INSTRUMENT FOR MEASURING THE INTENSITY OF SPARKING AT THE COMMUTATOR IN DC MACHINES

Z. F. Idrisov, G. P. Trofimov, A. Ya. Tsirulik, V. S. Stukach

(Presented by the scientific seminar of the Department of Electrical Machines)

In view of the subjectivity of visual assessment of sparking at the commutator, attempts are being made to create objective methods for measuring sparking by means of instruments that would measure the parameters of objective signals generated by sparking. Devices are known that assess the level of sparking from the magnitude of the parameters of high-frequency voltage components measured on brushes of opposite polarity. Devices that use photoelements as sensors assess the sparking level from the magnitude of the photocurrent.

The most advanced are instruments of the II type (their latest modification is IE-1), created under the direction of M. F. Karasev. In these instruments, voltage pulses at the trailing edge of the brush are used as the signals carrying information about sparking. All the devices listed are based on one principle—the principle of averaging the parameters of discharges over all lamellae.

In article [1] the inadequacy of this principle was substantiated, and it was shown that the sparking level is determined by the parameters of the most intense spark-arc discharges, while the parameter of spark-arc discharges that unambiguously determines their energy, for a known inductance of the armature-winding section, can be the discharge duration, determined by the duration of the discharge-voltage pulses.

The authors have developed a method for measuring the intensity of sparking at the commutator, based on measuring the duration of voltage pulses at the trailing edge of the brush, classifying the pulses according to the duration criterion in accordance with the function relating discharge duration to sparking intensity, which was found by the authors [1], and evaluating the sparking level from the pulses of maximum duration. To implement this method, a sparking meter has been developed.

As a sparking sensor, one may use an additional brush (the method of M. F. Karasev) or a contactless capacitive sensor according to the method developed in article [2]. The instrument uses the second method. It is simple and convenient for implementation under the conditions of a plant test station, and does not require preparing the machine for the measurement process.

The functional diagram of the instrument is shown in Fig. 1, and the timing diagrams explaining the operating principle are shown in Fig. 2. With the aid of a broadband non-inverting amplifier, the signal from the capacitive sensor, with an amplitude of 100–120 mV (depending on the dimensions of the sensor and the dis-

from the brush) is amplified to several volts for reliable control of the sawtooth-voltage generator. The amplifier must be broadband so as not to increase the duration of the fronts of the monitored signal and thereby introduce an additional error into determining the duration of the arc that arises.

Signals from the sensor may be of either positive or negative polarity, whereas positive pulses are required to control the sawtooth-voltage generator (SVG). The polarity of the signal is changed by an inverter; selection of the positive pulse is performed by switch B, which makes it possible to measure sparking separately for undercommutation and overcommutation. The amplitude of the SVG output voltage, proportional to the duration of the microarc, is compared in the amplitude comparator AK with a reference voltage level, for example \(U_{\text{op1}}\). All signals exceeding this level cause the comparator to operate.

Fig. 1. Functional diagram of the instrument with one channel for discriminating discharge duration

The very first pulse from AK switches the flip-flop into a state in which the indicator switch closes and lights the lamp. Simultaneously, a transition from the other output triggers the monostable multivibrator, which, after the regeneration time \(t_{\text{reg}}\), returns the flip-flop to its initial state. Such automatic resetting of the flip-flop makes it possible to exclude indication of random noise signals acting on the flip-flop both through the preceding devices and through the power-supply circuit. If the signal was not random, then after the time interval \(T\) of one revolution of the machine the flip-flop will again switch and the lamp will light. The return of the flip-flop to its initial state could be carried out after its operation on each pulse.

Fig. 2. Timing diagrams of instrument operation

However, if an arc arises only on one or even several commutator bars, then the lamp burning time relative to the period will be small and the light from the lamp may go unnoticed. Intensification of the luminous flux from the lamp by increasing the brightness of the flashes

may cause the lamps to fail. Therefore it is desirable that \(t_{\mathrm{reg}}\) constitute a large part of the period \(T\). In this case it is necessary to take into account that devices of the multivibrator type, after the end of the interval \(t_{\mathrm{reg}}\), have a considerable subsequent stage of recovery \(t_{\mathrm{r}}\) of the stationary state, amounting to one third of the time \(t_{\mathrm{reg}}\) [3]. Obviously, the inequality

\[ t_{\mathrm{reg}} + t_{\mathrm{r}} < T . \tag{1} \]

must be satisfied.

During the interval \(t_{\mathrm{reg}}\), signals from the sensor are not recorded. The instrument indicates the presence of at least one pulse whose duration exceeds the reference duration set by the reference voltage \(U_{\mathrm{op1}}\). For this reason the duration \(t_{\mathrm{reg}} + t_{\mathrm{r}}\) is immaterial. After the waiting multivibrator and the trigger return to the initial state, the instrument will record the very first signal from the next passing plate, if the duration of the arc occurring on it exceeds the specified value.

Thus, the next indicated pulse may be the one which in Fig. 2 has duration \(t_3\). The maximum sparking intensity can be found by varying the calibrated reference voltage from the maximum negative level to zero. At the instant when the lamp lights up, \(t_3\) is determined from the readings of the corresponding scale \(U_{\mathrm{op}}\).

Fig. 3. Schematic circuit diagram of the instrument

Let us note the features of the schematic circuit (Fig. 3). A two-stage amplifier with series voltage feedback is assembled on transistors \(T_1\) and \(T_2\). The gain, determined by the ratio

\[ K = \frac{R'}{R''}, \]

is adjustable, which is convenient when setting up the system, since the amplitude of the signal from the sensor varies depending on the location

its installation on the brush or on the parameters of the type of machine under investigation. The signal from the sensor, obtained from the brushes of series-P machines, was 100–120 mV; consequently, for reliable control of the GPN it is sufficient to provide \(K>50\). The dc feedback, implemented by means of resistor \(R\) and capacitor \(C\), serves to stabilize the operating point. Series voltage feedback makes it possible to increase the input resistance of the amplifier, which is very important, since the signal under investigation is taken from a high-impedance sensor.

The inverter \(T_3\), with a unity transmission coefficient, has no special features.

The GPN is made on transistors \(T_4\) and \(T_5\) according to a circuit with compensating positive feedback [3]. Its output signal controls an amplitude comparator on transistor \(T_6\). A diode-regenerative comparator circuit [3] was selected, which showed reliable operation under real conditions. The reference level \(U_{\text{op}}\) at the base of diode \(D_4\) is set by potentiometer \(R_{\text{p}}\), supplied from the common stabilized source \(E_k\). Capacitor \(C_{\text{sh}}\) shunts part of \(R_{\text{p}}\), eliminating its influence on the operation of the AK.

The trigger on transistors \(T_7\) and \(T_8\) operates with fairly deep transistor saturation (\(N>4\)) in order to avoid false operation under conditions of a significant interference level.

The monostable multivibrator on transistors \(T_9\) and \(T_{10}\) is triggered by the differentiated positive transition from the collector of \(T_7\). The regeneration time is determined by the relation [3, 4]

\[ t_{\text{reg}} \simeq 0.7 \cdot R_6 \cdot C^* . \tag{2} \]

To reduce the recovery time, an emitter follower on transistor \(T_{11}\) is built into the monostable multivibrator. In this case

\[ t_{\text{v}} \simeq 4 \cdot r_{\text{out}} \cdot C^*, \tag{3} \]

where

\[ r_{\text{out}} \text{ is the output resistance of the follower.} \]

The values of \(R_6\) and \(r_{\text{out}}\) in a specific device are easy to estimate. Then, knowing the rotation period of the machine rotor, one can find \(C^*\) from relation (1). Under laboratory or factory operating conditions the ambient temperature is usually \(25 \pm 10^\circ\text{C}\). In this case the temperature dependences of \(t_{\text{reg}}\) and \(t_{\text{v}}\) may be neglected.

If necessary, the effect of temperature on the main component can be estimated from the equation [4]

\[ t_{\text{reg}} R_6 \cdot C \cdot \ln \frac{2E_k + I_{\text{ko}} \cdot R_6}{E_k + I_{\text{ko}} \cdot R_6}. \tag{4} \]

where

\[ I_{\text{ko}} \text{ is the thermal current of the collector junction, which is the main} \]

destabilizing factor in multivibrators. The temperature dependence of \(I_{\text{ko}}\) is known [4].

The indicator switch on transistor \(T_{12}\) is controlled by the transition from the left arm of the trigger and remains closed for the time \(t_{\text{reg}}\). To simplify the circuit, the indicator lamp of type TN-0.3 is connected directly into the collector circuit; the high-voltage transistor MP26B (\(T_{12}\)) is supplied from a 70 V source (at a nominal line voltage of 220 V).

Under the conditions of the manufacturer’s test station, during acceptance tests it is more convenient to monitor the intensity of sparking of machines with a multichannel instrument, the number of channels of which

corresponds to the number of sparking classes according to GOST. Each channel is set to a duration corresponding to a definite sparking class on the GOST scale. Depending on the intensity of sparking, a different number of channels will operate, and from the state of the corresponding indicator lamps one can determine the sparking intensity of the machine under test.

The implemented three-channel version of the instrument has shown reliable operation both under laboratory and factory conditions.

If it is necessary to diagnose the distribution of discharge durations, further improvement of the developed instrument is required.

REFERENCES

1. A. Ya. Tsirulik, V. S. Stukach, Z. F. Idrisov. Investigation of the relation between the parameters of spark-arc discharges at the commutator and the sparking intensity in points. Izvestiya TPI. Present collection.

2. V. S. Stukach, A. Ya. Tsirulik. A method for oscillographing the voltage at the trailing edge of brushes of commutator electrical machines. Izvestiya TPI, vol. 242, Tomsk, 1972.

3. Calculation and design of transistor pulse devices. Edited by M. D. Shterk. Moscow, “Sovetskoe radio,” 1964.

4. I. P. Stepanenko. Fundamentals of the theory of transistors and transistor circuits. Moscow, “Energiya,” 1967.