Commutation and brush loss

The paper discusses the brush loss in direct-current machines and calls attention to the relation between the commutating characteristics of the machines and this loss. Examples are given of commercial brush drop tests and reasons suggested for their great variation. These tests as usually made on machines are practically worthless as a measure of the actual brush loss. Part of this loss is really a commutation loss, but there is no easy way of separating it from the brush loss proper. The relation between the short circuit or reactance voltage, brush drop and commutation is illustrated by several examples. A machine with difficult commutating characteristics must have a high resistance brush or it will be very sensitive to sudden changes in load, brush spacing, vibration, etc. The actual current density in parts of the brush is shown to often be two or three times the apparent density. When local currents are eliminated by using narrower brushes, resistance leads or commutating poles, higher apparent densities are permissible. On the assumption of constant brush resistance, the brush loss is calculated and the increase in loss, due to local currents shown. Actually the increase in loss is much greater than indicated by these calculations due to the fact that the brush resistance decreases with increasing current. At present no way of calculating or measuring the actual loss at the brushes is known. At best only a rough approximation can be obtained. Probably in a majority of cases the most satisfactory method of dealing with this loss is to obtain figures for it from brush loss tests on slip rings or short-circuited commutators. The increase in loss actually occurring can be regarded as a commutation or load loss and as part of the stray losses of the machine.