Vibrator Power Supply Design - Basic Power Transformer Characteristics - III
By planright Aug 17th, 2009
Magnetic leakage in a transformer is still another factor which must be considered with regard to its effect upon vibrator operation. This factor has some detrimental effects upon sine-wave AC transformers, such as increasing the regulation of power transformers and impairing the frequency characteristics of audio-frequency amplifier transformers. However, in vibrator transformers magnetic leakage is particularly serious. In effect, it introduces into the primary circuit an inductive reactance through which the load-current passes and in which induced transients occur when the vibrator contacts break this load-current. These transient voltages create arcs across the contact interfaces at the "break" and naturally have some destructive effect upon the performance and life of the contacts. When the timing capacitor is in the secondary circuit this leakage inductance interposed between the secondary circuit and the primary prevents the perfect reflection of the timing capacitance into the primary circuit. The larger values of leakage inductance reduces the magnetic coupling and the effectiveness of the timing capacitor. This same condition aids the creation of spurious damped oscillations during both the "on-contact" and "off-contact" periods of the vibrator. These oscillations have varying frequencies and amplitudes, and may be difficult to eliminate.
Leakage inductance results from imperfect coupling between the primary and the secondary, or secondaries, of the transformer. This lack of coupling results primarily from the mechanical arrangement of the coils, the relative number of turns employed, and the spacing of the coils with respect to each other. The fact that most vibrator transformers utilize both a center-tapped primary and secondary results in a considerable amount of leakage. Only one half of the primary and one half of the secondary carry load current at any point in the cycle. The other two sections of the windings act somewhat as space-occupying elements, which may separate the active windings and thus reduce their effective coupling. By employing the proper subdivision of the primary and secondary windings into individual coils, and by so arranging them mechanically and interconnecting them electrically so as to secure the closest coupling between the active coils, the leakage inductance can be held to a minimum for a given design. This method is comparatively expensive and slow for mass-produced units and is seldom used. The usual method of layer-winding of both secondaries and both primaries in sequence lends itself readily to mass production on automatic winding machines and to low production costs.
This method involves higher leakage inductance values, and an unbalance in the amount of leakage existing between the two halves of the cycle. Transformer designs having comparatively large cross-sectional areas of core and a small number of primary turns will have a lower value of leakage inductance than designs using a comparatively small core area and a large number of primary turns.
The two methods just described are illustrated in Figure 23. Illustrations "A" and "B" show the method of winding and assembly of the coils in individual "pies" so as to obtain the maximum coupling between the active coils. The connections to the coils can be so phased that the output voltage polarity, either with a self-rectifying vibrator or with a tube rectifier, will be as desired when using the closely coupled coils. Illustration "A" shows the simpler of the two arrangements, but this would not result in quite as low a leakage reactance as would "B," where the secondary is split and placed on either side of the primary.
Illustration "C" shows an arrangement of layer-wound coils where the primary and secondary sections are interleaved so as to secure close spacings and coupling between the active coils. This method requires that the electrical connections be properly phased in order to secure this close coupling, and thus all end-taps from the coils must be brought out from the transformer for external connection. Illustration "D" shows the usual method of manufacturing commercial vibrator transformer coils. Primary P-2 is a continuation of primary coil P-l, except that the start of the former is connected to the finish of the latter and often brought out as a single connection. The secondary sections are wound in a similar manner. This makes for the most economical coil winding arrangement, when layer-wound coils are being considered.
While leakage inductance can be calculated from the known physical dimensions and winding data, it is usually desirable to actually measure this characteristic from a sample of the production transformer. Since the total effective leakage inductance is the figure affecting the performance of the vibrator, this can best be determined from the primary side of the transformer. The test can be made by short-circuiting the secondary winding and measuring the resulting primary inductance on an AC bridge. Since a perfect transformer would reflect the short-circuit directly across the primary, the lack of perfect coupling is reflected by the inductance measured when the short-circuit is present across the opposite winding.
It may be of design interest to measure the leakage inductance between the active halves of the two windings, rather than between the entire windings. This emphasizes the amount of the unbalance which affects the two halves of the operating cycle. This test is made in the same manner, the short-circuit being placed across one active winding and the inductance measured across the other. If the measuring equipment requires that the measurement be made across the secondary, with the primary short-circuited, in order to use values that fall within the equipment range, this can easily be converted into primary values. The primary value is related directly to the secondary value by the square of the transformer turn-ratio.
Thus
Coil Construction
It is a universal practice to wind the coils without bobbin support, using an inner-tube support wound from heavy Kraft paper and cemented together. Kraft or glassine paper is used between layers for layer insulation. The thickness of the paper depends upon the size of wire and upon the electrical barrier insulation needed. The type of insulation used on the wire is a plain enamel in practically all instances. Occasionally, heavy or double enamel is used for special applications. A table, Figure 24, lists the usual permissible turns of enameled round copper magnet wire that can be wound per linear-inch of paper-layer coils for each size of wire. A guide to the thickness and type of paper inter-layer insulation commonly employed with the various sizes of wire is also listed. This table can be used as a guide for calculating copper requirements in various designs of reactor, solenoid, and relay coils. The same table shows the resistance of the various wire sizes listed by ohms per 1000 ft. of wire, so that by figuring the mean-length-of-turn for a coil, the resistance can be approximated.
Another table is furnished, Figure 25, in which the lamination type numbers and essential dimensions for the most commonly used small laminations adapted for vibrator use are given. A third table, which will probably be most useful, is given in Figure 26. This represents the present commercial winding information for various sizes of enameled-copper round magnet wire for the various lamination sizes, and is an excellent time-saver. The laminations listed are not all that are available, but are representative of those commonly used.
Complete data in the form of pamphlets or books are usually available from the various magnetic steel manufacturers upon request. It is suggested that steel manufacturer's representatives be contacted for additional information other than that included in this text.
No exact correlation has been attempted between the winding data furnished in the table of Figure 24 and that listed in the table of Figure 26. As noted on the tables, the sources of the information are different and one is a theoretical compilation while the other is the result of a practical commercial application of experimental and shop experience. The winding information relates particularly to automatic winding machines. It should be noted that in several cases, a number of laminations have the same coil and winding lengths, and thus the same winding data. The difference lies, of course, in the other dimensions of the lamination, such as width of window, center-leg, etc.
In addition to the foregoing, there are a number of other factors which have to be taken into consideration in designing transformers. These factors have been dictated by standard commercial practice in producing low-cost transformers in high-volume production. They have been considered in making the sample transformer designs included in this book, and a brief statement of each follows.
Through a number of years of experience in designing vibrator transformers, it has been found, with minor exceptions, that the most economical and practical use of winding space will result in the use of three layers of wire for the primary winding. Inasmuch as the primary winding of a vibrator transformer is center-tapped, this results in the necessity of bringing the center-tap out from the middle of a layer. The most common practice encountered in bringing this center-tap out from the middle of the layer is to lay in a tab of conductive material and fasten a flexible lead to this tab on the outside of the coil. Since some space is necessary for this tab to be laid in, it is common practice not to use the full number of turns which could have been wound on the tapped layer. In addition to this factor and to aid in producing transformers at low cost, the transformer companies make a practice of cutting the wire off even with the coil when the transformer is wound, and then bring the leads out for the start and finish of the winding by taking one turn off both the start and finish layers. As a result of this practice, a good design will allow approximately one turn per layer for the start and finish leads and the center-tap so that in effect the number of turns which can be placed in a primary layer is one less than shown in the table.
Current transformer manufacturing practices utilize machine winding for all secondaries. In order to bring out a center tap from the secondary winding, the transformer manufacturers have resorted to the practice of completely machine winding the secondary and then reaching in the winding with a pair of tweezers and pulling off one marked turn at the exact center of the coil, which produces the center tap. As a result, it is necessary to design the transformers so the secondary is in even-numbered layers. This will make the electrical center of the coil on the edge. For example, the number of turns required for the secondary winding should be calculated and divided by the number of turns which can be placed in one layer. If the number of layers required comes out to an odd figure, it will be necessary to reduce the number of turns per layer to make the layers come out an even number. The total secondary coil build can be calculated from this information.
The final factor which will influence the transformer design is the core-stacking factor. In actual practice this factor has been found to be between the limits of .88 to .95, or 88 % to 95 %. For the purpose of most design work, it will be found that a value of .91 or .92 will represent a satisfactory value for commercial practice.
Core Steel Grades
The transformer core steel most often used in vibrator transformers are limited to four grades. These grades are usually designated by their specified core-loss characteristics as expressed in watts-per-pound, at a flux-density of 10,000 Gauss and 60 cycles-per-second frequency. The grades have also been identified by their silicon content, but this was not entirely satisfactory since various grades may have the same silicon content and yet have different characteristics. 10,000 Gauss flux-density amounts to 10,000 lines per square centimeter, which is equivalent to 64,500 lines per square inch. The usual grades are listed below, with manufacturer's trade names for comparison:
The gauges listed above are for the thickness of the steel sheet used and are, in order, 0.014", 0.0188", and 0.025" thick respectively. As a general rule, only Types III and IV are used in #29 gauge. I, II and are used in #26 gauge, and I and II in #24 gauge sizes. The cost of the material or laminations plus the assembly labor versus the savings in core losses, etc., will determine to a large extent which grade and gauge of steel is used. It has been a more or less general practice to use #24 gauge of Types I or II for the medium or high-powered 115 cycle vibrator applications. On low output 115 cycle applications, where high efficiency or other factors warrant it, #26 or #29 gauge Types III or are often used. For higher frequency vibrators, Types III or IV in #29 gauge are necessary.
Leakage Inductance
Magnetic leakage in a transformer is still another factor which must be considered with regard to its effect upon vibrator operation. This factor has some detrimental effects upon sine-wave AC transformers, such as increasing the regulation of power transformers and impairing the frequency characteristics of audio-frequency amplifier transformers. However, in vibrator transformers magnetic leakage is particularly serious. In effect, it introduces into the primary circuit an inductive reactance through which the load-current passes and in which induced transients occur when the vibrator contacts break this load-current. These transient voltages create arcs across the contact interfaces at the "break" and naturally have some destructive effect upon the performance and life of the contacts. When the timing capacitor is in the secondary circuit this leakage inductance interposed between the secondary circuit and the primary prevents the perfect reflection of the timing capacitance into the primary circuit. The larger values of leakage inductance reduces the magnetic coupling and the effectiveness of the timing capacitor. This same condition aids the creation of spurious damped oscillations during both the "on-contact" and "off-contact" periods of the vibrator. These oscillations have varying frequencies and amplitudes, and may be difficult to eliminate.
Leakage inductance results from imperfect coupling between the primary and the secondary, or secondaries, of the transformer. This lack of coupling results primarily from the mechanical arrangement of the coils, the relative number of turns employed, and the spacing of the coils with respect to each other. The fact that most vibrator transformers utilize both a center-tapped primary and secondary results in a considerable amount of leakage. Only one half of the primary and one half of the secondary carry load current at any point in the cycle. The other two sections of the windings act somewhat as space-occupying elements, which may separate the active windings and thus reduce their effective coupling. By employing the proper subdivision of the primary and secondary windings into individual coils, and by so arranging them mechanically and interconnecting them electrically so as to secure the closest coupling between the active coils, the leakage inductance can be held to a minimum for a given design. This method is comparatively expensive and slow for mass-produced units and is seldom used. The usual method of layer-winding of both secondaries and both primaries in sequence lends itself readily to mass production on automatic winding machines and to low production costs.
This method involves higher leakage inductance values, and an unbalance in the amount of leakage existing between the two halves of the cycle. Transformer designs having comparatively large cross-sectional areas of core and a small number of primary turns will have a lower value of leakage inductance than designs using a comparatively small core area and a large number of primary turns.
The two methods just described are illustrated in Figure 23. Illustrations "A" and "B" show the method of winding and assembly of the coils in individual "pies" so as to obtain the maximum coupling between the active coils. The connections to the coils can be so phased that the output voltage polarity, either with a self-rectifying vibrator or with a tube rectifier, will be as desired when using the closely coupled coils. Illustration "A" shows the simpler of the two arrangements, but this would not result in quite as low a leakage reactance as would "B," where the secondary is split and placed on either side of the primary.
Illustration "C" shows an arrangement of layer-wound coils where the primary and secondary sections are interleaved so as to secure close spacings and coupling between the active coils. This method requires that the electrical connections be properly phased in order to secure this close coupling, and thus all end-taps from the coils must be brought out from the transformer for external connection. Illustration "D" shows the usual method of manufacturing commercial vibrator transformer coils. Primary P-2 is a continuation of primary coil P-l, except that the start of the former is connected to the finish of the latter and often brought out as a single connection. The secondary sections are wound in a similar manner. This makes for the most economical coil winding arrangement, when layer-wound coils are being considered.
While leakage inductance can be calculated from the known physical dimensions and winding data, it is usually desirable to actually measure this characteristic from a sample of the production transformer. Since the total effective leakage inductance is the figure affecting the performance of the vibrator, this can best be determined from the primary side of the transformer. The test can be made by short-circuiting the secondary winding and measuring the resulting primary inductance on an AC bridge. Since a perfect transformer would reflect the short-circuit directly across the primary, the lack of perfect coupling is reflected by the inductance measured when the short-circuit is present across the opposite winding.
It may be of design interest to measure the leakage inductance between the active halves of the two windings, rather than between the entire windings. This emphasizes the amount of the unbalance which affects the two halves of the operating cycle. This test is made in the same manner, the short-circuit being placed across one active winding and the inductance measured across the other. If the measuring equipment requires that the measurement be made across the secondary, with the primary short-circuited, in order to use values that fall within the equipment range, this can easily be converted into primary values. The primary value is related directly to the secondary value by the square of the transformer turn-ratio.
Thus
Coil Construction
It is a universal practice to wind the coils without bobbin support, using an inner-tube support wound from heavy Kraft paper and cemented together. Kraft or glassine paper is used between layers for layer insulation. The thickness of the paper depends upon the size of wire and upon the electrical barrier insulation needed. The type of insulation used on the wire is a plain enamel in practically all instances. Occasionally, heavy or double enamel is used for special applications. A table, Figure 24, lists the usual permissible turns of enameled round copper magnet wire that can be wound per linear-inch of paper-layer coils for each size of wire. A guide to the thickness and type of paper inter-layer insulation commonly employed with the various sizes of wire is also listed. This table can be used as a guide for calculating copper requirements in various designs of reactor, solenoid, and relay coils. The same table shows the resistance of the various wire sizes listed by ohms per 1000 ft. of wire, so that by figuring the mean-length-of-turn for a coil, the resistance can be approximated.
Another table is furnished, Figure 25, in which the lamination type numbers and essential dimensions for the most commonly used small laminations adapted for vibrator use are given. A third table, which will probably be most useful, is given in Figure 26. This represents the present commercial winding information for various sizes of enameled-copper round magnet wire for the various lamination sizes, and is an excellent time-saver. The laminations listed are not all that are available, but are representative of those commonly used.
Complete data in the form of pamphlets or books are usually available from the various magnetic steel manufacturers upon request. It is suggested that steel manufacturer's representatives be contacted for additional information other than that included in this text.
No exact correlation has been attempted between the winding data furnished in the table of Figure 24 and that listed in the table of Figure 26. As noted on the tables, the sources of the information are different and one is a theoretical compilation while the other is the result of a practical commercial application of experimental and shop experience. The winding information relates particularly to automatic winding machines. It should be noted that in several cases, a number of laminations have the same coil and winding lengths, and thus the same winding data. The difference lies, of course, in the other dimensions of the lamination, such as width of window, center-leg, etc.
In addition to the foregoing, there are a number of other factors which have to be taken into consideration in designing transformers. These factors have been dictated by standard commercial practice in producing low-cost transformers in high-volume production. They have been considered in making the sample transformer designs included in this book, and a brief statement of each follows.
Through a number of years of experience in designing vibrator transformers, it has been found, with minor exceptions, that the most economical and practical use of winding space will result in the use of three layers of wire for the primary winding. Inasmuch as the primary winding of a vibrator transformer is center-tapped, this results in the necessity of bringing the center-tap out from the middle of a layer. The most common practice encountered in bringing this center-tap out from the middle of the layer is to lay in a tab of conductive material and fasten a flexible lead to this tab on the outside of the coil. Since some space is necessary for this tab to be laid in, it is common practice not to use the full number of turns which could have been wound on the tapped layer. In addition to this factor and to aid in producing transformers at low cost, the transformer companies make a practice of cutting the wire off even with the coil when the transformer is wound, and then bring the leads out for the start and finish of the winding by taking one turn off both the start and finish layers. As a result of this practice, a good design will allow approximately one turn per layer for the start and finish leads and the center-tap so that in effect the number of turns which can be placed in a primary layer is one less than shown in the table.
Current transformer manufacturing practices utilize machine winding for all secondaries. In order to bring out a center tap from the secondary winding, the transformer manufacturers have resorted to the practice of completely machine winding the secondary and then reaching in the winding with a pair of tweezers and pulling off one marked turn at the exact center of the coil, which produces the center tap. As a result, it is necessary to design the transformers so the secondary is in even-numbered layers. This will make the electrical center of the coil on the edge. For example, the number of turns required for the secondary winding should be calculated and divided by the number of turns which can be placed in one layer. If the number of layers required comes out to an odd figure, it will be necessary to reduce the number of turns per layer to make the layers come out an even number. The total secondary coil build can be calculated from this information.
The final factor which will influence the transformer design is the core-stacking factor. In actual practice this factor has been found to be between the limits of .88 to .95, or 88 % to 95 %. For the purpose of most design work, it will be found that a value of .91 or .92 will represent a satisfactory value for commercial practice.
- planright's blog
- 373 reads
Post new comment