Commutation Properties of Carbon Brushes in Low-Power DC Machines

A. I. Skorospelkin, E. K. Damm, L. Ya. Zinger

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

The question of the commutating ability of carbon brushes is one of the central questions in the theory and practice of commutation, but, despite the considerable number of published works [1–5], it remains open. The investigations carried out by us made it possible to note certain features in the manifestation of the commutation properties of carbon brushes at the completion of the commutation cycle in machines with a relatively small reactive e.m.f. of commutation.

The experiments were carried out on an installation described in detail in [6].

Fig. 1

The commutator \(K\) (Fig. 1), two opposite plates of which are short-circuited, is set in rotation by the DC motor \(D\). During the time when both working segments are short-circuited by the brushes, a current pulse from a special pulse installation \(IU\) is passed through the circuit thus formed. Since the \(IU\) has a large output resistance that varies with time, the pulse is supplied to \(K\) through the pulse transformer \(IT\). All measures were taken to reduce the leakage inductance of the secondary winding of \(IT\).

A stepwise adjustable linear inductance \(L\) was included in the brush circuit. At the moment of starting the \(IU\), the working segments have reliable contact.

contact with the brushes; the duration of the pulses was set such that, when the lamella emerged from under brush 1, a certain current was interrupted, the magnitude of which could be regulated by acting on the VT. Brush 2 is wider than brush 1; therefore the circuit was always opened by the latter. The contact of the opposite lamella with brush 2 can, without introducing significant errors into the subsequent reasoning, be regarded as an additional constant resistance of the circuit. Switch \(P\) serves to change the polarity of brush 1. The curves of the voltage between the working lamella and the brush and of the circuit current were observed on the screen of a two-beam oscilloscope. A typical oscillogram is shown in Fig. 2.

Fig. 2

The current and voltage curves can be divided into three time intervals. In the interval \(ab\) the current decreases smoothly, since the contact area of the lamella with the brush decreases and the transition resistance increases in accordance with the volt-ampere characteristic of the contact. The slope of the current curve in this case depends on the form of the volt-ampere characteristic, and also on the portion of this characteristic corresponding to the interval \(ab\) of the current curve.

In the present case, the interval \(bv\) of the curves under consideration is of interest. During a time approximately equal to 5–10% of the period of contact of the lamella with the brush, the brush manages to commutate a very large current. The transition voltage drop between the lamella and the brush during this time is much greater than that obtained from the volt-ampere characteristic of the contact at the same current densities. If the volt-ampere characteristic of the contact (Fig. 3) is plotted from the oscillograms presented above, then in the interval 2–3 it will have a sharp rise corresponding to the interval \(bv\) in Fig. 2.

Fig. 3

The characteristic shown in Fig. 3 does not reflect the fundamental properties of the sliding contact. The increase in voltage at the trailing edge of the brush can be explained by a gradual relaxation of the contact pressure at the edges of the brush.

In Fig. 4 (left) are shown the dependences of the current \(I_k\), commutated by the brush owing to an increase in voltage at the trailing edge, on the inductance of the circuit for brushes of both polarities and at different rotational speeds of the commutator.

At first the inductance of the secondary winding of the VT, \(L_t\), was unknown; therefore curves were constructed

\[ L = L_c - L_{\mathrm{r}} = f(I_k), \]

where \(L_c\) is the total inductance of the circuit,
\(L\) is the adjustable inductance (Fig. 1).

The dependences \(L = f(1/I_k)\) are rectilinear (Fig. 4, right). By extending these straight lines to their intersection with the ordinate axis, one can determine the inductance \(L_{\mathrm{r}}\) and represent the obtained dependences in the axes \(L_c\) and \(I_k\).

Fig. 4

The curves \(L_c = f(I_k)\) have a hyperbolic form, and for the current \(I_k\) it is valid that

\[ L_c I_k = C = \mathrm{const}. \tag{1} \]

The magnitude of this constant characterizes the commutating capacity of the brush contact at the completion of commutation. \(L_c I_k\) has the dimension of flux linkage \((\mathrm{V}\cdot\mathrm{s})\).

If the area of the shaded region under the voltage-drop curve (Fig. 2) is determined graphically, then with sufficient accuracy we always have

\[ \int_{t_0}^{t_k}(\Delta u - u_0)\,dt = L_c I_k. \tag{2} \]

If it is assumed that, at the completion of commutation during a rapid change of current, the alternating component of the transient voltage is equal to the self-induction emf of the commutated circuit, then expression (2) agrees with the obvious conclusion from the theory of electric circuits that the total change in the flux linkage of the inductive section of the circuit over a definite time interval is equal to the integral of the voltage on this section over the same time interval. In this case, the constant component of the transient voltage, equal to \(U_0\) (Fig. 2), together with the voltage drop across the active resistance of the circuit, balances the emf of the secondary winding of the transformer \(E\).

On the section \(vg\) of the curves in Fig. 2, the lamella no longer has mechanical contact with the brush.

Thus, the section \(vg\) determines the breaking current of the commutated circuit \(I_p\). If it is assumed that the intensity of sparking is proportional

of the energy dissipated by the electric arc at a constant discharge repetition frequency, then for a constant sparking intensity we have

\[ \frac{L_c I_p^2}{2}=C_1=\text{const}. \]

There is no reason to suppose that \(C_1\) depends on the brush grade.

From expression (1) it follows that at large values of \(L_c\), \(I_k \approx 0\), \(I_0 \approx I_p\), and, for a constant sparking intensity, one can obtain the dependence

\[ \frac{L_c I_0^2}{2}=C_2=\text{const}. \]

However, \(C_2\) does not reflect the commutating properties of the brush contact at the completion of commutation, since in this case these properties do not manifest themselves. Consequently, when evaluating the commutating properties of the brush contact at the completion of commutation, the currents \(I_k\) and \(I_p\) should be distinguished.

Since the quantity

\[ \int_{t_0}^{t_k}(\Delta U-U_0)\,dt \]

decreases as the rotational speed of the commutator increases, the value \(L_c I_k\) should likewise decrease. This proposition is very clearly confirmed by experiment. Curves 3 and 4 (Fig. 4) were obtained at a rotational speed of 1000 rpm, and curves 1 and 2 at 2000 rpm. A comparison of the curves shows that the commutating properties at the trailing edge are somewhat better for cathode-polarized brushes and decrease for brushes of both polarities as the rotational speed increases.

In Fig. 5 an experimentally obtained dependence \(I_k=f(n)\) is presented for \(L_c=40\ \mu\text{H}=\text{const}\) for an EG-74 brush of anodic polarity; it shows that \(L_c I_k\) is practically inversely proportional to the rotational speed. Experiments showed that, under identical conditions, the value \(L_c I_k\) for brushes of grades EG-4 and EG-74 differs little and is approximately one and a half times smaller for brushes of grade MG-4.

Fig. 5

Conclusions

1. The commutating properties of carbon brushes manifest themselves in a dual manner. During the commutation period proper, the commutating action of the brush is determined by its volt-ampere characteristic. At the completion of commutation, the commutating capability of the brush is determined by the additional voltage that the brush introduces into the commutated circuit as a result of the gradual relaxation of the contact.

2. The commutating capability of a brush at the completion of commutation can be evaluated by the value \(C=L_c I_k\).

1. For a constant sparking intensity at a constant frequency of discharge occurrence, the following regularity holds:

\[ \frac{L_c I_p^2}{2} = \mathrm{const}. \]

1. The quantity \(L_c I_k\) is practically inversely proportional to the rotational speed of the commutator.

REFERENCES

1. Z. B. Vartanov. Pulse volt-ampere characteristics of brush contact. Vestnik elektropromyshlennosti, 1957, No. 2.

2. P. S. Lifshits. On the assessment of the commutating properties of electrobrush materials. Elektrichestvo, 1958, No. 1.

3. K. Binder. Boundary conditions of arc formation during commutation. ETZ — A, Bd 81, 1960.

4. F. Shreter. Commutating capability of a carbon brush. ETZ — B, 1962, No. 2.

5. M. F. Karasev. Commutation of DC commutator machines. Gosenergoizdat, 1961.

6. A. I. Skorospeshkin, L. Ya. Zinner, E. K. Damm, V. P. Karataev. Apparatus for an experimental sliding brush–commutator contact. Izvestiya TPI, vol. 212 (in press).