Vibrator power supply design - Timing Capacitor Considerations - I
By planright Jul 1st, 2009
An ideal condition, where the time efficiency is unity and no interval exists during the switching period of the vibrator, is shown in Figure 41, a reproduction of Figure 5. The contacts break at point (2) at the same instant that the opposite pair of contacts make at point (3), etc. This ideal arrangement, of course, cannot exist and is given for illustration only.
Figure 43 illustrates the change in the wave-form that can be obtained by properly proportioning a resistive load to the prevailing transformer characteristics. The wave-form varies during the "t2" and "t4" intervals according to the value of the load resistance. Such an effect occurs when a heavy "filament" or "heater" tube load is added to a vibrator power transformer. For the values of load normally occurring as "plate" or "B-circuit" loads and similar applications, the effect upon the circuit is not sufficient to prevent the occurrence of disastrous transients. When the self-rectifying vibrator is used, the load is disconnected before the interrupter contacts open and no control is present. When the interrupter vibrator and a rectifying tube is used, the threshold voltage action of the rectifier will produce similar results.
The suppression of these transient voltage conditions requires the use of "buffer" capacitors across the vibrator contacts and will provide reasonable control, thus preventing voltage breakdowns in the transformer insulation or external elements. In the early development stages the capacity of these capacitors was arbitrarily chosen, mainly by guess work and experimentation and without regard to the theory involved. Their use in this manner resulted in rapid material transfer of the vibrator contacts which eventually caused the contacts to "lock" together and shorted the vibrator, or destroyed the contacts.
Further investigation and development resulted in the use of a single capacitor connected across the full winding, which provided the same control of the transient conditions without the disastrous contacting action. Figure 44, a reproduction of Figure 8, illustrates the results obtained with the use of a timing capacitor in securing the desired control. For purpose of illustration, assume that the capacitor has been connected across the primary winding, and has a capacitance of Ci mfds., and that the inductance of the primary winding under the influence of the DC magnetization existing at point (2) in the cycle is Li henries. Upon the break of the contacts at point (2), a shock excitation of the LC circuit occurs, resulting in the start of a highly damped oscillation. If the succeeding half-cycle did not take place, the oscillographic trace of such an oscillation would appear similar to that shown in the dashed wave. The frequency of oscillation depends upons the equation:
The inductance of a transformer of a given design, and the vibrator characteristics are fairly constant. Therefore, in order to change the frequency of oscillation, a change in the capacitance must be made. By properly choosing the value of this capacitance, the slope of the portion of wave forming the first !4 cycle of the oscillation can be so adjusted that the wave will exactly close the gap between points (2) and (3) as shown in the illustration. Thus, when the inertia contact pair closes at point (3), the transformer primary voltage has been reversed during interval "ti" and the contacts close with zero voltage difference between them. The transient at the break has been eliminated and the cause of the transient at the make has been removed. Correspondingly, the same condition occurs when the inertia contacts open at (4). The process is a continuous repetition of the above cycle.
The determination of the value of capacitance that will satisfy the above conditions by the described method, requires an experimental arrangement with the specified transformer and vibrator combination and a variable timing capacitor. It also requires a suitable cathode-ray oscilloscope for observing the wave-form. The oscilloscope should be connected across the entire primary winding, so that any suppression of transient phenomena will be avoided. The connection to the primary, rather than to the secondary, involves working with lower voltages and eliminates a large portion of the effects of the leakage reactance of the transformer.
Occasionally it has been advocated that this capacitance value can be determined by meter readings only, but experience has demonstrated that that method is faulty and inaccurate and should only be used when equipment for wave-form observation is not available.
This method provides an ammeter in the battery lead, the operation of the power supply on no-load, and a variable timing capacitor. The indication of the meter is observed as the value of capacitance is changed, and the point at which a minimum reading on the meter is observed is noted. One disadvantage with this method is the low sensitivity of the indicating ammeter. The change in current is very slight over a rather wide percentage change in capacitance at the minimum point. This is the equivalent of a poor selectivity-curve in a radio receiver or the nose of a probability curve, and the difficulty of deciding the exact center of the curve is easily understood. A partial solution is the determination of equal points on either side and using an average value between them. Also, it is impossible to determine if the vibrator is "balanced," a condition which also affects the capacitance value. The "balance" will be discussed at a later point, and this connection will be further explained.
It is often very desirable to be able to predict the approximate value of the timing capacitor that is to be used with a given transformer design. Provided that the B-H curves used in making the necessary transformer calculations are reasonably representative of the lamination steel and that the Vibrator characteristics are near the average values given on the Data Sheets, this prediction can be made satisfactorily. At least the approximation will serve as a guide to the later accurate experimental determination.
Referring again to Figure 42, the time interval "t," for a vibrator of frequency "f." can be shown as
where the time efficiency, wt, is expressed as a decimal value. The above determines the time interval "ta," during which the voltage-reversal must take place, in terms of the known characteristics of the vibrator, i.e., the frequency and the time efficiency. In the case under consideration, the voltage must change from a maximum (equal to the battery voltage) of one polarity to a maximum of the opposite polarity in the time interval "ta." Therefore, the required timing capacitance for a given set of circuit conditions will be:
It can be assumed that, at no-load, Ei equals the battery voltage Ej. Equation (28) shows that the peak value of the magnetizing current must be determined for any given input voltage, in order to calculate the approximate value of timing capacitance required for that voltage. This follows the determination of the flux-density for the voltage concerned, using the B-H curve for the grade of steel involved, and the calculation of the current using one-half of the number of primary turns and the length of magnetic-flux path.
The use of the foregoing formula will sometimes result in an approximate value of timing capacitance which is different than the optimum value as found by the use of a standard transformer, standard vibrator and oscilloscope. Therefore, this calculated value may have to be modified by a factor which will vary with transformer design. This factor can be pre-determined only by experience. Consequently, the calculated value from the formula must be used only as an approximation and not as the final optimum value.
The foregoing discussion has been relative to the selection of the proper value of timing capacitance to secure complete reversal of the input voltage during the switching intervals ("tj" or "t") of the vibrator. If this has been accomplished, the wave-form will appear as illustrated in
Figure 45. Note that the vertical portions of the wave completely close the gaps between points (2) and (3) and between points (4) and (1). This is known as 100% closure of the wave-form and is a theoretical, or ideal, condition that is not possible or even desirable in actual practice. The attainment of such a condition would require an exact "balance" between the pull and the inertia sides of the vibrator (i.e. the values of time intervals would have to be such that ti =t3 and t2=t4). Also, the magnetizing action of the two halves of the primary would have to be identical and an exact value of timing capacitance would be required. However, this wave-form is the basic standard of reference.
It is interesting to note the change in wave-form when various values of timing capacitance are used. Figure 46 illustrates the change that would appear if too small a value of capacitance is used. The frequency of oscillation increases, reducing the time "t0" for one complete cycle, and causes the peak of the first !4 cycle to pass before the contacts close. This is known as "over-closure." Figure 47 illustrates the use of an excessively large value of timing capacitance. The frequency of oscillation has been greatly reduced, increasing the time "t0" for one complete cycle. This causes the amount of voltage reversal that can occur in the time interval "t2" to be a very small percentage of the needed amount. This is known as "under-closure," or "short-closure."
Both of the above conditions are to be avoided if good performance of the vibrator is to be secured. Over-closure induces poor starting characteristics with the resultant abnormal currents and induced voltages. Also, since any contact erosion (wear) will increase "ts" and "t4," the condition will rapidly get worse with age. Extreme under-closure induces transient voltage and current peaks at the make of the contacts, when the contact pressure is just beginning to build up, and causes rapid erosion, or "transfer" (metallic deposition of the material of one contact upon the other), of the contacts causing short vibrator life. If transfer of the contact material is severe and of the right character, "locking" of the contact pairs is inevitable, with failure resulting. In the past, some vibrator manufacturers have recommended excessive values of timing capacitance to counteract for the poor vibrator starting characteristics. However, increased experience and improved mechanisms have eliminated this problem. A wave-form condition having bad under-closure also will result in the generation of a greater amount 6f "hash" interference, as might be expected, because of the larger transients.
As a result of the above described effects of over-closure and under-closure, a rather general practice has been adopted, which is a compromise with the ideal condition as illustrated in Figure 45. For the average vibrator characteristics, a waveform standard is selected which has less than 100% closure, but which also has sufficient closure to off-set the effects of extreme under-closure. An example of such a wave-form standard is shown in Figure 48, where a closure of approximately 65 % has been illustrated. This permits the voltage-reversal to cross the zero, or neutral, line and increase to a small amount in the opposite direction before the contacts close. It must be remembered, however, that this closure of 65% has been selected on the basis of average vibrator characteristics. In actual practice, considering the production tolerances in vibrators, transformers, and condensers, it is highly improbable that this exact condition will be observed on any combination of these three units taken from production lots. However, these small variations do not seriously affect the overall performance and an adjustment to obtain a 65% closure for each individual power supply is not required.
Examination of Equation (28) will quickly indicate that the value of timing capacitance for 100 % closure will vary inversely as the input voltage, or vibrator frequency, and directly as the magnetizing current. If the vibrator frequency and time efficiency remained constant with changes in input voltage, then the magnetizing current would increase at the same proportional rate as does the input voltages, and the capacitance value required would remain constant for all input voltages. This would be true in practice were it not for the curvature of the B-H curve of the core steel, since the flux-density varies directly with the input voltage. However, at the flux-densities usually employed, the magnetizing current increases at a faster rate than does the input voltage. This infers that a greater value of timing capacitance will be required at the maximum input voltage than will be required at the rated input voltage, in order to maintain 100% closure in both cases. This has been proven in actual practice. An exception is the unusual instance of operation at comparatively low flux-densities.
This would indicate that it is highly desirable for the transformer to be designed so the operating variations in flux-density will follow the straight line portion of the B-H curves as nearly as possible, thus limiting the timing capacitance shift to an acceptable minimum.
Operation of auto-radio power units with a maximum flux-density around 65,000 lines per square inch usually results in a very acceptable shift in closure between 6 and 8 volts input. An increase to a maximum of 75,000 lines will result in a much greater shift in closure between input voltages of 6 and 8 volts and a consequent deterioration of vibrator performance. For example, with a 60% closure at 6 volts, approximately 100% closure would result at 8 volts.
The above presupposes that the time efficiency and the frequency do not change with changes in input voltage. Actually, an increase in voltage generally will cause a slight increase in time efficiency and frequency. The increase in frequency will slightly decrease the required capacitance through direct application to Equation (28). An increase in frequency will also slightly decrease the required capacitance by an indirect means, since it will reduce the flux-density and the resultant magnetizing current. While slight, these changes are in the desired direction. The slight increase in time efficiency will result in a much greater percentage reduction in the values of "t2" and "tt" off-contact times. Equation (28) shows that the value used is (1 — bit) and not merely o>(. Consequently the timing capacitance is reduced proportionately. However, these reductions are not sufficient to off-set the undesirable shift previously mentioned. Vibrators, transformers and timing capacitances for use in production must all be manufactured with certain minimum tolerances of" performance in order to be made economically. As the tolerances are made wider the cost per unit is generally reduced, and where cost is a major factor, tolerances are-set at the maximum allowable limits. However, with three components so interdependent upon each other for the proper-performance of the complete unit, there-must be rather close tolerances maintained if good average performance is to be secured. The effects of variations in the characteristics of the vibrators and transformers from the nominal averages affects the choice of the nominal value of timing capacitance. Since the capacitor is usually selected last among the three components, it is wise to secure an average condition to be used as a basis for such selection. This involves at least several representative sample transformers from all of the suppliers who are expected to furnish production units, and at least six vibrators from the principal source, or sources, of supply. These should be representative of production units such as will be supplied on the order. As was previously indicated, in the sample transformer design notes, the test for selection of the value for the timing capacitance may be coordinated with the test of the sample transformers and vibrators for correctness of the transformer design.
As a further check to determine if the selected combination of transformer and timing capacitance is completely satisfactory for the operation of the vibrator throughout its life, it is advantageous to use some partially wornout vibrators along with the new vibrators in these tests. Good starting should result at all voltages and serious over-closure should not result at high battery voltages. These same worn vibrators are valuable' as a check of the complete receiver for sufficient "hash" elimination filtering and shielding, since the older vibrator will probably create a greater amount of such undesirable "hash" interference.
Regardless of the final application of the power supply, the fact still remains that some wear of the vibrator will necessarily occur with age. The amount of wear will depend upon the design of the mechanism as well as upon the load being commutated and the service conditions, so the application must be considered when the choice of a capacitance value is being made. As wear occurs, the value of "t1" and "t4" intervals will increase; thus the value of original capacitance selected must be larger than is absolutely essential for the original conditions in order to provide a sufficient value for the worn condition. Mobile equipment, such as auto receivers, have a wider range of input voltages than do farm or portable receivers operating without a charger. Thus the spread of wave-form closure to be expected, or which must be tolerated, will be greater in the case of the former than it will be for the latter.
As a result of the foregoing discussion regarding timing capacitance value determination, it can readily be seen that a number of factors must be considered in making this determination. The P. R. Mallory Company has adopted a standard transformer for this purpose and uses a comparison method. This transformer has been so designed as to incorporate the maximum allowable capacitance shift between nominal and maximum input voltages, both load and no load. The buffer capacitance selected is optimum for this transformer and has been determined by years of experimenting and life testing with the transformer. In addition, duplicates of this standard transformer and timing capacitance are used for all factory-adjusting of vibrators. As a result, the use of this standard transformer and timing capacity provides a fast and accurate method of determining the timing capaci¬tance values required for new transformer design. The correct timing capacitance value required for any newly-designed transformer can readily be determined by observing the vibrator characteristics of any production vibrator when operated with the standard transformer and its timing capacitor. Then the newly-designed transformer is substituted for the standard transformer and the timing capacitance varied until the wave-form characteristics have been duplicated.
Timing Capacitor Considerations
The timing capacitor, or "buffer" capacitor, is the third important element to be considered in designing a vibrator power supply. The proper value of the timing capacity must be established if satisfactory vibrator performance and life is to be obtained. In fact, the vibrator, the transformer, and the timing capacitor are so interdependent on each other that one cannot be considered without the other two. The purpose of this chapter is to explain the use of the timing capacitor, its proper location in the circuit and the method of determining its proper value. The wave-form is interrelated to the timing capacity and will also be discussed in this chapter.
An ideal condition, where the time efficiency is unity and no interval exists during the switching period of the vibrator, is shown in Figure 41, a reproduction of Figure 5. The contacts break at point (2) at the same instant that the opposite pair of contacts make at point (3), etc. This ideal arrangement, of course, cannot exist and is given for illustration only.
Figure 42, a reproduction of Figure 6, illustrates the respective time intervals involved in a representative commercial vibrator design. In this practical arrangement, a discontinuity in applied voltage occurs between each half-cycle. Symbols "ti" and "t3" represent the "on-contact" time intervals, while "t2" and "t" represent the "off-contact" or "switching" time intervals. The upper, or first, pulse shown represents the "pull" side contacting interval, while the lower pulse represents the "inertia" contacting interval. The "pull" contacts make at (1) and break at (2), and the "inertia" contacts make at (3) and break at (4), after which the cycle is repeated. The time interval "t" is the length of time required for one complete cycle of operation.
The wave-form shown in Figure 42 can be reproduced experimentally by operating the vibrator into a center-tapped resistor having a comparatively low resistance value, and with provision made for obtaining the correct driving coil voltage. This arrangement offers a satisfactory laboratory expedient in testing vibrators, and will be discussed later. However, it should be noted that the required input voltage for securing proper reed amplitude of the vibrator is dependent upon the type of driving system employed. For a vibrator unit using a "separate-driver" system the nominal battery voltage of the unit is employed. In a "shunt-coil" driver system it is necessary to double the nominal battery voltage of the unit, since with this system the auto-transformer action, that is normally present in the transformer primary, will not be available. When the center-tapped primary of the transformer is connected to the vibrator and battery, an unstable inductive component is introduced into the circuit. Magnetic flux is built up in the transformer core while the contacts are closed during "ti." Since the polarity of the magnetic-flux during the next half-cycle, "t3," is the reverse of that during "ti," the energy previously stored during "ti" must be dissipated during the interval "t2," or a sudden reversal would take place when the inertia contacts close at (3). If no control is placed in the circuit, an excessively high induced voltage would occur at the break of the contacts at (2) when the collapse of the magnetic-flux occurs. A second, but lower, transient voltage would occur at the make of the contacts at (3). The above considerations are all based on a transformer under a "no load" condition.
Figure 43 illustrates the change in the wave-form that can be obtained by properly proportioning a resistive load to the prevailing transformer characteristics. The wave-form varies during the "t2" and "t4" intervals according to the value of the load resistance. Such an effect occurs when a heavy "filament" or "heater" tube load is added to a vibrator power transformer. For the values of load normally occurring as "plate" or "B-circuit" loads and similar applications, the effect upon the circuit is not sufficient to prevent the occurrence of disastrous transients. When the self-rectifying vibrator is used, the load is disconnected before the interrupter contacts open and no control is present. When the interrupter vibrator and a rectifying tube is used, the threshold voltage action of the rectifier will produce similar results.
The suppression of these transient voltage conditions requires the use of "buffer" capacitors across the vibrator contacts and will provide reasonable control, thus preventing voltage breakdowns in the transformer insulation or external elements. In the early development stages the capacity of these capacitors was arbitrarily chosen, mainly by guess work and experimentation and without regard to the theory involved. Their use in this manner resulted in rapid material transfer of the vibrator contacts which eventually caused the contacts to "lock" together and shorted the vibrator, or destroyed the contacts.
Further investigation and development resulted in the use of a single capacitor connected across the full winding, which provided the same control of the transient conditions without the disastrous contacting action. Figure 44, a reproduction of Figure 8, illustrates the results obtained with the use of a timing capacitor in securing the desired control. For purpose of illustration, assume that the capacitor has been connected across the primary winding, and has a capacitance of Ci mfds., and that the inductance of the primary winding under the influence of the DC magnetization existing at point (2) in the cycle is Li henries. Upon the break of the contacts at point (2), a shock excitation of the LC circuit occurs, resulting in the start of a highly damped oscillation. If the succeeding half-cycle did not take place, the oscillographic trace of such an oscillation would appear similar to that shown in the dashed wave. The frequency of oscillation depends upons the equation:
The inductance of a transformer of a given design, and the vibrator characteristics are fairly constant. Therefore, in order to change the frequency of oscillation, a change in the capacitance must be made. By properly choosing the value of this capacitance, the slope of the portion of wave forming the first !4 cycle of the oscillation can be so adjusted that the wave will exactly close the gap between points (2) and (3) as shown in the illustration. Thus, when the inertia contact pair closes at point (3), the transformer primary voltage has been reversed during interval "ti" and the contacts close with zero voltage difference between them. The transient at the break has been eliminated and the cause of the transient at the make has been removed. Correspondingly, the same condition occurs when the inertia contacts open at (4). The process is a continuous repetition of the above cycle.
The determination of the value of capacitance that will satisfy the above conditions by the described method, requires an experimental arrangement with the specified transformer and vibrator combination and a variable timing capacitor. It also requires a suitable cathode-ray oscilloscope for observing the wave-form. The oscilloscope should be connected across the entire primary winding, so that any suppression of transient phenomena will be avoided. The connection to the primary, rather than to the secondary, involves working with lower voltages and eliminates a large portion of the effects of the leakage reactance of the transformer.
Occasionally it has been advocated that this capacitance value can be determined by meter readings only, but experience has demonstrated that that method is faulty and inaccurate and should only be used when equipment for wave-form observation is not available.
This method provides an ammeter in the battery lead, the operation of the power supply on no-load, and a variable timing capacitor. The indication of the meter is observed as the value of capacitance is changed, and the point at which a minimum reading on the meter is observed is noted. One disadvantage with this method is the low sensitivity of the indicating ammeter. The change in current is very slight over a rather wide percentage change in capacitance at the minimum point. This is the equivalent of a poor selectivity-curve in a radio receiver or the nose of a probability curve, and the difficulty of deciding the exact center of the curve is easily understood. A partial solution is the determination of equal points on either side and using an average value between them. Also, it is impossible to determine if the vibrator is "balanced," a condition which also affects the capacitance value. The "balance" will be discussed at a later point, and this connection will be further explained.
It is often very desirable to be able to predict the approximate value of the timing capacitor that is to be used with a given transformer design. Provided that the B-H curves used in making the necessary transformer calculations are reasonably representative of the lamination steel and that the Vibrator characteristics are near the average values given on the Data Sheets, this prediction can be made satisfactorily. At least the approximation will serve as a guide to the later accurate experimental determination.
Referring again to Figure 42, the time interval "t," for a vibrator of frequency "f." can be shown as
where the time efficiency, wt, is expressed as a decimal value. The above determines the time interval "ta," during which the voltage-reversal must take place, in terms of the known characteristics of the vibrator, i.e., the frequency and the time efficiency. In the case under consideration, the voltage must change from a maximum (equal to the battery voltage) of one polarity to a maximum of the opposite polarity in the time interval "ta." Therefore, the required timing capacitance for a given set of circuit conditions will be:
It can be assumed that, at no-load, Ei equals the battery voltage Ej. Equation (28) shows that the peak value of the magnetizing current must be determined for any given input voltage, in order to calculate the approximate value of timing capacitance required for that voltage. This follows the determination of the flux-density for the voltage concerned, using the B-H curve for the grade of steel involved, and the calculation of the current using one-half of the number of primary turns and the length of magnetic-flux path.
The use of the foregoing formula will sometimes result in an approximate value of timing capacitance which is different than the optimum value as found by the use of a standard transformer, standard vibrator and oscilloscope. Therefore, this calculated value may have to be modified by a factor which will vary with transformer design. This factor can be pre-determined only by experience. Consequently, the calculated value from the formula must be used only as an approximation and not as the final optimum value.
The foregoing discussion has been relative to the selection of the proper value of timing capacitance to secure complete reversal of the input voltage during the switching intervals ("tj" or "t") of the vibrator. If this has been accomplished, the wave-form will appear as illustrated in
Figure 45. Note that the vertical portions of the wave completely close the gaps between points (2) and (3) and between points (4) and (1). This is known as 100% closure of the wave-form and is a theoretical, or ideal, condition that is not possible or even desirable in actual practice. The attainment of such a condition would require an exact "balance" between the pull and the inertia sides of the vibrator (i.e. the values of time intervals would have to be such that ti =t3 and t2=t4). Also, the magnetizing action of the two halves of the primary would have to be identical and an exact value of timing capacitance would be required. However, this wave-form is the basic standard of reference.
It is interesting to note the change in wave-form when various values of timing capacitance are used. Figure 46 illustrates the change that would appear if too small a value of capacitance is used. The frequency of oscillation increases, reducing the time "t0" for one complete cycle, and causes the peak of the first !4 cycle to pass before the contacts close. This is known as "over-closure." Figure 47 illustrates the use of an excessively large value of timing capacitance. The frequency of oscillation has been greatly reduced, increasing the time "t0" for one complete cycle. This causes the amount of voltage reversal that can occur in the time interval "t2" to be a very small percentage of the needed amount. This is known as "under-closure," or "short-closure."
Both of the above conditions are to be avoided if good performance of the vibrator is to be secured. Over-closure induces poor starting characteristics with the resultant abnormal currents and induced voltages. Also, since any contact erosion (wear) will increase "ts" and "t4," the condition will rapidly get worse with age. Extreme under-closure induces transient voltage and current peaks at the make of the contacts, when the contact pressure is just beginning to build up, and causes rapid erosion, or "transfer" (metallic deposition of the material of one contact upon the other), of the contacts causing short vibrator life. If transfer of the contact material is severe and of the right character, "locking" of the contact pairs is inevitable, with failure resulting. In the past, some vibrator manufacturers have recommended excessive values of timing capacitance to counteract for the poor vibrator starting characteristics. However, increased experience and improved mechanisms have eliminated this problem. A wave-form condition having bad under-closure also will result in the generation of a greater amount 6f "hash" interference, as might be expected, because of the larger transients.
As a result of the above described effects of over-closure and under-closure, a rather general practice has been adopted, which is a compromise with the ideal condition as illustrated in Figure 45. For the average vibrator characteristics, a waveform standard is selected which has less than 100% closure, but which also has sufficient closure to off-set the effects of extreme under-closure. An example of such a wave-form standard is shown in Figure 48, where a closure of approximately 65 % has been illustrated. This permits the voltage-reversal to cross the zero, or neutral, line and increase to a small amount in the opposite direction before the contacts close. It must be remembered, however, that this closure of 65% has been selected on the basis of average vibrator characteristics. In actual practice, considering the production tolerances in vibrators, transformers, and condensers, it is highly improbable that this exact condition will be observed on any combination of these three units taken from production lots. However, these small variations do not seriously affect the overall performance and an adjustment to obtain a 65% closure for each individual power supply is not required.
Examination of Equation (28) will quickly indicate that the value of timing capacitance for 100 % closure will vary inversely as the input voltage, or vibrator frequency, and directly as the magnetizing current. If the vibrator frequency and time efficiency remained constant with changes in input voltage, then the magnetizing current would increase at the same proportional rate as does the input voltages, and the capacitance value required would remain constant for all input voltages. This would be true in practice were it not for the curvature of the B-H curve of the core steel, since the flux-density varies directly with the input voltage. However, at the flux-densities usually employed, the magnetizing current increases at a faster rate than does the input voltage. This infers that a greater value of timing capacitance will be required at the maximum input voltage than will be required at the rated input voltage, in order to maintain 100% closure in both cases. This has been proven in actual practice. An exception is the unusual instance of operation at comparatively low flux-densities.
This would indicate that it is highly desirable for the transformer to be designed so the operating variations in flux-density will follow the straight line portion of the B-H curves as nearly as possible, thus limiting the timing capacitance shift to an acceptable minimum.
Operation of auto-radio power units with a maximum flux-density around 65,000 lines per square inch usually results in a very acceptable shift in closure between 6 and 8 volts input. An increase to a maximum of 75,000 lines will result in a much greater shift in closure between input voltages of 6 and 8 volts and a consequent deterioration of vibrator performance. For example, with a 60% closure at 6 volts, approximately 100% closure would result at 8 volts.
The above presupposes that the time efficiency and the frequency do not change with changes in input voltage. Actually, an increase in voltage generally will cause a slight increase in time efficiency and frequency. The increase in frequency will slightly decrease the required capacitance through direct application to Equation (28). An increase in frequency will also slightly decrease the required capacitance by an indirect means, since it will reduce the flux-density and the resultant magnetizing current. While slight, these changes are in the desired direction. The slight increase in time efficiency will result in a much greater percentage reduction in the values of "t2" and "tt" off-contact times. Equation (28) shows that the value used is (1 — bit) and not merely o>(. Consequently the timing capacitance is reduced proportionately. However, these reductions are not sufficient to off-set the undesirable shift previously mentioned. Vibrators, transformers and timing capacitances for use in production must all be manufactured with certain minimum tolerances of" performance in order to be made economically. As the tolerances are made wider the cost per unit is generally reduced, and where cost is a major factor, tolerances are-set at the maximum allowable limits. However, with three components so interdependent upon each other for the proper-performance of the complete unit, there-must be rather close tolerances maintained if good average performance is to be secured. The effects of variations in the characteristics of the vibrators and transformers from the nominal averages affects the choice of the nominal value of timing capacitance. Since the capacitor is usually selected last among the three components, it is wise to secure an average condition to be used as a basis for such selection. This involves at least several representative sample transformers from all of the suppliers who are expected to furnish production units, and at least six vibrators from the principal source, or sources, of supply. These should be representative of production units such as will be supplied on the order. As was previously indicated, in the sample transformer design notes, the test for selection of the value for the timing capacitance may be coordinated with the test of the sample transformers and vibrators for correctness of the transformer design.
As a further check to determine if the selected combination of transformer and timing capacitance is completely satisfactory for the operation of the vibrator throughout its life, it is advantageous to use some partially wornout vibrators along with the new vibrators in these tests. Good starting should result at all voltages and serious over-closure should not result at high battery voltages. These same worn vibrators are valuable' as a check of the complete receiver for sufficient "hash" elimination filtering and shielding, since the older vibrator will probably create a greater amount of such undesirable "hash" interference.
Regardless of the final application of the power supply, the fact still remains that some wear of the vibrator will necessarily occur with age. The amount of wear will depend upon the design of the mechanism as well as upon the load being commutated and the service conditions, so the application must be considered when the choice of a capacitance value is being made. As wear occurs, the value of "t1" and "t4" intervals will increase; thus the value of original capacitance selected must be larger than is absolutely essential for the original conditions in order to provide a sufficient value for the worn condition. Mobile equipment, such as auto receivers, have a wider range of input voltages than do farm or portable receivers operating without a charger. Thus the spread of wave-form closure to be expected, or which must be tolerated, will be greater in the case of the former than it will be for the latter.
As a result of the foregoing discussion regarding timing capacitance value determination, it can readily be seen that a number of factors must be considered in making this determination. The P. R. Mallory Company has adopted a standard transformer for this purpose and uses a comparison method. This transformer has been so designed as to incorporate the maximum allowable capacitance shift between nominal and maximum input voltages, both load and no load. The buffer capacitance selected is optimum for this transformer and has been determined by years of experimenting and life testing with the transformer. In addition, duplicates of this standard transformer and timing capacitance are used for all factory-adjusting of vibrators. As a result, the use of this standard transformer and timing capacity provides a fast and accurate method of determining the timing capaci¬tance values required for new transformer design. The correct timing capacitance value required for any newly-designed transformer can readily be determined by observing the vibrator characteristics of any production vibrator when operated with the standard transformer and its timing capacitor. Then the newly-designed transformer is substituted for the standard transformer and the timing capacitance varied until the wave-form characteristics have been duplicated.
Figure 49 illustrates an interesting condition of operation, which may be confused by some observers and should therefore be explained. This wave-form is that representing an interrupter type of vibrator with the output loaded, and with insufficient timing capacitance provided. It represents the loaded condition of Figure 46. The short peak and dip in the wave-form at points (1) and (3) are the result of loading being applied through the rectifier tube when the voltage has risen to a value exceeding that of the output (filter) condenser. Since no energy is being supplied by the battery as yet, the load reduces the voltage, forming the dip, until the contacts subsequently close. This waveform can be confused with contact "chatter" on the make, but can be distinguished quickly by the simple expedient of removing the rectifier tube. If the disturbance persists, the cause probably lies in the vibrator action. If the wave-form changes to that similar to Figure 46, the difficulty is over-closure. Greater difficulty in making this determination is involved with a self-rectifying vibrator. Sharp and erratic peaks will usually occur at the make, although the over-closure loop may be present. This prevails on both load and. no-load, since the load cannot be applied until after the rectifier contacts close. A positive check requires that the output load circuit, following the smoothing filter, must be opened in order to secure a no-load condition.
- planright's blog
- 530 reads
Post new comment