SOME RESULTS OF A STUDY OF THE CHARACTERISTICS AND COMMUTATION OF A TRANSVERSE-FIELD EMU WITH A SMOOTH ARMATURE

A. I. Skorospeshkin, G. G. Konstantinov

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

In the design and operation of EMUs, commutation questions present serious difficulties. The commutation process largely determines the quality of operation of the amplifier. This article presents some results of investigations of the principal characteristics and commutation of two transverse-field EMUs with smooth armatures.

We manufactured two transverse-field EMUs with smooth armatures. One of the amplifiers has the following data:

\[ P = 0.7\ \text{kW},\quad U = 30\ \text{V},\quad n = 2900\ \text{rpm}; \]

the other:

\[ P = 0.7\ \text{kW},\quad U = 60\ \text{V},\quad n = 2900\ \text{rpm}. \]

In the first amplifier a simple lap, single-layer winding with one-turn sections was used, and in the second a single-layer winding with two-turn sections was used.

Because of the design features of an EMU with a smooth slotless armature (the winding is laid on the surface of the armature stack and fastened with epoxy resin), its air gap increases considerably, and the amplification factor correspondingly decreases.

To increase the amplification factor of an EMU with a smooth armature, a longitudinal biasing winding was introduced. The longitudinal biasing winding is placed in large slots and is connected into the transverse circuit of the EMU (Fig. 1). In this case the m.m.f. of excitation is the resultant of two m.m.f.’s

\[ F_{\mathrm{v}} = F_{\mathrm{u}} + F_{\mathrm{pod}}, \tag{1} \]

where \(F_{\mathrm{u}}\) is the m.m.f. of the control winding,

\(F_{\mathrm{pod}}\) is the m.m.f. of the longitudinal biasing winding.

Calculations and experimental investigations made it possible to determine the maximum possible distribution of these m.m.f.’s. It is approximately within the following limits: 68 percent falls on the longitudinal biasing winding and 32 percent on the control winding.

With a further increase in the percentage ratio of the m.m.f. of the longitudinal biasing winding, self-excitation of the EMU begins.

Figure 2 gives the dependence of the e.m.f. of the output stage on the control current, \(E_3 = f(i_{\mathrm{u}})\). From Fig. 2 it is seen that one and the same output e.m.f. can be obtained by varying the control current over wide limits at different ratios of the m.m.f. of the control winding and the longitudinal biasing winding.

In the course of the tests the following data were obtained for the principal characteristics of the EMU:

Fig. 1. Diagram of an EMU with a smooth armature:
CU — control winding,
\(P_{od}\) — longitudinal magnetizing winding,
\(P_{oq}\) — transverse magnetizing winding,
\(K\) — compensating winding,
\(R_n\) — load resistance,
\(R_1\) and \(R_2\) — shunt resistances.

Fig. 2. Resultant characteristic
\(E_z = f(i_y)\) at various percentage ratios of the m.m.f. of the longitudinal magnetizing winding.

a) for the first amplifier

\[ K_y = 350,\quad T = 0.2\ \text{s},\quad K_d = 1750. \]

b) for the second amplifier

\[ K_y = 2400,\quad T = 0.23\ \text{s},\quad K_d = 10\,400, \]

where \(K_y\) is the amplification factor,

\(T\) is the speed of response of the amplifier,

\(K_d\) is the quality factor.

The low value of the quality factor \(K_d\) of the first amplifier is explained by the considerable magnitude of the nonmagnetic layer (insulation of the armature packet, the diameter of the armature-winding conductors, the layer of epoxy resin plus the air gap), equal to \(2.4\ \text{mm}\), and by the large number of turns of the transverse magnetizing winding \((\beta = 3)\). In the second amplifier the magnitude of the nonmagnetic layer has been brought down to \(1.57\ \text{mm}\) by using wire of smaller diameter in the armature winding, and the number of turns of the transverse magnetizing winding has been reduced \((\beta = 1.5)\),

where \(\beta = \dfrac{W_{\text{trans}}}{W_a}\) is the ratio of the number of turns of the transverse magnetizing winding to the armature winding per pole.

In amplifiers with smooth armatures the winding is cooled better than when it is placed in the slots of a toothed armature. Therefore the specific thermal loading in an EMU with a smooth armature can be increased.

In Fig. 3 the external characteristics, in relative units, are plotted for the second model of an EMU with a smooth armature and for the production EMU-12A amplifier with the following data:

\[ P = 1.2\ \text{kW},\quad U = 115\ \text{V},\quad n = 2900\ \text{rpm}. \]

Fig. 3. External characteristics of an EMU at various degrees of compensation: — external characteristics of the amplifier with a smooth armature; — · — external characteristics of the production EMU-12A.

It is evident from the figure that, in the case of unit compensation, the external characteristic of the amplifier with a smooth armature has a greater slope. This is explained by the increased magnitude of the voltage drop in the armature as compared with a conventional EMU.

As for the adjustment of the compensation, in an amplifier with a smooth armature it has less influence than in a conventional EMU. This is indicated by the characteristics obtained at $\varepsilon = 0.95$ and $\varepsilon = 0.98$ for the corresponding amplifiers.

In studying the EMU with a smooth armature, the main attention was devoted to the investigation of commutation. The aim was to determine how the EMU with a smooth armature would behave with respect to commutation under various overloads and current densities, and with various degrees of compensation. The influence of additional poles on commutation was also investigated.

The commutation study was carried out both separately in the transverse and longitudinal circuits and with the simultaneous operation of the transverse and longitudinal brushes.

In this process, the current and field curves were oscillographed by means of special turns, bifilar with respect to the sections under investigation. Since the character of the current and field curves in the above-mentioned cases of operation of the amplifier circuits turned out to be approximately the same, Figs. 4 and 5 show the current and field curves for the case of simultaneous operation of the transverse and longitudinal circuits. In Fig. 4, curves $a$ and $b$ are the currents of the longitudinal circuit under brushes of positive and negative polarity; curves $c$ and $d$ are the currents of the transverse circuit under brushes of positive and negative polarity.

Fig. 4. Current curves of the longitudinal $a$ and $b$ and transverse $c$ and $d$ circuits under brushes of both polarities.

Fig. 5. Field curve during simultaneous operation of the brushes of the transverse and longitudinal circuits.

As can be seen from Fig. 4, the current curves in the transverse circuit have a somewhat retarded character; in the longitudinal circuit they approach a straight-line form.

On the field curve (Fig. 5):

I — region of a large groove and
II — region of an intermediate groove.

On the field curve, pulsations caused by the slotting of the stator are visible.

The current and field curves presented were obtained for the second transverse-field EMU with unity compensation, rated load, and a collector temperature of \(+75^\circ\mathrm{C}\). The collector temperature was measured with a thermocouple.

Commutation in an EMU with a smooth armature is exceptionally good. At current densities in the brushes above \(40\ a/\mathrm{cm}^2\), sparking is completely absent.

Studies of commutation showed that the need for interpoles on the longitudinal axis is eliminated.

In one of the amplifiers the interpoles were removed; in the other they were left in place. No substantial difference in the current curves was observed.

Unlike conventional transverse-field EMUs, in amplifiers with a smooth armature the degree of compensation has little effect on the character of commutation.

Thus, for example, in the second model, with overcompensation of 20 percent and undercompensation of 50 percent, no noticeable changes in the character of the current curves were observed. This can apparently be explained by the large magnitude of the nonmagnetic layer and the insignificant armature field in the commutation zone.

Sparkless commutation in an EMU with a smooth armature is explained by the improved commutation conditions in comparison with a conventional EMU. Thus, according to [1], the magnetic permeance of the winding of a smooth armature is approximately 10 times smaller than that of an EMU with a toothed armature. Accordingly, the value of the reactive e.m.f. is also reduced by a factor of 10. The values of the reactive e.m.f. calculated for both EMUs with a smooth armature are given below:

First amplifier:

\(e_{rq}=0.026\ \mathrm{V},\) \hspace{4em} \(e_{rd}=0.053\ \mathrm{V}.\)

Second amplifier:

\(e_{rq}=0.086\ \mathrm{V},\) \hspace{4em} \(e_{rd}=0.11\ \mathrm{V},\)

where

\(e_{rq}\) — value of the reactive e.m.f. along the transverse axis;

\(e_{rd}\) — value of the reactive e.m.f. along the longitudinal axis.

The conditions in amplifiers with smooth armatures that favorably affect commutation undoubtedly include the following: identity of all commutation cycles, and absence of pulsations of the magnetic flux in the commutation zone.

The mechanical-strength tests carried out on these EMUs showed the high reliability of smooth armatures. As a result of prolonged tests, no changes in or damage to the armature were found.

On the whole, the following conclusions may be drawn from the results of the investigations:

1. In its principal characteristics (\(K_u\), \(T_1\), external characteristics), an amplifier with a smooth armature is on the same level as standard transverse-field EMUs.

2. In terms of commutation, in comparison with a conventional EMU, it is characterized by sparkless commutation under considerable overloads and high current densities under the brushes (above \(40\ a/\mathrm{cm}^2\)).

3. Amplifiers with smooth armatures have high mechanical reliability.

REFERENCES

1. E. Arnold and I. La-Cour. Direct-Current Machines, GONTI, vol. 1, 1931.