mathrm{a}}$, and when $I_{\mathrm{a}}$ is less than $1\,a$, the electromagnetic energy of the commutated section is evidently insufficient to produce visible sparking.

It is also seen from Fig. 2 that the boundary of the sparkless zone of $M_2$ without taking into account the commutating properties of the brushes (curve 2, Fig. 2), as the armature current decreases, does not go to zero. The most probable reason for this may be the influence of the main field on the commutation zone, especially strong at small armature currents.

It should be noted that adjustment of the commutation of machine $M_1$ (selection of the best grade of brushes, the shape of the commutating field, etc.) was not carried out. Carrying it out might possibly somewhat broaden the zone. However, the character of the zone would remain the same, since with ordinary commutation it is, in principle, very difficult, and in most cases impossible, to achieve the ideal case—the parallelism of the zone boundaries.

Figure 3 shows the sparkless zones of both machines at an armature current equal to $6\,a$ and a variable rotational speed from 500 to 7000 rpm. It is seen that in machine $M_1$ the sparkless zone is absent at rotational speeds greater than 4000 rpm, while in $M_2$ it is practically independent of speed.

A feature of the boundary of the zone of machine $M_2$ (curve 2, Fig. 3), taken without taking into account the commutating properties of the brushes, is the presence of a critical speed $n_{\mathrm{cr}}$. For $n < n_{\mathrm{cr}}$, sparkless operation is ensured by the commutating properties of the brushes.

In the VMC circuit (Fig. 1), no measures were provided to eliminate sparking at the leading edge of the brushes. Observation of it showed that in both machines it is the same (not more than $1\frac{1}{4}$ points) and occurs at an armature current greater than $6\,a$.

Analysis showed that, from the curves of the currents and voltages of the diodes, it is possible to form a sufficiently complete picture of commutation in the machine.

Fig. 3. Boundaries of the sparkless zones of the experimental machines at constant armature current \((I_{\mathrm{a}})\) and variable rotational speed \((n)\): 1—under ordinary commutation, 2—with VMC without allowance for the commutating properties of the brushes, 3—with VMC with allowance for the commutating properties of the brushes.

Thus, for example, from the diode currents one can judge the change of currents in all the commutated sections of the machine; from the reverse voltage on the diodes, the magnitude of the commutating emf in a section; and from the shape of these curves, the instants at which the sections are closed and opened by the brush, and, consequently, the operation of the brush contact, etc.

In general, the character of the current and voltage curves of the diodes, observed with an electronic oscillograph, fully confirmed the expected operating principle of the VMC circuit. We note only one characteristic feature of the shape of the currents in the commutated sections lying in the same slots. Because of the inaccuracy in the arrangement of the commutator bars, and because of the very nature of the brush contact, commutation in them does not occur simultaneously.

The change of current in the section that entered commutation earlier is sharply slowed when the second section enters commutation. The current of the second section, conversely, changes more rapidly. The two currents are rapidly equalized, and commutation of the current in the sections of a slot ends almost simultaneously. The equalization of the currents in the commutated sections of a slot can be explained only by the difference in the voltage drops across their active resistances and in the brush contacts, and by the simultaneous action of mutual induction.

In the VMC circuit (Fig. 1) this phenomenon is not dangerous, and at times is even useful, since it “corrects” the nonsimultaneity of commutation of the slot sections caused by instability of the contact at the entering edge of the brush.

On the basis of the results obtained, the following conclusions may be drawn:

1. The experimental verification of the VMC circuit fully confirmed its expected operating principle and its high commutation stability under considerable overloads and over a wide range of changes in rotational speed.

2. The commutating properties of the brushes in the VMC circuit noticeably enlarge the sparkless operating zone, especially when \(n<n_{\mathrm{cr}}\).

1. Compared with ordinary commutation, the VMC scheme is less sensitive to disturbances of brush contact that consist in a reduction of the effective period during which a section is short-circuited, since with VMC the time of actual current commutation in the section may be considerably shorter than the period of its short-circuiting by the brushes.

2. The design features of the VMC scheme make it possible to carry out a sufficiently simple study of the commutation process in all armature sections and to exert an effective influence on it.

REFERENCES

1. O. A. Bratkovskii, V. V. Ivashin, I. A. Miloradov, Circuits for valve-mechanical commutation of direct-current machines (see the present collection).