Figs. 2,826 to 2,829.—General Electric diagrams of connections. A, ammeter; C.B, circuit breaker; C.P, candle power; C.T, current transformer; D.R, discharge resistance; F, fuse; F.S, field switch; L, lamp; O.C, overload coil; P.P, pressure plug; P.R, pressure receptacle; R.C, reactance; rheo, rheostat; R.P, synchronizing plug, running; R.S, resistance; S, switch; S.I, synchronous indicator: S.P, synchronizing plug, starting; S.R, synchronizing receptacle: V, voltmeter.

Fig. 2,830.—Diagram of switchboard connections for General Electric automatic voltage regulator with two exciters and two alternators.

Figs. 2,831 and 2,832.—Garton-Daniels alternating current lightning arrester; diagram showing connections. A lightning discharge takes the path indicated by the dotted line, across the upper air gap A, through resistance rod B, C, D, across copper strip R on the base, thence flowing to ground through the movable plunger M, lower on gap N, and ground binding post L. The discharge path is practically straight, contains an air gap, distance of but 3/32 inch, a series resistance averaging but 225 ohms. The lightning discharge does not flow through the flexible lead connecting band D on the lower end of the resistance rod with the top of the movable plunger. These two points are electrically connected by the heavy copper strip R, and lightning discharges generally, if not always, take the path across this copper strip in preference to flowing through the inductance of the one turn of flexible cable. When a discharge occurs from line to ground through any lightning arrester, the air gaps arc over, and so there is offered a path from line to ground for the line current. This flow of line current following the lightning discharge to ground may vary anywhere from a small capacity current where the arrester is installed on an ungrounded circuit, a moderately heavy flow on a partially grounded circuit, to a very heavy flow on a grounded circuit—either a circuit operated as a dead grounded circuit, or a circuit which has become accidentally grounded during a storm. The path taken by this flow of line current from line to ground may be traced by following the path shown by the dashed line. It, as seen, crosses upper air gap A, flows through section B of the resistance rod to band C. Leaving band C it flows through the magnet winding H, thence to band D on the resistance rod, through flexible lead to upper end of movable plunger, through movable plunger, across lower air gap N, to ground binding post L, thence to ground. The function of the short length of resistance rod CD is as follows: It has an ohmic resistance of about 30 ohms but is non-inductive. Magnet winding H, connected to bands C and D on the ends of this short length of rod has an ohmic resistance of 3 ohms, but is highly inductive. Lightning discharges being of high frequency take the higher resistance but non-inductive path CD in their passage from line to ground. The flow of normal current from line to ground being of a very low frequency, 25 or 60 cycles in ordinary alternating current circuits, zero in direct current circuits—takes the low resistance path through coil H in its path to ground. Section CD of the rod is used therefore simply to shunt the inductance of winding H to high frequency lightning discharges, leaving the lightning discharge path in the arrester a non-inductive highly efficient path. In all Garton-Daniels A. C. lightning arresters operating on non-grounded or partially grounded circuits, the action of the air gaps and series resistance are together sufficient to extinguish the flow of normal current to ground at the zero point of the generator voltage wave. If, however, as frequently happens, the line grounds accidentally during a storm, then the arrester does not have to depend for its proper operation on the arc extinguishing properties of the air gaps and resistance, but the heavier flow of line current through the arrester energizes the movable plunger, which raises upward in the coil, opening the circuit between the discharge point M and the lower end of the plunger. To limit the flow of line current to ground the resistance rod B is provided, there being approximately 225 ohms between the discharge point A and clamp C in the 2,500 volt arrester. This feature is particularly effective where the circuit is temporarily or accidentally grounded. The series resistance prevents a heavy short circuit through the arrester and limits the current to a value that is readily broken by the cut out and is not enough to impede the passage of the discharge.

Fig. 2,833.—Diagram of switchboard connections for General Electric automatic voltage regulator with three exciters and three alternators.

Figs. 2,834 and 2,835.—Front and rear views showing General Electric automatic voltage regulator mounted on switchboard panel.

Fig. 2,836.—Method of supporting the framework of a switchboard.

Fig. 2,837.—Diagram showing elementary connections of General Electric automatic regulator for direct current. It consists essentially of a main control magnet with two independent windings and a differentially wound relay magnet. One winding, known as the pressure winding, of the main control magnet is connected across the dynamo terminals, the other across a shunt in one of the load mains. The latter is the "compensating winding" and it opposes the action of the pressure winding so that as the load increases, a higher pressure at the dynamo is necessary to "over compound" for line drop. In ordinary practice, the voltage terminals are connected to the bus bars, and the compensating shunt inserted in one of the principal feeders of the system. In operation the shunt circuit across the dynamo field rheostat is first opened by means of a switch provided for that purpose on the base of the regulator and the rheostat turned to a point that will reduce the generator voltage 35 per cent below normal. The main control magnet is at once weakened and allows the spring to pull out the movable core until the main contacts are closed. This closes the second circuit of the differential relay, thus neutralizing its windings. The relay spring then lifts the armature and closes the relay contacts. The switch in the shunt circuit across the dynamo field rheostat is now closed, practically short circuiting the rheostat, and the dynamo voltage at once rises. As soon as it reaches the point for which the regulator has been adjusted, the main control magnet is strengthened, which causes the main contacts to open, which in turn open the relay contacts across the rheostat. The rheostat is now in the field circuit, the voltage at once falls off, the main contacts are closed, and relay armature released, and shunt circuit across the rheostat again completed. The voltage then starts to rise and this cycle of operation is continued at a high rate of vibration, maintaining not a constant but a steady voltage at the bus bars. When neither the compensating winding nor pressure wires are used, there will be no "over compounding" effect due to increase of load and a constant voltage will be maintained at the bus bars. The compensating winding on the control magnet, which opposes the pressure winding is connected across an adjustable shunt in the principal feeder circuit. As the load increases the voltage drop across the shunt increases and the effect of the compensating winding becomes greater. This will require a higher voltage on the pressure winding to open the main contacts and the regulator will therefore cause the dynamo to compensate for line drop, maintaining at the bus bars a steady voltage without fluctuations, which rises and falls with a load on the feeders, giving a constant voltage at the lamps or center of distribution. The compensating shunt may be adjusted so as to compensate for any desired line drop up to 15 per cent; it is preferably placed in the principal lighting feeder, but may be connected to the bus bars so that the total current will pass through it. The latter method, however, is sometimes desirable, as large fluctuating power loads on separate feeders might disturb the regulation of the lighting feeders. Adjustment is made by sliding the movable contact at the center of the shunt. This contact may be clamped at any desired point and determines the pressure across the compensating winding of the regulator's main control magnet. Where pressure wires are run back to the central station from the center of distribution they may be connected directly to the pressure winding of the main control magnet, and it is unnecessary to use the compensating shunt. The pressure wires take the place of the leads from the control magnet to the bus bars and maintain a constant voltage at the center of distribution.

Fig. 2,838.—Diagram showing connections of General Electric automatic voltage regulator for direct current as connected for maintaining balanced voltage on both sides of a three wire system using a balancer set. In operation, should the voltage on the upper bus bars become greater than that on the lower ones, the middle and upper contacts on the regulator will close, thus opening the relay contacts to the left and closing those to the right. This inserts all the resistance in the field of balancer A, and short circuits the resistance in the field of balancer B. A will then be running as a motor, and B as a dynamo, thereby equalizing the two voltages until that on the lower bus bars becomes greater than that of the upper ones; then the regulator contacts operate in the opposite direction and balancer A is run as a dynamo, and balancer B as a motor. This cycle of operation is repeated at the rate of from three to four hundred times per minute, thus maintaining a balanced voltage on the system.

Fig. 2,839.—Diagram of connections of General Electric voltage regulators for one or more alternators using one exciter.

Figs. 2,840 and 2,841.—General Electric equalizer regulator designed to equalize the load on two machines, and diagram of connections.

Fig. 2,842.—Connection of General Electric equalizing regulator for equalizing loads on an engine driven dynamo and rotary converter running in parallel. Should the load on the dynamo become greater than that on the rotary converter, the middle and upper contacts on the regulator close, and thus by means of the relay switch and control motor, cause the feeder regulator to boost the voltage on the rotary until the loads again become equal. Should the load on the rotary converter become greater than that on the generator, the regulator contacts operate in the reverse direction and the feeder regulator is caused to buck the rotary voltage.

Fig. 2,843.—Method of synchronizing with one lamp; dark lamp method. Assuming A to be in operation, B, may be brought up to approximately the proper speed, and voltage. Then if B, be run a little slower or faster than A, the synchronizing lamp will glow for one moment and be dark the next. At the instant when the pressures are equal and the machines in phase, the lamp will become dark, but when the phases are in quadrature, the lamp will glow at its maximum brilliancy. Since the flickering of the lamp is dependent upon the difference in frequency, the machines should not be thrown in parallel while this flickering exists. The nearer alternator approaches synchronism, in adjusting its speed, the slower the flickering, and when the flickering becomes very slow, the incoming machine may be thrown in the moment the lamp is dark by closing the switch. The machines are then in phase and tend to remain so, since if one slow down, the other will drive it as a motor.

Fig. 2,844.—Diagram of connections of General Electric automatic voltage regulator for several alternators running in parallel with exciters in parallel.

Fig. 2,845.—Method of synchronizing with two lamps; dark lamp method. The two synchronizing lamps are connected as shown, and each must be designed to supply its rated candle power at the normal voltage developed by the alternators. Now since the alternators are both running under normal field excitation the left hand terminals of each of them will alternately be positive and negative in polarity, while the right hand terminals are respectively negative and positive in polarity. If, however, the alternators be in phase with each other, the left hand terminals of both of them will be positive while the right hand terminals are negative, and when the left hand terminals of both machines are negative the right hand terminals will be positive. Hence, when the machines are in phase there will be no difference of pressure between the left hand terminals or between the right hand terminals of the two machines. Hence, if the synchronizing lamps be connected as shown, both will be dark. The instant there is a difference of phase, both lamps will glow attaining full candle power when the difference of phase has reached a maximum. As the alternators continue to come closer in step, the red glow will gradually fade away until the lamps become dark. Then the switch may be closed, thereby throwing the two machines in parallel. If the intervals between the successive lighting up of the lamps are of short duration it is advisable to wait until these become longer even though the other conditions are satisfied, because where the phases pass each other rapidly there is a greater possibility of not bringing them together at the proper instant. An interval of not less than five seconds should therefore be allowed between the successive lighting up of the lamps, before closing the switch.

Fig. 2,846.—Inductor type synchroscope. This type is especially applicable where pressure transformers are already installed for use with other meters. As it requires only about ten apparent watts it may be used on the same transformers with other meters. There are three stationary coils, N, M and C, and a moving system, comprising an iron armature, A, rigidly attached to a shaft suitably pivoted and mounted in bearings. A pointer is also attached to the shaft. The moving system is balanced and is not subjected to any restraining force, such as a spring or gravity control. The axes of the coils N and M are in the same vertical plane, but 90 degrees apart, while the axis of C is in a horizontal plane. The coils N and M are connected in "split phase" relation through an inductive resistance P and non-inductive resistance Q, and these two circuits are parallel across the bus bar terminals 3 and 4 of the synchroscope. Coil C is connected through a non-inductive resistance across the upper machine terminals 1 and 2 of the synchroscope. In operation, current in the coil C magnetizes the iron core carried by the shaft and the two projections, marked A and "iron armature." There is however, no tendency to rotate the shaft. If current be passed through one of the other coils, say M, a magnetic field will be produced parallel with its axis. This will act on the projections of the iron armature, causing it to turn so that the positive and negative projections assume their appropriate position in the field of the coil M. A reversal of the direction in both coils will obviously not affect the position of the armature, hence alternating current of the same frequency and phase in the coils C and M cause the same directional effect upon the armature as if direct current were passed through the coils. If current lagging 90 degrees behind that in the coils M and C be passed through the coil N, it will cause no rotative effect upon the armature, because the maximum value of the field which it produces will occur at the instant when the pole strength of the armature is zero. The two currents in the coils M and N produce a shifting magnetic field which rotates about the shaft as an axis. As all currents are assumed to be of the same frequency, the rate of rotation of this field is such that its direction corresponds with that of the armature projections at the instant when the poles induced in them by the current in the coil C are at maximum value, and the field shifts through 180 degrees in the same interval as is required for reversal of the poles. This is the essential feature of the instrument, namely, that the armature projections take a position in the rotating magnetic field which corresponds to the direction of the field at the instant when the projections are magnetized to their maximum strength by their current in the coil C. If the frequency of the currents in the coils which produce the shifting field be less than that in the coil which magnetized the armature, then the armature must turn in order that it may be parallel with the field when its poles are at maximum strength.

Fig. 2,847.—Brilliant lamp method of synchronizing. The synchronizing lamps are connected as shown, and must be of the alternator voltage. When the voltages are equal and the machines in phase, the difference of pressure between a and a given point is the same as that between a' and the same point; this obtains for b and b'. Accordingly, a lamp connected across a b' will burn with the same brilliancy as across a' b; the same holds for the other lamp. When the voltages are the same and the phase difference is 180° the lamps are dark, and as the phase difference is decreased, the lamps glow with increasing brightness until at synchronism they glow with maximum brilliancy. Hence the incoming alternator should be thrown in at the instant of maximum brilliancy.

Fig. 2,848.—Synchronizing with high pressure alternators; dark and brilliant lamp methods. In both methods the primaries of the transformers are connected in the same way across the terminals of the alternators as shown. In the dark lamp method, the connections between the secondary coils of the transformers must be made so that when each is subjected to the same conditions the action of the one coil opposes that of the other as in the dark lamp method; then, if the transformers be both of the same design, there will be no voltage across the lamps when the alternators are in phase with each other. If the ratio of each transformer is such as to give, for example, 100 volts across its secondary terminals, then the two incandescent lamps since they are joined together in series must each be designed for 100 volts. One 200 volt lamp could be used in either method in place of the two 100 volt lamps. When, therefore, the alternators are directly opposite in phase to each other, both the lamps will burn brightly; as the alternators come together in phase the lamps will produce less and less light, until when the machines are exactly in phase no light will be emitted at all, at which instant the incoming alternator should be thrown in. It must be evident, if the transformer secondary connections are arranged as in the brilliant lamp method, so that they do not oppose each other, the lamps will be at maximum brilliancy when the alternators are in phase and dark when the phase difference is 180°, assuming of course equalized voltage.

Fig. 2,849.—Diagram of Lincoln Synchronizer. In construction, a stationary coil F, has suspended within it a coil A, free to move about an axis in the planes of both coils and including a diameter of each. If an alternating current be passed through both coils, A, will take a position with its plane parallel to F. If now the currents in A and F be reversed with respect to each other, coil A will take up a position 180° from its former position. Reversal of the relative directions of currents in A and F is equivalent to changing their phase relation by 180°, and therefore this change of 180° in phase relation is followed by a corresponding change of 180° in their mechanical relation. Suppose now, instead of reversing the relative direction of currents in A and F, the change in phase relation between them be made gradually and without disturbing the current strength in either coil. It is evident that when the phase difference between A and F reaches 90°, the force between A and F will become reduced to zero, and a movable system, of which A may be made a part, is in condition to take up any position demanded by any other force. Let a second number of this movable system consist of coil B, which may be fastened rigidly to coil A, with its plane 90° from that of coil A, and the axis of A passing through diameter of B. Further, suppose a current to circulate through B, whose difference in phase relation to that in A, is always 90°. It is evident under these conditions that when the difference in phase between A and F is 90°, the movable system will take up a position, such that B is parallel to F, because the force between A and F is zero, and the force between B and F is a maximum; similarly when the difference in phase between B and F is 90°, A will be parallel to F. That is, beginning with a phase difference between A and F of zero a phase change of 90° will be followed by a mechanical change on a movable system of 90°, and each successive change of 90° in phase will be followed by a corresponding mechanical change of 90°. For intermediate phase relation, it can be proved that under certain conditions the position of equilibrium assumed by the movable element will exactly represent the phase relations. That is, with proper design, the mechanical angle between the plane of F and that of A and also between the plane of F and that of B, is always equal to the phase angle between the current flowing in F and those in A and B respectively. As commercially constructed coil F consists of a small laminated iron field magnet with a winding whose terminals are connected with binding posts. The coils A and B are windings practically 90° apart on a laminated iron armature pivoted between the poles of the magnet. These two windings are joined, and a tap from the junction is brought out through a slip ring to one of two other binding posts. The two remaining ends are brought out through two more slip rings, one of which is connected to the remaining binding post, through a non-inductive resistance, and the other to the same binding post through an inductive resistance. A light aluminum hand attached to the armature shaft marks the position assumed by the armature.

Fig. 2,850.—General Electric field of synchronous condenser provided with amortisseur winding. Hunting is accompanied by a shifting of flux across the face of the pole pieces due to the variation in the effect of armature reaction on the main field flux as the current varies and the angular displacement between the field and armature poles is changed. Copper short circuited collars placed around the pole face have currents induced in them by this shifting flux, which have such a direction as to exert a torque tending to oppose any change in the relative position of the field and armature. This action is similar to that of the running torque of an induction motor and the damping device has been still further developed until in its best form it resembles the armature winding of a "squirrel cage" induction motor. The pole pieces are in ducts, and low resistance copper bars placed in them with their ends joined by means of a continuous short circuiting ring extending around the field. Such a device has proven very effective in damping out oscillations started from any cause, the same winding doing duty as a damping device and to assist the starting characteristics.

Fig. 2,851.—Method of synchronizing three phase alternators with, three lamps, being an extension of the single phase method.

Fig. 2,852.—General Electric 500 kw., horizontal mixed pressure Curtis turbine connected to a 500 kw. dynamo. In a Curtis turbine it is not necessary to use the whole periphery of the first stage for low pressure steam nozzles. A section can be partitioned off and equipped with special expanding nozzles to receive steam at high pressure direct from the boilers. Such nozzles deliver their steam against the same wheel as do the low pressure nozzles, but occupy only a small portion of its periphery. The steam is expanded in these nozzles from high pressure all the way down to the normal pressure of the first stage, and in such expansion acquires a high velocity and consequently contains a great deal of energy—much more than does an equal quantity of low pressure steam. In consequence of this, high pressure steam is used with a far lower water rate than is obtained with low pressure steam, or with high pressure steam reduced to low pressure in a reducing valve. This construction is called "mixed pressure." Its function is the same as that of the reducing valve, that is, it makes up for a deficiency of low pressure steam by drawing direct on the boilers. With this construction, the full power of the turbine can be developed with: All low-pressure steam, all high pressure steam, or, any necessary proportion of steam of each pressure. Furthermore, the transition from all low pressure to all high pressure, through all the conditions intermediate between these extremes, is provided for automatically by the turbine governor; a deficiency of low pressure steam causes the high pressure nozzles to open automatically.

Fig. 2,853.—Diagram of connections for synchronizing two compound wound three phase alternators. A and A' are the armatures of the two machines, the fields of which are partly separately excited, the amount of excitation current being controlled by the series compounding rheostats B and B', which form a stationary shunt. It is assumed that the alternator A is connected to the bus bars 1, 2, and 3, by the switch 1S. If an increase make it necessary to introduce the alternator A', it is first run up to speed and excited to standard pressure by its exciter, and then the double plug switch 3S is closed, connecting the primary of the station transformer T and T' with the bus bars through the secondary coil, so that the synchronizing lamps light up when the secondary circuit is closed through the single pole switch 4S. The primary of the station transformer T is thus excited through the double pole switch 5S, connecting it with the outer terminals of the armature A'. The two alternators will now work in opposition to each other upon the synchronizing lamps, the transformer T being operated by the new alternator A' through the switch 2S, and the transformer T' being operated by the working alternator A, from the bus bars. If the new alternator be not in step with the working alternator, the synchronizing lamps will glow, growing brighter and dimmer alternately with greater or lesser rapidity. In this case, the armature speed of the new alternator must be controlled in such a manner that the brightening and dimming will occur more and more slowly, until the lamps cease to glow or remain extinguished for a decided interval of time. The extinction of the light is due to the disappearance of the secondary current, and indicates that the alternators are in step. The switch 2S should now be thrown, thus coupling the two machines electrically, and both of them will continue to operate in step. The double pole equalizer switch 6S should now be closed, connecting the two field windings in parallel and equalizing the compounding, so that any variations of load will affect the two alternators equally. After the alternators have been connected in parallel, the switches 4S and 5S, may be opened leaving the switch 3S closed, to operate the switchboard lamps K, K, as pilot lights from the bus bars.

Fig. 2,854.—Performance curves of Westinghouse air blast 550 kw, 10,500 volt transformer, 3,000 alternations.

Fig. 2,855.—General arrangement of air blast transformers and blowers.

Fig. 2,856.—Small Curtis turbine generator set as made by the General Electric Co., in sizes from 5 kw., to 300 kw. It can be arranged to operate either condensing or non-condensing, and at any steam pressure above 80 lbs. for the smaller sizes and 100 lbs. for the larger. There are only two main bearings. A thrust bearing, consisting of roller bearings and running between hardened steel face washers located at either end of the main bearings is provided solely for centering the rotor so as to equalize the clearance. A centrifugal governor is provided (in the smaller sizes) completely housed, and mounted directly on the main shaft end. It controls a balanced poppet valve through a bell crank. In the larger sizes (75 kw. and above) the governor is mounted on a vertical secondary shaft geared to the main shaft and controls a cam shaft which opens or closes a series of valves in rotation, admitting the steam to different sections of the first stage nozzles. In this way throttling of the steam is avoided. There is also an emergency governor which closes the throttle valve in the event of the speed reaching a predetermined limit. The speeds of operation range from 5,000 R.P.M. for the smallest size to 1,500 R.P.M. for the largest. The lubrication system is enclosed and is automatic. Air leakage where the shaft passes through the wheel casing is prevented by steam seal.

Fig. 2,857.—Cut off coupling for power transmission by line shafting. It is used to cut off a driving shaft from a driven shaft. Its use obviates the use of a quill, such as is shown in fig. 2,858.

Fig. 2,858.—Quill drive. This is the proper transmission arrangement substitute for heavy service, requiring large pulleys, sheaves, gears, rotors, etc. It is a hollow shaft supported by independent bearings. The main driving shaft running through the quill is thus relieved of all transverse stresses. The power is transmitted to the quill by means of a friction or jaw clutch. When the clutch is thrown out the pulley or sheave stands idle and the driving shaft revolves freely within the quill. As there is no contact between moving parts there is no wear. Jaw clutches should be used for drives demanding positive angular displacement. They can only be thrown in and out of engagement when at rest. All very large clutch pulleys, sheaves, or gears designed to run loose on the line shaft are preferably mounted on quills. The letters A, B, C, etc., indicate the dimensions to be specified in ordering a quill.

Fig. 2,859.—Single overhung tangential water wheel equipped with Doble ellipsoidal buckets. The central position of the front entering wedge or lip of the bucket is cut away in the form of a semi-circular notch, which allows a solid circular water jet to discharge upon the central dividing wedge of the bucket without being split in a horizontal plane.

Fig. 2,860.—Motor generator exciter set driven by a Pelton-Doble tangential water wheel. The water wheel runner is mounted on the shaft overhung and the jet is regulated by either a hand actuated or governor controlled needle nozzle. The speed of the water wheel is equivalent to the synchronous speed of the induction motor, hence, the latter floats on the line, and under certain conditions may perform the functions of an alternator by feeding into the circuit, should the water wheel tend to operate above synchronous speed. Should any interruption to the operation of the wheel occur, causing a diminution of speed, the induction motor would drop back to full load speed and take up the exciter load, resulting in no appreciable drop of exciter voltage. The only variation of speed possible is dependent upon the "slip" of the motor. Where two or more exciter sets are employed in the station, an advantageous arrangement embraces the installation of a water wheel driven motor generator set and an exciter set, consisting of merely the direct current generator and water wheel. The induction motor being electrically tied into the circuit, the possibility of a runaway of the water wheel is eliminated, since its speed can only slightly exceed the synchronous speed of the system.

Figs. 2,864 to 2,866.—Converter connections; fig. 2,864 double delta connection; fig. 2,865 diametrical connection; fig. 2,866 two circuit single phase connection. For six phase synchronous converter, two different arrangements of the connections are generally used. One is called the double delta, and the other the diametrical connection. Let the armature winding of the converter be represented by a circle as in figs. 2,864 and 2,865, and let the six equidistant points on the circumference represent collector rings, then the secondary of the supply transformers can be connected to the collector rings in a double delta as in fig. 2,864, or across diametrical pairs of pointer as in fig. 2,865. In the first instance, the voltage ratio is the same as for the three phase synchronous converter and simply consists of two delta systems. The transformers can also be connected in double star, and in such a case the ratio between the three phase voltage between the terminals of each star, and the direct voltage will be the same as for double delta, while the voltage of each transformer coil, or voltage to neutral, is 1 ÷ v3 times as much. With the diametrical connection, the ratio is the same as for the two ring single phase converter, it being analogous to three such systems. Hence six phase double delta E₁ = v3 E ÷ 2v2 = .612E. Six phase diametrical, E₁ = E ÷ v2 = .707E. The ratio of the virtual_voltage E₀ between any collector ring and the neutral point is always E₀ = (E ÷ 2) v2 = .354E. For single phase synchronous converters, consisting of a closed circuit armature winding tapped at two equidistant points to the two collector rings the virtual voltage is 1 ÷ v2 × the direct current voltage. While such an arrangement of the single phase converter is the simplest, requiring only two collector rings, it is undesirable, especially for larger machines, on account of excessive heating of the armature conductors. In fig. 2,866, which represents the armature winding of a single phase converter, the supply circuits from two secondaries of the step down transformers are connected to four collector rings, so that the two circuits are in phase with each other, but each spreads over an arc of 120 electrical degrees instead of over 180 degrees as in the single phase circuit converter. To distinguish the two types, it is generally called a two circuit single phase synchronous converter. The virtual voltage E₂ bears to the direct voltage the same relation as in the three phase converter, that is single phase two circuit, E₁ = v3 ÷ 2v2 =.612E.

Fig. 2,867.—Skeleton diagram showing wiring of alternator, exciter, transformer and converter. The cut also shows switchboard and connections.

Fig. 2,868.—General Electric synchronous converter with series booster. This type of converter generally consists of an alternator with revolving field mounted on the same shaft as the converter armature. The armature of the alternator, or booster, as it is usually called, is stationary and connected electrically in series between the supply circuit and the collector rings of the synchronous converter. The booster field has the same number of pole as the converter and is generally shunt wound. A change in the booster voltage will correspondingly change the alternating voltage impressed on the converter and this regulation can, of course, be made so as to either increase or decrease the impressed voltage by means of strengthening or weakening the booster field. The voltage variation can be made either non-automatic or automatic, and in the latter case, it becomes necessary to provide a motor operated rheostat controlled by suitable relays, or the booster can be provided with a series field. By means of a booster, it is possible to vary the direct voltage of the converter with a constant alternating supply voltage, and this voltage regulation is obtained without disturbance of the power factor or wave shape of the system. Synchronous converters are frequently installed in connection with Edison systems, where three wire direct current is required. The three wire feature is obtained either by providing extra collector rings and compensator, as with ordinary direct current generators, or also by connecting the neutral wire directly to the neutral point of the secondary winding of step down transformers, if such be furnished.

Fig. 2,869.—Wiring diagram for General Electric synchronous converter with series booster as illustrated in fig. 2,868.

Fig. 2,870.—Wiring diagram for three wire synchronous converter with delta-Y connected step down transformer with the neutral brought out. It is evident that in this case each transformer secondary receives ? of the neutral current, and if this current be not so small, as compared with the exciting current of the transformer, it will cause an increase in the magnetic density.

Fig. 2,871.—Wiring diagram of three wire synchronous converter with distributed Y secondary. This system eliminates the flux distortion due to the unbalanced direct current in the neutral. Two separate interconnected windings are used for each leg of the Y. The unbalanced neutral current flowing in this system may be compared in action to the effect of a magnetizing current in a transformer. The effect of the main transformer currents in the primary and secondary is balanced with regard to the flux in the transformer core, which depends upon the magnetic current. When a direct current is passed through the transformer, unless the fluxes produced by the same neutralize one another, its effect on the transformer iron varies as the magnetizing current. For example, assume a transformer having a normal ampere capacity of 100 and, approximately, 6 amperes magnetizing current, and assume that three such transformers are used with Y connected secondaries for operating a synchronous converter connected to a three wire Edison system. Allowing 25 per cent. unbalancing, the current will divide equally among the three legs giving 8.33 amperes per leg, which is more than the normal magnetizing current. The loss due to this current is, however, inappreciable, but the increased core losses may be considerable. If a distributed winding be used, the direct current flows in the opposite direction, around the halves of each core thus entirely neutralizing the flux distortion. Whether the straight Y connection is to be used is merely a question of balancing the increased core loss of the straight Y connection against the increased copper loss and the greater cost of the interconnected Y system. The straight Y connection is much simpler, and it would be quite permissible to use it for transformers of small capacities where the direct current circulating in the neutral is less than 30 per cent. of the rated transformer current.

Fig. 2,872.—Wiring diagram showing arrangement of incandescent lamps for determining the proper phase relations in starting a rotary converter. The alternating current side of a three phase converter is shown at C. The three brushes, D, T and G pressing on its collector rings are joined in order to the three single pole switches H, L and B which can be made to connect with the respective wires M, R, and V, of the alternating current supply circuit. Across one of the outside switches, H, for example, a number of incandescent lamps are joined in series as indicated at E, while the three pole switch (not shown) in the main circuit, between the alternator and the single pole switches is open. If then the main switch just mentioned and the middle switch L be both closed, and the armature of the alternator be brought up to normal speed by running it as a direct current motor, the lamps at E will light up and darken in rapid succession; the lighting and darkening of the lamps will continue until, by a proper adjustment of the speed, the correct phase relations be established between the alternating current in the supply circuit and the alternating current developed in the armature of the converter. As this condition is approached, the intervals between the successive lighting up and darkening of the lamps will increase until they remain perfectly dark. There is then no difference of pressure between the supply circuit M R V and the rotary converter armature circuit, so the source of the direct current may at that instant be disconnected from the machine, and the switches H and B, closed. If the change over has been accomplished before the phase relations of the two circuits differed, the converter will at once conform itself to the supply circuit and run thereon as a synchronous motor without further trouble. The opening of the direct current circuit and the closing of the alternating current supply circuit may be done by hand, but preferably by employing a device that will automatically trip the circuit breaker in the direct current circuit at the instant the switches in the alternating current circuit are closed.

Fig. 2,873.—Diagram of motor converter. This machine which is only to be used for converting from alternating to direct current, consists of an ordinary induction motor with phase wound armature, and a dynamo. The revolving parts of both machines are mounted on the same shaft and from the figure it is seen that the armature of the motor and the armature of the dynamo are also electrically connected. The motor converter is a synchronous machine, but the dynamo receives the current from the armature of the motor at a frequency much reduced from that impressed upon the field winding of the motor. Assuming that the motor and the converter have the same number of pole, the motor will rotate at a speed corresponding to one-half the frequency of the supply circuit. The motor will operate half as a motor and half as a transformer, and the converter, half as a dynamo and half as a synchronous converter, in that one-half of the electrical energy supplied to the motor will be converted into mechanical power for driving the converter, while the other one-half is transferred to the secondary motor windings and thereby to the converter armature in the form of electrical power. The capacity of the motor is theoretically only half what it would be if it were to convert the whole of the electrical energy into mechanical power because the rating depends upon the speed of the rotating field and not on that of the rotor. If the two machines have a different number of pole, or are connected to run at different speeds, the division of power is at a different but constant ratio. The machine starts up as an ordinary polyphase induction motor and the field of the converter is built up as though it were an ordinary dynamo. Motor converters are occasionally used on high frequency systems, as their commutating component is of half frequency, and thus permits better commutator design than a high frequency converter. The advantage of this type of machine is that for phase control it requires no extra reactive coils, the motor itself having sufficient reactance. It is, however, larger than standard converters, but smaller than motor generators, as half the power is converted in each machine. Its efficiency is less than for synchronous converters, and the danger of reaching double speed in case of a short circuit on the direct current side is very great. It has been used abroad to some extent for 60 cycle work, in preference to synchronous converters, but with the present reliable design of 60 cycle converters, and the general use of 25 cycles, where severe service conditions are met, as in railroading, motor converters should not be recommended.

Fig. 2,874.—Sectional view of General Electric vertical synchronous converter. In this construction, the field frame carrying the poles is mounted on cast iron pedestals and is split vertically. This allows the two halves of the frame to be separated for inspection or repairs of the armature. The armature, including commutator and collector rings, is mounted on a vertical stationary shaft, which is rigidly supported from the foundation. The thrust of the armature is carried on a roller bearing attached to the top of the shaft and upper side of the armature spider. The under side of the lower plate of the roller bearing is made spherical and fits into a corresponding spherical cup on the end of the shaft, making the bearing self aligning. The armature spider has a babbitted sleeve along the fit of the vertical shaft, which acts as a guide bearing and has to take only the thrust due to the unbalancing effect of the rotating parts. A circulating pump furnishes oil to the roller bearing, the oil draining off through the guide bearing. A marked advantage of this type of construction is the accessibility of the commutator for adjustment of the brushes, etc., as there is no pit or pedestal bearing to interfere.

Fig. 2,875.—Resistance measurement by "drop" method. The circuit whose resistance is to be measured, is connected in series with an ammeter and an adjustable resistance to vary the flow of current. A voltmeter is connected directly across the terminals of the resistance to be measured, as shown in the figure. According to Ohm's law I = E ÷ R, from which, R = E ÷ I. If then the current flowing in the circuit through the unknown resistance be measured, and also the drop or difference of pressure, the resistance can be calculated by above formula. In order to secure accurate determination of the resistance such value of current must be used as will give large deflections of the needle on the instruments employed. A number of independent readings should be taken with some variation of the current and necessarily a corresponding variation in voltage. The resistance should then be figured from each set of readings and the average of all readings taken for the correct resistance. Great care must be taken, however, in the readings, and the instruments must be fairly accurate. For example, suppose that the combined instrument error and the error of the reading in the voltmeter should be 1 per cent., the reading being high, while the corresponding error of the ammeter is 1 per cent. low. This would cause an error of approximately 2 per cent. in the reading of the resistance. In making careful measurements of the resistance, it is also necessary to determine the temperature of the resistance being measured, as the resistance of copper increases approximately .4 of 1 per cent. for each degree rise in temperature. Use is made of this fact for determining the increase in temperature of a piece of apparatus when operating under load. The resistance of the apparatus at some known temperature is measured, this being called the cold resistance of the apparatus. At the end of the temperature test the hot resistance is taken. Assume the resistance has increased by 15 per cent. This would indicate a rise in temperature of 37½ degrees above the original or cold temperature of the apparatus. Suppose then that in measuring the cold resistance, results are obtained which are 2 per cent. low, and that in measuring the hot resistance, there be 2 per cent. error in the opposite direction. This would mean that a total error of 4 per cent. had been made in the difference between the hot and cold resistances, or an error of 10 degrees. The correct rise in temperature is, therefore, about 27½ instead of 37½ degrees. In other words, an error of 2 per cent. in measuring each resistance has caused an error of approximately 36½ per cent. in the measurement of the rise in temperature. The constant .4 which has been used above is only approximate and should not be used for exact work. For detail instructions of making calculations of resistance and temperature, see "Standardization Rules of the A.I.E.E."

Figs. 2,876 to 2,879.—How to connect instruments for power measurement. There are several ways of connecting an ammeter, voltmeter and wattmeter in the circuit for the measurement of power. A few of the methods are discussed below. With some of the connections it is necessary to correct the readings of the wattmeter for the losses in the coil, or coils, of the wattmeter, or for losses in ammeter or voltmeter. This is necessary since the instruments may be so connected that the wattmeter not only measures the load but includes in its indications some of the instrument losses. If the load measured be small, or considerable accuracy is required, these instrument losses may be calculated as follows: Loss in pressure coils is E² ÷ R, in which E is the voltage at the terminals of the pressure coil and R is the resistance. Loss in current coil is I² R in which I is the current flowing and R the resistance of the current coil. In general let E_v = voltage across terminals of the voltmeter; E_w = voltage across the terminals of the pressure coil of the wattmeter; I_w = current through current coil of wattmeter; I_a = current through current coil of ammeter; R_v = resistance of pressure coil of voltmeter; R_w = resistance of pressure coil of wattmeter; R¹_w = resistance of current coil of wattmeter; R_a = resistance of current coil of ammeter. Then the losses in the various coils will be as follows: E²_v ÷ R_v = loss in pressure coil of voltmeter. E²_w ÷ R_w = loss in pressure coil of wattmeter. I²_w ÷ R_v = loss in current coil of wattmeter. I²_aR_a = loss in current coil of ammeter. If connection be made as in fig. 2,876, the correct power of the circuit will be the wattmeter reading W-(E²_v ÷ R_v + E²_w ÷ R_w) in which E_v = E_w. In fig. 2,877, the power is W-E²_w ÷ R_w. In fig. 2,878, the power is W-I²_wR¹_w, or the correct power is the wattmeter reading minus the loss in the current coil of the wattmeter. In fig. 2,879, the power is W-(E²_w ÷ R_w + I²_aR_a)· The usual method of connection is either as in fig. 2,876 or fig. 2,877. In either case the current reading is that of the load plus the currents in the pressure coils of the voltmeter and wattmeter. Unless the current being measured, however, is very small, or extreme accuracy is desired, it is unnecessary to correct ammeter readings. In fig. 2,877 a small error is introduced due to the fact that the actual voltage applied to the load is that given by the voltmeter minus the small drop in voltage through the current coil of the wattmeter. If an accurate measure of the current in connection with the power consumed by the load be required, the connections shown in fig. 2,879 are used, and if extreme accuracy is required, the wattmeter reading is reduced by the losses in the ammeter and in the pressure coil of the wattmeter. The loss in the pressure coil of a wattmeter or voltmeter may be as high as 12 or 15 watts at 220 volts. The loss in the current coil of a wattmeter with 10 amperes flowing may be 6 or 8 watts. It can be easily seen that if the core or copper losses of small transformers are being measured, it is quite necessary to correct the wattmeter readings, for the instrument losses. In measuring the losses of a 25 or 50 H.P. induction motor, the instrument losses may be neglected. A careful study of the above will show when it becomes necessary to correct for instrument losses and the method of making these corrections. Connections are seldom used which make it necessary to correct for the losses in the current coils of either ammeter or wattmeter, as the losses vary with the change in the current. On the other hand, the voltages generally used are fairly constant at 110 or 220, and when the losses of the pressure coils at these voltages have once been calculated, the necessary instrument correction can be readily made.

Fig. 2,881.—Three phase motor test; voltmeter and ammeter method. If it be desired to determine the approximate load on a three phase motor, this may be done by means of the connections as shown in the figure, and the current through one of the three lines and the voltage across the phase measured. If the voltage be approximately the rated voltage of the motor and the amperes the rated current of the motor (as noted on the name plate) it may be assumed that the motor is carrying approximately full load. If, on the other hand, the amperes show much in excess of full load rating, the motor is carrying an overload. The heat generated in the copper varies as the square of the current. That generated in the iron varies anywhere from the 1.6 power, to the square. This method is very convenient if a wattmeter be not available, although, it is, of course, of no value for the determination of the efficiency or power factor of the apparatus. This method gives fairly accurate results, providing the load on the three phases of the motor be fairly well balanced. If there be much difference, however, in the voltage of the three phases, the ammeter should be switched from one circuit to another, and the current measured in each phase. If the motor be very lightly loaded and the voltage of the different phases vary by 2 or 3 per cent., the current in the three legs of the circuit will vary 20 to 30 per cent.

Fig. 2,882.—Three phase motor test by the two wattmeter method. If an accurate test of a three phase motor be required, it is necessary to use the method here indicated. Assume the motor to be loaded with a brake so that its output can be determined. This method gives correct results even with considerable unbalancing in the voltages of the three phases. With the connections as shown, the sum of the two wattmeter readings gives the total power in the circuit. Neither meter by itself measures the power in any one of the three phases. In fact, with light load one of the meters will probably give a negative reading, and it will then be necessary to either reverse its current or pressure leads in order that the deflection may be noted. In such cases the algebraic sums of the two readings must be taken. In, other words, if one read plus 500 watts and the other, minus 300 watts, the total power in the circuit will be 500 minus 300, or 200 watts. As the load comes on, the readings of the instrument which gave the negative deflection will decrease until the reading drops to zero, and it will then be necessary to again reverse the pressure leads on this wattmeter. Thereafter the readings of both instruments will be positive, and the numerical sum of the two should be taken as the measurement of the load. If one set of the instruments be removed from the circuit, the reading of the remaining wattmeter will have no meaning. As stated above, it will not indicate the power under these conditions in any one phase of the circuit. The power factor is obtained by dividing the actual watts input by the product of the average of the voltmeter readings × the average of the ampere readings × 1.73.

Fig. 2,883.—Three phase motor test; one wattmeter method. This method is equivalent to the two wattmeter method with the following difference. A single voltmeter (as shown above) with a switch, A, can be used to connect the voltmeter across either one of the two phases. Three switches, B, C and D, are employed for changing the connection of the ammeter and wattmeter in either one of the two lines. With the switches B and D in the position shown, the ammeter and wattmeter series coils are connected in the left hand line. The switch C must be closed under these conditions in order to have the middle line closed. Another reading should then be taken before any change of load has occurred, with switch A thrown to the right, switch B closed, switch D thrown to the right and switch C opened. The ammeter and the current coil of the wattmeter will then be connected to the middle line of the motor. In order to prevent any interruption of the circuit, the switches B, D and C should be operated in the order given above. With very light load on the motor the wattmeter will probably give a negative deflection in one phase or the other, and it will be necessary to reverse its connections before taking the readings. For this purpose a double pole, double throw switch is sometimes inserted in the circuit of the pressure coil of the wattmeter so that the indications can be reversed without disturbing any of the connections. It is suggested, before undertaking this test, that the instructions for test by the two wattmeter and by the polyphase wattmeter methods be read.

Fig. 2,884.—Three phase motor, one wattmeter and Y box method. This method is of service, only, provided the voltages of the three phases are the same. A slight variation of the voltage of the different phases may cause a very large error in the readings of the wattmeter, and inasmuch as the voltage of all commercial three phase circuits is more or less unbalanced, this method is not to be recommended for motor testing. With balanced voltage in all three phases, the power is that indicated by the wattmeter, multiplied by three. Power factor may be calculated as before.

Fig. 2,885.—Test of three phase motor with neutral brought out; single wattmeter method. Some star connected motors have the connection brought out from the neutral of the winding. In this case the circuit may be connected, as here shown. The voltmeter now measures voltage between the neutral and one of the lines, and the wattmeter the power in one of the three phases of the motor. Therefore, the total power taken by the motor will be three times the wattmeter readings. By this method, just as accurate results can be obtained as with the two wattmeter method. The power factor will be the indicated watts divided by the product of the indicated amperes and volts.

Fig. 2,886.—Temperature test of a large three phase induction motor. Temperature tests are usually made on small induction motors by belting the motor to a generator and loading the generator with a lamp bank or resistance until the motor input is equal to the full load. If, however, the motor be of considerable size, such that the cost of power becomes a considerable item in the cost of testing, the method here shown may be employed. For this purpose, however, two motors, preferably of the same size and type, are required. One is driven as a motor and runs slightly below synchronism, due to its slip when operating with load. This motor is belted to a second machine. If the pulley of the second machine be smaller than the pulley of the first machine, the second machine will then operate as an induction generator, and will return to the line as much power as the first motor draws from the line, less the losses of the second machine. By properly selecting the ratio of pulleys, the first machine can be caused to draw full load current and full load energy from the line. In this way, the total energy consumed is equivalent to the total of the losses of both machines, which is approximately twice the losses of a single machine. The figure shows the connection of the wattmeters, without necessary switches, for reading the total energy by two wattmeter method. Detailed connection of the wattmeter is shown in fig. 2,883. It is usual, in making temperature tests, to insert one or more thermometers in what is supposed to be the hottest part of the winding, one on the surface of the laminae and one in the air duct between the iron laminae. The test should be continued until the difference in temperature between any part of the motor and the air reaches a steady value. The motor should then be stopped and the temperature of the armature also measured. For the method of testing wound armature type induction motors of very large size, see fig. 2,890. For the approved way of taking temperature readings and interpreting results, see the "Standardization Rules of the A.I.E.E."

Fig. 2,887.—Alternator excitation or magnetization curve test. The object of this test is to determine the change of the armature voltage due to the variation of the field current when the external circuit is kept open. As here shown, the field circuit is connected with an ammeter and an adjustable resistance in series with a direct current source of supply. The adjustable resistance is varied, and readings of the voltmeter across the armature, and of the ammeter, are recorded. The speed of the generator must be kept constant, preferably at the speed which is given on the name plate. The excitation or magnetization curve of the machine is obtained by plotting the current and the voltage.

Fig. 2,888.—Three phase alternator synchronous impedance test. In determining the regulation of an alternator, it is necessary to obtain what is called the synchronous impedance of the machine. To obtain this, the field is connected, as shown above. Voltmeters are removed and the armature short circuited with the ammeters in circuit. The field current is then varied, the armature driven at synchronous speed, and the armature current measured by the ammeters in circuit. The relation between field and armature amperes are then plotted. The combination of the results of this test, with those obtained from the test shown in fig. 2,887, are used in the determination of the regulation of an alternator. Engineers differ widely in the application of the above to the determination of regulation, and employ many empirical formulae and constants for different lines of design.

Fig. 2,889.—Three phase alternator or synchronous motor temperature test. In this test two alternators or synchronous motors of same size and type are used, and are belted together, one to be driven as a synchronous motor and the other as an alternator. The method employed is to synchronize the synchronous motor with the alternator or alternators on the three phase circuit, and then connect to the line by means of a three pole single throw switch. The alternator is then similarly synchronized with the alternator of the three phase circuit and thrown onto the line. By varying the field of the alternator it can be made to carry approximately full load, and the motor will then be also approximately fully loaded. The usual method is to have the motor carry slightly in excess of full load, and the alternator slightly less than full load. Under these conditions the motor will run a little warmer than it should with normal load, while the alternator will run slightly cooler. Temperature measurements are made in the same way as discussed under three phase motors. The necessary ammeters, voltmeters and wattmeters for adjusting the loads on the motors and generator are shown in above figure. If pulleys be of sufficient size to transmit the full load, with, say one per cent. slip, the pulley on the motor should be one per cent. larger in diameter than the pulley on the alternator, so as to enable the alternator to remain in synchronism and at the same time deliver power to the circuit. With very large machines under test, it is inadvisable to use the above method as it is sometimes difficult to so adjust the pulleys and belt tension that the belt slip will be just right to make up for the difference in diameter of the pulleys, and very violent flapping of the belt results. To meet such cases, various other methods have been devised. One which gives consistent results is shown in fig. 2,890.

Fig. 2,890.—Three phase alternator or synchronous motor temperature test. Supply the field with normal field current. The armature is connected in open delta as illustrated, and full load current sent through it from an external source of direct current, care being taken to ground one terminal of the dynamo so as to avoid danger of shock due to the voltage on the armature winding. The field is then driven at synchronous speed. If the armature be designed to be connected star for 2,300 volts, the voltage generated in each leg of the delta will be 1,330 volts, and unless one leg of the dynamo were grounded, the tester might receive a severe shock by coming in contact with the direct current circuit. The insulation of the dynamo would also be subjected to abnormal strain unless one terminal were grounded. By the above method the field is subjected to its full copper loss and the armature to full copper loss and core loss. Temperature readings are taken as per standardization rules of the A.I.E.E. This method may also be used with satisfactory results on large three phase motors of the wound rotor type. If the alternator pressure be above 600 volts, a pressure transformer should be used in connection with the voltmeter.

Fig. 2,891.—Direct motor or dynamo magnetization test. The object of this test is to determine the variation of armature voltage without load, with the current flowing through the field circuit. The armature should be driven at normal speed. The adjustment resistance in the field circuit is varied and the voltage across the armature measured. The curve obtained by plotting these two figures is usually called magnetization curve of the dynamo. It is usual to start with the higher resistance in the field circuit so that very small current flows, gradually increasing this current by cutting out the field resistance. When the highest no load voltage required is reached, the field current is then diminished, and what is called the descending (as opposed to the ascending) magnetization curves are obtained. The difference in the two curves is due to the lag of the magnetization behind the magnetizing current, and is caused by the hysteresis of the iron of the armature core.

Fig. 2,892.—Shunt dynamo external characteristic test. The external characteristic of a shunt dynamo is a curve showing the relation between the current and voltage of the external circuit. This is obtained by the connection as here shown. The shunt field is so adjusted that the machine gives normal voltage when the external circuit is open. The field current is then maintained constant and the external current varied by varying the resistance in the circuit. By plotting voltage along the vertical, against the corresponding amperes represented along the horizontal, the external characteristic is obtained.

Fig. 2,893.—Load and speed test of direct current shunt motor. The object of this test is to maintain the voltage applied to the motor constant, and to vary the load by means of a brake and find the corresponding variation in speed of the machine and the current drawn from the circuit. If the motor be a constant speed motor, the field resistance is maintained constant. The above indicates the method of connecting instruments for the test alone; for starting the machine the ordinary starting box, should, of course, be inserted.

Fig. 2,894.—Temperature test of direct current motor or dynamo; loading back method. In making temperature tests on a small dynamo it is usual to drive the dynamo with a motor and load the dynamo by means of a lamp bank or resistance, the voltage across the dynamo being maintained constant, and the current through the external circuit adjusted to full load value. The temperatures are then recorded, and when they reach a constant value above the temperature of the atmosphere, the test is discontinued. Similarly, in making a test on a small motor, the motor is loaded with a dynamo and the load increased until the input current reaches the normal full load value of the motor, the test being conducted as for a small dynamo. When, however, the apparatus, either motor or dynamo, reaches a certain size, it becomes necessary, in order to economize energy, to use what is called the loading back method, as here illustrated. The motor is started in the usual way, with the dynamo belted to it, the circuit of the dynamo being open. The field of the dynamo is then adjusted so that the dynamo voltage is equal to that of the line. The dynamo is then connected to the circuit and its field resistance varied until it carries normal full load current. Under these conditions, if the motor and dynamo be of the same size and type, the motor will carry slightly in excess of full load, the difference being approximately twice the losses of the machines. Under these conditions the total power drawn from the line is equal to twice the loss of either machine. Temperature readings are taken as in other temperature tests.

Fig. 2,895.—Compound dynamo external characteristics test; adjustable load. The object of this test is to determine the relation between armature voltage and armature current. Shunt field is adjusted to give normal secondary voltage when the external circuit is open. The load is then applied by means of an adjustable resistance or lamp bank, and readings of external voltage and current recorded. If the machine be normally compounded, the external voltage will remain practically constant throughout the load range. If the machine be under-compounded, the external voltage will drop with load, while if over-compounded, there will be a rise in voltage with increase in load.

Figs. 2,896 and 2,897.—Transformer core loss and leakage, or exciting current test. With the primary circuit open, the ammeter indicates the exciting or no load current. It should be noted that all instruments are inserted on the low voltage side, for both safety of the operator and because the measurements are more accurate. The no load primary current, if the ratio of transformation be 10: 1, will be one-tenth of the measured secondary current. The wattmeter connected, as shown, measures the sum of the losses, in the transformer, in the pressure coil of the wattmeter, and in the voltmeter. On all standard makes of portable instruments, the resistance of the wattmeter pressure coil and of the voltmeter is given, and the loss in either instrument is the square of the voltage at its terminals, divided by its resistance. Subtracting these losses from the total indicated upon the wattmeter, gives the true core or iron loss. It should be noted that in this diagram is shown an auxiliary transformer with a number of taps for obtaining the exact rated voltage of the transformer under test. In fig. 2,897 is shown, in general, the same connections as in fig. 2,896, except that the auto-transformer has been replaced by a resistance. If the line voltage available be not much in excess of the rated voltage of the transformer under test, very little error is introduced by the use of the resistance method. However, if the difference be 10 per cent. or more the auxiliary transformer shown in fig. 2,896 should be used. Measurements made under the resistance method always give lower results than those obtained with the auxiliary transformer.

Fig. 2,898.—Diagram of connections for calibrating a wattmeter. The calibration of a portable wattmeter is accomplished with direct current of constant value which is passed through the series winding by connecting the source thereof with the current terminals. A direct current voltage which may be varied throughout the range of the wattmeter is also applied to the instrument between the middle and right hand pressure terminals A and E the wiring in the meter between these terminals being such that its differential winding is then cut out of circuit. The method of procedure consists in comparing the deflections on the wattmeter at five of six approximately equidistant points over its scale with the corresponding products of volts and amperes used to obtain them. The changes in the wattmeter deflections are effected by merely varying the voltage, the value of the current being maintained constant at a value which represents the full current capacity of the meter.

Fig. 2,899.—Transformer copper loss by wattmeter measurement and impedance. At first glance, this method would seem better than the calculation of loss after measurement of the resistance. However, it should be noted that the wattmeter is, in itself, subject to considerable error under the low power factor that will exist in this test. The secondary of the transformer is short circuited, and a voltage applied to the primary which is just sufficient to cause full load primary current. If full current pass through the primary of the transformer with the secondary short circuited, the secondary will also carry full load current. With connections as shown, and with the full load current, the voltmeter indicates the impedance volts of the transformer. This divided by the rated voltage gives what is called the per cent. impedance of the transformer. In a commercial transformer of 5 kw., this should be approximately 3 per cent. The iron loss of the transformer under approximately 3 per cent. of the normal voltage will be negligible, and the losses measured will be the sum of the primary and secondary copper losses. As in the discussion of the core loss measurements, the wattmeter readings must be corrected for the loss in its pressure coil, the method of correction being the same as that discussed under the core loss measurement. If the impedance volts, as measured, be divided by the primary current, the impedance of the transformer is obtained. The reciprocal of this quantity is known by the term "admittance." When two or more transformers are connected in parallel they divide the load in proportion to their admittance. It is, therefore, important that the users of transformers know the impedance of the apparatus used, in order to determine whether two or more transformers will operate satisfactorily in parallel. For discussion of wattmeter error on low power factor, see note on page 2,075. For accurate measurement of impedance, the voltmeter should be connected directly across the terminals of the transformer rather than as shown in the diagram.

Fig. 2,900.—Temperature test of transformer with non-inductive load. The figure shows the simplest way of making the test. Connect the primary of the transformer to the line as shown, and carry normal secondary load by means of a bank of lamps or other suitable resistance, until full load secondary current is shown by the ammeter in the secondary circuit. The transformer should then be allowed to run at its rated load for the desired interval of time, temperature readings being made of the oil in its hottest part, and also of the surrounding air. Where temperatures of the coil rather than temperatures of the oil are desired, it is necessary to use the resistance method. This is obtained by first carefully measuring the resistance of both primary and secondary coils at the temperature of the room, and then, after the transformer has been under heat test for the desired time, disconnect it from the circuit and again measure the resistance of primary and secondary. For proper method of calculating the temperature rise from resistance measurements, the reader is referred to the standardization rules of the A.I.E.E. In making resistance measurements of large transformers by the drop method care should be taken to allow both ammeter and voltmeter indications to settle down to steady values before readings are taken. This may require several minutes. Each time the current is changed it is necessary in order to obtain check values on resistance measurements, to wait until the current is again settled to its permanent value before taking readings. All resistance measurements must be taken with great care, as small errors in the measurement of the resistance may make very large errors in the determination of the temperature rise. The method above described is satisfactory for small transformers. Where large units are to be tested, the cost of current for testing becomes an important item. The "bucking test" as in fig. 2,901, is more economical.

Fig. 2,902.—Transformer insulation test. In applying a 10,000 volt insulation test between the primary and secondary of a transformer, the testing leads should be disconnected from the transformer under test, and a spark gap introduced as shown, with the test needle set at a proper sparking distance for 10,000 volts. A high resistance should be connected in the secondary before closing its circuit, and the voltage gradually increased by cutting out this secondary resistance until a spark jumps across the spark gap. When the spark jumps across the spark gap, the voltmeter reading should be recorded and the testing transformer disconnected. The spark gap should then be increased about 10 per cent. and the high tension leads connected to the transformer under test as indicated in the diagram. In order to equalize the insulation strains, all primary leads should be connected together, all secondary leads not only connected together, but to the core as well. All resistance in the rheostat in the low tension circuit should then be inserted and the switch closed. Gradually cut out secondary resistance until the voltmeter shows the same voltage as was recorded previously when the spark jumped across the gap, and apply this voltage to the transformer for one minute. Insulation tests for a period of over one minute are very unadvisable, as transformers with excellent insulation may be seriously damaged by prolonged insulation tests. The longer the strain to which any insulation is subjected, the shorter the subsequent life of the insulation. Also the greater the applied voltage above the actual operating voltage of the apparatus, the shorter the subsequent life of the insulation. In testing small transformers, the spark gap may be omitted and the voltage of the low pressure coil of the testing transformer measured. This multiplied by the ratio of transformation gives the testing voltage.

Fig. 2,903.—Transformer insulation test as made when a special high tension transformer be not available. In this method a number of standard transformers, connected as shown may be employed, but great care should be taken to have such transformer cases thoroughly insulated from the ground and from one another, in order to minimize the insulation strains in the testing transformers. Care should be taken to insert in the circuit of each testing transformer a fuse, not in excess of the transformer capacity, which will blow, in case of a break down in the apparatus under test. In testing insulation between secondary and core, disconnect the primary entirely, apply one terminal of the testing transformer to the secondary terminals of the transformer under test, and the other terminal of the testing transformer to the core of the transformer under test. This test should also not be in excess of one minute.

Fig. 2,904.—Transformer internal insulation test, sometimes called double normal voltage test, from the fact that most transformers are tested with double normal voltage across their terminals. If either the primary or secondary of the transformer be connected to some source of current with voltage double that of the voltage of the transformer under test, the insulation between adjacent turns, and also the insulation between adjacent layers will be subjected to twice the normal operating voltage. It is good practice to employ high frequency for this test in order to prevent an abnormal current from passing through the transformer. Sixty cycle transformers are usually tested on 133 cycles, and 25 cycle transformers on 60 cycle circuits for this double normal voltage test. It is necessary to insert the resistance in the circuit of the transformer and bring the voltage up gradually, the same as applying other high insulation tests in order to prevent abnormal rises in pressure at the instant of closing the circuit.

Fig. 2,905.—Transformer insulation resistance test. The insulation, besides being able to resist puncture, due to increased voltage, must also have sufficient resistance to prevent any appreciable amount of current flowing between primary and secondary coils. It is, therefore, sometimes important that the insulation resistance between primary and secondary be measured. This can be done, as here shown. Great care should be taken to have all wires thoroughly insulated from the ground, and to have an ammeter placed as near as possible to the terminals of the transformer under test, in order that current leaking from one side of the line to the other, external to the transformer, may not be measured. Great care is required in making this measurement, in order to obtain consistent results.]

Ques. What are the usual remedies applied to a voltmeter to correct a 3 or 4 per cent. error?

Ans. They consist of straightening the pointer, varying the tension of the spiral springs, renewing the jewels in the bearings, altering the value of the high resistance, and, in the case of a direct current instrument, strengthening the permanent magnet.

Ques. How is the permanent magnet strengthened?

Ans. After detaching it from the instrument, wrap around several turns of insulated wire, and pass through this wire for a short time 3 or 4 amperes of direct current in such a direction as to reinforce the magnet magnetism.

Ques. How may the value of the high resistance of a voltmeter be altered?

Ans. Determine the resistance of the voltmeter and add or subtract, according as the reading is high or low, a certain length of wire whose resistance is in per cent. of the voltmeter resistance the same as the per cent. of error.

NOTE.—The complete calibration of a two scale voltmeter does not, as might be supposed, necessitate that the readings on both scales be checked with standards, for since the resistance corresponding to the one scale is always some multiple of the resistance of the other, the constants of the two scales are proportional. For instance, if S = the reading at the end of the high scale of the voltmeter; S¹ = the reading at the end of the low scale of the voltmeter; R = the resistance in the meter corresponding to the high scale; R¹ = the resistance in the meter corresponding to the low scale; K = the constant for the high scale, and K¹ = the constant for the low scale. Then

SK ÷ R = S¹K¹ ÷ R¹

from which

K¹ = SKR ÷ S¹R

That is to say, if the respective resistances corresponding to the two scales be known, and the constant of the high scale be determined by comparison with a standard, then by aid of these known values and the maximum readings on the two scales, the constant of the low scale may be calculated. It is also possible to calculate the constant of the high scale if the constant of the low scale be known, together with the values of the resistances corresponding to the two scales; for from the equation previously given.

K = RS¹K¹ ÷ R¹S

Fig. 2,911.—Characteristic curve of shunt dynamo. Suppose in making the test, the deflections on the meters for the first readings be 63 volts and 0 amperes, the plotting of these values will give the first point on the curve. Similarly, the second readings with main circuit closed and maximum resistance in the water rheostat may be assumed to be 62.5 volts and 7.5 amperes, which plotted gives the second point B. A still further lowering of the plate will permit a stronger current in the main circuit, and the value of this together with its corresponding voltage will give a third point for the curve. Neither for this reading, however, nor for the following readings of the test should the field rheostat be altered. When six or eight points ranging from zero to a maximum current have been obtained and plotted, a curved line should be drawn through them such as shown through ABCDEFG0, the characteristic curve of the dynamo. While the curve may be sketched in free hand, it should preferably be drawn by the aid of French curves. In case the French curve cannot be exactly made to coincide with all the points as for instance C and D, it should be run in between giving an average result, and smoothing out irregularities, or small errors due to the "personal equation." The meter of course must be correct or calibrated and the readings corrected by the calibration coefficient.

Fig. 2,912.—Characteristic curves for a compound dynamo. If the machine be over compounded, the characteristic curve has the form of the curve A B, which curve was obtained from a machine over-compounded from 118 to 123 volts, and designed to give 203 amperes at full load. The preliminary arrangements for testing a compound dynamo are similar to those for a shunt generator, and if the shunt across the series field winding be already made up and in position, the readings are taken precisely in the same manner. It is generally considered sufficient if observations be recorded at zero, ¼, ½, ¾ and full load. If it be desired to ascertain the effect which residual magnetism has upon the field magnets the current is decreased after the full load point is reached without opening the circuit, and readings are taken in succession at ¾, ½, ¼ and zero load giving in this case the curve B C D E S. It is thus seen that residual magnetism exerts no small effect upon the voltage obtained at the different loads, for had there been no residual magnetism in the field magnets the curve B C D E S would have coincided with the curve A B. The curve A B, and the straight line A X drawn through the points A and B, are almost identical, and as A X represents the theoretical characteristic curve for the machine, it is seen that compounding is practically perfect. In order to insure such accurate results being obtained, providing the machinery be correctly designed, requires considerable care in taking the readings; for example, each step or load on the ascending curve should not be exceeded before the corresponding deflection is taken, else the residual magnetism will cause the pressure reading to be higher than it actually should be, and the following pressure readings will also be affected in the same manner. In case the shunt to be employed across the series field has not been made up, it is advisable to perform a trial test before taking the readings for the curve as previously described. The trial test consists in taking two readings,—one at no load and the other at full load, the shunt being so adjusted as to length and section that the desired amount of compounding will be obtained in the latter reading with normal voltage at no load. If the first trial fail to produce the desired result by giving too low a voltage at full load, the length of the shunt across the series field should be increased, or its section should be reduced by employing a less number of strips in its makeup; again, if the voltage at full load be higher than that desired, there must be made a decrease in length or an increase of section in the shunt employed.