HAWKINS
ELECTRICAL GUIDE
NUMBER
SIX
QUESTIONS
ANSWERS
&
ILLUSTRATIONS

Figs. 1,585 to 1,588.—Synchronous motor principles: I. A single phase synchronous motor is not self-starting. The figures show an elementary alternator and an elementary synchronous motor, the construction of each being identical as shown. If the alternator be started, during the first half of a revolution, beginning at the initial position ABCD, fig. 1,585, current will flow in the direction indicated by the arrows, passing through the external circuit and armature of the motor, fig. 1,586, inducing magnetic poles in the latter as shown by the vertical arrows. These poles are attracted by unlike poles of the field magnets, which tend to turn the motor armature in a counter-clockwise direction. Now, before the torque thus set up has time to overcome the inertia of the motor armature and cause it to rotate, the alternator armature has completed the half revolution, and beginning the second half of the revolution, as in fig. 1,587, the current is reversed and consequently the induced magnetic poles in the motor armature are reversed also. This tends to rotate the armature in the reverse direction, as in fig. 1,588. These reversals of current occur with such frequency that the force does not act long enough in either direction to overcome the inertia of the armature; consequently it remains at rest, or to be exact, it vibrates. Hence, a single phase synchronous motor must be started by some external force and brought up to a speed that gives the same frequency as the alternator before it will operate. A single phase synchronous motor, then, is not self-starting, which is one of its disadvantages; the reason it will operate after being speeded up to synchronism with the alternator and then connected in the circuit is explained in figs. 1,589 to 1,592.

Figs. 1,589 to 1,592.—Synchronous motor principles: II. The condition necessary for synchronous motor operation is that the motor be speeded up until it rotates in synchronism, that is, in step with the alternator. This means that the motor must be run at the same frequency as the alternator (not necessarily at the same speed). In the figures it is assumed that the motor has been brought up to synchronism with the alternator and connected in the circuit as shown. In figs. 1,589 and 1,590 the arrows indicate the direction of the current for the armature position shown. The current flowing through the motor armature induces magnetic poles which are attracted by the field poles, thus producing a torque in the direction in which the armature is rotating. After the alternator coil passes the vertical position, the current reverses as in fig. 1,591, and the current flows through the motor armature in the opposite direction, thus reversing the induced poles as in fig. 1,592. This brings like poles near each other, and since the motor coil has rotated beyond the vertical position the repelling action of the like poles, and also the attraction of unlike poles, produces a torque acting in the direction in which the motor is rotating. Hence, when the two armatures move synchronously, the torque produced by the action of the induced poles upon the field poles is always in the direction in which the motor is running, and accordingly, tends to keep it in operation.

Figs. 1,593 and 1,594.—Synchronous motor principles: III. The current which flows through the armature of a synchronous motor is that due to the effective pressure. Since the motor rotates in a magnetic field, a pressure is induced in its armature in a direction opposite to that induced in the armature of the alternator, and called the reverse pressure, as distinguished from the pressure generated by the alternator called the impressed pressure. At any instant, the pressure available to cause current to flow through the two armatures, called the effective pressure, is equal to the difference between the pressure generated by the alternator or impressed pressure and the reverse pressure induced in the motor. Now if the motor be perfectly free to turn, that is, without load or friction, the reverse pressure will equal the impressed pressure and no current will flow. This is the case of real synchronous operation, that is, not only is the frequency of motor and alternator the same, but the coils rotate without phase difference. In figs. 1,593 and 1,594, the impressed and reverse pressures are represented by the dotted arrows P_i and P_r, respectively. Since in this case these opposing pressures are equal, the resultant or effective pressure is zero; hence, there is no current. In actual machines this condition is impossible, because even if the motors have no external load, there is always more or less friction present; hence, in operation there must be more or less current flowing through the motor armature to induce magnetic poles so as to produce sufficient torque to carry the load. The action of the motor in automatically adjusting the effective pressure to suit the load is explained in figs. 1,595 and 1,596.

Figs. 1,595 and 1,596.—Synchronous motor principles: IV—A synchronous motor adjusts itself to changes of load by changing the phase difference between current and pressure. If there be no load and no friction, the motor when speeded up and connected in the circuit, will run in true synchronism with the alternator, that is, at any instant, the coils A B C D and A°B°C°D° will be in parallel planes. When this condition obtains, no current will flow and no torque will be required (as explained in figs. 1,593 and 1,594). If a load be put on the motor, the effect will be to cause A°B°C°D° to lag behind the alternator coil to some position A"B"C"D" and current to flow. The reverse pressure will lag behind the impressed pressure equally with the coil, and the current which has now started will ordinarily take an intermediate phase so that it is behind the impressed pressure but in advance of the reverse pressure. These phase relations may be represented in the figure by the armature positions shown, viz.: 1, the synchronous position A°B°C°D° representing the impressed pressure, 2, the intermediate position A'B'C'D', the current, 3, the actual position A"B"C"D" (corresponding to mechanical lag), the reverse pressure. From the figure it will be seen that the current phase represented by A'B'C'D' is in advance of the reverse pressure phase represented by A"B"C"D". Hence, by armature reaction, the current leading the reverse pressure weakens the motor field and reduces the reverse pressure, thus establishing equilibrium between current and load. As the load is increased, the mechanical lag of the alternator coil becomes greater and likewise the current lead with respect to the reverse pressure, which intensifies the armature reaction and allows more current to flow. In this way equilibrium is maintained for variations in load within the limits of zero and 90° mechanical lag. The effect of armature reaction on motors is just the reverse to its effect on alternators, which results in marked automatic adjustment between the machines especially when a single motor is operated from an alternator of about the same size. In other words, the current which weakens or strengthens the motor field, strengthens or weakens respectively the alternator field as the load is varied.

Figs. 1,597 and 1,598.—Synchronous motor principles: V. The effectiveness of armature reaction in weakening the field is proportional to the sine of the angle by which the current lags behind the impressed pressure. If a motor be without load or friction, its armature will revolve synchronously (in parallel planes) with the alternator armature. In the figures let ABCD represent an instantaneous position of the motor armature when this condition obtains; it will then represent the phase relationship of impressed and reverse pressures for the same condition of no load, no friction, operation. Now, if a light load be placed on the motor for the same instantaneous position of alternator armature, the motor coil will drop behind to some position as A", fig. 1,597 (part of the coil only being shown). The reverse pressure will also lag an equal amount and its phase with respect to the impressed pressure will be represented by A". The armature current will ordinarily take an intermediate phase, represented by coil position A'B'C'D', inducing a field strength corresponding to the 9 lines of force OF, O'F', etc. The current being in advance of the phase of the reverse pressure A", the armature reaction weakens the field, thus reducing the reverse pressure and allowing the proper current to flow to balance the load. The amount by which the field is weakened may be determined by resolving the induced magnetic lines OF, O'F', O"F", etc., into components OG, GF, O'G', G'F', O"G", G"F", etc., respectively parallel and at right angles to the lines of force of the main field. Of these components, the field is weakened only by OG, O'G', O"G", etc. Since by construction, angle OFG = AOA', and calling OF unity length, OG = sine of angle by which the current lags behind the impressed pressure. The construction is shown better in the enlarged diagram. For a heavier load the armature coil will drop back further to some position as A"', fig. 1,598, and the lag of the current increase to some intermediate phase as A"B"C"D". By similar construction it is seen that the component OG (fig. 1,597) has increased to OJ (fig. 1,598), this component thus further weakening the main field, by an amount proportional to the sine of the angle by which the current lags behind the impressed pressure. The increased current which is now permitted to flow, causes the induced field to be strengthened (as indicated by the dotted magnetic lines M, M', M", etc.), thus increasing the torque to balance the additional load.

Figs. 1,599 and 1,600.—Synchronous motor principles: VI. A single phase synchronous motor has "dead centers," just the same as a one cylinder steam engine. Two diagrams of the motor are here shown illustrating the effect of the current in both directions. When the plane of the coil is perpendicular to the field, the poles induced in the armature are parallel to field for either direction of the current; that is to say, the field lines of force and the induced lines of force acting in parallel or opposite directions, no turning effect is produced, just as in analogy when an engine is on the dead center, the piston rod (field line of force) and connecting rod (induced line of force) being in a straight line, the force exerted by the steam on the piston produces no torque.

Figs. 1,601 to 1,604.—Synchronous motor principles VII. An essential condition for synchronous motor operation is that the mechanical lag be less than 90°. Figs. 1,601 and 1,602 represent the conditions which prevail when the lag of the motor armature A'B'C'D' is anything less than 90°. As shown, the lag is almost 90°. The direction of the current and induced poles are indicated by the arrows. The inclination of the motor coil is such that the repulsion of like poles produces a torque in the direction of rotation, thus tending to keep motor in operation. Now, in figs. 1,603 and 1,604, for the same position of the alternator coil ABCD, if the lag be greater than 90°, the inclination of the motor coil A'B'C'D' is such that at this instant the repulsion of like poles produces a torque in a direction opposite to that of the rotation, thus tending to stop the motor. In actual operation this quickly brings the motor to rest, having the same effect as a strong brake in overcoming the momentum of a revolving wheel.

Figs. 1,605 to 1,608.—Synchronous motor principles: VIII. If the torque and current through the motor armature be kept constant, strengthening the field will increase the mechanical lag, and the lead of the current with respect to the reverse pressure. In the figures, let A be an instantaneous position of the alternator coil, A°, synchronous position of motor coil, A', position corresponding to current phase, A", actual position or mechanical lag of motor coil behind alternator coil necessary to maintain equilibrium. In fig. 1,606, let A' and A" represent respectively the relation of current phase and mechanical lag corresponding to a certain load and field strength. For these conditions OG, O'G', O"G", etc., will represent the components of the induced lines of force in opposition to the motor field, that is, they indicate the intensity of the armature reaction at the instant depicted. Now, assume the field strength to be doubled, as in fig. 1,608, the motor load and current being maintained constant. Under these conditions, the armature reaction must be doubled to maintain equilibrium; that is, the components OG, O'G', etc., fig. 1,608, must be twice the length of OG, O'G', etc., fig. 1,605. Also since the current is maintained constant, the induced magnetic lines OF, O'F' are of same length in both figures. Hence, in fig. 1,608 the plane of these components is such that their extremities touch perpendiculars from G, G', etc., giving the other components FG, F'G', etc. The plane A', normal to OF, O'F', etc., gives the current phase. By construction, the phase difference between A° and A' is such that sin A°OA' (fig. 1,608) = 2 × sin A°OA' (fig. 1,606). That is, doubling the field strength causes an increase of current lag such that the sine of the angle of this lag is doubled. Since the intensity of the armature reaction depends on the lead of the current with respect to the reverse pressure, the mechanical lag of the coil must be increased to some position as A" (fig. 1,608), such as will give an armature reaction of an intensity indicated by the components OG, O'G', etc.

Fig. 1,609.—Diagram illustrating method of representing the performance of synchronous motors. The V shaped curve is obtained by plotting the current taken by motor under different degrees of excitation, the power developed by the motor remaining constant. The current may be made to lag or lead while the load remains constant, by varying the excitation. By varying the excitation, a certain value may be reached which will give a minimum current in the armature; this is the condition of unity power factor. If now the excitation be diminished the current will lag and increase in value to obtain the same power; if the excitation be increased the current will lead and increase in value to obtain the same power. The results plotted for several values of the excitation current will give the V curve as shown. This is an actual curve obtained by Mordey on a 50 kw. machine running unloaded as a motor. Other curves situated above this one may be obtained for various loadings of the motor.

Fig. 1,610.—Westinghouse self-starting synchronous motor. Motors of this type are suitable for constant speed service where starting conditions are moderate, such as driving compressors, pumps, and large blowers. Synchronous motors can be made to operate not only as motors but as synchronous condensers to improve the power factor of the circuit. The field is provided with a combined starting and damper or amorlisseur winding so proportioned that the necessary starting torque is developed by the minimum current consistent with satisfactory synchronous running without hunting. The armature slots are open and the coils form wound, impregnated, and interchangeable. Malleable iron finger plates at each end of the core support the teeth. Ventilating finger plates assembled with the laminations form air ducts. The frames are of cast iron, box section with openings for ventilation; shoes and slide rails permit adjustment of position. The brush holders are of the standard sliding shunt type. Two or more brushes are provided for each ring.

Fig. 1,611.—Mechanical analogy illustrating "hunting." The figure represents two flywheels connected by a spring susceptible to torsion in either direction of rotation. If the wheels A and B be rotating at the same speed and a brake be applied, say to B, its speed will diminish and the spring will coil up, and if fairly flexible, more than the necessary amount to balance the load imposed by the brake; because when the position of proper torque is reached, B is still rotating slightly slower than A, and an additional torque is required to overcome the inertia of B and bring its speed up to synchronism with A. Now before the spring stops coiling up the wheels must be rotating at the same speed. When this occurs the spring has reached a position of too great torque, and therefore exerting more turning force on B than is necessary to drive it against the brake. Accordingly B is accelerated and the spring uncoils. The velocity of B thus oscillates above and below that of A when a load is put on and taken off. Owing to friction, the oscillations gradually die out and the second wheel takes up a steady speed. A similar action takes place in a synchronous motor when the load is varied.

Fig. 1,612.—Diagram illustrating the use of a synchronous motor as a condenser. If a synchronous motor be sufficiently excited the current will lead. Hence, if it be connected across an inductive circuit as in the figure and the field be over excited it will compensate for the lagging current in the main, thus increasing the power factor. If the motor be sufficiently over excited the power factor may be made unity, the minimum current being thus obtained that will suffice to transmit the power in the main circuit. A synchronous motor used in this way is called a rotary condenser or synchronous compensator. This is especially useful on long lines containing transformers and induction motors.

Figs. 1,613 to 1,625.—Disassembled view of Western Electric three phase squirrel cage skeleton frame induction motor.

Figs. 1,626 to 1,628.—General Electric base construction for polyphase induction motors. The base is made of cast iron. Adjusting gear is provided to slide the motor along the base as shown in the illustrations, the movement being from 6 to 12 inches according to size. With this design of base, motors are securely held in position under all conditions and may be run with an upward pull on the belt. Close fitting guides moving in an accurately machined slot on the base preserve a correct alignment of the motor when adjustment of the latter is required. The same base can be used whether the motor be supported from the wall or ceiling or located on the floor. A single adjusting screw is placed under the center line of the motor frame, which produces an even and balanced draw in either direction on all parts of the motor when the belt tension is altered. This screw can be located at either end of the base. The base can be omitted when the motor is direct connected or when provision for belt adjustment is not required.

Fig. 1,629.—Richmond three phase induction motor on base fitted with screw adjusting gear for shifting the position of the motor on the base to take up slack of belt.

Figs. 1,630 to 1,641.—Terminals for General Electric polyphase induction motors. In order to prevent any mechanical strain on the leads being transmitted to the motor windings, the terminal cables are clamped in insulated bushings with a connector for each cable.

Figs. 1642 and 1643.—Western Electric end flange rivets and punchings of riveted frame induction motor. The riveted frame is constructed of two cast iron flanges between which the stator laminations of sheet steel are securely clamped and riveted under hydraulic pressure. This construction exposes the laminations directly to the air and improves the radiation, thus insuring high overload capacity and low operating temperatures. The field slots are overhung or partially closed, affording mechanical protection to the coils.

Figs. 1,644 to 1,649.—Construction of General Electric drawn shell fractional horse power motors. The distinguishing feature of drawn shell motors is the field construction which consists of a steel shell or cylinder supporting and clamping together the stator or field punchings. This method avoids the cast frame work outside the active magnetic material. A disc is first punched or "blanked" out of soft steel, fig. 1,644, this disc being faced into the shape, fig. 1,645, with one end closed. The other end of the shell is then cut out, leaving the small flange as in fig. 1,646. It is now ready to receive the core punchings. In the next operation a suitable number of spacing rings, fig. 1,647, are forced into the shell and seated against the retaining lip, which may be seen in fig. 1,646. The field punchings or laminæ, fig. 1,648, are now assembled, after which a second and equal set of spacing rings are put into place to center the active field iron. The open edge of the shell is then rolled over the punchings under heavy pressure, thus preparing the field structure for the machining and fitting of the end heads and base. Fig. 1,649 shows a section of the completely assembled field structure, the parts being cut away to indicate the relation between the field punchings, spacing rings and shell. After the spacing rings at both frame ends have been turned true and grooved, the bearing heads, fig. 1,649, are ready for fastening in place by four fillister headed screws. A complete wound field is shown in fig. 1,858, with flat base casting attached.

Fig. 1,650.—Ideal fifteen horse power two phase induction motor. The armature core is supported by a cast iron frame carried on a base, with sliding ways and screw adjustment for tightening the belt. The armature core is provided with ventilating apertures, with metal spacers between each tooth. The revolving field is a steel casting with radially projecting poles, to which the pole shoes are bolted. The overhanging pole tips retain the field coils. All coils of the smaller sizes are wound with insulated copper wire of square section, and of the larger sizes, with flat copper, wound on edge, each turn being insulated by sheet insulation. Motors of this type are adapted for use in small power plants and isolated plants. The relatively high speed for which they are designed, reduces considerably the weight and overall dimensions, and likewise the cost. The exciter is belt driven. The normal kw. capacity of the exciter usually exceeds the kw. required for the excitation under normal load conditions to permit of station lighting. All exciters are built as compound wound dynamos, capable of delivering the exciter current up to 125 volts, which is sufficient margin in the field to control the alternating current line voltage on circuits of unusually low power factor.

Fig. 1,651.—Western Electric core construction and method of winding field of skeleton frame induction motor. The coils are wound on forms to give them exact shape and dimensions required. They are pressed into hot moulds to remove any irregularities and then the coils are impregnated with hot cement, to bind the layers together in their permanent shape. The portion of the coil which fits into the slot is wrapped with varnished cloth and a layer of dry tape is wound over the entire coil. The coils are then impregnated with an insulating compound and baked, the process being repeated six times. Coils for 1,100 and 2,200 volt motors have an extra covering of insulation and double the amount of impregnating and baking. The coils may be furnished with special insulation and treatment for exceptionally severe service conditions, such as exposure to excessive moisture, extreme heat, acid or alkaline fumes, etc. The coils are accessible and for the final finish are sprayed with black varnish.

Fig. 1,652.—Armature of Allis-Chalmers squirrel cage induction motor. The frame casting is of the box type and has large cored openings for ventilation. Lugs are cast on the interior surface of the frame to support the core, leaving a large air space between.

Fig. 1,653.—Sectional view showing parts of Reliance polyphase induction motor. A special feature of the squirrel cage armature construction is the multiplicity of short circuiting rings. The holes in the rings are bored slightly smaller than the diameter of the copper rods, and the force fit gives good contact. The rings having been forced in place are dip soldered in an alloy of tin of high melting point. The motor parts are: 1, end yoke; 2, shaft; 3, armature short circuiting rings; 4, oil ring; 5, self-aligning bearing bushing; 6, spider; 7, armature bars; 8, field coils; 9, field lamination end plate; 10, field laminations; 11, eye bolt; 12, stator locking key; 13, armature laminations; 14, armature lamination end plate; 15, armature locking key; 16, dust cap; 17, oil well cover; 18, oil throws; 19, field frame; 20, squirrel cage armature.

Fig. 1,654.—Tesla's rotating magnetic field. The figure is from one of Tesla's papers as given in The Electrician, illustrating how a rotating magnetic field may be produced with stationary magnets and polyphase currents. The illustration shows a laminated iron ring overwound with four separate coils, AA, and BB, each occupying about 90° of the periphery. The opposite pairs of coils AA and BB respectively are connected in series and joined to the leads from a two phase alternator, the pair of coils AA being on one circuit and the coils BB on the other. The resultant flux may be obtained by combining the two fluxes due to coils AA and BB, taking account of the phase difference of the two phase current, as in fig. 1,655.

Fig. 1,655.—Method of obtaining resultant flux of Tesla's rotating magnetic field. The eight small diagrams here seen show the two components and resultant for eight equivalent successive instants of time during one cycle. At 1, the vertical flux is at + maximum and the horizontal is zero. At 2, the vertical flux is still + but decreasing, and the horizontal is + and increasing, the resultant is the thick line sloping at 45° upwards to the right. At 3, the vertical flux is zero, and the horizontal is at its + maximum, and similarly for the other diagrams. Thus at 8, the vertical flux is + and increasing, while the horizontal is-and decreasing, the resultant is the thick line sloping at 45° upwards to the left. At points 2, 4, 6, and 8 the increasing fluxes are denoted by full and the decreasing by dotted lines. The laminated iron of the ring is indicated by the circles, and the result is that at the instants chosen the flux across the plane of the ring is directed inwards from the points 1, 2, 3, 4, etc., on the inner periphery of the iron. There will, therefore, appear successively at these points effective north poles, the corresponding south poles being simultaneously developed at the points diametrically opposite. These poles travel continuously from one position to the next, and thus the magnetic flux across the plane of the ring swings round and round, completing a revolution without change of intensity during the cycle time of the current.

Fig. 1,656.—Arago's rotations. The apparatus necessary to make the experiment consists of a copper disc M, arranged to rotate around a vertical axis and operated by belt drive, as shown. By turning the large pulley by hand, the disc M may be rotated with great rapidity. Above the disc is a glass plate on which is a small pivot supporting a magnetic needle N. If the disc now be rotated with a slow and uniform velocity, the needle is deflected in the direction of the motion, and stops at an angle of from 20° to 30° with the direction of the magnetic meridian, according to the velocity of the rotation of the disc. If the velocity increase, the needle is ultimately deflected more than 90° and then continues to follow the motion of the disc.

Fig. 1,657.—Explanation of Arago's rotations. Part of fig. 1,656 is here reproduced in plan. Faraday was the first to give an explanation of the phenomena of magnetism by rotation in attributing it to the induction of currents which by their electro-dynamic action, oppose the motion producing them; the action is mechanically analogous to friction. In the figure, let AB be a needle oscillating over a copper disc, and suppose that in one of its oscillations it goes in the direction of the arrow from M to S. In approaching the point S, for instance, it develops there a current in the opposite direction, and which therefore repels it; in moving away from M it produces currents which are of the same kind, and which therefore attract, and both these actions concur in bringing it to rest. Again, suppose the metallic mass turn from M towards S, and that the magnet be fixed; the magnet will repel by induction points such as M which are approaching A, and will attract S which is moving away; hence the motion of the metal stops, as in Faraday's experiment. If in Arago's experiment the disc be moving from M to S, M approaches A and repels it, while S, moving away, attracts it; hence the needle moves in the same direction as the disc. If this explanation be true, all circumstances which favor induction will increase the dynamic action; and those which diminish the former will also lessen the latter.

Fig. 1,658.—Experiment made by Faraday being the reverse of Arago's first observation. Faraday assumed that since the presence of a metal at rest stops the oscillations of a magnetic needle, the neighborhood of a magnet at rest ought to stop the motion of a rotating mass of metal. He suspended a cube of copper by a twisted thread, which was placed between the poles of a powerful electromagnet. When the thread was left to itself, it began to spin round with great velocity, but stopped the moment a powerful current was passed through the electromagnet.

Fig. 1,659.—Production of a rotating magnetic field by two phase currents. The figure represents an iron ring, wound with coils of insulated wire, and supplied with two phase currents at the four points A, B, C, and D. The action of the two phase current on the ring in producing a rotating magnetic field is explained in the accompanying text.

Fig. 1,660.—Production of rotating magnetic field in a two pole two phase motor. The poles are numbered from 1 to 4 in a clockwise direction. Phase A winding is around poles 1 and 3, and phase B winding, around poles 2 and 4. In each case the poles are wound alternately, that is, if 1 be wound clockwise, 3 will be wound counter clockwise, thus producing unlike polarity in opposite poles. Now during one cycle of the two phase current, the following changes take place, starting with pole 1 of N polarity and 3, of S polarity:

Fig. 1,661.—Diagram showing resultant poles due to two phase current.

Fig. 1,662.—Diagram of two phase, six pole field winding. There are six coils in each phase, as shown. The coils of each phase are connected in series, adjacent coils being joined in opposite senses, thus, for each phase, first one coil is wound clockwise, and the next counter clockwise.

Fig. 1,663.—Diagram of two phase, eight pole field winding. The winding is divided into 16 groups (equal to the product of the number of poles multiplied by the number of phases). Each group such as at A comprises a number of coils in series, each coil being located in a separate pair of slots, the end of one being connected to the beginning of the next. When the currents are in the same direction, the currents circulate in the same direction in two adjacent groups, a pole then with this arrangement being formed by two groups, both phases contributing to the formation of the pole. After ½ cycle when the current in each phase reverses, the pole advances the angular distance, covered by two groups; hence the field completes one revolution in eight alternations of current.

Figs. 1,664 to 1,683.—Sine curves of two phase current and diagrams showing the physical conception of a two phase rotating magnetic field. The alternating magnetizing current is assumed to be of such strength that, at its maximum strength, the field produced may be represented by 10 lines of force as indicated by the parallel lines. At the beginning of the rotation, fig. 1,664, phase A magnetization, according to sine curve is zero, indicated by the solid black poles, while phase B is of strength 10 with

current in the direction to produce a south pole at B. Similarly, in fig. 1,665, the strength of A is 4 lines, and of B, 9 lines, the resultant magnetization having rotated 22½°. The direction of the resultant magnetization is indicated by the arrow in each figure. It should be noted in fig. 1,669, that the polarity of B is reversed, the current curve now being above the zero line. By following the arrow through the successive positions the rotation of the resultant magnetization is clearly seen.

Fig. 1,684.—Moving picture method of showing motion of a rotary magnetic field. A number of sheets of paper are prepared, each containing a drawing of the motor frame and a magnetic needle in successively advancing angular positions, indicating resultant directions of the magnetism. The sheets are bound together so that the axis of the needle on each sheet coincides. When passing the sheets in one way the revolving field will be seen to rotate in one direction, while, when moving the sheets backward, the rotation of the magnetic field is in the opposite direction, showing that the reversal of the order of the coils has the effect of reversing the rotation of the magnetic field.

Fig. 1,685.—Production of a rotating magnetic field by three phase current. A ring wound as shown is tapped at points A, B, and C, 120° apart, and connected with leads to a three phase alternator. As described on page 1,304, a rotating magnetic field is produced in a manner similar to the two phase method.

Fig. 1,686.—Diagram of three phase, four pole Y connected field winding.

Fig. 1,687.—Production of a rotating magnetic field in a two pole three phase motor. In order to obtain a uniformly rotating magnetic field, it is necessary to arrange the phase windings in the direction of rotation, in the sequence ACB, not ABC as indicated on the magnets. Thus poles 1 and 4 are connected in series to phase A, 2 and 5 in series to phase C, and 3 and 6 in series to phase B. The different phase windings are differently lined, and it should be noted that they have a common return wire, though this is not absolutely necessary. Since the phases of the three currents differ from each other by one-third of a period or cycle, each of the phase windings will therefore set up a field between its poles, which at any instant will differ, both in direction and magnitude, from the fields set up by the other phase windings. Hence, the three phase windings acting together will produce a resultant field, and if plotted out, the directions of this field for various fractions of the period is such that in one complete period the resultant field will make one complete round of the poles in a clockwise direction, as indicated by the curved arrow. The positions of the resultant field during one complete period may be tabulated as follows:

Fig. 1,688.—Production of three phase rotating magnetic field with winding on laminated iron ring. The winding is divided into twelve sections, which are connected in three groups, A, B, and C, of four sections each, the sections in each group being evenly placed round the ring with the sections of the two other groups between them. One end of each group is to be connected to the line wire and the other end to the common junction J, from which it follows that the winding given is an example of "star" winding. With three phase currents the winding will give at every instant four N poles and four S poles round the ring, and in actual working these poles will be on the inner periphery because of the presence of an inner ring or cylinder of good magnetic iron placed, with the requisite clearance to allow of rotation, as close as is mechanically possible to the outer ring. Each one of these eight poles will make a complete revolution round the ring in four times the periodic time of the currents supplied. Thus, if the supply current has a frequency of 50, a complete revolution of the field will take place in .08 (=⁴/₅₀) of a second, which corresponds to an angular velocity of 750 revolutions per minute in place of 3,000 revolutions per minute, which would be the angular velocity with a bipolar field at this periodicity. Similarly a continuously wound Gramme ring tapped at twelve points, joined in three groups of four each to the supply mains, would give an eight pole rotary field. In this case the grouping would be a "mesh" grouping, with each side of the mesh formed of four coils in parallel.

Figs. 1,689 to 1,708.—Sine curves of three phase current and diagrams showing the physical conception of a three phase rotating magnetic field. The diagrams are constructed in the same manner as explained in figs. 1,664 to 1,683. It should be noted that the phase windings are arranged in the direction of rotation in the sequence ACB, phase C being wound in opposite

sense to A and B, as indicated by the curves, in that north poles are produced at A and B when the respective curves are above the zero line, a south pole being produced at C when its curve is above the zero line. The rotation of the resultant magnetization is clearly seen by following the arrow through its successive positions.

Fig. 1,709.—Diagram of three phase, six pole field winding. There are 18 groups, and the sequence of phases is ABC in a counter clockwise direction. For a Y connection, the middle phase is reversed, so that a pole will be formed by the three consecutive phases when the current is in the same direction in A and C, and opposite in B. The beginning of the middle coil C, and not the end, as with the other two, is connected to the common point O. In this case the pole shifts a distance equal to three groups for each alternation, so that one revolution of the field requires three cycles.

Fig. 1,710.—Copper cylinder and rotating magnet illustrating the principle of operation of an induction motor. The "rotating magnetic field" which is necessary for induction motor operation is for simplicity here produced by rotating a magnet as shown. In starting, the cylinder being at rest, any element as AB, as it is swept by the field will cut magnetic lines, which will induce a current upward in direction as determined by applying Fleming's rule (fig. 132, page 133). The inductive action is strongest at the center of the field hence as AB passes the center the induced pressure along AB is greater than along elements more or less remote on either side. Accordingly a pair of eddy currents will result as shown (see fig. 291, page 271). Applying the right hand rule for polarity of these eddy currents (see fig. 119, page 117) it will be seen that a S pole is induced by the eddy on the side of the cylinder receding from the magnet, and a N pole by the eddy on the side toward which the magnet is approaching. The cylinder, then, is attracted in the direction of rotation of the magnet by the induced pole on the receding side, and repelled in the same direction by the induced pole on the approaching side. Accordingly, the cylinder begins to rotate. The velocity with which it turns depends upon the load; it must always turn slower than the magnet, in order that its elements may cut magnetic lines and induce poles to produce the necessary torque to balance the load. The difference in speed of the magnet and cylinder is called the slip. Evidently the greater the load, the greater is the slip required to induce poles of sufficient strength to maintain equilibrium. The figure is drawn somewhat distorted, so that both eddies are visible.

Figs. 1,711 to 1,718.—Parts of Allis-Chalmers polyphase induction motor with squirrel cage armature.

Fig. 1,719.—General Electric vertical type induction motor; sectional view showing oiling system. It is provided with ball thrust bearings and top and bottom guide bearings, and a continuous flow of oil is maintained through all the bearings by means of a pump which is made integral with the motor. The ball thrust bearings are designed to support the weight of the armature only. In cases where the armature is direct connected a flexible coupling should be used to prevent additional weight coming on the thrust bearings. In operation, when the motor starts, the oil, revolving with the pan, flows against the stationary nozzle and is forced by its velocity at a high pressure through the oil pipe into the reservoir on top. It then flows down through the ball bearing and upper guide bearing, through a slot in the armature spider into the lower guide bearing and thence into the oil pan. Thus a continuous stream of oil is delivered through all bearings.

Fig. 1720.—Triumph back geared polyphase induction motor. A great many applications, especially for direct attachment, require the use of either a very slow or special speed motor. As these are quite costly, the preferable arrangement, and one equally as satisfactory, is the use of a standard speed motor combined with a back geared attachment. Rawhide pinions are furnished whenever possible, insuring smooth running with a minimum of noise.

Figs. 1,721 to 1,735.—Parts of General Electric small polyphase induction motors. A, armature; B, key for armature shaft; C, oil ring; D, bearing lining; E, bearing head, pulley end; F, cap screw for bearing heads; G, field, complete with winding, terminal plate and leads; H, motor leads; I, terminal connector for motor leads; J, soft rubber bushing for motor leads; K, terminal plate; L, screw for terminal board; M, field coils; N, wooden top sticks for field coils; O, oil filler; P, bearing head opposite pulley end; Q, screw for oil well cover; R, oil well cover; S, socket pipe plug for bearing head; U, motor base; V, yoke for motor base; W, motor base adjusting screw; X, bolt for motor base and frame (short); Y, cap screw for bearing head; Z, internal directive fan; Aa, pulley.

Fig. 1,736.—Sector method of measuring the slip of induction motors. A black disc having a number of white sectors (generally the same as the number of poles of the induction motor) is fastened with wax to shaft of the induction motor, and is observed through another disc having an equal number of sector shaped slits (that is a similar disc with the white sectors cut out) and attached to the shaft of a small self-starting synchronous motor, which is fitted with a revolution counter that can be thrown in or out of gear at will; then the slip (in terms of N_r) = N ÷ (N_s ÷ N_r), in which: N = number of passages of the sectors; N_s = number of sectors; N_r = number of revolutions recorded by the counter during the interval of observation. For large values of slip, the observations may be simplified by using only one sector (N_s = 1), then N will equal the slip in revolutions.

Fig. 1,737.—Detail of Westinghouse squirrel cage armature for induction motor. This is an example of cast on construction similar to that of Morse-Fairbanks (see figs. 1,752, 1,753 and 1,915). The inductors are embedded in a special cement.

SO CALLED SQUIRREL CAGE

Figs. 1,738 to 1,744.—Evolution of the squirrel cage armature. The early experiments of Arago, Herschel, Babbage and Baily demonstrated that a mass of copper or any conducting metal, if placed in a revolving magnetic field, will be urged to revolve in the direction of the revolving field. They used discs, but Ferraris used a copper cylinder as shown in figs. 1,710 and 1,738; this was the first squirrel cage armature. Figs. 1,739 to 1,744 show the gradual development of the primitive device shown in fig. 1,738; fig. 1,739, Ferraris' cylinder with slots restricting the path of induced currents; fig. 1,740, Dobrowolsky's so called squirrel cage which he embedded in a solid iron core, as in fig. 1,741; fig. 1,742, design with insulated bars and laminated core to prevent eddy currents in the core; fig. 1,743, laminated core with ventilating ducts; fig. 1,744, modern squirrel cage armature representing the latest practice as built by Mechanical Appliance Co. The core is built up of discs punched from No. 29 gauge electrical sheet, insulated from each other and firmly clamped between end plates locked on the shaft. The slots in the discs are of the same general form as those in the core. Heavy fibre end pieces, punched to match the discs are placed at each end of the core, to prevent the bars coming in contact with the sharp edges of the teeth. The winding is made up of rectangular copper bars, passing through slots in the core, and short circuited on each other by means of copper end rings of special design. The bars are pressed into holes punched in the end rings, and the contact is then protected from corrosion by being dipped in a solder bath. The bars are insulated from the iron of the core by fibre cell projecting beyond the end of the slot. To secure ventilation the short circuiting rings are set some distance from the end of the core. In this way the bars between the core and the ring act as the vanes of a pressure blower, forcing a large volume of air through the field coils and ventilating openings.

Fig. 1,745.—Mechanical Appliance Co. solid core discs as used on small and medium size induction motors.

Fig. 1,746.—Allis-Chalmers squirrel cage armature construction. The core laminæ are mounted on a cast iron spider having arms shaped to act as fan blades for forcing air through the motor. The spider is pressed on to the shaft. In the smallest sizes the punchings are mounted directly on the shaft, which is properly machined to hold them firmly. Copper bars are used as inductors in the larger sizes, and copper rods in the smaller sizes. The ends of the inductors are turned down somewhat smaller than the body and fit in holes in the end rings. The shoulder thus formed fits firmly against the end rings. Good electrical contact is obtained by expanding the inductors in the end ring holes. In large armatures both bars and end rings are of rectangular cross section, the bars and rings being fastened by machine steel cap screws.

Fig. 1,747.—Triumph squirrel cage armature. In construction thin sheet steel laminations, japanned, are built up to form the core, and are rigidly clamped together by heavy malleable iron end plates. Semi-enclosed slots are punched in the outer periphery to receive the windings, so that none of the centrifugal force is carried by the inductors. These inductors are set edge on, and are riveted and soldered into resistance rings. These rings are punched to receive the inductors in such a manner that there is an unbroken strip of metal completely surrounding them. Moreover, the short circuiting rings are set some distance from the end of the core, so that the inductors between the core and ring act as vanes to force air through the coils for ventilation.

Fig. 1,748.—General Electric soldered form of end ring construction on squirrel cage armatures. The armature inductors or copper bars laid in the core slots are short circuited by these end rings, which are also made of copper. For the smaller sizes the rings are thin, but of considerable radial depth and are held apart by spacing washers. They have rectangular holes punched near their outer peripheries through which the bars pass. Lips are formed on the rings, as shown, to which the bars are soldered.

Fig. 1,749.—General Electric welded form of end ring construction on squirrel cage armatures. Space limitations make it difficult to provide multiple soldered rings of sufficient area for large motors; hence, on such machines welding is resorted to, as shown. The ring in welded construction is placed beneath the bars at each end of the armature. Short radial bars are welded to the edges of these rings and to the inductors or squirrel cage bars, thereby making good electrical contact.

Figs. 1,750 and 1,751.—Built up core construction with discs punched in one piece. The spider proper consists of a hub provided with four radial arms, which fit the inner diameter of the disc. The hub is bored out so that it fits very tightly on the shaft, and a key is provided to avoid any chance of turning. The core disks are clamped firmly in place by two heavy cast iron end plates which are pressed up and held by the bolts. These bolts pass under the discs, so that there is no danger of their giving rise to eddy currents. The key not only prevents the discs turning on the spider but also ensures the alinement of discs, which is necessary to make the teeth form smooth slots when the core is assembled.

Fig. 1,752.—Fairbanks-Morse squirrel cage armature with cast-on rings showing inspection grooves. The method consists in fusing the ends of the inductors into an end ring of a special composition, thereby producing a perfect electrical and mechanically strong joint. In this process the armature with its bars in place is put into a mould and the molten metal poured around the inductors, melting their ends and effectually fusing them into the body of the ring. The ring is then turned down to finished size and polished. An inspection groove is cut as shown to indicate that the fusion is complete and the joint perfect.

Fig. 1,753.—Section of Fairbanks-Morse "cast-on" joint showing union of end ring and inductor. The view shows the V-shape inspection groove as described in fig. 1,752.

Fig. 1,754.—Richmond field construction for polyphase induction motors, showing style of winding for use with squirrel cage and wound armature types.

Fig. 1,755.—Western Electric squirrel cage armature of high speed induction motor for centrifugal pump service. This armature is an example of heavy duty construction. The inductors are welded to the short circuiting end rings, the latter being located beneath the inductors, as shown. Fan vanes are provided at one end for ventilation. In the field construction, the core laminations are assembled in a closed box frame, and clamped by heavy rings while under hydraulic pressure. The stator coils are form wound and subjected to a special insulating process, which renders them especially impervious to moisture, and capable of operating without breakdown in locations which are too damp for ordinary motors. The bearing brackets are of rigid mechanical construction, and the pulley end bracket and bearings of all sizes are split to facilitate removal of the rotor and complete inspection. These machines range in size from 50 to 200 horse power, the rugged construction adapting them to heavy and severe service, such as is met with in mining, the construction of dams, canals, aqueducts, tunnels, etc.

Fig. 1,756.—Wagner squirrel cage armature for polyphase induction motor, as employed on motors of from 5 to 25 horse power. The features of construction as seen in the illustration are bar inductors, ventilating passages through the core laminæ, riveted connection between inductors and end rings ventilating vanes on end plate, extra large end rings. The object of making the rings unusually large is to make the resistance of the rings lower than is desirable for some classes of service, in order to obtain motors having minimum slip, increased efficiency, and maximum overload capacity under normal operation. When the torque required by some very unusual and entirely abnormal installation exceeds that of the average conditions, it is an easy matter to reduce the section of the end rings, by turning them down in a lathe, thereby increasing the resistance and starting torque.

Fig. 1,757.—Richmond squirrel cage armature. The copper bars are double riveted at either end to the resistance rings, then dipped into a solder bath.

Fig. 1,758.—Field construction of Crocker-Wheeler induction motor with magnetic bridge. Steel bridges are inserted in the grooves where the coils are placed, to protect them from dirt and mechanical injury and at the same time provide a path for the magnetic flux which has a more uniform reluctance, thereby insuring a better distribution of the flux in the air gap and at the same time retaining open slot construction from which the coils can be readily removed.

Fig. 1,759.—Western Electric squirrel cage armature. The inductors consist of solid copper bars embedded in the slots of a laminated core, with their projecting ends securely fitted and soldered to heavy copper rings.

Fig. 1,760.—Holzer Cabot combination polyphase induction motor set, consisting of wound frame and three rotors: 1, squirrel cage armature, 2, wound armature, 3, rotating field. The set is intended for school demonstration of induction motor phenomena. The motor operating with the squirrel cage armature has an inherent constant speed characteristic and on brake tests will show its exceptionally strong starting torque and ability to take excessive overloads. This motor can be used as a generator also, in the sense that if connected to the line and driven above synchronous speed by some external means, it will act as an asynchronous generator and return power to the line. For variable speed service, an armature having a winding upon it similar to that on the frame must be used. External resistances inserted in the armature circuit may be used to produce, first, a reduction of starting current, second, an increase of starting torque, or third, a variation of speed. Thus an extensive list of experiments can be performed with this phase wound armature directly along the line of present engineering practice. The phase wound armature can be used as an alternator in the same sense as mentioned above for the squirrel cage machine. For synchronous motor and three phase operation the revolving field with projecting poles and slip rings would be used, the field being excited from a direct current supply.

Fig. 1,761.—Westinghouse auto-starter. Polyphase induction motors may be started by connecting them directly to the circuit with an ordinary switch, and the smaller motors are started in this way in practice. In the larger motors, however, the starting torque at normal voltage is several times its full load torque; therefore, they are started on a reduced voltage, and the full pressure of the circuit is not applied until they have practically reached their operating speed. The figure shows connections with a two phase alternating current circuit. The auto-starter consists of two auto-transformers T and T', each having only a single winding for both primary and secondary, which are tapped at certain points by switches, thus dividing the winding into a number of loops, so that one of several voltages may be applied for starting, and the starting torque thus adjusted to the work that has to be performed. At the highest points tapped by the switches S, and S', the full pressure, and at the lowest points, the lowest pressure, is applied to the motor by the operation of the main switch M. This switch has four blades and three positions. When thrown to the left as indicated, it connects the auto-transformers T and T', across the circuits A and B respectively, so that the pressure across the transformer coils, as determined by the position of the switches S and S', is applied to the motor circuits A and B. The intermediate position of the switch M interrupts both circuits. To start the motor, the switch M is thrown to the left and a reduced pressure applied; after the motor has started and come up to speed the switch M is thrown to the right, thus cutting out the transformer and connecting the motor directly to the circuit. The starting device can be located at a point remote from the motor, thus eliminating danger from fire due to possible sparks, in case where it is necessary to install the motors in grain elevators, woolen mills, or in any place exposed to inflammable gases, or floating particles of combustible matter. This feature is also valuable in cases where motors are suspended from the ceiling, or installed in places not easily accessible.

Fig. 1,762.—Auto-transformer or compensator connections for three phase induction motor. In operation when the double throw switch is thrown over to starting position, the current for each phase of the motor flows through an auto-transformer, which consists of a choking coil for each phase, arranged so that the current may be made to pass through any portion of it (as 1, 2, 3) to reduce the voltage to the proper amount for starting. After the motor has come up to speed on the reduced voltage, the switch is thrown over to running position, thus supplying the full line voltage to the motor. [7]In actual construction fuses are usually connected, so that they will be in circuit in the running position, but not in the starting position, where they might be blown by the large starting current.]

Fig. 1,763.—View of armature interior of Wagner polyphase induction motor with wound armature, showing the centrifugal device which at the proper speed short circuits all the coils, transforming the motor to the squirrel cage type. The winding is connected with a vertical "commutator" so called. Inside the armature are two governor weights, which are thrown outwards by the centrifugal force when the machine reaches the proper speed, thus pushing a solid copper ring (which encircles the shaft) into contact with the inner ends of the "commutator" bars, thus completely short circuiting the armature winding.

Fig. 1,764.—Western Electric wound armature for internal resistance induction motor. In starting the inductors are short circuited through a resistance which is gradually cut out as the motor comes up to speed.

Figs. 1,765 to 1769.—Western Electric wound armature for external resistance, or slip ring induction motor, showing brush rigging, slip rings and bar winding.

Fig. 1,770.—Richmond slip ring motor.

Fig. 1,771.—Richmond slip ring armature as used on motor in fig. 1,770.

Fig. 1,772.—Western Electric riveted frame slip ring induction motor for variable speed service; adapted either to continuous or intermittent operation.

Figs. 1,773 to 1,778.—Sprague skeleton type motor frame with various types of armature. Fig. 1,777, plain squirrel cage armature; fig. 1,778, internal resistance armature; fig. 1,773 slip ring armature. In the construction of the plain squirrel cage armature, fig. 1,777, copper bars are inserted in the slots of the core, and are insulated from the core by enclosing tubes which project about one-half inch beyond the iron at each side. The bars are short circuited at their ends by copper rings. These rings are thin, but of considerable radial depth and are held apart by spacing washers. They have rectangular holes punched near their outward periphery, through which the armature bars pass, and to which they are soldered. The internal resistance armature, fig. 1,778, is provided with a phase winding, starting (internal) resistance, and switch located on the shaft. The starting resistance is designed to give approximately full load torque with full load current at starting. A greater torque than full load torque can be obtained for starting, if required, by cutting out resistance. The resistance consists of cast iron grids enclosed in a triangular cover which is bolted to the end plates holding the armature laminæ together, and is short circuited by sliding laminated spring metal brushes along the inside surface of the grids. The brushes are supported by a metal sleeve sliding on the shaft which is operated by a lever secured to the bearing bracket and located just above the bearing. A rod passing through the end of the shaft operates the short circuiting arrangement in sizes up to about 25 horse power. The external resistance or slip ring armature, fig. 1,773, is similar in construction to fig. 1,778, with the exception that slip rings are provided because of the external location of the resistance. These rings connect the inductor through brushes to a controlling and external resistance, two or more carbon brushes being provided for each ring, as in fig. 1,776.

Fig. 1,779.—External resistance or slip ring induction motor connections. The squirrel cage armature winding is not short circuited by copper end rings, but connected in Y grouping and the three free ends connected to three slip rings, leads going from the brushes to three external resistances, arranged as triplex rheostat having three arms rigidly connected as shown, so that the three resistances may be varied simultaneously and in equal amounts.

Fig. 1,800.—Allis-Chalmers phase wound external resistance type or slip ring armature construction. The winding is for three phases and the terminals are brought out to three slip rings. The front bracket is slightly modified to make room for these rings on the inside. For starting duty sufficient resistance is supplied to reduce the starting current taken by the motor to 1¼ times the normal full load current. In the running position the resistance is all cut out of the circuit. For speed regulation sufficient resistance is supplied to reduce the speed 50% on normal full load torque.

Figs. 1,801 to 1,828.—Disassembled view of Western Electric three phase external resistance or slip ring mill type induction motor. It is adapted to severe working conditions, such as are met with in steel mills, crane and hoist service, etc. Designed for 220 or 440 volt, 25 cycle circuits. The frame is divided horizontally into an upper and a lower steel casting, both of which are bolted together at the corners by four heavy bolts. The lower casting is provided with four feet for bolting the motor to its foundation. The end of the upper frame which covers the slip rings is equipped with malleable iron covers held in place by lock bolts. The field and armature are of the usual construction. One end of the armature winding is protected against mechanical injury by the slip rings which are of heavy construction and of practically the same diameter as the armature, and the other end by a detachable flange of the same diameter as the outside of the winding. The slip rings are mounted on the same spider as the armature, so that the shaft can be removed without disturbing any of the connections. The brushes are equipped with riveted pigtails, and held in brass brush boxes machined to gauge. Heavy coiled clock springs are used to maintain an even pressure of the brushes on the slip rings. The armature leads are brought out through holes in the upper half of the frame, and the field leads are brought through a block, which fits in an opening in the upper edge of the lower half.

Fig. 1829.—General Electric single phase induction motor. It is suitable for constant speed service where full load torque at starting does not exceed 140 per cent., and in general is adapted to drive all geared and belted machinery requiring constant speed with light or moderate starting torque.

Fig. 1,830.—Simplified diagram showing the principle of phase splitting for starting single phase induction motors. By the use of an auxiliary set of coils connected in parallel with the main coils and having in series a resistance or condenser as shown, the single phase current delivered by the alternator is "split" into two phases, which are employed to produce a rotating field on which the motor is started.

Figs. 1,831 to 1,850.—Parts of Sprague single phase clutch type induction motor. The armature is of the high resistance smooth core squirrel cage type, the core laminæ being assembled upon a steel sleeve. On starting the armature revolves freely around the shaft on roller bearings until it accelerates to about 75% of its rated speed, when a centrifugal clutch engages with an outer shell keyed directly on the shaft, thus throwing on the load. This type of motor is adapted to drive all belted, geared, or direct connected machinery requiring constant speed with moderate starting torque, such as generators, blowers, line shafting in machine shops and factories, drill presses, laundry machinery, baking machinery, and the like. When greater torque is required at the moment of starting type RI motors should be used, or clutch couplings may be installed between the motor and the machine it is to drive. The parts are as follows; A, field frame; B, field coils; C, terminal block; D, terminal block screws; E, connectors; F, bearing head pulley end; G, bearing head opposite pulley end; H, motor clamping bolts; I, oil well cover; J, oil well plug; K, drain plug; L, oil filter; M, cap bolts; N, bearing lining; O, oil ring; P, belt tightener screw; Q, armature core; R, latch; S, driving shell; T, driving shell set screw; U, clutch ring; V, clutch ring spring; W, spring adjusting screw; X, nut for belt tightener screw; Y, shaft; Z, driving shell key; Aa, armature bearing; Ba, pulley; Ca, pulley set screw; Da, pulley key; Ea, sliding base; Fa, yoke.

Fig. 1,851.—General Electric high resistance clutch type smooth core squirrel cage armature of single phase induction motor. The core laminæ are slotted near the circumference to retain the bar inductors, which extend beyond the core at either end where they are permanently connected to heavy short circuiting rings.

Figs. 1,852 to 1,855.—Parts of General Electric centrifugal clutch pulley as used on clutch type, single phase induction motor. A, clutch; B, friction band; C, adjusting spring; D, outer clutch shell with pulley sleeve; E, solid removable pulley; F, internal mechanism comprising parts A, B, and C; G, outer shell and pulley comprising parts D and E.

Figs. 1,856 and 1,857.—Partly assembled clutch pulley. F, internal mechanism comprising parts A, B, C, of fig. 1,852. G, outer shell and pulley, comprising parts D and E of fig. 1,852.

Fig. 1,858.—Switch end view of General Electric drawn shell type fractional horse power single phase motor.

Figs. 1,859 to 1,862.—Detail construction of clutch parts of General Electric drawn shell type fractional horse power single phase motor. The starting switch, which is assembled within the motor frame, consists essentially of three parts: a rotating member mounted on the armature and provided with two spring controlled pivoted levers in contact with an insulated collector ring.

Fig. 1,863.—Single phase fan motor with shading coils for starting. In addition to the main field coils, one tip of each pole piece is surrounded by a short circuited coil of wire or frame of copper, as indicated in the figure. This coil, or copper frame, is called a shading coil and it causes a phase difference between the pulsating flux that emanates from the main portion of each polar projection and the pulsating flux which emanates from the pole tip, thus introducing a two phase action on the armature which is sufficiently pronounced to start the motor.

Fig. 1,864.—Diagram showing action of shading coil in alternating current motor. The extremities of these pole pieces are divided into two branches, one of which a copper ring called a shading coil is placed as shown, while the other is left unshaded. The action of the shading coils is as follows: Consider the field poles to be energized by single phase current, and assume the current to be flowing in a direction to make a north pole at the top. Consider the poles to be just at the point of forming. Lines of force will tend to pass downward through the shading coil and the remainder of the pole. Any change of lines within the shading coil generates an e.m.f., which causes to flow through the coil a current of a value depending on the e.m.f. and always in a direction to oppose the change of lines. The field flux is, therefore, partly shifted to the free portion of the pole, while the accumulation of lines through the shading coil is retarded.

Figs. 1,865 and 1,866.—Fort Wayne split phase factional horse power induction motor with stationary armature. The object of placing the squirrel cage armature winding on the stationary part or frame is to decrease the radial depth of the latter more than would be possible with the usual arrangement where the armature forms the rotating part. The small radial depth of the stationary armature makes possible a revolving field of maximum diameter giving in turn an exceptionally large air gap area, which reduces the magnetizing current, hence improves the power factor of the motor.

Figs. 1,867 and 1,868.—General Electric disassembled clutch as used on clutch type, single phase (KS) induction motor. In starting, the armature revolves freely on the shaft until approximately 75 per cent. of normal rated speed is reached. The load is then picked up by the automatic action of a centrifugal clutch, which rigidly engages an outer shell, keyed directly to the shaft. The brass friction band of the clutch is permanently keyed to the pulley end of the armature.

Fig. 1,869.—Wagner single phase variable speed commutator motor. The commutator is of the regular horizontal type and the brushes remain in contact all the time. As the torque of alternating motors varies directly as the square of the applied pressure, wide speed variation may be obtained by varying the voltage applied at the motor terminals.

Figs. 1,870 and 1,871.—Diagrams illustrating construction and operation of Wagner "unity power factor" single phase motor. In the field construction, fig. 1,870, two windings are used. The main winding 1 produces the initial field magnetization as heretofore; the auxiliary winding 2 controls the power factor or "compensates" the motor. The main structural departure is in the armature, the construction of which is more clearly indicated in fig. 1,871. Here again two windings are employed. The main or principal winding 4 is of the usual well known squirrel cage type and occupies the bottom of the armature slots. The second or auxiliary winding 3 is of the usual commuted type, is connected to a standard form of horizontal commutator and occupies the upper portion of the armature slots. Between the two is placed a magnetic separator in the form of a rolled steel bar. Two sets of brushes are provided, as indicated in the diagram of connections shown in fig. 1,870. The main pair of brushes 5-6 is placed in the axis of the main field winding 1 and is short circuited. The auxiliary pair of brushes 7-8 is placed at right angles to the axis of the main field winding and is connected in series with it at starting. The auxiliary field winding 2 is permanently connected to one auxiliary brush 7, and is adapted to be connected to the other auxiliary brush 8 by means of the switch 9. The purpose of the peculiar armature construction illustrated in fig. 1,871 and of the brush arrangement and connections shown in fig. 1,870 is to accentuate, at starting, the effect of the squirrel cage along the axis 5-6 of the main field winding 1, while suppressing it as far as possible along the axis 7-8 at right angles to main winding. The magnetic separator placed above the squirrel cage winding 4 tends to suppress the effect of that winding along all axes, by making it less responsive to outside inductive effects. But the influence of the separator is nullified along the axis of the main field winding by the presence of the short circuited brushes 5-6, while no means are provided for nullifying its effects along the axis at right angles to that of the main field winding. Thus the main field winding 1 will be able to induce heavy currents in both armature windings because of the short circuited brushes in the axis 5-6, and in spite of the magnetic separator; while the armature winding 3, connected in series with 1, will not be able to produce heavy currents in the squirrel cage winding 4 along the axis 7-8 because of the magnetic separator between 3 and 4, which shunts or side tracks the inducing magnetic flux. In operation, at starting, switch 9 of fig. 1,870 is open, the commuted winding 3 along the axis 7-8 being connected in series with the main field winding 1 and across the mains. The winding 1 induces a large current in the armature windings 3 and 4 along the axis 5-6, and the winding 3 produces a large flux along the axis 7-8. The armature currents in the main axis co-acting with the flux threading the armature along the auxiliary axis yield the greater part of the starting torque. As the motor speeds up, the squirrel cage gradually assumes those functions which it performs in the ordinary single phase, squirrel cage motor, developing a magnetic field of its own along the axis 7-8 and a correspondingly powerful torque, which increases very rapidly as synchronism is approached, but falls suddenly to zero at or near actual synchronism. It is known that the magnetizing currents circulating in the bars of the squirrel cage of a single phase motor have, at synchronism, double the frequency of the stator currents; the fluxes they produce must therefore also be of double frequency. Now, the magnetic separator is made of solid steel, and, while this separator forms a sufficiently effective shunt for the fluxes of line frequency induced from the field, it is quite ineffective as a shunt for the double frequency fluxes produced by the armature. With respect to the squirrel cage, the effect of this magnetic separator diminishes with increasing speed, and at synchronism the machine operates practically in the same manner as if the magnetic separator did not exist.

Fig. 1,872.—Diagram of ring armature in alternating field illustrating the principles of commutator motors.

Fig. 1,873.—Wagner single phase repulsion induction commutator motor. Its working principle is repulsion start and induction operation. Starting with the machine at rest, brushes in pairs cross connected through a low resistance conductor, bear upon the commutator, temporarily short circuiting the armature winding then developing a strong starting torque on the repulsion principle. On attaining full load speed the individual segments of the commutator are all positively connected together by the operation of an automatic centrifugal governor, thereby transforming the armature winding to the squirrel cage form, the motor then continuing as an induction motor. The governor at the same time removes the brushes from contact with the commutator to save wear. If the power service should fail for any reason, the motor returns to the starting condition, and picks up its load when the power comes on again without attention of the operator.

Figs. 1,874 and 1,875.—Armature of Wagner single phase repulsion-induction commutator motor as seen from the commutator and rear ends, showing the vertical commutator and type of governor employed on the smaller sizes. The operation is explained in fig. 1,873.

Fig. 1,876.—General Electric single phase compensated repulsion motor. The frame is of the riveted form and the field winding consists of distributed concentric coils, each being separately insulated and taped up to each core slot. The compensating winding (depending usually on the size of frame), forms either the center portion of the main winding or a separate winding concentric therewith. The polar groupings are arranged for a frequency of 25 and 60. There are four terminal leads permitting interchangeability of operation on 110 or 220 volt circuits. By connecting adjacent pairs of these terminals in multiple, motors of this type are made adaptable for 110 volt service; for double this pressure the four leading in wires are connected in series. The motor will operate satisfactorily where the arithmetical sum of voltage and frequency variation does not exceed 10 per cent.; that is, the voltage may be 10 per cent. high if the frequency remain at normal, or the frequency may be 10 per cent. high assuming no variation in voltage. A decrease of 5 per cent. in frequency accompanied by a similar increase in voltage is permissible or, as above stated, any similar combination whose arithmetical sum is within 10 per cent. of normal. The armature winding is of the series drum type connected to a commutator carrying two sets of brushes, each set being displaced electrically from the other by 90 degrees. The first set, known as the energy brushes, is permanently short circuited and disposed at an angle to the lines of field or primary magnetization, as in an ordinary repulsion motor. The second set, or compensating brushes, is connected to a small portion of the primary winding included in the field circuit, so as to impress upon the armature an electromotive force, which serves both to raise the power factor and at the same time maintain approximately synchronous speed at all loads. The armature laminations are built up on a cast iron sleeve having the same inside bore as the commutator. In case the shaft become damaged or worn, it can be readily pressed out and replaced without disturbing the commutator or windings. The motor is connected to run counter clockwise. Clockwise rotation is obtained by interchanging the leads to the compensating brushes and slightly shifting the brush holder yoke. This type motor may be thrown on the line without the use of a rheostat, and is suitable for operating refrigerating machines, air compressors, house pumps or similar apparatus where a float switch or pressure regulator is used to close or open the supply circuit.

Fig. 1,877.—Armature of General Electric single phase compensated repulsion motor, assembled ready for dip and banding.

Fig. 1,878.—Cast brush rigging of General Electric single phase compensated repulsion motor as used for the 3 and 5 horse power motors.

Fig. 1,879.—Field of Sprague single phase compensated repulsion motor. The frame is of the skeleton form which exposes the core, giving effective heat radiation. The single phase field winding is of the distributed concentric type. To facilitate connection to circuits of either 110 or 220 volts, four plainly tagged leads are brought out to the back of the removable terminal board.

Figs. 1,880 to 1,884.—Assembly and disassembled view of short circuiting device as used on Bell single phase repulsion induction motor. The armature, which is wound in a similar manner to those used in direct current motors, has a commutator, and brushes, which being short circuited on themselves, allow great starting torque, with small starting current. The motor starts by the repulsion principle, and on reaching nearly full speed, a centrifugal governor pushes the copper ring against the commutator segments, thereby short circuiting them, and the motor then operates on the induction principle.

Fig. 1,885.—Section of ring armature of commutator motor showing local current set up by transformer action of the alternating flux.

Fig. 1,886.—General Electric 5 H.P., 6 pole adjustable speed single phase compensated repulsion motor. This type is suitable for service requirements demanding the use of a motor whose speed can be adjusted over a considerable range, this speed at a fixed controller setting remaining practically unaffected by any load within the motor's rated capacity. With the controller on the high speed points, the motor possesses an inherent speed regulation between no load and full load of approximately 6 per cent. At the low speed points, under similar load conditions, the speed variation will be approximately 20 per cent. To secure adjustable speed control, the armature circuits employ transformers, whose primaries are excited by the line circuit. The secondaries of these transformers are divided into two sections; the first or "regulating" circuit is placed across the energy brushes; the other section, since it is connected in series with the compensating winding, maintains the high power factor and speed regulation obtained in the constant speed type. The speed range is 2:1, approximately one-half of this range being below and one-half above synchronous speed.

Fig. 1,887.—Diagram of single phase series commutator motor. It is practically the same as the series direct current motor, with the exception that all the metal of the magnetic circuit must be laminated.

Fig. 1,888.—Diagram of neutralized series motor; conductive method. In the simple series motor, there will be a distortion of the flux as in the direct current motor. As the distorting magnetic pressure is in phase with that of the magnets, the distortion of the flux will be a fixed effect. If the poles be definite as in direct current machines, this distortion may not seriously affect the running of the motor, but with a magnetizing system like that universally adopted in induction motors the flux will be shifted as a whole in the direction of the distortion, which will produce the same effect as if in the former case the brushes had been shifted forward, whereas for good commutation they should have been shifted backward. As in direct current machines, this distortion is undesirable since it is not conducive to sparkless working, and also reduces to a more or less extent the torque exerted by the motor. The simplest remedy is to provide neutralizing coils displaced 90 magnetic degrees to the main field coils as shown. The neutralizing current is obtained by the method of connecting the neutralizing coils in series in the main circuit.

Fig. 1,889.—Diagram of neutralized series motor; inductive method. Although the conductive method of neutralization is employed in nearly all machines, it is possible merely to short circuit the neutralizing winding upon itself, instead of connecting it in series with the armature circuit. In this case the flux due to the armature circuit cannot be eliminated altogether, as sufficient flux must always remain to produce enough pressure to balance that due to the residual impedance of the neutralizing coil. It would be a mistake to infer, however, that on this account this method of neutralization is less effective than the conductive one, since the residual flux simply serves to transfer to the armature circuit a drop in pressure precisely equivalent to that due to the resistance and local self-induction of the neutralizing coil in the conductive method.

Fig. 1,890.—Diagram of simple shunt commutator motor. Owing to its many inherent defects it is of little importance.

Fig. 1,891.—Compensated shunt induction single phase motor. The transformer shown in the arrangement is capable of being replaced by a coil placed on the frame having the same axis as the field winding, so that the flux produced by the field winding induces in the coil a pressure in phase with the supply pressure. Such a coil will now be at right angles to the circuit to which it is connected. In a similar manner a coil at right angles to the armature circuit, that is, the circuit parallel to the stator axis, if connected in series with that circuit, will also serve to compensate the motor.

Fig. 1,891.—Compensated shunt induction single phase motor. The transformer shown in the arrangement is capable of being replaced by a coil placed on the frame having the same axis as the field winding, so that the flux produced by the field winding induces in the coil a pressure in phase with the supply pressure. Such a coil will now be at right angles to the circuit to which it is connected. In a similar manner a coil at right angles to the armature circuit, that is, the circuit parallel to the stator axis, if connected in series with that circuit, will also serve to compensate the motor.

Fig. 1,892.—Fynn's shunt conductive single phase motor. In order to supply along the stator axis a constant field, suitable for producing the cross flux to which the torque is due by its action on the circuit perpendicular to the stator axis, the "armature circuit," as it may be called, has a neutralizing coil in series with it, so that the armature circuit and neutralizing coil together produce no flux. In addition to this, there is a magnetizing coil along the same axis, which is connected across the mains and so produces the same flux as the primary coil in a shunt induction machine. Fynn has proposed a number of methods of varying the speed and compensating this machine. It is, however, complicated in itself, and is only suited for very low voltages, so that on ordinary circuits it would need a separate transformer.

Fig. 1,893.—Effect of alternating field on copper ring. If a copper ring be suspended in an alternating field so that the plane of the ring is oblique to the lines of force, it will turn until its plane is parallel to the lines of force, that is, to the position in which it does not encircle any lines of force. The turning moment acting upon the ring is proportional to the current in it, to the strength of the field, and to the cosine of the angle ß. Hence it is proportional to the product sin ß cos ß. The tendency to turn is zero both at 0° and at 90°; in the former case because there is no current, in the latter because the current has no leverage. It is a maximum when ß = 45°. Even in this position there would be no torque if there were no lag of the currents in the ring, for the phase of the induced pressure is in quadrature with the phase state of the field. When the field is of maximum strength there is no pressure, and when the pressure reaches its maximum there is no field. If there be self-induction in the ring causing the current to lag, there will be a net turning moment tending to diminish ß. The largest torque will be obtained when the lag of the current in the ring is 45°.

Figs. 1,894 to 1,908.—Parts of General Electric single phase compensated repulsion motor. The field frame employs the riveted form of construction, so that the ends of the laminations are exposed directly to the air, insuring low operating temperatures and high overload capacity. The field winding consists of a main winding of the distributed concentric type and a compensating winding. The series type of winding is employed, and the completed rotor is treated with a special insulating compound, which renders the coils moisture proof under ordinary conditions. On motors of more than 2 horse power capacity a ventilating fan is attached to the rotor which provides a continuous supply of cool air while the motor is in operation. Two types of brush holder yoke are used. The smaller motors use a moulded yoke of insulting compound, reinforced by a cast iron L section ring embedded in the moulded structure. Cast iron yokes are used on larger motors. The brushes are of carbon with copper pigtails, which carry all the current. The brushes in this machine remain permanently in contact with the commutator. The parts are: A, field; B, field winding; C, line terminal; D, tube terminal; E, compensating terminal; F, terminal board; G, brush yoke; H, brush holder; I, carbon brush; J, brush stud; K, short circuit connection; L, armature; M, commutator, N, shaft; O, fan; P, commutator end shield; Q, pulley end shield; R, oil well cover; S, oil plugs; T, oil gauge; U, bearing lining; V, oil ring; W, pulley; X, pulley set screw; Y, commutator end shield holding bolts; Z, pulley end shield holding bolts; AA, base; BB, float bolts; CC, belt tightener screw.

Fig. 1,909.—Fynn's compensated shunt induction motor. This is a combination of the compensated shunt induction motor with the ordinary squirrel cage form. In one form, in addition to the ordinary drum winding on the armature, there is another three phase winding into the "star," of which the drum winding is connected. This second winding is connected to three slip rings which are short circuited when the machine is up to speed. Upon the commutator are placed a pair of brushes connected to an auxiliary winding placed on the frame in such a position that the flux from the primary coil induces in it a pressure of suitable phase to produce compensation. The same pair of brushes is also used for starting.

Fig. 1,910.—Diagram of connection of Sprague single phase compensated repulsion motor. To reverse direction of rotation interchange leads C₁ and C₂ and slightly shift the brush holder yoke. Brushes E₁ and E₂ are permanently short circuited. This diagram of connections applies also to fig. 1,911.

Fig. 1,911.—Diagram of connections of Sprague variable speed single phase compensated repulsion motor and controller. The controller is designed to give speed reduction and speed increase as resistance or reactance is inserted in the energy and compensating circuits. With the exception of the leads brought out from these circuits, the constant speed and variable speed motors are identical. The standard controller gives approximately 2:1 speed variation.

Fig. 1,912.—Diagram of connections of Sprague reversing type of single phase compensated repulsion motor. As shown, there is a special reverse field winding having terminals for connection to a four pole double throw switch.

Fig. 1,913.—Fairbanks-Morse squirrel cage armature, showing ball bearings.

Fig. 1,914.—Fairbanks-Morse 20 horse power squirrel cage induction motor connected to a 20 inch self-feed rip and chamfering saw. The absence of commutator and brushes on the squirrel cage armature eliminates sparking and therefore renders this type of motor particularly adapted for use in places where sparking would be dangerous, such as in wood working plants, textile mills, etc.

Fig. 1,915.—Method of casting end rings on squirrel cage armatures of Fairbanks-Morse induction motors. The metal being fused to the bars at a temperature in excess of 1,832 degrees Fahr., it is readily seen that the destructive effect of any subsequent heating is eliminated. While giving the most intimate contact at the joints, a multiplicity of joints is avoided as well as solder.

Fig. 1,916.—Diagram of elementary transformer with non-continuous core and connection with single phase alternator. The three essential parts are: primary winding, secondary winding, and an iron core.

Fig. 1,917.—Diagram of elementary transformer with continuous core and connections with alternator. The dotted lines show the leakage of magnetic lines. To remedy this the arrangement shown in fig. 1,918 is used.

Fig. 1,918.—Cross section showing commercial arrangement of primary and secondary windings on core. One is superposed on the other. This arrangement compels practically all of the magnetic lines created by the primary winding to pass through the secondary winding.

Fig. 1,919.—Wagner transformer coil formed, ready for taping. These are known as "pan cake" coils. They are wound with flat cotton covered copper strip. In heavy coils, several strips in parallel are used per turn in order to facilitate the winding and produce a more compact coil.

Fig. 1,920.—Wagner coils with insulation ready for core assembly. The flat coils, sometimes called pancake coils are wound of flat, cotton covered, copper strip with ample insulation between layers. In heavy coils several flat strips in multiple are used per turn in order to facilitate the winding and produce a more compact coil. In many cases normal current flow per high tension coil is very low and could be carried with a very small cross sectional area of copper; however, flat strip is almost always used on account of the increased mechanical stability thus obtained.

Figs. 1,921 and 1,922.—Assembled coils of Westinghouse 10 and 15 kva. transformers; views showing ventilating ducts.

Fig. 1,923.—Rear view of Fort Wayne distributing transformer, showing hanger irons for attaching to pole cross arm.

Fig. 1,924.—Diagram of elementary step up transformer. As shown the primary winding has two turns and secondary 10 turns, giving a ratio of voltage transformation of 10 ÷ 2 = 5. Since only ⅕ as much current flows in the secondary winding as in the primary, the latter requires heavier wire than the former.

Fig. 1,925.—Diagram of elementary step down transformer. As shown the primary winding has 10 turns and the secondary 2, giving a ratio of voltage transformation of 2 ÷ 10 = .2. The current in the secondary being 5 times greater than in the primary will require a proportionately heavier wire.

Figs. 1,926 and 1,927.—Core type transformer. It consists of a central core of laminated iron, around which the coils are wound. A usual form of core type transformer consists of a rectangular core, around the two long limbs of which the primary and secondary coils are wound, the low tension coil being placed next the core.

Fig. 1,928.—Shell type transformer. In construction the laminated core is built around and through the coils as shown. For large ratings this type has some advantages with respect to insulation, while for small ratings the core type is to be preferred in this respect. The shell arrangement of the core gives better cooling; with this arrangement minimum magnetic leakage is easily obtained.

Fig. 1,929.—View illustrating the construction of cores and coils of Maloney transformers.

Fig. 1,930.—Maloney mica shield between primary and secondary coils, showing lapping feature which prevents the wrinkling and cracking of the mica.

Figs. 1,931 and 1,932.—The Berry combined core and shell transformer. It consists of a number of inner and outer vertical and radial laminated iron blocks built up of the usual thin sheet iron, with the coils between. The magnetic circuit is completed at the top and bottom by other laminated blocks placed horizontally, and the whole is held together between top and bottom cast iron frame plates by a bolt passing right down the center. Fig. 1,931 gives a general view, W being the winding, and B, B, B, etc., the outer laminated blocks. The construction will be better understood from fig. 1,932, where it may be supposed that the top cap and laminated cross pieces have been removed. Here I, I, I and O, O, O are respectively the inner and outer radial vertical blocks, P the primary, and S, S the secondary; the latter being in two sections with the primary sandwiched between, as an extra precaution against shock. It will be evident that this form of transformer possesses excellent ventilation; and this is still further enhanced by opening out the winding at intervals to leave ventilating apertures, as at A, A, A. Fig. 1,932 shows only 6 sets of radial blocks, but the usual plan is to provide 24 or 36, according to the size of the transformer.

Fig. 1,933.—Plan of core of General Electric combined core and shell transformer. The core used contains four magnetic circuits of equal reluctance, in multiple; each circuit consisting of a separate core. In this construction one leg of each circuit is built up of two different widths of punchings forming such a cross section that when the four circuits are assembled together they interlock to form a central leg, upon which the winding is placed. The four remaining legs consist of punchings of equal width. These occupy a position surrounding the coil at equal distances from the center, on the four sides; forming a channel between each leg and coil, thereby presenting large surfaces to the oil and allowing its free access to all parts of the winding. The punchings of each size transformer are all of the same length, assembled alternately, and forming two lap joints equally distributed in the four corners of the core, thereby giving a magnetic circuit of low reluctance.

Figs. 1,934 and 1,935.—Top view showing core and coils in place, and view of coils of Westinghouse distributing transformer. The coils are wound from round wire in the smaller sizes of transformers and from strap copper in the larger sizes. Strap wound coils allow a greater current carrying conductor section than coils wound from large round wire, as there is little waste space between the different turns of the conductor. The coils are arranged concentrically with the high tension winding between the two low tension coils, this arrangement giving the fine regulation found in these transformers. The low tension coils are wound in layers which extend across the whole length of the coil opening in the iron, while the high tension coils are wound in two parts and placed end to end. This construction reduces the normal voltage strains to a value which will not give trouble under any condition of service. The magnetic circuit is built up of laminated, alloy steel punchings, each layer of laminæ being reversed with reference to the preceding layer and all joints butted. This gives a continuous magnetic circuit of low reluctance, low iron loss and low exciting current. When assembled, the magnetic circuit consists of four separate parallel circuits encircling the coils and protecting the windings from mechanical injury. Separate high and low tension terminal blocks of glazed porcelain are mounted upon extensions of the upper end frames. All danger of confusing the leads or inadvertently making an electrical connection between the high and low tension sides of the transformer is thus averted. The high tension winding has four leads brought to the studs in the terminal block. Adjustable brass connectors or links between the studs provide for series or multiple connections between two points of the high tension winding. The position of the studs and the length of the links are so proportioned that wrong connections on the block are impossible. Barriers on the porcelain block separate the studs and prevent danger of arcing. Leads with means of preventing creeping of oil by capillary action are attached to these studs and brought out of the core through porcelain bushings.

Figs. 1,936 and 1,937.—Core and shell types of three phase transformer. In the core type, fig. 1,936, there are three cores A, B, and C, joined by the yokes D and D'. This forms a three phase magnetic circuit, since the instantaneous sum of the fluxes is zero. Each core is wound with a primary coil P, and a secondary coil S. As shown, the primary winding of each phase is divided into three coils to ensure better insulation. The primaries and secondaries may be connected star or mesh. The core B has a shorter return path than A and C, which causes the magnetizing current in that phase to be less than in the A and C phases. This has sometimes been obviated by placing the three cores at the corners of an equilateral triangle (as in figs. 1,939 and 1,940), but the extra trouble involved is not justified, as the unbalancing is a no load condition, and practically disappears when the transformer is loaded. The shell type, fig. 1,937, consists practically of three separate transformers in one unit. The flux paths are here separate, each pair of coils being threaded by its own flux, which does not, as in the core type, return through the other coils. This gives the shell type an advantage over the core type, for should one phase burn out, the other two may still be used, especially if the faulty coils be short circuited. The effect of such short circuiting is to prevent all but a very small flux from threading the faulty coil.

Fig. 1,938.—Interior of General Electric oil cooled 500 kva. 33,000 volt outdoor transformer showing lifting arrangement.

Figs. 1,939 and 1,940.—Triangular arrangements of cores of three phase transformer. Fig. 1,939, form with three cornered yokes at bottom and top of cores; fig. 1,940, form with circular yokes. While these designs give perfect symmetry for the three phases, there is some trouble in the mechanical arrangement of the yokes. If these be stamped out triangularly and inserted horizontally between the three cores, it is necessary to interpose a layer of insulation, otherwise there would be objectionable eddy currents formed in the stampings.

Fig. 1,941.—View showing mechanical construction of coil and core of Moloney pole type ½ to 50 kw. transformer. Moloney standard transformers of these sizes are regularly wound for 1,100 to 2,200 primary volts. For 1,100 volts the primary coils are connected in parallel by means of connecting links; for 2,200 volts, they are connected in series. The porcelain primary terminal board is provided with two connecting links so that connections can be made for either 1,100 or 2,200 volts.

Fig. 1,942.—Fort Wayne transformer coils and core complete.

Fig. 1,943.—Top view of Fort Wayne (type A) transformer cover removed, showing assembly of coils and core and disposition of leads.

Fig. 1,944.—General Electric 10 kva., (type H) transformer removed from tank. That part of the steel core composing the magnetic circuit outside of the winding is divided into four equal sections. Each section is located a sufficient distance from the winding so that all portions of the winding and core are equally exposed to the cooling action of the oil. On all except the very smallest sizes the winding is divided by channels and ducts through which a continual circulation of oil is maintained. The result is uniform temperature throughout the transformer, thus eliminating the detrimental effects of unequal expansion in the coils with consequent rubbing and abrasion of the insulation.

Fig. 1,945.—Cover construction of Wagner 350 kva., oil filled 1,100-2,200 volt transformer. In transformers with corrugated cases, the base and top ring are cast to the corrugated iron sheets.

Figs. 1,946 to 1,948.—Methods of connecting the low tension sides of Westinghouse transformers using the connectors illustrated in figs. 1,949 to 1,953.

Figs. 1,949 to 1,953.—Westinghouse low tension transformer connectors for connecting the low tension leads to the feeder wires. The transformers of the smaller capacities have knuckle joint connectors and those of the larger sizes have interleaved connectors. These connectors form a mechanically strong joint of high current carrying capacity. Since the high tension leads are connected directly to the cut out or fuse blocks, connectors are not required on these leads. The use of these connectors allows a transformer to be removed and another of the same or a different capacity substituted usually without soldering or unsoldering a joint. The connectors also facilitate changes in the low tension connections.

Fig. 1,954.—Method of bringing out the secondary leads in Wagner central station transformers. Each primary lead is brought into the case through a similar bushing. Observe the elimination of all possibility of grounding the cable on the case or core.

Figs. 1,955 and 1,956.—Westinghouse transformer terminal blocks for high and low tension conductors.

Figs. 1,957 to 1,960.—Porcelain bushing for Westinghouse transformers.

Figs. 1,961 to 1,963.—Fuse blocks for Westinghouse transformers. The fuses furnished with the transformers are mounted in a weather proof porcelain fuse box of special design. The stationary contacts are deeply recessed in the porcelain and are well separated from each other. The contacts are so constructed that the plug is held securely in place by giving it a partial turn after inserting it. When the plug is in position, the fuse is in sight and its condition can be noted which eliminates all danger of pulling the fuse while same is still intact and the transformer is under load.

Fig. 1,964.—Forced draught or "air blast" transformer. As is indicated by the classification, this type of transformer is cooled by forcing a current of air through ducts, provided between the coils and between sectionalized portions of the core. The cold air is forced through the interior of the core containing the coils by a blower, the air passing vertically through the coils and out through the top. Part of the air is sometimes diverted horizontally through the ventilating ducts provided in the core, passing off at one side of the transformer. The amount of air going through the coils, or through the core, may be controlled independently by providing dampers in the passages.

Fig. 1965.—Looking down into a Wagner central station transformer, showing the connection board, which provides facility for varying the ratio of transformation and also for interchanging the primaries.

Fig. 1966.—Section through Westinghouse ½ kilovolt ampere type S transformer. Fig. 1967.—Section through Westinghouse 50 kilovolt ampere type S transformer showing large oil ducts.

Fig. 1,968.—Wagner 300 kva, 4,400 volt three phase oil cooled transformer. In this type of transformer the case is filled with oil and fluted so as to increase the cooling surface, an oil drain valve is screwed to a wrought iron nipple cast into the base, the duct to which is in such a position as to make it possible not only to drain all of the oil from the transformer, but when desirable, to draw off a small quantity from the bottom. Should any moisture be in the oil it is therefore drawn off first.

Fig. 1,969.—Section through Fort Wayne (type A) transformer showing interior of case, core conductors, and insulation, also division of laminæ.

Fig. 1,970.—Water cooled transformer with internal cooling coil, that is, with cooling coil within the transformer case. In this type, the cooling coil, through which the circulating water passes, is placed in the top of the case or tank, the latter is filled with oil so that the coil is submerged. The oil acts simply as a medium to transfer the heat generated by the transformer to the water circulating through the cooling coil. In operation a continual circulation of the oil takes place, as indicated by the arrows, due to the alternate heating and cooling it receives as it flows past the transformer coils and cooling coil respectively.

Fig. 1,971.—Water cooled transformer with external cooling coil. In this arrangement the cooling coil is placed in a separate tank as shown. Here forced circulation is employed for both the heat transfer medium (oil) and the cooling agent (water), two pumps being necessary. The cool oil enters the transformer case at the lowest point and absorbing heat from the transformer coils it passes off through the top connection leading to the cooling coil and expansion tank. Since the transformer tank is closed, an expansion tank is provided to allow for expansion of the oil due to heating. The water circulation is arranged as illustrated.

Fig. 1,972.—Interior of General Electric water cooled 140,000 volt transformer showing cooling coil.

Fig. 1,973.—Assembled coils of General Electric water cooled 500 kva., 66,000 volt transformer.

Fig. 1,974.—Three Westinghouse 20 kva, outdoor transformers, for irrigation service. These are mounted on a drag so that they may be readily transported from place to place. 33,000 volts high tension; 2,200 and 440 volts low tension, 50 cycles. These outdoor transformers are of the oil immersed, self-cooling type and have been developed to meet the requirements for transformers of capacities greater or of voltages higher than are usually found in distribution work. They are in reality distributing transformers for high voltage, outdoor installations, single or three phase service, for voltages up to 110,000. Where the magnitude of the load does not warrant an expensive installation, transformers of the outdoor type are particularly applicable. The cost of a building and outlet bushings which is often the item of greatest expense is eliminated where outdoor type transformers are installed.

Fig. 1,975.—Efficiency curve of Westinghouse 375 kw., transformer. Pressure 500 to 15,000 volts; frequency 60. Efficiencies at different loads: full load efficiency, 98%; ¾ full load efficiency, 98%; ½ full load efficiency, 97.6%; ¼ full load efficiency, 96.1%; regulation non-inductive load, 1.4%; load having .9 power factor, 3.3%.

Figs. 1,976 and 1,977.—Westinghouse double pole fuse box; views showing box open with tubes in place, and with tubes removed.

Fig. 1,978.—Diagram illustrating connections and principles of auto-transformers as explained in the accompanying text.

Figs. 1,979 and 1,980.—Two winding transformer and single winding or auto-transformer. Fig. 1,979 shows a 200:100 volt transformer having a 10 amp. primary and a 20 amp. secondary, the currents being in opposite directions. If these currents be superposed by using one winding only, the auto-transformer shown in fig. 1,980 is obtained where the winding carries 10 amp. only and requires only one-half the copper (assuming the same mean length of turn). If R be the ratio of an auto-transformer, the relative size of it compared with a transformer of the same ratio and output is ((R - 1) / R):1. For example, a 10 kw. transformer of 400 volts primary and 300 volts secondary could be replaced by an auto-transformer of 10 × (1.33 - 1) / 1.33 = 2.5 kw.; or, in other words, the amount of material used in a 2½ kw. transformer could be used to wind an auto-transformer of 400:300 ratio and 10 kw. output.

Fig. 1,981.—Elementary diagram illustrating the principles of constant current transformer as used for series arc lighting.

Fig. 1,982.—Mechanism of General Electric air cooled constant current transformer. It operates on the principle explained in the accompanying text and is built to supply 25 to 100 arc lamps at 6.6 to 7.5 amperes. The transformers are interchangeable and will operate on 60 or 125 cycles. The relative positions of the two coils may be changed in order to regulate the strength of the current more closely, by shifting the position of the arc carrying the counterbalance by means of the adjusting screw on it. A dash pot filled with special oil prevents sudden movements of the secondary coil and keeps the current through the lamps nearly constant, when they are being cut in or out of the circuit. In starting up a constant current transformer, it is necessary to separate the two coils as far as possible and then close the primary circuit switch and allow the two coils to come together. If the primary circuit be thrown directly on the generator the heavy rush of current which will follow due to the two coils being too close together might injure the lamps.

Fig. 1,983.—General Electric air cooled constant current transformer. View showing external appearance with case on.

Fig. 1,984.—Single phase transformer connection with constant pressure main.

Fig. 1,985.—Usual method of single phase transformer connections for residence lighting with three wire secondaries. A balancing transformer is connected to the three wire circuit near the center of distribution as shown.

Fig. 1,986.—Diagram of single phase transformer having primary and secondary windings in two sections, showing voltages per section with series connections.

Fig. 1,987.—Diagram of single phase transformer with primary and secondary windings of two sections each, showing voltages per section with parallel connection.

Fig. 1,988.—Diagram showing unlike single phase transformers in parallel.

Fig. 1,989.—Method of comparing instantaneous polarity. Two of the terminals are connected as shown by a small strip of fuse wire, and then touching the other two terminals together. If the fuse blows, then the connections must be reversed; if it does not, then they may be made permanent.

Figs. 1,990 and 1,991.—Methods of altering the secondary connections of a transformer having two sections in the secondary to obtain a different voltage. Fig. 1,990 shows the two sections in parallel giving say 100 volts; fig. 1,991 shows the two sections in series giving 200 volts.

Figs. 1,992 to 1,995.—Three phase transformer connections. Fig. 1,992 delta connection; fig. 1,993 star connection; fig. 1,994 delta star connection; fig. 1,995 star-delta connection.

Fig. 1,996.—Two phase transformer connections. Two single phase transformers are used and connections made just as though each phase were an ordinary single phase system.

Fig. 1,997.—Two phase transformer connections, with secondaries arranged for three wire distribution, the primaries being independently connected to the two phases. In the three wire circuit, the middle or neutral wire is made about one-half larger than each of the two outer wires. In fig. 1,996 it makes no difference which secondary terminal of a transformer is connected to a given secondary wire, so long as no transformers are used in parallel. For example, referring to the diagram, the left hand secondary terminal of transformer, A, could just as well be connected to the lower wire of the secondary phase, A, and its right hand terminal connected to the upper wire, the only requirement being that the two pairs of mains shall not be "mixed"; that is, transformer, A, must not be connected with one secondary terminal to phase, A and the other to phase, B. In the case shown by fig. 1,997, there is not quite so much freedom in making connections. One secondary terminal of each transformer must be connected to one of the outer wires and the other two terminals must be both connected to the larger middle wire of the secondary system. It makes no difference, however, which two secondary terminals are joined and connected to the middle wire so long as the other terminal of each transformer is connected to an outer wire of the secondary system.

Fig. 1,998.—Three wire connections for transformer having two secondary sections on different legs of the core. If the secondary terminals be connected up to a three wire distribution, as here shown diagrammatically, it is advisable to make the fuse, 2, in the middle wire, considerably smaller than necessary to pass the normal load in either side of the circuit, because, should the fuse, 1, be blown, the secondary circuit through the section, Sa, will be open, and the corresponding half of the primary winding, Pa, will have a much higher impedance than the half of the primary winding, Pb, the inductance of which is so nearly neutralized by the load on the secondary winding, Sb. The result will be that the voltage of the primary section, Pa, will be very much greater than that of the section, Pb, and as the sections are in series the current must be the same through both halves of the winding; the drop or difference of pressure, therefore, between the terminals of Pa will be much higher than that between the terminals of Pb, consequently, the secondary voltage of Sb will be greatly lowered and the service impaired. As the primary winding, Pa, is designed to take only one-half of the total voltage, the unbalancing referred to will subject it to a considerably higher pressure than the normal value; consequently, the magnetic density in that leg of the transformer core will be much higher than normal, and the transformer will heat disastrously. If the fuse, 2, in the middle wire be made, say, one-half the capacity of each of the other fuses, this condition will be relieved by the blowing of this fuse, and as the lamps in the live circuit would not be anywhere near candle power if the circuit remained intact, the blowing of the middle fuse will not be any disadvantage to the user of the lamps. Some makers avoid the contingency just described by dividing each secondary coil into two sections and connecting a section on one leg in series with a section on the other leg of the core, so that current applied to either pair of the secondary terminals will circulate about both legs of the core.

Figs. 1,999 to 2,002.—Three phase delta, and star connections using three transformers. There are two ways of connecting up the primaries and secondaries, one known as the "delta" connection, and illustrated diagrammatically by fig. 1,999, and the other known as the "star" connection, and illustrated by fig. 2,001. In both diagrams the line wires are lettered, A, B and C. Fig. 2,000 shows the primaries and secondaries connected up delta fashion, corresponding to fig. 1,999, and fig. 2,002 shows them connected up star fashion, corresponding to fig. 2,001. In both of the latter sketches the secondary wires are lettered to correspond with the respective primary wires. When the primaries are connected up delta fashion, the voltage between the terminals of each primary winding is the same as the voltage between the corresponding two wires of the primary circuit, and the same is true of the secondary transformer terminals and circuit wires. The current, however, flowing through the transformer winding is less than the current in the line wire, for the reason that the current from any one line wire divides between the windings of two transformers. For example, in figs. 1,999 and 2,000, part of the current from the line wire, A, will flow from A to B through the left hand transformer, and part from A to C through the right hand transformer; if the current in the line wire, A, be 100 amperes, the current in each transformer winding will be 57.735 amperes. When transformers are connected up star fashion, as in figs. 2,001 and 2,002, the current in each transformer winding is the same as that in the line wire to which it is connected, but the voltage between the terminals of each transformer winding is 57.735 per cent. of the voltage from wire to wire on the circuit. For example, if the primary voltage from A to B is 1,000 volts, the voltage at the terminals of the left hand transformer (from A to J) will be only 577.35 volts, and the same is true of each of the other transformers if the system is balanced. These statements apply, of course, to both primary and secondary windings, from which it will become evident that if the three transformers of a three phase circuit be connected up star fashion at the primaries, and delta fashion at the secondaries, the secondary voltage will be lower than if both sides are connected up star fashion. For example, if the transformers be wound for a ratio of 10 to 1, and are connected up with both primaries and secondaries alike, no matter whether it be delta fashion or star fashion, the secondary voltage will be one-tenth of the primary voltage; but if the primaries be connected up star fashion on a 1,000 volt circuit, and the secondaries be connected up delta fashion, the secondary voltage will be only 57.735 volts, instead of 100 volts. The explanation of the difference between the voltage per coil in a delta system and that in a star system is that in the former each winding is connected directly across from wire to wire; whereas in the star system, two windings are in series between each pair of line wires. The voltage of each winding is not reduced to one-half, however, because the pressures are out of phase with each other, being 120°, or one-third of a cycle, apart; consequently, instead of having 500 volts at the terminals of each coil in fig. 2,001 the voltage is 577.35. The same explanation applies to the current values in a delta system. The current phase between A and B, in fig. 1,999, is 120° removed from that in the winding between A and C; consequently the sum of the two currents, in the wire, A, is 1.732 times the current in each wire; or, to state it the opposite way, the current in each winding is 57.735% of the current in the wire, A. It will be well for the reader to remember that in all cases pressures differing in phase when connected in series, combine according to the well-known law of the parallelogram of forces; currents differing in phase, and connected in parallel, combine according to the same law.

Fig. 2,003.—Diagram showing three wire secondary connections General Electric (type H) transformer. As will be seen, the method adopted consists of distributing equally, on each side of the primary coil, both halves of the secondary winding, so that each secondary throughout its length is closely adjacent to the entire primary winding. In order to insure the exact equality of resistance and reactance in the two secondary windings necessary to obtain perfect regulation of the two halves, the inside portion of the secondary winding on one side of the primary coil is connected in series with the outside portion of that on the other side. As a result, the drop of voltage in either side of the secondary under any ordinary conditions of unbalanced load, does not exceed the listed regulation drop. This particular arrangement is used because it is the simplest and best method for this construction.

Figs. 2,004 to 2,011.—Connections of standard transformers. All stock transformers are wound for some standard transformation ratio, such as 10 to 1, but various leads are brought out by means of which ratios of 5, 10 and 20 to 1 may be obtained for one transformer. The figures show the voltage combinations possible with a standard transformer.

Fig. 2,012.—Method of determining core loss. Connect voltmeter and wattmeter as shown in the illustration to the low tension side of the transformer. By means of a variable voltage transformer bring the applied voltage to the point for which the transformer is designed. The wattmeter indicates directly the core loss, which includes a very small loss due to the current in the copper.

Cautions.—1. Make sure of the voltage and frequency. The manufacturers' tabulated statements refer to a definite voltage and frequency and these have a decided influence upon the core loss. 2. The high tension circuit must remain open during the test.

Fig. 2,013.—Method of determining copper loss. Connect ammeter and wattmeter to high tension side of transformer short circuit secondary leads, as shown in the illustration, and by means of a variable voltage, adjust current to the full load value for which the transformer is intended. The wattmeter reading shows the copper loss at full load. The full load primary current of any transformer is found from the following equation.

• full load current = full load watts ÷ primary volts

EXAMPLE: To find proper full load current on a five kw. 2,200 volt transformer, divide 5,000 watts by 2,200 volts, the full load current will then be 2.27 amperes. A slight variation in primary current greatly increases or decreases the copper loss.

Remarks.—Copper loss increases with temperature because the resistance of the metal rises. Do not overload the current coil of the wattmeter. For greater accuracy the I²R drop of potential method should be used.

Fig. 2,014.—Diagram of connections for regulation test. Connect transformer under test to high tension supply circuit. A second transformer with same or other known change ratio is also to be connected up, as illustrated. By means of a double pole double throw switch, the voltmeter can be made to read the pressure on the secondary of either transformer. Supposing the same change ratio it is evident that if both remain unloaded the voltmeter will indicate the same pressure. A gradually increasing lamp load up to the limit of the transformer capacity, will be attended by a drop in pressure at the terminals. This drop can be read as the difference of the voltmeter indications, and when expressed in per cent. of secondary voltage stands for "regulation." Remarks: The auxiliary transformer is necessary in order to make sure of the high tension line voltage. A large transformer under test may cause primary drop in taking power. This must be set down against it in testing regulation. The second transformer gives notice of such drop, whatever be the cause. Figs. 2,012 to 2,014 used by courtesy of the Moloney Electric Co.

Fig. 2,015.—Wagner central station core type transformer repair unit consisting of one half set of primary and secondary windings together with the section of the iron core upon which the coils are wound.

Fig. 2,016.—Moloney tubular air draft oil filled transformer. The case is made of cast iron, with large steel tubes passing from the bottom through the top. In operation the air in the tubes becomes hot and expands; a draft is thus produced which carries away considerable heat.

Fig. 2,017.—Moloney pressure transformer adapted for switchboard work in connection with voltmeters, wattmeters the, etc., in sizes from 25 to 500 watts.

Fig. 2,018.—Moloney current transformer switchboard or indoor type. It is used ordinarily for insulating an ammeter, a current relay, the current coil of a watt meter or watt hour meter from a high tension circuit, for reducing the line current to a value suitable for these instruments.

Fig. 2,019.—Diagram showing a method of operating a three phase motor on a two phase circuit, using a transformer having a tap made in the middle of the secondary winding, so as to get the necessary additional phase. While this does not give a true balanced three phase secondary, it is close enough for motor work. In the above arrangement, the main transformer supplies 54 per cent. of the current and the other with the split winding 46 per cent.

Fig. 2,020.—Three phase motor transformer connections; the so-called Delta connected transformers.

Fig. 2,021.—Three phase motor connections using two transformers.

Fig. 2,022.—Delta-star connection of three transformers for low pressure, three phase, four wire system.

Fig. 2,023.—Diagram of transformer connections for motors on the monocyclic system.

Fig. 2,024.—Moloney flaming auto type arc lamp transformer for 110 volts primary to 55 volts secondary. A hook in bottom of case provides means for suspension of lamp. The transformer may be operated on circuits from 100 to 120 volts primary, 50 to 60 volts secondary. The secondary capacity is 8 to 12 amperes.

Fig. 2,030.—Installation of a transformer on pole; view showing method of attachment and disposition of the primary and secondary leads, cutouts, etc.

Fig. 2,031.—Diagram of static booster or regulating transformer. It is used for regulating the pressure on feeders. In the figure, B are the station bus bars, R the regulable transformer, F the two wire feeders, and T a distant transformer feeding into the low pressure three wire distributing network N. The two ends of the primary, and one end of the secondary of R, are connected to the bus bars as shown. The other end of the secondary, as well as a number of intermediate points, are joined up to a multiple way switch S, to which one of the feeder conductors is attached, the other feeder main being connected to the opposite bus bar. As will be evident from the figure, by manipulating S extra volts may be added to the bus bar pressure at will, and the drop along F compensated for. R is a step transformer, the total secondary difference of pressure being comparatively small. The above device possesses rather serious drawbacks, in that the switch S has to carry the main current, and that the supply would be stopped if the switch got out of order. Kapp improved on the arrangement by putting the switch in the primary circuit.

Figs. 2,032 and 2,033.—Gramme ring dynamo and alternator armatures illustrating converter operation. The current generated by the dynamo is assumed to be 100 amperes. Now, suppose, an armature similar to fig. 2,032 to be revolving in a similar field, but let its windings be connected at two diametrically opposite points to two slip rings on the axis, as in fig. 2,032. If driven by power, it will generate an alternating current. As the maximum voltage between the points that are connected to the slip rings will be 100 volts, and the virtual volts (as measured by a voltmeter) between the rings will be 70.7 (= 100 ÷ √2), if the power applied in turning this armature is to be 10 kilowatts, and if the circuit be non-inductive, the output in virtual amperes will be 10,000 ÷ 70.7 = 141.4. If the resistances of each of the armatures be negligibly small, and if there be no frictional or other losses, the power given out by the armature which serves as motor will just suffice to drive the armature which serves as generator. If both armatures be mounted on the same shaft and placed in equal fields, the combination is a motor dynamo. In actual machines the various losses are met by an increase of current to the motor. Since the armatures are identical, and as the similarly placed windings are passed through identical magnetic fields, one winding with proper connections to the slip rings and commutator will do for both. In this case only one field is needed; such a machine is called a converter.

Fig. 2,034.—Diagram of ring wound single phase rotary converter. It is a combination of a synchronous motor and a dynamo. The winding is connected to the commutators in the usual way, and divided into two halves by leads connecting segments 180° apart to collector rings. A bipolar field is shown for simplicity; in practice the field is multipolar and energized by direct current.

Fig. 2,035.—Diagram of two phase rotary converter. This is identical with the single phase machine with the exception that another pair of collector rings are added, and connected to points on the winding at right angles to the first, giving four brushes on the alternating side for the two phase current. The pressure will be the same for each phase as in the single phase rotary. Neglecting losses the current for each phase will be equal to the direct current × 1 ÷ / √2 = direct current × .707.

Fig. 2,036.—Diagram of three phase rotary converter. In this type, the winding is tapped at three points 120° distant from each other, and leads connected with the corresponding commutator segments.

Figs. 2,037 to 2,041.—Various rotary converter and transformer connections. Fig. 2,037 two phase connections; fig. 2,038 three phase delta connections; fig. 2,039 three phase Y or star connections; fig. 2,040 six phase delta connections; fig. 2,041 six phase Y connections.

Fig. 2,042.—Westinghouse rotary converter armature coils. These are wound from bar copper and are interchangeable. The armature coils are heavily insulated to withstand the tests specified in the standardization rules of the American Institute of Electrical Engineers.

Fig. 2,043.—Westinghouse rotary converter armature spider. It is made of cast iron or cast steel. The dovetail grooves are machined in the feet or ends of the arms and in these slots the laminations forming the armature coil engage.

Fig. 2,044.—Equalizer connections of Westinghouse rotary converter. The armature coils are cross connected at points of equal voltage and taps are led out from the winding at suitable points to the slip rings. This construction insures a uniform armature saturation below each pole piece and eliminates one cause of sparking at the commutator.

Figs. 2,045 and 2,046.—Westinghouse pole construction for converters. Fig. 2,045, pole without windings; fig 2,046, pole with windings. Poles are built up of steel laminations held together with rivets. Projections on the inner ends of the poles form seats for the field coils and hold them in position. Copper dampers set in slots in the pole faces insure stable operation. Rotary converters for railway service are almost invariably compound wound. The series windings are formed of bare copper strap. The shunt windings are of insulated copper strap or wire. Spaces between coil turns and sections are provided for ventilation.

Fig. 2,047.—Westinghouse rotary converter brush rigging showing method of bracing the brushes. The brushes are supported by a rigid cast iron rocker ring which fits accurately in the frame. A handwheel worm and screw arrangement for shifting the brushes is provided. Cast iron arms bolted to, but insulated from the rings, carry the rods on which the brush holders are mounted. Brush holders are of brass cast in one piece, of the sliding type and have braided copper shunts. Brush tension is adjustable.

Fig. 2,048.—Westinghouse commutating pole rotary converter. The construction details are substantially the same as for the railway converter, with exception of the commutating poles. The application of commutating pole converters is particularly desirable where special requirements such as great overload capacity or large capacity and low voltage enable them to show to the greatest advantage. Commutating poles as applied to rotary converters fulfill the same functions as in the more familiar applications to dynamos and motors. That is, the commutating pole insures sparkless commutation from no load to heavy overloads with a fixed brush position. Brush shifting devices are not furnished on commutating pole converters. Commutating pole rotary converters for railway service are normally arranged for automatic compounding which is effected by the proper combination of series excitation and inductance between the generator and the rotary converter. This inductance is normally included in the transformer but in special cases may be partly in a transformer and partly in a separate reactance. It is possible to produce by this means a slight increase in the direct current voltage provided the voltage drop in the alternating current line be not excessive. Usually it is so arranged that the compounding that can be obtained is just sufficient to overcome the alternating current line voltage drop. The standard Westinghouse method of starting is alternating current self-starting. With this method of self-starting, the brushes of a commutating pole rotary converter must be lifted from the commutator during the starting operation to prevent sparking. A mechanical device, as shown in fig. 2,050, is provided which accomplishes this. With direct current or motor starting a brush lifting device is not necessary.

Fig. 2,049.—Commutating pole of Westinghouse commutating pole rotary converter. The commutating poles are similar in general construction to the main poles. The coils are of bare copper strap wound on edge. Ventilating spaces are provided between the pole and coil and between turns. The copper winding is bare except for a few turns at each end. Insulating bolts retain the turns in their proper position.

Fig. 2,050.—Westinghouse brush lifting device for commutating pole rotary converter. A rack is attached to each brush as shown. Into this rack the spring hinged lifting hook of the raising device engages only when the lifting lever is shifted toward the raised position. The lifting arrangement is independent of the brushes during normal running, so it can in no way affect the operation of the machine. Each brush is merely raised and lowered within its own holder so the brush position or commutation is not altered.

Figs. 2,051 to 2,053.—Woodbridge split pole rotary converter. Each pole is split into three sections and provided with windings as indicated in fig. 2,051. When excited as in fig. 2,052, the commutator voltage is at its highest value; when excited as in fig. 2,053, the commutator voltage is low. The change in commutator voltage for constant collector ring voltage is in virtue of the property of rotary converters that the ratio of these two voltages is a function of the width of the pole arc.

Fig. 2,054.—General Electric regulating pole rotary converter. The field structure is divided into two parts, a main pole and a regulating pole. The ratio between the voltages on the direct current and alternating current sides may be readily varied by varying the excitation of the regulating poles, the only auxiliary apparatus required being a field rheostat for controlling the exciting current. Where automatic regulation is required, machines may be provided with compound windings, or automatic field regulators may be used responsive to either voltage or current. These converters are adapted for a variety of purposes where a variable conversion ratio is required, either to maintain constant D. C. voltage with varying A. C. voltage or to vary the D. C. voltage as required. Converters may be operated inverted where it is required to furnish constant or variable A. C. voltage from a D. C. source. Where converter and inverted converter operation are desired, an opposite direction of rotation is required for the inverted operation. Converters of this type are built in capacities from 300 kw. up to 3,000 kw., and constructed to give a voltage range between 240 and 300 volts, to cover the usual lighting circuit requirements. In design, they are similar to standard rotary converters, with the exception that the regulating poles are located next to the main pole pieces and a slightly different form of pole piece bridge is used for the main poles, in order to allow the auxiliary poles to be readily removed or assembled.

Fig. 2,055.—Detail of Westinghouse commutating pole rotary converter brush, showing rack. The brush lifting mechanism and its operation is explained in fig. 2,050.

Fig. 2,056.—Diagram of field of regulating pole converter illustrating principles explained in the accompanying text.

Fig. 2,057.—Voltage diagram for regulating pole converter illustrating principles explained in the accompanying text.

Figs. 2,058 and 2,059.—Diagrams illustrating the effect on the alternating current voltage due to varying the regulating field strength (of a machine proportioned according to fig. 2,060), from a density equal to that in the main poles to the same density reversed, the main field strength remaining constant. The D. C. voltage in this case varies from 30 per cent. above that produced by the main field alone to 30 per cent. below, or from 325 to 175 volts, while the A. C. voltage varies only from 200 to 175 volts. To keep the A. C. voltage constant with such a machine the main field must be strengthened as the regulating field is weakened or reversed to reduce the D. C. voltage. This strengthening increases the core loss particularly on low direct current voltages, which however, are rarely required, hence a machine proportioned as in fig. 2,060, would not be operated through so wide a range as 175 to 325 volts. Assume that the range is 240 to 300 volts, and that at the highest voltage, both main and regulating fields have the same density, presenting to the armature practically one continuous pole face of uniform flux intensity. The diagram of A. C. component voltages to give constant A. C. resultant voltage across the rings for the case, is shown in fig. 2,059. At 300 volts D. C., the main field produces an A. C. voltage OA, and the regulating field, a voltage AB, with a resultant OB, equal to about 200 volts A. C. At 270 volts D. C., the main field produces an A. C. voltage OA, and a regulating field voltage AB, giving a resultant A. C. voltage OB, equal to 200 volts. Similarly, at 240 volts D. C., the main field produces an A. C. voltage OA, and the regulating field (now reversed) produces the reverse voltage AB, giving the resultant OB again equal to 200 volts. It will be noted that, theoretically the main field strength must be increased about 15 per cent. above its value at 300 volts D. C. in order to keep the D. C. voltage at 250 volts.

Fig. 2,060.—Diagram illustrating placement of regulating poles. In practice machines are not built as indicated diagrammatically in fig. 2,056, that is, with regulating poles spaced midway between the main poles, because a better construction is obtained by placing the regulating pole closer to the corresponding main pole, as shown above.

Fig. 2,061.—Westinghouse 300 kw., 1,500 volt, three phase, 25 cycle, commutating pole rotary converter. The illustration shows clearly the commutating, and main poles and the relative sizes, also arrangement of the terminal connections.

Fig. 2,062.—Mechanical oscillator and speed limit device of Westinghouse commutating pole rotary converter. It automatically prevents the armature of the converter remaining in one position and thus not allowing brushes to wear grooves in both commutator and collector rings. The oscillator is a self-contained device carried at one end of the shaft. The operating parts consist of a hardened steel ball and a steel plate with a circular ball race, backed by a spring. The machine is so installed with a slight inclination toward the end carrying the oscillator. As the armature revolves the ball is carried upward and owing to the convergence of the steel race and shaft face, the spring is compressed. The reaction of the spring forces the armature away from its natural position and allows the ball to drop back to the lowest point of the race.

Fig. 2,063.—Westinghouse 2,000 kw., 270 volt, direct current, 6 phase, 167 R.P.M., synchronous booster rotary converter, having a voltage range from 230 to 310 volts. It consists of a standard rotary converter in combination with a revolving armature alternator mounted on the same shaft with the rotary converter and having the same number of poles. By varying the field excitation of the alternator, the alternating current voltage impressed on the rotary converter can be increased or decreased as desired. The direct current voltage delivered by the converter is thereby varied accordingly. The principle of operation of the booster converter is therefore very simple and easily understood. It is simply a combination of two standard pieces of electrical apparatus, accordingly there are incorporated in it no details of construction essentially different from those encountered in standard rotary converters and alternators. The only novelty is in their combination. The frames may be supported either from the rotary converter frame, as in the small units, or from the bed plate, as in the larger ones. A synchronous booster converter can be built, if necessary, with a vertical shaft to satisfy special floor space and head room requirements.

Fig. 2,064.—Armature of Westinghouse synchronous booster converter. Heavy cast yokes form the frames. They are proportioned to rigidly support the laminated steel field poles. The poles are fastened to the frame with through bolts. A lifting hook is provided on all frames. The bed plates are in one piece for the smaller machines but two piece bed plates are used for the larger ones. The bearings are ring oiling and have babbitt wearing surfaces that are renewable.

Fig. 2,065.—Westinghouse field frames and bearings for synchronous booster rotary converter. The frames consist of heavy cast yokes. The poles are fastened to the frames with through bolts. The bed plate is in one piece for the small machines and in two pieces for the large ones.

Fig. 2,066.—General Electric motor generator set consisting of 2,300 volt synchronous motor and 550 volt dynamo.

Fig. 2,067.—General Electric motor generator set consisting of 230 volt induction motor and 125 volt dynamo.

Fig. 2,068.—General Electric generator set, as installed for the Cleveland Electric Illuminating Company, Cleveland, Ohio. It consists of 11,431 volt motor and 275 volt generator. Speed, 360 revolutions per minute.

Fig. 2,069.—General Electric frequency changer set, consisting of a 11,000 volt synchronous motor with direct connected exciter and a 2,300 volt alternator.

Fig. 2,070.—General Electric four unit frequency changer set consisting of a 11,000 volt synchronous motor, 13,200 volt alternator, 250 volt exciter, and 440 volt starting induction motor. Where parallel operation is required between synchronous motor driven frequency changers, a mechanical adjustment is necessary between the fields or armatures of the alternator and motor to obtain equal division of the load. The adjustment can be obtained by shifting the keyway, or by special cradle construction. In the latter method, one machine is bolted to a cradle fastened to the base. By taking out the bolts, the frame can be turned around through a small angle relative to the cradle and therefore to the armature frame of the other machine, when the bolts can be replaced.

Fig. 2,071.—Diagram of "Cascade" motor generator set or motor converter, as it is called in England where it is used extensively for electric railway work. In the diagram of motor armature winding, some of the connections are omitted for simplicity. The windings are Y connected, and as they are fed by wires joined to the slip rings at the right and center, the rest of the power passes to the converter windings back to rotor winding and out to the slip rings so that part of the power enters the rotor and part through the converter.

Fig. 2,072.—General Electric shunt wound booster set. Sets of this class are used in railway stations to raise the pressure of the feeders extending to distant points of the system, for storage battery charging and regulation, and in connection with the Edison three wire lighting system. The design of the various sets is closely dependent upon their application. Booster sets are constructed in either series or shunt wound types and they may be arranged for either automatic or hand regulation, depending on the nature of the service required. Where there are a number of lighting feeders connected and run at full load for only a short time each day it will generally be economical to install boosters rather than to invest in additional feeder copper. It is important, however, to consider each case where the question of installing a booster arises, as a separate problem, and to determine if the value of the power lost represents an amount lower than the interest charge on the extra copper necessary to deliver the same voltage without the use of a booster.

Fig. 2,073.—General Electric 9.5 kw. 1,000 R.P.M. balancer set. This is a form of direct current compensator, the units of which it is composed may be alternately motor or generator, and the secondary circuit is interconnected with the primary. Balancer sets are widely used to provide the neutral of Edison three wire lighting systems. They are also installed for power service in connection with the use of 250 volt motors on a 500 volt service or 125 volt motors on a 250 volt service. Although the balancer set may be manufactured for more than one intermediate voltage, the General Electric Company recommends the three wire system as it is simpler. The motors may be more equally divided and the reserve margin of the motor is more conservative for that class of work requiring increase in torque with decrease in speed. The voltage of the system being given, the capacity of a three wire balancer set is fixed by the maximum current the neutral wire is required to carry. This figure is a more definite specification of capacity than a statement in per cent. of unbalanced load. As designed for power work and generally for lighting service, the brushes of each machine are set at the neutral point in order to get the best results for operating alternately either as a generator or motor. Where the changes of balance are so gradual as to permit of hand adjustment if desired, a considerable increase in output is obtainable.

Figs. 2,074 and 2,075.—General Electric charging sets. Fig. 2,074 set consists of dynamo and direct current motor; fig. 2,075 set consists of a dynamo and alternating current motor. Fig. 2,074 set is equipped with 125 volt, shunt wound dynamo and 230 or 550 volt motor and range in capacity from .125 kw. to 13 kw., the speed varying from 2,250 R.P.M. in the .125 kw. set to 925 R.P.M. in the 13 kw. set with 230 volt motor. Both motor and dynamo have the same type and size of frame; these are bolted together and form a compact and symmetrical outfit, no base being necessary. Sets of the type shown in fig. 2,075 range in capacity from .2 kw. to 10 kw. and are equipped with 125 volt shunt wound dynamo and 110, 220, 440 or 550 volt two or three phase motors. If desired they can be furnished with single phase motors wound for 110 or 220 volts. The speed of this type is 1,800 R.P.M. When a motor generator set is used to charge only one battery, the insertion of a resistance between the charging dynamo and the battery is not necessary, inasmuch as all adjustments of voltage can be made by varying the field strength of the dynamo, and, therefore, there are no large losses due to resistance since the loss in the dynamo field rheostat is very small. When a motor generator set is used to charge two or more batteries of different capacities, or voltages, or which are in different conditions of charge, it is necessary to insert a resistance in series with each battery, in order that the current may be properly adjusted for each particular battery.

Figs. 2,076 to 2,078.—Diagrams showing alternating currents, and partial and complete rectification.

Fig. 2,079.—Mechanical rectifier. The rectifier consists of two castings M and S with teeth which fit together as shown, being insulated so they do not come in contact with each other. Every alternate tooth, being of the same casting, is connected together, the same as though joined by a conducting wire. There are as many teeth as there are poles. The part M of the rectifier is connected to one of the collector rings by F, and the part S to the other ring by G.

Figs. 2,080 and 2,081.—Two views of Nodon valve. This is an electrolytic rectifier in which the cathode is a rod of aluminum alloy held centrally in a leaden vessel which forms the anode and contains the electrolyte, a concentrated solution of ammonium phosphate. Only a short portion at the lower end of the cathode is utilized, the rest, which is rather smaller in diameter, being protected from action by an enclosing glass sleeve. The current density at the cathode ranges from 5 to 10 amp. per sq. dm. In the larger sizes, the cells are made double, and a current of air is kept circulating between the walls by means of a motor driven fan. In order to utilize both halves of the supply wave, the Gratz method of connection is adopted. The maximum efficiency is obtained at about 140 volts, and the efficiency lies between 65 and 75 per cent., and is practically independent of the frequency between the limits of 25 ~ and 200 ~. Above a pressure of 140 volts, the efficiency falls off very rapidly, owing to breakdown of the film. The pressure difference is high, being over 90 per cent. at full load. Temperature largely influences the action of the valve, and should never exceed 122° Fahr.

Fig. 2,082.—Oscillograph record from Nodon valve showing original supply voltage and the corresponding pulsating current at the terminals of such a valve.

Fig. 2,083.—Performance curves of five ampere Nodon valve. Constant secondary voltage test. Loaded on non-inductive resistances. Frequency 50. Maximum power factor on valve .7.

Fig. 2,084.—Mohawk electrolytic rectifier and switchboard; diagram showing connections for charging storage battery. Operating instructions: After assembling battery as in fig. 2,085, the film must be formed on the aluminum alloy electrodes so that the rectifier will pass current only in the right direction. Open switch B, close switch T to the right; discharge lever can be in any position; charging regulator lever must be to the extreme left, the zero position; now close main switch M. Moving regulator lever R from the zero position to the first button or contact, let it remain there for a time, not less than five minutes; this is important, as the proper rectification of the current depends on the film formed on the aluminum rods. The ammeter after the first rush of current may not show any current as passing, or it may show a reverse current. In the latter case, leave the contact finger on the first button until the needle comes back to zero. This may take some time, but the needle will eventually come back; it also indicates that the film is properly formed when the needle returns to zero. Move regulator R to the extreme right step by step and note that the ammeter continues to return to zero, which indicates that the film on rectifier electrodes is formed properly. Move regulator R to zero, close switch T to the left in normal charging position. Close charging switch B. To regulate the flow of current through the battery move charging lever R to the right slowly until ammeter indicates the correct charging current. After the batteries are charged and ready for use, discharge lever can be moved to connect either set of storage batteries to the load terminal. The voltage of the batteries can be read at any time, by pressing the strap key. The discharge lever connects the batteries to the volt meter and it is possible by moving it to measure the voltage of either set of battery, charging or discharging. Trouble in the rectifier demonstrates itself by the solution becoming heated. The condition of the rectifier can be tested any time in a few seconds by opening switch B and closing switch T to the right. If the rectifier be in proper condition the ammeter will read zero. And if it be not rectifying and permitting A. C. current to flow through the rectifier, the ammeter will read negative or to the left of the zero. An old solution that is heating and not rectifying properly will turn a reddish brown color.

Figs. 2,085 and 2,086.—Mohawk electrolytic rectifier. To put in commission, clean out the jar. Fill with distilled or rain water. Add six pounds of electro salts, stir and after all salts are dissolved place the cover in position. The specific gravity of the solution should be 1.125. The middle iron electrode must hang straight down in the solution and not touch either of the other aluminum alloy electrodes. The aluminum alloy electrodes are mounted on an insulated bracket that slides up and down on a ¼" rod. This rod screws in the hole taped in the middle of the cover. The electrodes give the best results only when perfectly smooth. Should they get rough, covered with a deposit or a white coating remove from the solution, and clean with fine sand paper. Finish with fine sand paper. Form the film again and the electrodes will be as good as new. Clean iron electrode occasionally.

Fig. 2,087.—The Fleming oscillation valve. It depends for its action on the well-known Edison effect in glow lamps. The valve consists of a carbon filament glow lamp with a simple central horseshoe filament. Around this filament inside the exhausted bulb is fixed a small cylinder of nickel, which is connected by means of a platinum wire sealed through the bulb to a third terminal. The valve is used as follows: The carbon loop is made incandescent by a suitable battery. The circuits in which the oscillations are to be detected is joined in series with a sensitive mirror galvanometer, the nickel cylinder terminal and the negative terminal of the filament of the valve being used. The galvanometer will then be traversed by a series of rapid discharges all in the same direction, those in the opposite direction being entirely suppressed.

Fig. 2,088.—The Churcher valve. This is of the modified Nodon type. It differs from the latter in that it has two cathodes of aluminum and an anode of lead or platinum, suspended in the one cell. This permits the complete utilization of both halves of the supply wave with one cell instead of the four required in the Gratz method. The connections of such a cell are shown in the figure. The secondary of the transformer carries a central tapping, and is connected through the direct current load to the central anode, while each of the cathodes is connected to the ordinary terminals of the transformer itself. The practical limits of the cell are 50 volts direct current, or 130 volts at the transformer terminals AB. F, is the anode; C, cathode I; D, cathode II.

Figs. 2,089 and 2,090.—The De Faria valve. This is an aluminum lead rectifier. The cathode is a hollow cylinder of aluminum placed concentrically in a larger cylinder of lead, and the whole immersed in electrolyte of sodium phosphate in an ebonite containing vessel. Cooling is effected by promoting automatic circulation of the electrolyte by providing the lead cylinder with holes near its extremities; the heated electrolyte then rises in the lead cylinder, passes out at the upper holes, is cooled by contact with the walls of the containing vessel, and descends outside the lead cylinder. It is claimed that this cooling action is sufficient to allow of a current density of 8 amp. per sq. dm. of aluminum.

Fig. 2,091.—75 light Westinghouse-Cooper Hewitt mercury vapor rectifier constant current regulating transformer. View showing assembly in case.

Fig. 2,092.—75 light Westinghouse-Cooper Hewitt mercury vapor rectifier constant current regulating transformer with case removed. The transformer is of the repulsion coil type, oil cooled and oil insulated. It is so arranged as to give a constant secondary current and to insulate the arc lines from the primary circuit. The regulating transformer contains two stationary secondary coils and two moving primary coils balanced against each other. Each secondary coil of the 75 light regulator is wound in two parts, owing to the use of two rectifier bulbs in series in outfits of this capacity. The repulsion between the primary and secondary coils changes the distance between them according to the variation of load, and the induced current in the secondary is thus kept constant. An increase in current causes the primary and secondary coils to separate, and a decrease in current permits them to approach each other, until the normal balance is restored. The moving coils are hung from sheave wheels having roller bearings and are balanced so that they are sensitive to the slightest impulse tending to separate them or draw them closer together. (See figs. 1,981-2, and 2,111.) The windings are insulated for a voltage considerably in excess of that existing in normal service. Several taps are provided to take care of different voltages and wave forms. A combination of taps will be found which will be suitable for any wave form coming within the American Institute of Electrical Engineer's limits for a sine wave. The secondary coils are also provided with taps for 85 per cent. of normal load, so that less than normal load can be taken care of at a good power factor. Any part of the full load can be carried temporarily with the full load connections of the transformer, but at permanent light loads the power factor and efficiency will be improved by using the 85 per cent. connections. Standard regulating transformers are wound for 6.6 and 4 amperes, and for primary circuits of 220, 440, 1,100 2,200, 6,600 and 13,200 volts. Regulators can be specially wound for 5.5 amperes. For three phase circuits three regulators can be used, one on each phase, or they can be furnished in pairs with an auxiliary auto-transformer to give a balanced load. The regulators can be connected, in cases where the unbalancing is not objectionable, to separate phases.

Fig. 2,093.—Diagram of connections of Westinghouse-Cooper Hewitt mercury vapor rectifier arc light circuit.

Fig. 2,094.—Cooper Hewitt mercury vapor rectifier. The mercury vapor rectifier as developed by Peter Cooper Hewitt for changing alternating current into direct current is the result of a series of careful experiments and investigations of the action going on in his mercury vapor lamp for electric lighting used on direct current circuits only. While many attempts have been made to produce an alternating current lamp, up to the present time, they have been unsuccessful. The difficulty of operating a lamp on the alternating current circuit lies in the fact that while a current will flow freely through it in one direction, when the current reverses the negative electrode or cathode acts as an electric valve and stops the current, thus breaking the circuit and putting out the light. By following up this new electrical action, Hewitt applied the principle in the construction of a vacuum tube with suitable electrodes, and by using two electrodes of iron or graphite for the positive or incoming current and one of mercury for the negative or where the current leaves the tube, the circuits could be arranged so that a direct current would flow from the mercury electrode and be used for charging storage batteries, electro-chemical work or operating direct current flame arc lamps. As shown in the figure, the rectifier consists essentially of a glass bulb into which are sealed two iron or graphite anodes and one mercury cathode, and a small starting electrode. The bulb is filled with mercury vapor under low pressure. The action of this device depends on the property of ionized mercury vapor of conducting electricity in one direction only. In operation no current will flow until the starting or negative electrode resistance has been overcome by the ionization of the vapor in its neighborhood. To accomplish this, the voltage is raised sufficiently to cause the current to jump the gap between the mercury cathode and the starting cathode, or by bringing the cathode and starting electrode together in the vapor by tilting and then separating them, thus drawing out the arc. When this has been done, current will only flow from the anode to the mercury cathode, and not in the reverse direction. In order to maintain the action, a lag is produced in each half wave by the use of a reactive or sustaining coil; hence the current never reaches its zero value, otherwise the arc would have to be restarted. There are two kinds of losses in the tube: 1, arcing, or leakage from one anode to the other, and 2, the mercury arc voltage drop. This drop does not depend on the load, the energy represented by the drop being converted into heat, which is dissipated at the surface of the containing vessel. According to Steinmetz, the limit of voltage must be very high, as 36,000 volts has been rectified. The current output is limited principally by the leading-in wires to the electrodes, it being a difficult problem to seal into the glass container the large masses of metal required for the conduction of large currents. Frequency has but little influence. The direct current voltage ranges from 20 to 50 per cent. that of the arc supply. The life of the valve depends somewhat upon its size, being longer in the small sizes and never, with fair usage, less than 1,000 hours.

Fig. 2,095.—Three phase mercury arc rectifier. The rectifier bulb is provided with three positive electrodes or anodes, a negative electrode or cathode, and a starting anode, as shown. The three phase leads are connected to the anodes at the top of the bulb, a branch from one phase being brought down to the starting anode, a resistance being placed in the circuit to prevent excessive current on account of the proximity of the two lower electrodes. Since there is always a pressure on one of the three anodes in the right direction, a reactance coil is not necessary. The apparatus is started in the usual way by tilting.

Figs. 2,096 to 2,098.—Westinghouse diagrams showing comparative efficiencies of different systems of series arc lighting.

Fig. 2,099.—Westinghouse-Cooper Hewitt mercury vapor rectifier bulb. It consists essentially of a hermetically sealed glass bulb filled with highly attenuated vapor of mercury, and provided with electrodes. Its operation is fully explained in the accompanying text.

Fig. 2,100.—Westinghouse-Cooper Hewitt mercury vapor rectifier bulb and box. The life of the bulb is materially increased by operating it at certain temperatures and for this reason the bulbs in the arc light rectifier outfits are immersed in oil and mounted in the same tank with the regulating transformer. Two bulbs in series are used with the 75 light outfit. The bulb is mounted on tilting trunnions in a box which can be lifted out through a door in the top of the tank without disconnecting any leads. The containing box has contacts on the bottom so arranged that when it is lowered into place, the bulb is automatically connected in circuit. To replace a defective bulb it is only necessary to lift out the containing box by its handle through the door in the top of the case and mount a new bulb in it, after which the box can be lowered into place. It is desirable to have at each installation, a spare bulb box in which a bulb can be kept, connected ready for use. If this be done, it is only necessary, in case of trouble with the bulb, to withdraw the old bulb and box and replace them with the spare set. This avoids having the lamps out of service.

Fig. 2,101.—General Electric 50 light double tube combined unit series mercury arc rectifier outfit; front view. This unit consists of a constant current transformer, reactance, tube tank and exciting transformer mounted on a common base; also a static discharger and pilot lamp mounted on top of the transformer. This arrangement makes the rectifier outfit, with the exception of the switchboard panel, complete in itself.

Fig. 2,102.—General Electric 2,200 volts, 60 cycle, primary, 6.6 ampere, secondary, 75 light, double tube mercury arc rectifier outfit with automatic shaking device, the case being removed to show parts. The constant current transformer is air cooled. The winding which supplies energy for the exciter transformer is located at the top of, and around the core of the constant current transformer. The exciting transformer is mounted on the base of the constant current transformer inside of the casing. It supplies low pressure currents to the starting anodes of the rectifier tube. This current establishes an auxiliary arc when the tube is shaken, which is necessary in order to start the rectifier. The exciting transformer is wound for 110 volts and it consumes about 200 volt-amperes. The direct current reactance is mounted on the base of the transformer and enclosed in the same casing. It is connected in series with the lamp load and its function is to reduce the pulsations of the circuit to a value most satisfactory for operation. The tube tank for holding the oil is mounted on the same base as the transformer. It is provided with a cooling coil; a tube carrier is provided for raising or lowering the tube in the tank. A thermometer is provided to gauge the temperature of the oil in the tank. The static dischargers consisting of horn gaps in series with resistance, are connected between the anodes and the cathode in order to protect the tubes and other apparatus from excessive electrical strains. The horn gaps open the circuit after discharge, and in case the resistance becomes damaged the discharge passes across the spark gap provided, thereby shunting the resistance.

Figs. 2,103 to 2,106.—Diagram of current waves and impressed pressure of Westinghouse-Cooper Hewitt mercury vapor rectifier. The whole of the alternating current wave on both sides of the zero line is used. The two upper curves in the diagram show the current waves in each of the two positive electrodes, and the resultant curve III represents the rectified current flowing from the negative electrode. Curve IV shows the impressed alternating current pressure. It is evident that if the part of the wave below the zero line were reversed, the resulting current would be a pulsating direct current with each pulsation varying from zero to a positive maximum. Such a current could not be maintained by the rectifier, because as soon as the zero value was reached the negative electrode resistance of the rectifier would be re-established and the circuit would be broken. To avoid this condition, reactance is introduced into the circuit, which causes an elongation of current waves so that they overlap before reaching the zero value. The overlapping of the rectifier current waves reduces the amplitude of the pulsations and produces a comparatively smooth direct current as shown in curve III. In this way the whole of the alternating current is transformed to direct current because each of the alternations in both directions is alternately rectified.

Fig. 2,107.—General Electrical rectifier tube as used on series mercury arc rectifier. The tube rectifies the alternating current into direct current. It consists of an exhausted glass vessel containing one anode or positive terminal in each of the two upper arms, two mercury starting anodes and a cathode or negative terminal of mercury at the bottom of the tube. It is submerged in oil and supported in a removable carrier. The tube is put into operation by slightly shaking it. In the combined unit set, this shaking is accomplished by an electromagnet mounted above the tube tank and operated from a pull button switch on the panel. An automatic shaker is sometimes installed which will automatically start the tube when the set is started, or if its operation should become interrupted while in service. The energy for the operation of this magnet (110 volts alternating current) is obtained from the small auxiliary winding on the main transformer which also supplies energy to the exciting transformer. The oil in which the tube is placed is cooled by a circulation of water through the cooling coils on the inside of the tube tank. The amount of water necessary for cooling the rectifier tubes varies according to local conditions, depending upon the temperature of the water and that of the air in the station but under the most favorable conditions no water is required. As rectifiers are commonly installed in steam driven stations, the drip from the tube tanks is usually piped to the boiler supply thereby eliminating any loss for cooling water.

Fig. 2,108.—Elementary diagram of mercury arc rectifier connections. A.A., graphite anodes; B, mercury cathode; C, small starting electrode; D, battery connection; E and F, reactance coils; G and H, transformer terminals; J, battery.

Fig. 2,109.—Diagram showing connections of General Electric combined unit mercury arc single tube rectifier outfit with remote controlled non-automatic shaking device.

Fig. 2,110.—Diagram showing connections of General Electric combined unit mercury arc double tube rectifier outfit with remote controlled non-automatic shaking device.

Fig. 2,111.—Diagram showing connections of General Electric series mercury arc rectifier.

Figs. 2,112 and 2,113.—Diagram of current waves showing effect of reactance coil. If the alternating current wave could be rectified without the use of the reactance coil, the direct current produced would consist of a series of impulses which would rise and fall from the zero line as illustrated in fig. 2,112. The action of the reactance coil not only maintains the current through the tube while the supply current is passing through zero, but helps to smooth out the pulsations of the direct current which is passing out of the cathode terminal of the tube to the batteries, or other direct current apparatus put in its circuit. The smoothing out effect of the reactance is shown in fig. 2,113. It will be seen from the diagram that the current does not drop down to zero and the pulsations of the direct current are greatly reduced. The waves A, A, etc., are from the positive waves of the alternating current supply, while B, B, are from the negative waves, and together they form the rectified current, flowing in the same direction to the external direct current circuit shown at B in the diagram, fig. 2,108.

Figs. 2,114 to 2,116.—General Electric mercury arc rectifier outfit, or charging set. The cut shows front, rear, and side views of the rectifier, illustrating the arrangement on a panel, of the rectifier tube with its connection and operating devices.

Figs. 2,117 and 2,118.—General Electric mercury arc bulbs or tubes for 200 and 10,000 volt circuits.

Fig. 2,119.—General Electric series mercury arc rectifier outfit; view showing method of replacing a tube. The illustration also shows tube carrier and drip tray.

Fig. 2,120.—General Electric 100 volt mercury arc rectifier tube. A,A, anodes; B, cathode; C, starting anode; D, tube or bulb.

Fig. 2,121.—Diagram showing essential features of Premier Ampero electromagnetic rectifier. Details of construction and principles of operation are given in the accompanying text.

Fig. 2,122.—Diagram of General Electric (Batten) electromagnetic rectifier. It is desirable for light and occasional service, where direct current is required but only an alternating supply is available, being used for charging storage batteries, exciting spark coils, performing electrolytic work, etc. The rectifier consists of a step down static transformer T, by means of which the circuit pressure is reduced to about 50 volts; also, a polarized relay R, the contact tongue C of which moves to one side or the other in sympathy with the alternations of the current in the primary winding P, the secondary current induced in the winding S being thereby rendered direct in the outer circuit. T', T' are the main terminals which are connected to the alternating current supply through the wires W. Lamps inserted at L are used as resistances in the primary circuit, the reduction of the voltage already alluded to being effected by this means. In charging storage batteries where a low pressure is required, a lamp (or lamps) should be connected in the secondary circuit as shown, S B being the storage battery, and L' L' the lamp resistances in series therewith, the battery has one end of the secondary S connected to its middle. Thus the alternating current leaving the transformer by the wire 1, passes by flexible connection 2, to the vibrating contact tongue C of the relay, the latter causing the currents in either direction to flow through the two halves H, H' of the battery, whence the current re-enters the secondary of the transformer by the wire 3. The soft iron core of the relay is in two halves S' S' and the armature A, carrying C, vibrates between their polar extremities. M, M' are two permanent magnets with their like poles together at the center C' where A is pivoted. Supposing these poles are north as indicated, the extremities of A will be south. The south ends of M, M being in juxtaposition with the centers of the soft iron cores S', S' will render their extremities facing the ends of A of north polarity. The windings on S', S' are connected in series with each other, and in shunt with P across the main terminals T', T'. Then because of the polarization of A and S', S', the former will vibrate rapidly in sympathy with the alternations of the current. K is a condenser shunted by a lamp resistance L", this being found to improve the working of R.