On Valve-Mechanical Commutation of DC Machines

V. V. Ivashin, I. A. Miloradov

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

In DC machines, causes of sparking of an electromagnetic nature often occupy the leading place.

The relatively weak commutating properties of brushes are often unable to ensure good commutation. They manifest themselves the more weakly, the greater the voltages in the commutated circuit, that is, the greater the power of the machine and the greater its load.

The use of valve devices makes it possible, to a considerable degree, to suppress sparking of an electromagnetic nature and, consequently, to increase the commutational stability of DC machines.

The method of commutation set forth below, using semiconductor diodes, is free from a number of shortcomings inherent in known circuits [1, 2, 3] and makes it possible, by a simple means, substantially to improve the commutation of electrical machines.

On the side of the trailing edge of the main brush 1 (Fig. 1a), which carries out the short-circuiting of the section and conducts the main current, an auxiliary brush 2 is installed. The auxiliary brush runs onto the commutator plate when the main brush runs off it, that is, both brushes are overlapped by one commutator plate.

The auxiliary brush is connected with the main brush through the semiconductor valve 3.

Let us suppose that commutation has a retarded character, that is, by the moment when the main brush runs off the commutator plate, the current in the section has not had time to change to the magnitude of the current of the parallel branch.

Fig. 1a. Commutation circuit

Fig. 1b. Change of the commutator-plate current with time.

This means that, at this instant of time ($t_2$ in Fig. 1b), the current $I_2$ flows through the commutator segment. In a normal machine this would inevitably lead to sparking. But at the instant $t_1$ the auxiliary brush runs onto the commutator segment, and the current, under the action of the voltage drop at the trailing edge of the brush, passes into the valve and into the auxiliary brush, which at this time shunt the main brush.

At the same time, the current in the trailing edge of the main brush decreases to zero and becomes equal to zero precisely at the instant when the commutating circuit is opened by the main brush.

The current in the valve and in the auxiliary brush simultaneously increases to a value equal to the current $I_2$, i.e., to the magnitude of the current that the main brush would have been forced to interrupt in the absence of the auxiliary branch.

Beginning at the instant $t_2$, the valve is connected in series with the commutator segment.

Subsequently, the current of the segment, varying according to the natural law, passes through zero (the instant of time $t_3$), and the valve opens the commutating circuit. The auxiliary brush leaves the commutator segment somewhat later than the instant at which the valve is blocked (point $t_4$ in Fig. 1b).

Thus, both the main and the auxiliary brushes disconnect circuits that have already been de-energized.

If the commutating properties of the brush are neglected, sparkless operation of the machine is ensured for all values of the commutating emf for which the instant when the current of the commutator segment passes through zero lies in the interval $t_2—t_4$.

Thus, the width of the zone of sparkless operation is determined primarily by a number of design parameters: the dimensions of the brushes and their mutual arrangement.

If the brushes used possess good commutating properties, this should only further improve commutation.

It is obvious that, for normal operation of the circuit, the auxiliary brush must be narrower than the insulation width between the commutator segments, since otherwise this brush will in turn short-circuit the sections and normal operation of the machine will become impossible.

Practical implementation of commutation according to the scheme described is possible only if the brush–commutator assembly is changed. The commutator must have wide insulating gaps, for which idle commutator segments may be used, or else the commutator segments must have a more complex shape than usual. It is possible to use a second, auxiliary commutator, electrically connected with the main one but having wider insulating gaps.

In this case it is obvious that the total power of the two commutators when this commutation scheme is used is, in general, equal to the power of the commutator with the ordinary method of commutation, i.e., the overall dimensions of the machine do not increase.

The auxiliary commutator may be made with the same number of segments as the main one, and also with a smaller number of segments, if the aim is to improve the commutation conditions only of those sections that commutate last in phase.

Since the magnitude of the commutating emf in the present case does not determine commutation to the same degree as in normal machines, the number of segments of both commutators can be considerably reduced, which will greatly simplify the design and manufacturing technology of the machine.

Experiments and calculations show that, in order to obtain a satisfactory zone of sparkless operation (15–20 percent), the power of the valve

and the auxiliary commutator (in the case of a two-commutator design) amounts to no more than 5–10 percent of the machine power. The remaining 90–95 percent of the current load is carried by the main commutator and the main brushes.

Experimental verification of the commutation scheme set forth above was carried out on a series П motor with a power of 4.7 kW, 1500 rpm, 26 A. The armature winding was wave-wound. \(2p = 4\).

The machine was first re-equipped in such a way as to bring out explicitly the advantages of the commutation scheme under study. The number of sections of the armature winding was reduced, and the number of turns in the sections was increased threefold. Only one brush bracket of each polarity was left on the commutator.

Correspondingly, the number of commutator segments was also reduced. This was achieved by short-circuiting segments of the normal commutator with one another. In this case every third segment was left idle, since the interlamellar voltage increased sharply.

In the final form the number of turns in a section was \(W_c = 24\) (as compared with four turns in the factory machine), so that the commutating emf at rated load on the middle line of the sparkless-operation zone was of the order of 15–20 V.

The investigations were conducted on the basis of a circuit using a second, auxiliary commutator. The number of segments of the auxiliary commutator was equal to the number of segments of the main one, but the insulating gaps on the auxiliary commutator were wider.

In the experiments, the currents of the arc arrester, of the section, and of the valves were oscillographed, as were also the voltages across the valves. The quality of commutation was assessed from the width of the sparkless-operation zone, obtained by the method of V. T. Kas’yanov.

The modifications made in the experimental machine led to the result that the machine completely ceased to operate under the normal commutation scheme, i.e., it sparked at all loads and excitation currents.

Only when valves were used was the operability of the machine restored. Moreover, in the rated mode the width of the sparkless-operation zone was approximately 15 percent, whereas the best machines in factory construction have a zone no wider than 7–9 percent.

Figure 2-a shows the sparkless-operation zone of the experimental machine, obtained at constant rotational speed and constant excitation current. The zone is plotted in the axes \(\pm i_n = f(i_{\mathrm{я}})\) and \(\pm \dfrac{i_n}{i_{\mathrm{я}}} = f(i_{\mathrm{я}})\).

Fig. 2-a. Sparkless-operation zone.

The zone in the axes $\pm i_p = f(i_y)$ (curves 1) has the form of a bell widening as the load increases. This is due to the fact that the width of the zone is determined by the time interval $t_2—t_4$ (in Fig. 1b), which remains constant at $n=\mathrm{const}$. A change in the commutation time by one and the same amount, determined by the segment $t_2—t_4$, as the load increases requires a practically linear increase in the commutating emf. Consequently, the magnitude of the excitation current determined by the sparkless-operation zone also increases with increasing machine load.

If the zone is plotted in the axes

\[ \pm \frac{i_p}{i_y} = f(i_y) \]

(curves 2), then the upper and lower boundaries of the zone, beginning with a certain load, run approximately parallel to one another. The widening of the zone at loads close to no-load is explained by the influence of the commutating properties of the brush, which are relatively stronger at low loads, and by the action of the main field of the machine.

Figure 2b shows superposed oscillograms: the current and voltage of the valve for a brush of one polarity, the current of the commutator bar and the voltage on the valve for brushes of different polarities.

In the oscillograms of the reverse voltage of the valve, dips caused by the presence of mutual-induction coupling between sections of the armature winding are clearly visible: the beginning of commutation of a section under a brush of one polarity coincides in time with dips in the voltage curve of the valve for a brush of the other polarity.

Fig. 2b, c. Oscillograms taken on an experimental machine.
$i_p$ — commutator-bar current,
$i_v$ — valve current,
$U_{\mathrm{pr}}$ — forward voltage drop across the valve,
$U_{\mathrm{rev}}$ — reverse voltage drop across the valve.

It is obvious that an additional condition for normal operation of the circuit is constancy of the sign of the reverse voltage on the valves. Indeed, if the reverse voltage changes sign, the valve will open. This may lead to sparking of the auxiliary brush if the change in sign of the reverse voltage coincides with the instant at which the auxiliary-brush circuit is switched off.

The reason for a change in the sign and magnitude of the reverse voltage on the valve may be, in addition to the influence of mutual induction, the entry of part of the section under the main pole of the machine, following in rotation after the commutation zone and having (in motors) a polarity opposite to that of the interpole.

Thus, the circuit under investigation undoubtedly improves commutation. Other conditions being equal, it makes it possible to obtain significantly-

but it increases the commutating e.m.f. Since the commutating properties of the brushes do not play an essential role, brushes with small values of transition voltage may be used.

These brushes allow higher current densities, which makes it possible to reduce the number of brushes and decrease the dimensions of the commutator.

From a fundamental point of view, the described circuit can be applied to improve the commutation of DC machines of any rating. However, apparently, its use is more justified in high-rating machines and in machines that must have increased commutational stability.

REFERENCES

1. N. S. Klimov, “On improving commutation conditions in DC machines,” Elektrichestvo, No. 8, 1959.

2. I. A. Didorenko, I. M. Kovtun, “Expansion of the sparkless operating zone of a DC machine,” VEP, No. 10, 1957.

3. G. T. Vasiliev, “Electric machine,” USSR Author’s Certificate No. 79604, class 21 d¹, 39.