Examination of the two lower diagrams will disclose two types of recommended power supplies with which to operate the equipment. The one at the left has been in successful use for many years, and is to be highly recommended. However, it involves the use of storage batteries with the resultant care and attention that is required. The one at the right is a simplified version of an AC power supply which can be used where batteries are not desired. The version shown will supply only one input voltage without adjustment of the AC voltage; this might be corrected by having taps on the secondary of the rectifier transformer, and have these taps selected by a tap switch in the same manner as in Diagram No. 1.

In considering Diagram No. 1, the basic source of energy is a bank of storage-battery cells, four being used in this nominally 6-volt application. Heavy leads are brought from the four-volt, 6-volt, and 8-volt terminals, and carried to one section of a two-circuit, four-position, heavy-duty tap switch SWi as shown, the negative lead being taken to a D.P.S.T. switch SWV, which serves as an off-on switch. The movable blades of these switches are connected to the input terminals of the equipment by heavy cable leads so that the primary circuit resistance will be held to a minimum value. The total series resistance of the input circuit from the battery terminal to battery terminal, including the voltage selective switch, the off-on switch, the fuse, the connecting cables and the wiring but not including the transformer or the vibrator, should not exceed .070 ohms for the 6-volt equipment. Measurement of such resistance values requires the use of a low range bridge, such as a Kelvin type or careful meter measurements. Reference to Figure 14 will provide information which will assist in the selection of cable and wire sizes.

In order to secure the required voltages at the measuring point in the input circuit, it is usually necessary to provide a charger of some sort with which to raise the battery voltage above its nominal value. The diagram shows a Variac-type of variable-voltage transformer which controls the AC input to a transformer-dry disc rectifier combination, that in turn supplies pulsating DC current to charge the various sections of the battery. In order to avoid separate Variac adjustments at each voltage, series resistors Re and R7 are provided which can be adjusted for the charging requirements. Switch SW7 is provided with an additional circuit which opens the charging circuit when the on-off switch is opened. This prevents excessive variation in the battery voltage while the vibrator circuit is off. The charging circuits are kept isolated from the power circuits by means of the second section of the switch SW«, and separate leads to the battery, and thus impose the least amount of 120-cycle ripple upon the input voltage.

A very definite reason for the extreme precautions that must be taken to avoid excessive primary circuit resistance should be discussed before continuing. Since the current delivered by the battery, or power unit, to the transformer is a series of pulses interspersed with zero-current periods, the resultant voltage drop in the primary circuit is also a series of pulses. With the voltmeter connected at the primary center-tap and the reed terminal of the vibrator, the effect of any voltage drop is to cause a slight error in the reading of the meter. The error is greater with a low time efficiency vibrator design, since the no-current periods will be a greater percentage of the total time and the reading of the DC meter will be high over this longer time, resulting in a higher average reading. The true effective voltage during the "on-contact" time will be lower than the meter reading indicates. If the circuit resistance is allowed to be made high, the value of the IR drop pulses becomes excessive, resulting in a greater error occurring. Thus, if the meter reading is set to a fixed rated value under these latter conditions, the output will be lower than would be the case if the maximum resistance were held to the recommended low value.

These conditions have very definitely been the cause of considerable difficulty in the receiving inspection departments of many purchasers until they were corrected. One practice that has been found to be more or less common is the use of a series rheostat in the battery lead as a control on the meter reading of the voltmeter at the test equipment. Where the test requirements called for a voltage of 6.4 volts, for example, this voltage was secured from four cells of battery and a dropping resistor. In some cases the control was in the form of a potentiometer across the fourth cell, which was an improvement, assuming that a low value of resistance was used. Unless the current through the potentiometer was approximately 10 times that through the load, however, a considerable error resulted, and such currents rapidly discharged the battery.

By the system used in power supply No. 1, certain limitations are incurred in the range of voltages that can be specified for testing. This is because of the difficulty in maintaining battery voltages reasonably constant during the test. Therefore, certain basic test voltages have been developed for such circuits. Starting of new vibrators is made at a maximum voltage of 4.5 volts, and of course, if the units start at a lower voltage they are still satisfactory. The running and load test is usually made at an input of 6.3 or 6.4 volts, as this voltage can be maintained rather easily, and must be adjusted rather closely in order to determine the output accurately. The over-voltage test is made at 8.0 volts, as representative of the upper limit of voltage provided in an automobile by the voltage-regulating system.

Power supply No. 2 permits the adjustment of the input voltage to any desired value within the range of the unit. This is advantageous if voltages are desired which are in the range between those available in the battery system. However, to make the ease of testing equivalent to that of No. 1, three such rectifier units should be used, each adjusted to the desired voltage. Taps will provide different voltages, as mentioned previously, but these will have fixed ratios to the nominal test voltage. (A 3-phase AC input to the power supply will furnish a much steadier DC output with less filtering than will the single-phase construction shown.) Three heavy-duty battery-chargers, such as are supplied for auto-battery charging, could be used with a Variac type of input control and with high capacitance capacitors connected across the outputs. These capacitors must have in the neighborhood of 5000 mfd. capacitance for each ampere of output to approximate the effect of a storage battery.

Power supply No. 2 can be stabilized somewhat, also, by connected a fairly low value of resistance across the output filter capacitor, C4. This can be in the nature of 5 ohms or so, for a 6-volt supply.

As was intimated, the principal tests consist of the observation of wave-form and the electrical performance of the vibrator at various input voltages. The waveform test provides an electrical measurement of the mechanical action of the mechanism.

The unit should start at the designated voltage, or below.

The unit should run at the rated input voltage with a satisfactory wave-form, and provide the minimum output or more.

The unit should run at the designated over-voltage input with a satisfactory wave-form.

The designated input voltages, the acceptable output voltage limits and the acceptable tolerances in wave-form are factors to be mutually determined by the supplier and the purchaser. The waveforms illustrated in the preceding chapters may be used as a guide but should not be set up as standards unless by previous agreement.

There has been no mention made relative to the design, type or construction of the transformer to be used in the test equipment. With the information and discussions that have been propounded in this text, the importance of this component in the test equipment can be readily understood. It would be impossible to include a recommendation for an exact design for this purpose. Again, a mutual arrangement with the vibrator supplier should determine the "standard" transformer for test use. Often a transformer of the same design as used in production can be selected as a "standard." Standard test specifications can be established by the use of a large number of vibrators that are known to be satisfactory for the purpose, and at the same time, a value of timing capacitance can be determined which will provide for average operating wave-forms.