In order to stabilize the circuit and suppress this spurious oscillation, a resistor is placed in series with the secondary timing capacitor. The value of this resistor must be such that the critical damping resistance of the unwanted LC circuit is exceeded, while the critical damping resistance of the desired LC circuit is not reached. The values have been checked experimentally for various designs of transformers and it has been found that a wide range of values can be used without exceeding these limits. Approximately 10,000 ohms is satisfactory for most applications, with values ranging from 5000 to 25,000 ohms being acceptable on the average. After the insertion of this damping resistor, the wave-form will again resume the appearance shown in Figures 48, 51, and 53. Under no circumstances should the resistance be placed in series with the portion of the timing capacitance placed in the primary circuit, because if a value large enough to accomplish the desired result is used, the advantage gained in suppressing the starting arc will be lost. Even when using the secondary capacitance alone, as in the case of 6-volt applications, it is wise to use the resistor in series with the capacitor, as the slight additional expense is offset by a more stable circuit and improved performance. This is especially true of the higher-flux-density designs of transformers.
Perhaps it should be pointed out at this point that when using the primary timing capacitor the wave-form will have one additional characteristic not illustrated. There will occur at the make point on each half of the cycle, points (1) and (3), a fine line peak rising a short distance above the horizontal top of the wave. This is caused by the in-rush of charging current into the primary capacitor upon the closing of the contacts. Should the damping resistor be used in series with this capacitor, these peaks would be reduced, or eliminated. However, since the purpose of the primary section is to eliminate the starting arc, the small amount of contact deterioration caused by the absence of the resistor is justified in order to accomplish this purpose.

Identification of faulty wave-forms and their interpretations is sometimes of inter¬est. One of the most commonly observed characteristics of this nature is illustrated in Figures 55 and 56. These wave-form illustrations show the effect, in an exaggerated manner, of vibrator contact chatter and bounce. Figure 55 shows the appearance of the wave-form when operating on a transformer-timing capacitor circuit. The sharp peaks shown at the make of the vibrator at points (1) and (3) represent a condition of chatter existing in the vibrator. The sharp "V" appearing in the latter portions of intervals "ti" and "t3" represent a bounce, or "hop-off," of the contacts after they have originally come to rest and are supplying load current. Chatter is distinguished from bounce by the relative length of contact-dwell occurring between contact openings and the rapidity at which the dwell periods occur. This is best illustrated by reference to Figure 56, which shows the same vibrator as in Figure 55, but now operating upon a center-tapped resistor. The elimination of the inductance and capacitance permits the transients to appear on the oscilloscope in a more pronounced manner, resulting in greater ease of identification of their source.

The final illustration of wave-forms is shown in Figure 57, covering the characteristics of worn vibrators and improper starting. The wave-form shows the respective contact-dwell periods, "ti" and "ta," as they are affected by low amplitude and erratic contacting action. Also, it will be noted that the switching intervals, "t2" and "t4," have greatly increased in respect to "ti" and "ta." These operating characteristics have resulted in bad over-closure of the wave-form, typical of so-called "single-footing." It is quite possible that the addition of a large amount of timing capacitance to the circuit would result in the unit returning to normal operation, unless the wear has been so great as to cause the unit to be completely worn out.

These calculated values are for 100 % closure. Thus, if 100 % closure was secured at 8.0 volts, (and all vibrator constants remained the same), 81.6% closure would result at 6.3 volts. To obtain the value required for a theoretical closure of 60% these values are divided by .6. Thus, a capacity of .0052 mfd. is required at 6.3 volts and .0063 mfd. is required at 8.0 volts. The compromise value would be .006 mfd. as compared to .006 mfd. as determined by the Standard Transformer Comparison method, which illustrates that the calculated value was very close to the required optimum.

Thus, if 100% closure is achieved at 8.0 volts, 81.8% would result at 6.3 volts. Again, translating these values to those required for 60% closure, it will be found that .0075 mfd. is required at 6.3 volts and .0092 mfd. at 8.0 volts. The compromise value of .008 mfd. compares with .008 mfd. found by the Standard Transformer Comparison method.

Thus, at the low flux-densities being used in this design there is practically no change in the required value of timing capacitor over the rather narrow range of input voltages considered. If any, the change is in the reverse direction. For 60 % closure, a value of .007 mfd. would apply as compared to the Standard Transformer Comparison method value of .005 mfd. In this case, a correction factor would have to be used with the calculated value.

The above indicates that a rather close spread of capacitance values exists in this design, and that for 250-cycle operation the value of capacitance is very small indeed, compared to 115-cycle operation. Then for 60% closure, the capacitance values are .0027 for 6.3 volts and .0037 for 8.0 volts; a compromise value of .003 mfd. could be used. This compares with the .003 mfd. value determined by the Standard Transformer Comparison method.