Still another factor that indirectly affects the choice of primary wire size is the fact that, at least for the usual 6-volt application, the number of layers of wire available for the primary winding is limited to three, or in a few cases four, for the wire diameters normally required. This bears directly upon the information contained in the preceding paragraphs. The fact that a large number of application designs require only three primary layers involves bringing the center-tap of the winding from the center of one layer, which is an undesirable, though a necessary procedure. Because of the greater number of layers involved in the secondary winding, and also because of the comparatively fragile wire necessary, it is common practice to design the secondary with an even number of layers. This permits the center-tap to be made at the end of a layer.
Other factors involved in the coil design may be outlined as follows. In many applications, it has been necessary to isolate the primary from the secondary electrostatically in order to prevent the transmission of "hash" interference from one circuit to the other. This is accomplished by the inclusion of a "static shield" between the respective coils which consists of a sheet of copper or brass foil the width of the coil and wound around the coil so as to overlap itself. This overlap is insulated, so that a short-circuited turn does not occur. The shield is usually grounded to the core, and from there to the receiver chassis. This device is moderately successful, but adds considerably to the cost and to the bulk of the coil without eliminating any of the major components of the interference-filter. It is no longer in general use.
Where some isolation is still desired, the present trend is to wind the secondary coil and make its connections in such a manner that some large measure of self-shielding results. This is done by what is known as the "inversion" of the secondary. To accomplish this, the two halves of the secondary are wound as in any other design, but instead of the center-tap occurring at the mid-point of the two windings, the "start" of the first half and the "finish" of the secondary half are connected together to form the "center-tap." Thus, the "finish" of the first half and the "start" of the second half become the end-taps of the winding. Figure 38a illustrates this arrangement. When the "center-tap" of the secondary winding is connected electrically to ground, the layer of the secondary adjacent to the primary is essentially placed at ground potential, and thus acts as a shield between the rest of the winding and the primary. In this case, however, the full difference of potential of the secondary exists between the two center layers of wire and sufficient insulation to withstand this voltage must be used.
Figure 38b illustrates the final form of the coil and core assembly of the "shell" type of transformer normally used in vibrator service. The laminations have been interleaved and held in place by insulated bolts prior to impregnation. The insulation around the bolts is usually a thin paper tube, or similar device, and is used to reduce the electrical conductivity between laminations and thus to reduce the eddy-current losses in the core.
Figure 39a illustrates the coil assembly features that have been previously discussed. The inner-tube coil support, which is designed to slip over the center-leg lamination assembly of the proper stack, is usually pre-formed and constructed before winding takes place. It may be made of spirally-wound cemented tubing, or of layer-wound heavy kraft paper cemented over each layer, the winding form furnishing an inside dimension which provides suitable small, mechanical clearances to permit the satisfactory assembly of the laminations. For automatic-winding machinery, the core tube would be made the full length of the machine capacity, and would be carried upon an arbor to be mounted in the machine. For individual winding machines, the core tube would be cut off to the proper length, shown on the diagram of Figure 39a as the "coil length."
Assuming in this instance that the wire size is such that the primary winding can be placed next to the core, the large primary wire would then be wound over the core tube, with the proper thickness of inter-layer paper placed between each successive layer. This paper thickness is selected by consulting the table in Figure 26 for the wire size required. If the voltage insulation between this first winding and the core is to be high, such as would be the case if the first winding were to be the secondary, additional insulating paper or cloth might be interposed between the Kraft core tube and the first layer of wire. The required clearances for the electrical insulation and for proper mechanical support at the ends of the coil necessitate that the winding length be less than the coil length. This is shown in Figure 39a as the dimension noted "winding length." Commercial practice relative to these dimensions are shown in the table given in Figure 26.
If the primary winding and leads have been applied smoothly, the addition of an inter-winding wrapper of several layers of the same thickness of Kraft paper being used between the primary layers will serve as both insulation and a mechanical support for the secondary winding. Possibly additional glassine, or even cloth, insulation may be desired for inter-winding electrical insulation, but in the average 6-volt application this is not required. If large wire size on the primary winding, bulky lead arrangement, or such similar problems prevent a smooth and even winding surface being provided for the secondary, the secondary must often be wound on a second core tube which will slip over the primary winding in making the assembly. This is an uneconomical procedure and is also wasteful of window space, and is to be avoided if at all possible.
Referring again to Figure 39a, it will be seen that the coil "build" occupies only a portion of the window height. The inner tube coil support occupies a more or less fixed dimension, depending upon the size of the lamination, and in some respects upon the size of the wire being supported. For the sizes of laminations under consideration, two thicknesses of tube are considered sufficient: .035" for the smaller lamination sizes and .050" for the larger ones. The table in Figure 26 furnishes a guide to these selections. The remainder of the "coil-build" dimension is made up of the layers of wire plus inter-layer insulation and inter-winding and outer wrappers. The total coil-build represents only a fraction of the window height, the remainder being clearance to accommodate bulging of the coil, inaccuracies of winding, and ease of assembly of the laminations. This clearance amounts to from roughly 10 % to 15 % of the window height. However, since the dimensions are smaller in the smaller laminations, the percentage for clearance in those instances will be higher to secure satisfactory mechanical clearance. It is suggested that the coil-build be limited, as shown in the tables of Figure 26, or that the clearance be kept at a minimum of 15 % in each instance.
During the process of winding the coil, required leads are often attached to the winding wire, in order to anchor them in the windings and furnish a strong assembly. With coils wound on automatic machinery, however, where the units are wound in multiple, this is usually a finishing operation after the coils are completed. The leads are brought out, or are attached to the coil, on the sides which do not fit into the lamination window, as is shown in Figure 39b. Often they are located on opposite sides of the coil instead of as shown, in order to reduce the bulging of the coil in its finished state. If minimum clearances must be maintained, to fit the unit into some tight dimension, the leads may be anchored at the corners of the coil so as not to increase the over-all dimension greatly.
The leads, especially on the primary, may be "self-leads" if the wire size is sufficiently large to provide satisfactory strength and handling ease. Otherwise, the leads are of insulated wire, usually flexible, which is anchored in the coil assembly.
During the winding of multiple coils the inter-layer insulating paper is fed into the machine in wide strips, or sheets, which are the full width of the core tube. These sheets are of sufficient length so that the paper extends around the completed layer with a satisfactory overlap. The lap in the insulation is always maintained on the two sides of the coil assembly which will be outside of the window. In this manner the overlap and the lead placement will not decrease the window space available for wire. The leading of the end-turn of one layer to the start of the next layer, which can cause some distortion of the coil, also should take place in this same general location.
Following the completion of the multiple winding, the outer-wrapper is anchored, and then the coils are separated by a set of cutting knives or saws. These knives are spaced correctly so as to separate the assemblies into coil lengths that will fit into the required window lengths. The leads are then attached and the coil is finish-wrapped. Then follows the impregnation process.
Impregnation
The primary purpose for impregnating insulated electrical windings is to remove any moisture entrapped in the absorbent insulating materials and carried in the air voids, and to erect a barrier against the return of moisture. The barrier is an impregnating compound which is itself a moisture impervious insulator as compared to the original insulation. Temperature, humidity, and possibly corrosive gases, can combine to cause failure of electrical windings through destruction of the copper wire or of breakdown of the insulation. This is caused by the chemical actions resulting from this combination, or by charring of insulation through over-heating, or both.
The removal of all moisture is an essential step in the impregnating process, and its importance cannot be overlooked. This process can be carried out in varying degrees of perfection, depending to a large extent upon the added cost that can be tolerated, as well as upon the equipment that is available and the time cycle permissible in the production schedule. The simplest procedure is that of baking the assembly in a ventilated oven just prior to impregnation. An improved method consists of making the heating cycle with the oven under a vacuum which greatly assists and speeds up the dehydration of the windings. The time involved depends upon the size and relative volume of the windings, but should be at least long enough to permit all of the winding to be heated sufficiently so that all entrapped or absorbed moisture is driven off.
The impregnating compound must have characteristics that will permit it to enter into the layers of the windings, fill all voids and saturate the absorbent insulating materials. This not only blocks the re-absorption of moisture, but improves the dielectric characteristics of the insulation, provides for improved heat transfer from the interior of the coils, and in some cases provides a certain amount of mechanical protection for the windings. The impregnant also serves as an anchor to hold the coil to the core assembly, and also as a medium for holding the laminations firmly in place. The latter prevents lamination vibration and hum.
There are many impregnating processes in use at the present time. The one used by any individual manufacturer depends upon the equipment available, the requirements of the customer's application, the allowable production cost, and the final assembly or mounting design. Obviously, if the core and coil assembly is to be mounted in a hermetically sealed container, the impregnation process need not provide as perfect a moisture barrier as would be required with an open mounting.
However, the usual 6-volt power unit will not use hermetically sealed transformers. The mounting will ordinarily consist of either of two types. In one, the transformer is anchored to the chassis and covered with a removable sheet-metal box, without any "potting" compound filling the unoccupied portions of the case. The other type of mounting consists of assembling the transformer in a sheet-metal or drawn-steel case and potting the unit with a high-melting-point compound. This anchors the core and coil assembly in place, together with the leads, in their proper location for assembly to the chassis. The encased assembly is then secured to the chassis. While both methods are common, the latter one provides for somewhat greater heat transfer to the outer-case from the transformer.
The most simple of impregnating methods consists of dipping the heated core and coil assembly into melted wax especially compounded for impregnating purposes. After a period of immersion, determined experimentally to be sufficiently long to permit complete saturation of the coil, the assembly is permitted to drain and cool. A refinement of this method consists of adding a means for the vacuum removal of air from the heated assembly, followed by pressure impregnation with the wax.
The above methods can be used employing a baking varnish for the impregnant. However, such processes require an oven-baking period, following the draining period, to dry the varnish throughout the depth of the coil. Most of the baking varnishes are of the oxidizing type and depend upon the oxidization of the carrier-oil to provide proper curing. Entrapment of solvent or uncured oil through incomplete curing can result in a rapid deterioration of that part of the coil in contact with it.
A newer type of varnish has received wide acceptance, being manufactured under various trade names. This varnish is of a different nature, being so compounded as to possess "polymerizing" characteristics. The characteristics are similar to those of bakelite, or other "heat-reactive" plastics, in that the curing is not the result of solvent evaporation or oxidization, but results from a chemical reaction that occurs with the application of heat for a definite length of time. The result is a superior penetration, complete curing throughout, high-strength coil, with good mechanical protection, and excellent protection from exposure to moisture and corrosion. Repeated dips and baking can provide sufficient protection to eliminate the need for added potting, etc.
Wax impregnation does not affect the flexibility of leads, or prevent the separation of the coil from the laminations, if required. Exposure to high temperatures, however, will result in gradual dripping of wax from the assembly. Ordinary insulating varnish impregnation has some stiffening effect upon the flexibility of leads, and if properly cured, results in a fairly permanent assembly of core and coil. Since the resulting finish of the polymerizing varnish is very hard and stiff, its effect upon the flexibility of leads is rather great. While very durable and reasonably tenacious, it is naturally subject to fracture as the result of flexing of any leads. Therefore, it is suggested that solder-lug terminal boards be employed with this type of impregnation. In specifying the type of impregnation for a given application, the foregoing characteristics as well as cost must be considered. Also, the decision must depend upon the climatic conditions to which the equipment will be exposed during service, and the expected life before servicing will be required. Except for service conditions to which they will be exposed, the impregnation of vibrator transformers is no different than that of other power transformers of small size.
Lamination Size
The required lamination size to accommodate the necessary primary turns can be estimated, as well as the approximate "stack" of laminations, to provide the calculated core area determined by equation (15). For pre-calculation purposes in determining the lamination size, the space required for the secondary can be estimated as being approximately equal to that required for the primary. The secondary winding can then be calculated, and the preliminary design can be considered completed.
Coil Resistance
One additional consideration is often desired, after the preliminary coil has been determined. This is the estimated DC resistance of the two windings. To make this estimate it is necessary to calculate a "mean-length-of-turn" for each of the primary and secondary windings. However, it is generally sufficient to arrive at an average for each winding, and thus neglect the inherent difference between the two halves of each occasioned by the different radius of the layers.
Figure 39b illustrates a cross-sectional view of a typical coil mounted upon a laminated core, with dimensions that will be used in determining the M.L.T. (mean-length-of-turn). The dashed lines refer to the M.L.T. of the primary and the secondary windings which are to be determined. If the dimensions shown are known, one method of procedure can be as follows:
Inner dimensions of core tube = X and Y
Thickness of core tube = T
Build of Primary Winding = P
Build of secondary winding =S
Thickness of primary wrapper =W,
Thickness of secondary wrapper = W2
Thickness added over leads =L
Inner-Circumference of Primary = Cir. i
Cir.!=2X+2Y+8T
Outer-Circumference of Primary =Cir.2 Cir.^Cir^+SP
Either equations (22) and (24) may be used, or (23) and (25) can be substituted; however, the latter seems to be a somewhat shorter method. Following the determination of the M.L.T. of both the primary and secondary, the total average length of wire used in each half of each winding can be calculated, as follows:
where M.L.T. is expressed in inches, and N is the total number of turns on either winding.
By referring to the table of Figure 26, the resistance for the wire size involved, expressed in ohms per 1000 ft. of length, can be determined. Incidentally, winding tension during the coil-winding operation will stretch the wire to some degree, depending largely upon the size of the wire and the machine requirements. This reduces the cross-section and affects the accuracy of the results of such calculations as are being discussed here.
Equation (27) allows the calculation of the average resistance of either winding, and is self-explanatory.
Transformer Design Procedure
Most of the present designs of automobile radio receivers can be classified into two groups; those with a single power output tube and those with a dual, or push-pull, power output system. The first group will normally have a power supply output voltage of 240 to 260 volts DC at a current of from 50 to 60 milliamperes, the voltage being that measured at the first filter capacitor. The second group will usually have a power supply output voltage of 260 to 275 volts DC at a current of from 60 to 75 milliamperes. There have been, and probably will be, receivers manufactured which have power requirements outside of these ranges, but the large majority of designs will fall within these limits. With these requirements in mind, it is possible to select several arbitrary values and specifications and develop sample designs. A brief summary of the design steps are as follows:
1. List the requirements, as discussed in Chapter IV.
2. Convert the output voltage and current requirements in terms of normal rated input voltage, if originally given at a value other than normal.
3. From the values obtained in (2), calculate the output watts now required. Estimate the efficiency of the power supply from the data given in Chapter V, and then figure the input watts necessary.
From this data figure the input current required at the rated input center-tap voltage.
4. Decide upon the approximate value of maximum flux density desired, for the highest input voltage expected to be encountered in service. (Generally 8.0 volts for 6-volt systems.)
5. Select a vibrator, from the Vibrator Characteristic Data Sheets (supplied separately) , whose characteristics and ratings will meet the requirements as determined above. Or, if a type of vibrator has previously been selected, see if it will be operating within its ratings when supplying the above requirements.
6. List the vibrator characteristics pertinent to the application.
7. Determine the wire sizes necessary to carry the currents previously calculated, applying the information given in Equations (17), (18), (19), (20) and (21), (Chapter VIII), relative to the RMS value of currents and the heating effect in center-tapped windings. Also refer to the discussion following Equation (21), on the current densities in the wire for various classes of designs. After the required area of the wires are determined, select the wire sizes accordingly, from the table of Figure 26.
8. Calculate the number of primary turns and the core size. Refer to Equation (15) (Chapter VIII), and to the table in Figure 26. Assume for the time being that the secondary coil-build will equal that of the primary coil. Using the preceding information, select a lamination size from the table of Figure 25 which will accommodate the required primary turns of specified wire size, and which will permit an approximately square cross-section of the center-leg to be used. As a guide, the allowable coil-build for any given lamination size, as shown at the bottom of the table in Figure 26, may be divided by 2, and the result used as the approximate build for either coil.
9. Calculate the secondary winding by first making preliminary estimates: Assume that the "regulation factor" will be about 0.70; also assume that the resistance of the secondary coil will be equal to that of the primary coil, when the latter is multiplied by the square of the turn-ratio.
Divide the required output voltage by the "regulation factor" and by the rated input voltage to secure an estimated turn-ratio.
Substitute the correct values in the Equation (14) to determine if the turn-ratio is approximately correct. The primary transformer resistance can be found by applying the data found in (7) and (8), and calculating the mean-length-of-turn of the winding by using the proper equations of (22), (23), (24) and (25).
The total primary regulating resistance is found by adding the primary coil resistance per half of the winding to the vibrator contact resistance, as discussed after Equation (14).
Correct the estimated turn-ratio if desirable. Then calculate the preliminary secondary winding by determining the number of turns required and the resultant coil build.
10. Determine the total transformer coil build and compare with the allowable build for the lamination size selected. If the design falls within the limits of less than 85 % build, it may be considered satisfactory for commercial production. If the build is over 85% but under 90%, the design is questionable. However, consideration of the actual clearance in inches, the size of the lamination and cross-section of the core, and the sizes of wire being used, will determine whether or not the design can be used.
If the cross-section is nearly square and wire sizes are small, the coil can be wound with less bulging and the layers will more nearly conform to the shape of the mandrel. If the lamination is large with a correspondingly large coil, the clearance must be greater to allow for mechanical tolerances in the coil assembly, etc. If a primary of large wire is placed over the secondary, this coil may be shaped somewhat after winding to reduce excessive bulging without probable damage to the wire or insulation. If the secondary is over the primary, this cannot be done satisfactorily.
11. If the result of (10) is an over-build, a redesign will be required. An increase in the size of the lamination to accommodate more copper in the window, or a decrease in the wire sizes to reduce the coil build in the present lamination are the first apparent measures to adopt in order to accomplish the redesign. Often a slight change in the turns used will permit the use of fewer secondary layers, which may reduce the build enough to be satisfactory with the present wire sizes. If necessary, the core stack can be adjusted somewhat to correct the flux density.
12. Calculate the new resistances of the two coils. Substitute these values in the output voltage Equation (14), to estimate if the final turn-ratio was correct. If necessary, or desirable, this factor can be re-figured.
13. Write up the sample specifications.
14. Calculate the peak magnetizing currents for the extremes of voltage input, or for the maximum and nominal conditions.
15. Calculate the approximate core losses.
16. Calculate the approximate copper losses, using the RMS values of the input and output currents involved.
Note: Items 14, 15 and 16, as given above, are supplementary calculations and are desirable only from the standpoint of comparison of the transformer design with other similar designs to determine which transformer has the better characteristics. The calculations can only be considered t approximate at best, because the performance curves and information used in calculating these results are determined on the basis of a 60-cycle sine-wave input, and, hence, cannot duplicate the action of the iron when used with the 115 or 250-cycle square wave voltage encountered in vibrator use.
Post new comment