UNIVERSITY OF CALIFORNIA 
 
 ANDREW 
 
 SMITH 
 
 HALLIDIC: 
 
AMERICAN 
 METER PRACTICE 
 
 BY 
 
 LYMAN C. REED 
 
 NEW YORK 
 
 McGRAW PUBLISHING COMPANY 
 
 1 1 4 Liberty Street 
 
 1903 
 

 COPYRIGHTED, 1903 
 
 BY THE 
 
 McGRAW PUBLISHING COMPANY 
 NEW YORK 
 
\ 
 
 PREFACE 
 
 THE subject of metering the output of central stations 
 has been to me one of the most interesting of the 
 many problems arising in the development of the supply 
 of current for various commercial needs; There are many 
 minor details omitted in the work, which would doubtless 
 prove of interest to the practical worker in meters, but, 
 however poorly set forth, the effort is made to outline the 
 underlying principles of operation and practice and leave 
 the minor details to be worked out to suit local conditions. 
 
 In describing only a few meters my object is to select one 
 each of well known and representative types and not 
 to weary the reader by reciting the same characteristics 
 held by a number of meters of the same type. 
 
 Any intention of slighting in any way many excellent 
 meters, herein mentioned but not described, is entirely 
 foreign to the purpose of this work. The meters selected 
 have each some distinguishing feature which makes them 
 representative of as many respective classes. 
 
 To the consumer of power who is not a technical man, 
 Chapter XIII, on How to Read Meters, will probably be 
 of most interest since it will enable him to figure out and 
 check up his meter bills. This knowledge should bring 
 about the very friendliest relations between the supplier 
 and consumer of electricity. 
 
 LYMAN C. REED. 
 
 3 
 
 141415 
 
CONTENTS 
 
 Chapter I Measurement of Power in Direct 
 
 Current Circuits 7 
 
 Chapter II Measurement of Power in Alternat- 
 ing Current Circuits 17 
 
 Chapter III Meter Selection ; Requisites of a 
 Good Meter ; Commercial Consider- 
 ations 34 
 
 Chapter IV Torque and Friction 46 
 
 Chapter V The Edison Chemical Meter 59 
 
 Chapter VI The Thomson Recording Wattmeter 67 
 
 Chapter VII The Duncan Recording Wattmeter 
 
 for Alternating Current 82 
 
 Chapter VIII The Duncan Wattmeter for Direct 
 
 Current 91 
 
 Chapter IX The Stanley Recording Wattmeter . 105 
 
 Chapter X The Guttman Wattmeter 113 
 
 Chapter XI The Westinghouse Induction Meter . 119 
 
 Chapter XII General Management of the Meter 
 Department ; Records ; Testing ; 
 General Policy 124 
 
 Chapter XIII Reading Meters 159 
 
 Chapter XIV Value of Losses in Meters Relative 
 
 to Income 163 
 
 Chapter XV Differential Rating 171 
 
 Chapter XVI Elements of Photometry 178 
 
 5 
 
AMERICAN METER PRACTICE. 
 
 CHAPTER I. 
 Measurement of Power in Direct Current Circuits. 
 
 Electrical measurements in the laboratory have been 
 carried on for many years with great accuracy. The con- 
 ditions are carefully studied when a determination of some 
 quantity is sought, and the errors which effect the accur- 
 acy of the measurement eliminated. The study of the 
 phenomena of electricity is rightly called an exact science, 
 but the perfecting of an accurate commercial meter has 
 been a slow and tedious process. At first the principles of 
 correct design were not embodied in the meters put on the 
 market, and no matter how perfect they were made me- 
 chanically a correct registration of the power consumed 
 could not be obtained. But, as a rule, the mechanical 
 features were more defective than the theoretical ones, and 
 the combination of these defective elements has given the 
 meter of commerce an unenviable reputation. 
 
 Such great inventors as Edison, Thomson, Shallenberger 
 and Duncan have spent much time in trying to get to- 
 gether various elements in suitable combination to meet 
 the requirements and accurately measure the current or 
 the power in the many various applications and uses to 
 which electricity is put. Edison gave us the chemical 
 ampere hour meter, which has done yeoman service in the 
 
8 AMERICAN METER PRACTICE. 
 
 metering of direct currents; and Thomson gave us his 
 recording wattmeter, which has gradually been developed 
 into an excellent meter, and which can be used for both 
 direct and alternating current. Shallenberger and Dun- 
 can produced at first ampere hour meters for alternating 
 current, but in later years wattmeters, also, for alternating 
 current. There is a large number of other meters, but 
 these four are the pioneers in the meter field. In each 
 meter different principles are involved, which meet more 
 or less successfully the requisite of a good commercial 
 meter. Heat and cold, moisture, dust, insects and vibra- 
 tion are some of the physical obstacles to be met and over- 
 come by the successful meter; and short circuits, over- 
 loads, light loads and no loads are some of the electrical 
 conditions under which it must operate. The wearing 
 qualities of all of the elements of the meter must be taken 
 into consideration. As in the one-horse chaise, each part 
 of the meter should be as strong as any other part. The 
 ability to withstand the test of time; the property of 
 being the same meter at the end of one year's use, is the 
 real test of the value of an instrument. The problem, 
 then, becomes one of combining certain elements in such 
 relation to each other as to form a perfect measuring 
 device, and the maintaining of this same relation in the 
 face of the various conditions which are partially enumer- 
 ated above. However clearly these principles were 
 recognized twenty years ago, the state of the art was such 
 that only very imperfect devices were produced as meters. 
 It has been only within the last few years that com- 
 mercial meters have been brought to a reasonable state of 
 efficiency. The vast interests involved have quickened 
 the managements of central stations to the importance of 
 having better meters and meter service, and the field of in- 
 
MEASUREMENT OF POWER. 9 
 
 vention is so attractive along these lines that we see each 
 year a gradual improvement in the design and performance 
 of the meters placed on the market. 
 
 The measurement of any form of transmitted energy is 
 always attended by a diminution of the energy trans- 
 mitted equal to the amount consumed in the measuring 
 device. 
 
 Hence all instruments that measure transmitted energy 
 are shunt instruments, the power shunted or diverted from 
 the transmission being usually a very small fraction of the 
 total power. We may class all such devices as shunt 
 dynamometers to distinguish them from absorption 
 dynamometers, which usually translate the energy to be 
 measured into heat or mechanical work and destroy its 
 original properties. 
 
 The measurement of electrical energy which is trans- 
 mitted is accomplished by means of shunt dynamometers. 
 Its translation into useful commercial forms is accom- 
 plished by absorption dynamometers of various cjasses, 
 such as motors and incandescent lamps. 
 
 For the sake of analysis the flow of electrical energy is 
 usually considered by treating its two components, quan- 
 tity and intensity, as individual entities, and instruments 
 are constructed which will measure each separately as 
 well as the product of the two. 
 
 In commercial circuits the voltage or intensity usually 
 remains constant, in which case a measurement of the 
 power flowing in a circuit may be obtained by the record- 
 ing of the variable quantity of current in amperes and 
 multiplying by the known voltage. If the amperes remain 
 constant, and the voltage vary, a record can be obtained 
 of the power flowing by recording the variable voltage and 
 multiplying by the known amperes of the circuit. If both 
 
10 AMERICAN METER PRACTICE. 
 
 the voltage and the amperage of the circuit be variable 
 their product gives the power flowing, which, for conveni- 
 ence, is usually indicated by one instrument called a watt- 
 meter. 
 
 In the first class of instruments, the ammeter class, the 
 loss in energy due to the insertion of the instrument in the 
 circuit is directly proportional to the current flowing. 
 
 The voltmeter, if on a constant potential circuit, has a 
 constant energy consumption The wattmeter has a con- 
 stant energy consumption in its potential circuit plus a 
 variable energy loss in the series circuit proportional to the 
 current flowing. 
 
 From the foregoing the deduction follows that on con- 
 stant potential circuits of known intensity the measure- 
 ment of the energy passing is the most economically carried 
 out by an ammeter, on circuits of constant and known 
 amperage with variable voltage by a voltmeter, and 
 where both current and voltage are variable by a watt- 
 meter. 
 
 These general statements hold true for the measurement 
 of the power in a direct current circuit, or a non-inductive 
 alternating circuit, but must be modified for inductive 
 alternating circuits containing reactance and capacity, 
 the effects of which will be brought out in Chapter II. 
 
 It is assumed that the reader is familiar with the well- 
 known forms of indicating ammeters, voltmeters and 
 wattmeters on the market, the descriptions of which are, 
 therefore, omitted. 
 
 Many of these instruments record the passing power of a 
 circuit with extreme accuracy, but being simply indicators 
 of the instantaneous values of the power they do not give a 
 summation of this power during the elapsed time.* 
 
 *See Flather's Dynamometers and Measurement of Power. 
 
MEASUREMENT OF POWER. 11 
 
 In order that an instrument may bring in the element of 
 time in the measurement of power, it must constantly 
 move some object, for one unit of energy passing, through 
 unit distance in unit time; and this movement must be 
 permanently recorded by means of a dial train, or some sim- 
 ilar summation device. Then, when n units of energy are 
 passing, the recording object will be moved through n 
 units of distance in unit of time. The application of this 
 principle to commercial recording meters has resulted in 
 
 FIG. 1. 
 
 a great number of types, the most prominent of which 
 are described in succeeding chapters. 
 
 The various classes of electrical energy to be measured 
 commercially resolve themselves into: 
 
 Direct current as distributed by two or three-wire 
 systems. 
 
 Alternating current, single and polyphase systems. 
 
 Direct current distributed by means of a five-wire 
 system is also used to some extent, but, as the principles 
 involved in its measurement are identical with those of the 
 two and three- wire systems, the five- wire system will not 
 be considered separately. 
 
 Two-wire direct current systems are usually confined to 
 500 volt, and arc circuits. The three- wire system is the 
 well known Edison system, the local distribution of which 
 may be either two or three- wire. 
 
 The amperes flowing in the simple two-wire circuit with 
 load at C, Fig. i, are the same in every part of the circuit. 
 
12 AMERICAN METER PRACTICE. 
 
 An ammeter placed at A reads the same as if placed at B, 
 but the energy in the circuit varies as the distance from D, 
 becoming less as C is approached, and is directly propor- 
 tional in a circuit of uniform resistance to the distance from 
 D. Hence the mean power in the circuit would be found at 
 some point between D and C. 
 
 It is customary in an ordinary two-wire service to meter 
 the energy passing at A, and, if the fall of potential, or loss 
 of energy, between A and C be small, the error is not great, 
 the allowable loss usually not exceeding two per cent., 
 hence one per cent, less energy than that recorded at A 
 would represent the mean energy passing in the given 
 circuit. 
 
 As the meter in commercial practice is placed at the en- 
 trance of the service into a building, the consumer usually 
 pays for about one per cent, more energy than he actually 
 receives at his translating devices. 
 
 Should the meter be a recording ammeter, one leg of the 
 circuit only passes through the instrument ; if a wattmeter 
 be used the other leg of the circuit must be brought into the 
 meter to enable the potential coils of the instrument to be 
 energized. 
 
 In a recording wattmeter, the product of the magnetiza- 
 tion set up by the series and shunt field coils must exert a 
 resultant torque on some movable member of the meter 
 which will cause it to move in a manner to record n units of 
 energy in n units of time. 
 
 The movable members of a recording instrument have, un- 
 fortunately, an appreciable amount of friction which intro- 
 duces an inertia factor operating as a counter torque to the 
 product of the energy torque by a function of time. This 
 friction of the moving parts is small, and, at full load of the 
 meter, the ratio between the counter torque, due to friction, 
 
MEASUREMENT OF POWER. 13 
 
 and the energy torque is negligible. As the load decreases, 
 the ratio between the counter frictional torque and the 
 energy torque approaches unity; in other words, there is a 
 certain point in every recording instrument where the 
 energy torque and frictional torque are equal. When this 
 point is reached the instrument remains at rest. 
 
 The counter frictional torque varies in different instru- 
 ments of the same and different makes from a fraction 
 of one per cent, to three or four per cent, of full load 
 energy torque, the large losses often being due to causes 
 arising in service. 
 
 For this reason, it is customary in commercial recording 
 instruments to compensate for this friction by introducing 
 
 >b 
 
 FIG. 2. 
 
 an additional torque which is a definite fraction of full load 
 value, and which compensates through all ranges of load 
 for the initial friction of the meter. 
 
 Theoretically this is a very pretty way of accomplishing 
 the result, but exactions of service introduce frictions 
 which this compensation does not entirely eliminate. 
 
 The power passing in a three-wire circuit is usually 
 measured by metering the two outside legs, although equally 
 good results can be obtained by metering one outside leg 
 and the neutral conductor. 
 
 In the three-wire circuit, Fig. 2, let A, B, D, C be the 
 
14 AMERICAN METER PRACTICE. 
 
 outside legs, E the neutral conductor, and b and c 
 translating devices. 
 
 When the currents in A B and D C are equal no current 
 flows in E ; when they are unequal E carries the difference 
 positive or negative. The energy flowing on each side of 
 the system can be measured by two two-wire meters, or one 
 three-wire meter. In either case, the series coils carry the 
 varying amperes of each side of the system, while the shunt 
 field coils are energized proportionately to the voltage of the 
 system. 
 
 If two two-wire meters be used, the same connections 
 are made as in a two-wire system. If a three-wire meter 
 be used, the two series fields are superposed to form a re- 
 sultant field proportional to the current flowing in both 
 sides of the system. The torque due to this resultant field 
 and the field set up by the potential circuit operates the 
 meter in such a manner as to record the total energy 
 passing in the three-wire circuit. 
 
 A form of meter which is not manufactured commerci- 
 ally, so far as the author is aware, can be constructed to 
 measure the total energy flowing in a three-wire circuit by 
 placing one series coil of an ordinary three-wire meter in 
 one outside leg of the circuit and a coil of half the number 
 of turns in series with the neutral. The power registered is 
 multiplied by two. 
 
 To illustrate, suppose A B, Fig 2, to carry 50 amperes, 
 positive, and D C 2$ amperes, negative, then the neutral 
 wire, E, carries 25 amperes, negative. As the neutral field 
 coil has half the number of turns, the sum of the magnetiza- 
 tions of the series fields is equal in effect to a positive cur- 
 rent of 37.5 amperes. Twice this amount is 75 amperes, or 
 the same as the series fields of the ordinary three-wire meter 
 would carry in the example given. 
 
MEASUREMENT OF POWER. 15 
 
 The PR losses in the fields of a meter at full load reach 
 an appreciable amount, but at light loads are not of much 
 consequence, hence, in the form of meter just given, on a 
 properly balanced three-wire service the PR losses would 
 be halved as the neutral would carry no current. On 
 unbalanced loads the loss in this form of meter would be 
 reduced by some quantity varying between the limits of 
 J and J, so that an average saving in PR losses in the 
 field coils for any meter would be 12 \ percent. In reality 
 this would be much larger, as any serious unbalancing 
 would scarcely take place except at light loads, when the 
 PR losses are negligible. 
 
 The shunt field coil, where a single meter is employed, 
 can' be fed from across the outside or between the neutral 
 and one outside. In either arrangement, unless the volt- 
 age be equal on both sides of the system, an error is intro- 
 duced into the record proportional to the difference of 
 voltage existing. To obtain a true record, if the voltage be 
 unequal, it is necessary to employ two two-wire meters or a 
 three-wire meter having two armatures fed respectively 
 from each side of the system. 
 
 The latter arrangement amounts to the combining of two 
 meters under one cover, resulting in a clumsy and expensive 
 device. In commercial practice the voltage on one side of 
 a three- wire distributing system is usually maintained 
 about equal to that on the other; the errors resulting 
 from inequalities of the sides supposedly balance, so that 
 it is the usual practice to meter with a three-wire meter 
 with the shunt field coils fed between the two outsides 
 or between one outside and neutral. 
 
 In meters of the class described, the torque urging the 
 rotary part at any given moment is proportional to the 
 energy of the current at that moment, while, to record 
 
16 AMERICAN METER PRACTICE. 
 
 correctly, the speed must be proportional to the energy, 
 and, consequently, to the torque. This result is secured by 
 providing a breaking device in which the resistance offered 
 to rotation is proportional to the speed. 
 
 This brake may be of various constructions, but the 
 usual form is that of a conducting non-magnetic metallic 
 disk revolving between the poles of a permanent magnet. 
 The eddy currents generated in the disk, by their reaction 
 on the permanent magnetic field, exert a resistance to 
 movement which varies in amount proportionately to the 
 speed of the torque exerted by the moving member to 
 which the disk is fixed. Hence, the power necessary to gen- 
 erate the eddy currents becomes proportional to the torque 
 exerted on the moving member or the watts passing in the 
 circuit. Other forms of brakes are employed, such as fans 
 and liquid brakes, but they have given place in the latest 
 types of meters to the now universal magnetic brake. 
 
CHAPTER II. 
 Measurement of Power in Alternating Current Circuits. 
 
 In passing to the measurement of alternating currents, 
 simple two and three-wire non-inductive single-phase 
 circuits will be first considered. 
 
 The power flowing in a non-inductive two-wire circuit is 
 at any instant the product of the instantaneous values of 
 the current and voltage. To get the mean power the in- 
 
 FIG. 3. 
 
 stantaneous values must be integrated through a complete 
 cycle. Graphically we may represent such a power as in 
 Fig. 3, wherein e represents the sine wave of varying voltage 
 values for a single period and i the varying ampere values 
 for the same period in phase with the voltage. The curve 
 p then represents the varying power values, and is obtained 
 by multiplying the corresponding ordinates of the voltage 
 and current curves. 
 
 17 
 
18 AMERICAN METER PRACTICE. 
 
 The product of the mean effective volts and amperes 
 gives the mean effective watts passing in the circuit or the 
 mean value of the power curve p, hence, for all non-in- 
 ductive circuits, Pe= true watts. 
 
 If the current flowing in curve i be passed through a coil 
 of wire and another coil be energized by a current pro- 
 portional to and varying with the voltage curve e, the two 
 coils, if so placed, will form a resultant magnetization for any 
 instant of time proportional to the product of the instanta- 
 neous values of e and i or the watts passing in the circuit. 
 
 If a movable member or armature be provided, of such 
 character that it is moved by the operation of this resultant 
 magnetization in such manner that its speed is propor- 
 tional to the watts passing in the circuit, a record of the 
 power passing is obtained. In recording the flow of 
 alternating current energy this principle has been used in 
 the development of two classes of meters, which for con- 
 venience are styled inductive and non-inductive. 
 
 A non-inductive meter can be used for either direct or 
 alternating current. A familiar type is the Thomson 
 recording wattmeter. In this type of meter the field 
 coils are in series with the current flowing, the shunt field 
 coils are energized proportionately to the voltage of the 
 circuit, and the torque exerted on the armature is for any 
 instant proportional to the product of the instantaneous 
 values of the current and E. M. P., hence the mean torque 
 is proportional to the product of effective volts by effective 
 amperes, or the true watts passing in the circuit. 
 
 The method of connecting this meter is the same as on 
 direct current circuits, and, in fact, is interchangeable with- 
 out recalibration, the error being for practical purpose 
 inappreciable. 
 
MEASUREMENT OF POWER. 19 
 
 The induction meter is either a "current" meter or 
 wattmeter, and operates by means of a "shifting" field 
 acting on a closed secondary armature, the eddy currents 
 induced in the armature reacting on the shifting or rotating 
 field in such manner as to produce a torque proportional to 
 the energy or current passing, according to whether it is a 
 "current " or wattmeter. 
 
 A well-known type of current meter is the Shallenberger 
 in which the series current passes through a primary coil 
 set at an angle with a closed secondary within it, the plane 
 of the two magnetizations being at a like angle and differ- 
 ing in phase. Mounted in inductive relation is a thin disk 
 armature in which eddy currents are generated by the 
 shifting or rotating field. The reaction of these resultant 
 fields tends to revolve the armature, the torque being 
 proportional to the square of the current. 
 
 A fan brake, whose retarding effect increases as the square 
 of the speed, is attached to the spindle on which the arma- 
 ture is carried, and enables the meter to record the current 
 flowing. 
 
 In the wattmeter, of which there are many types, the 
 series and shunt field coils are wound in inductive relation 
 to a movable closed secondary or armature in which eddy 
 currents are induced proportional to the magnetizations 
 set up in the series and shunt field coils. The shunt field 
 coils are displaced in phase from the series coils by having 
 a reactance placed in series with them, the magnetizations 
 thus creating a shifting or rotating resultant field. 
 
 The field set up by the induced eddy currents in the re- 
 volving secondary lags behind the resultant field set up by 
 the series and shunt field coils, creating a torque on the 
 secondary which is proportional to the square of the 
 energy flowing. The providing of a brake, whose retarding 
 
20 
 
 AMERICAN METER PRACTICE. 
 
 force varies as the square of the speed, enables the energy 
 passing to be obtained from the speed of the revolving 
 spindle to which the armature is attached. There are 
 many variations and different arrangements of the fore- 
 going general principles. 
 
 The power passing in an inductive single-phase circuit 
 may be graphically illustrated by Fig. 4, wherein the 
 current lags behind the voltage by 60. 
 
 FIG. 4. 
 
 The resultant power curve p represents the watts passing 
 
 /? A 
 
 in the circuit, and is written Pe = cos. 
 
 angle of lag. 
 
 The power passing in an alternating circuit at any 
 instant is equal to the product of the instantaneous values 
 of the current and voltage, etc. (see American Electrician, 
 March, 1902), and acts as a generator for a portion of each 
 period, the useful power being the difference between 
 positive and negative values of the curve p. 
 
 The resultant torque, therefore, exerted on the coils of a 
 meter in circuit with an inductive load is not proportional 
 to the product of the virtual volts and virtual amperes as 
 measured by an indicating instrument, but to this product 
 multiplied by the power factor of the circuit; the power 
 factor is the ratio between the true and apparent watts. 
 
MEASUREMENT OF POWER. 21 
 
 In the metering of three-wire single-phase systems, in- 
 ductive or non-inductive, the same general principles out- 
 lined in the foregoing hold true. 
 
 Multiphase systems are a combination of single-phase 
 systems differing in phase from each other, and are used in 
 distributions for light and power owing to the ease with 
 which motors can be operated thereon. 
 
 The systems in general use are the quarter-phase and 
 three-phase, the former composed of two single-phase 
 circuits differing by 90 and distributed by either four or 
 three-wires, the latter of three single-phase circuits differing 
 by 120 and distributed by either four or three- wires. 
 
 The quarter-phase system, employing four wires for its 
 distribution, is metered exactly as two separate single- 
 phase circuits would be, the total power in the two-phases 
 being the sum of the energy components of each phase. 
 When the two phases are balanced, that is, have the same 
 amount of energy passing in each phase, it is only neces- 
 sary to meter one phase and multiply the result by two to 
 get the total amount of energy passing. When the phases 
 are unbalanced, that is, have different loads on each phase, 
 two meters are necessary, one in each phase, to register the 
 energy passing. 
 
 In the commercial distribution of this system it is the 
 usual practice to take a lighting circuit into a customer's 
 premises from one phase only and a power circuit from 
 both phases. The motor furnishes a balanced load and 
 needs only one meter and the lights one meter. In this 
 way, by loading up the phases equally, the number of 
 meters needed for metering the lights can be kept down 
 to one meter per customer. 
 
 When three wires are used in the distribution of two- 
 phase currents, one wire acts as a common return for the 
 
22 
 
 AMERICAN METER PRACTICE. 
 
 other two, and the instantaneous value of the current on 
 the common return at any instant is the algebraic sum of 
 the instantaneous values of the currents of each phase. 
 
 The virtual resultant value would be the hypothenuse 
 of a right triangle which would represent the current in 
 phase and value. The resultant voltage is found in the 
 same way, hence the virtual watts would be the product 
 of the resultant current by the resultant voltage for a non- 
 inductive circuit. 
 
 As an example, suppose we have a two-phase system 
 carrying 10 amperes in each phase at a voltage of 100 v., 
 
 10 Amp. 
 
 TO. V. 
 
 FIG. 5. 
 
 FIG. 6. 
 
 the triangle of the current values, Fig. 5, would give a 
 resultant amperage of 14.14, the triangle of the voltage 
 values, Fig. 6, gives a resultant voltage of 141.4 volts. 
 The total watts passing in the circuit is the product of 
 these two resultants, or a watt value of 2,000 watts. 
 
 Calculating the watts separately for each phase we have 
 10 amp., x 100 v. = 1000 watts (i) 
 10 amp., x 100 v. = 1000 (2) 
 
 2000 watts 
 which is the same as the above. 
 
MEASUREMENT OF POWER. 
 
 23 
 
 The resultant voltage and resultant amperes being the 
 same in phase when the load is balanced, a meter, Fig. 7, 
 with its series field coils placed in the common return and 
 
 FIG. 7. 
 
 its shunt field coils energized by the resultant voltage of the 
 system, would give the true power passing in the circuit. 
 
 When the load is unbalanced the resultant ampere value 
 leads or lags behind the resultant voltage of the system 
 and may be displaced in phase either way by 45 on non- 
 inductive load. The true power in the circuit is then 
 obtained by multiplying by the cosine of angle of lag or 
 lead. 
 
 FIG. 8. 
 
 As a wattmeter records the product of the instantaneous 
 values and not the maximum values of current and voltage, 
 the record is a true one of the energy passing in the circuit. 
 
24 
 
 AMERICAN METER PRACTICE. 
 
 This connection, Fig. 7, will not record the true power 
 passing on an unbalanced inductive circuit, as an angle of 
 lag or lead of 45 could give a wattless register, allowing 
 70 per cent, of the energy passing to remain unrecorded. 
 On balanced circuits, however, feeding induction motors, 
 this connection readily takes the place of the two meters 
 shown in Fig. 8, where the series field coils are connected 
 in the outside legs of the circuit and the potential circuit fed 
 between the given leg in the meter and the common return. 
 
 When the induction type of meter is used, the potential 
 circuit is so connected as to differ in phase from the 
 A 
 
 FIG. 9. 
 
 series field circuit by 90, hence in Fig. 8 the potential 
 circuit for each meter would be fed from the given leg in the 
 meter to the other outside. This is the connection most 
 commonly used in metering two-phase circuits. The 
 elements of two meters are often combined under one 
 cover to act on a common armature. The elements are 
 connected in the same manner as though they were separate 
 meters, and need not be further elaborated. 
 
MEASUREMENT OF POWER. 
 
 25 
 
 Three-phase circuits are connected in "star" or "delta" 
 grouping; the former sometimes uses four wires in the dis- 
 tribution while the latter is always distributed by means of 
 three wires. 
 
 In Fig. 9, the "star" grouping with common centre at o 
 is shown, the voltage of the three legs being represented 
 in intensity and phase by the lines a, 6, c. The resultant 
 voltage of such a system for any instant of time may be 
 found graphically by producing the single-leg of opposite 
 
 D 
 
 FIG. 10. 
 
 sign to the other two backwards, and finding the resultant 
 by the parallelogram of forces. The diagonal will represent in 
 sign and phase the resultant of the voltage on the three legs. 
 In Fig. 10 the resultant is \J '3 times the voltage of one 
 phase and leading it by 30. If one phase be selected as the 
 leading one, the resultant voltage will be found to be always 
 leading it by 30. In the same Fig. 10 in a balanced 
 circuit the current will have a resultant coinciding with 
 
26 AMERICAN METER PRACTICE. 
 
 the voltage resultant, and the product of the two resultants 
 will, for non-inductive load, represent the true power 
 passing. 
 
 The algebraic sum of the resultants of the voltage taken 
 consecutively between the three legs of a system will equal 
 for any instant the resultant positive or negative value of 
 the voltage at that instant. Likewise, the resultant of the 
 current flowing for any instant is the algebraic sum of the 
 three resultant currents of the system found in like manner. 
 Therefore, the current passing over any two legs of a three- 
 wire three-phase circuit is the algebraic sum of the three 
 resultant currents of the system, hence the power passing 
 in the circuit is the product of the resultant current and 
 voltage. The algebraic product of two minus quantities is 
 always a plus quantity, hence the power passing is always 
 positive in character on non-inductive circuits. On in- 
 ductive circuits, where the current leads or lags behind the 
 voltage, a minus quantity of current flows for a portion of 
 each period. 
 
 Some simple rules deduced from the above may be of 
 service in readily computing the energy passing in poly- 
 phase systems. 
 
 In "star" or "Y" connected systems, or any system 
 employing a common neutral, the resultant voltage is 
 equal to the effective voltage X \/m, m being the number 
 of phases. The resultant amperes flowing in any system is 
 the average of the sum X \fm- The energy passing is, 
 therefore, the common resultant voltage X common re- 
 sultant amperage X cos. tp or lag angle. Suppose, for 
 example, the amperes in the different legs of a three-phase 
 unbalanced system are 10, 20 and 30 respectively, and the 
 effective volts of the circuit, taken with regard to a common 
 neutral point or conductor, 100 volts. The average value 
 
MEASUREMENT OF POWER. 27 
 
 of the amperes is 20, and the resultant is 20 X \/3 == 34-^ 
 amperes. The resultant voltage is 100 X v / 3~ ==I 73 v lts. 
 The cosine of angle of lag q> is taken as .5, corresponding 
 to an angle of 60. 
 
 The energy passing in such a circuit is 34.6 amperes 
 X 173 volts X .5 = 3000 watts. The same result is ob- 
 tained by taking each leg separately, multiplying a-mperes 
 X effective volts X cosine q> = 60 amperes X 100 V. 
 5 X3 OO watts. 
 
 The common neutral point or conductor is not always 
 available, hence the advantage of the first method where 
 in practice the resultant voltage is the voltage between 
 any two lines. 
 
 Where the current lag angles in the different branches of 
 the circuit are of different values the resultant amperes 
 can be plotted graphically, or the energy of each leg com- 
 puted by multiplying the apparent watts found by the 
 power factor, and adding the amounts thus obtained to 
 get the total energy of the circuit. 
 
 When the neutral point of a star connected system is 
 extended to the distribution, the system becomes three 
 independent circuits with a common neutral, the algebraic 
 sum of the currents on the neutral for any instant being 
 zero. 
 
 The energy in a three-phase circuit may be measured by 
 one, two or three meters, or their equivalent, of inductive 
 or non-inductive type. 
 
 The non-inductive type of meter has maximum torque 
 when the shunt field coils and series field coils are in 
 phase. Therefore, as the lag angle increases the torque 
 diminishes, until at quarter-phase the torque is zero. 
 Hence, it is impossible in meters of this character to com- 
 bine series fields of different phases to act in conjunction 
 
28 AMERICAN METER PRACTICE. 
 
 with shunt field coils in phase with only one of them to 
 measure the energy passing. The case where only one 
 meter is used to measure unbalanced polyphase energy 
 factors is confined to meters of the inductive type where 
 one or more measuring elements are combined to give a 
 resultant measuring value to a common movable armature. 
 In Fig. n, let A B C represent a three-phase relation 
 feeding circuit having a balanced load, the series coils of a 
 meter D are then placed in either one of the three legs and 
 the shunt field coil fed by the virtual voltage of the system 
 
 FIG. 11. 
 
 obtained from a star connected set of resistances having a 
 common neutral point. This connection applies to either 
 star or mesh groupings. The register of the meter is 
 multiplied by three to obtain the total power flowing in the 
 circuit. That is, the power passing in each leg of the circuit is 
 
 In this instance Fig. n represents the connection for a 
 non-inductive type of meter. If an inductive meter be 
 used the series and shunt field coils must be displaced 90 
 in phase, and the meter fed as shown in Fig. 12, wherein the 
 potential circuit is fed between C B whose resultant voltage 
 lags 90 behind current in A. 
 
MEASUREMENT OF POWER. 
 
 29 
 
 One meter is sufficiently accurate for balanced three- 
 phase loads, but, when the load becomes unbalanced, two 
 legs of the circuit must be metered. 
 
 If non-inductive meters be employed we connect them 
 as shown in Fig. 13, where the current in A lags on non- 
 
 FIG. 12. 
 
 inductive load 30 behind the voltage in shunt coil d\ like- 
 wise, current in b lags 30 behind the voltage in e. The sum 
 of the two registers of the meter is equal to the resultant 
 voltage of the system by the resultant amperes. If the 
 current be inductive, the current in a lags cos. (<p -f 30) 
 
 FIG. 13. 
 
 behind the voltage; likewise, current in b lags cos. (<p - 30) 
 behind the voltage in e. In other words, the true watts equal 
 the resultant current X resultant voltage X cos. <p. The 
 power passing is equal to the algebraic sum of the two 
 
30 
 
 AMERICAN METER PRACTICE. 
 
 registers. It is easily seen that when <p = 60, the expression 
 cos. (cp + 30) is equal to cosine 90, which gives a wattless 
 register. Under such conditions one meter registers the 
 total current. If the lag angle be greater than 60 one 
 meter runs backward. 
 
 Instead of the two meters just shown in Fig. 13, one 
 inductive meter may be used which, in general, consists of 
 two elements combined under one cover, in short, two 
 single-phase meters so arranged with regard to a common 
 armature that the power passing is recorded on a single 
 
 FIG. 14. 
 
 dial train. There are many types of this meter, but in 
 principle they bear sufficient resemblance to the one shown 
 diagrammatically in Fig. 14 for it to represent them. 
 
 A is a closed secondary or armature in inductive rela- 
 tion to the series and shunt field coils H, H', S, S', being 
 displaced in phase by 90. The maximum torque is there- 
 fore exerted when the phase displacement is 90, and 
 diminishes as the angle decreases, becoming zero when the 
 two circuits are in phase. This is the condition of the 
 circuit having a lag of 90 or wattless current. 
 
 This action is just the opposite of the action in the non- 
 inductive meter, where the torque diminishes as the series 
 and shunt field coils become displaced in phase. The 
 
MEASUREMENT OF POWER. 
 
 torque of both meters varies, however, with the cosine law,, 
 but in opposite directions. Mesh connected systems are 
 metered in the same manner; in fact, the three-wire two- 
 phase Y and mesh connections three-phase are all metered 
 alike. 
 
 D 
 
 The metering of such systems as the monocyclic is 
 worked out along the same general lines, but presents some 
 peculiarities which are interesting. 
 
 In the monocyclic system a regular single-phase circuit 
 is used in conjunction with a third wire called the teaser, 
 which differs from the single-phase circuit by 90 in phase. 
 
 FIG. 16. 
 
 In the generator, one end of the teaser coil is tapped on 
 to the middle of the single-phase coil, and the other end 
 leads to a collecting ring. The relation of the E.M.F. in 
 the resulting three-wire circuit may be illustrated diagram- 
 matically in Fig. 1 5 , wherein A C is voltage of single-phase, 
 B its middle joint, D B teaser winding, A D and D C the 
 resultant E.M.F. between the two wires and the teaser. 
 
32 AMERICAN METER PRACTICE. 
 
 Such a system is adapted to the distribution of light and 
 power, the lights being connected to the single-phase in 
 the usual way, and the motors as illustrated in Fig. 16. 
 
 As indicated, the primaries of the two transformers are 
 placed in series with the teaser connected to the point of 
 joining. The secondaries are cross connected, forming a 
 resultant lop-sided three-phase relation which is used to 
 operate a three-phase motor in the same manner that a 
 regular three-phase system would do. 
 
 Another way of connecting up two transformers to accom- 
 plish the same results is illustrated in Fig. 17, wherein a 
 small transformer is shown connected between the teaser 
 
 Motor 
 
 FIG. 17. 
 
 and the middle point of the large transformer, the relation 
 of the current in the secondaries then being the same as 
 in the generator. 
 
 To meter the power passing in the connection Fig. 16, 
 we proceed exactly as in unbalanced three-phase systems, 
 or two-phase systems employing a common return. Hence, 
 with non-inductive meters we place one in each outside 
 leg and feed the potential circuit between the given leg 
 and common return. 
 
 With inductive meters the regular polyphase meter used 
 for three and two-phase systems is placed in circuit. As 
 we have pointed out, this meter usually consists of two 
 
MEASUREMENT OF POWER. 33 
 
 elements combined under one cover and acting on a com- 
 mon armature. 
 
 When a service consisting of motor and lights is fed from 
 the connection shown in Fig. 17, the proper registration of 
 the energy passing is accomplished by using two non-induc- 
 tive meters connected as shown in Fig. 31, Chap. VI, or a 
 polyphase induction meter so designed as to have one of 
 its elements connected in circuit with the teaser circuit 
 as outlined diagrammatically Fig, 31, Chap, VI. 
 
CHAPTER III. 
 
 Meter Selection Requisites of a Good Meter Commercial 
 Considerations. 
 
 Considerations other than cost render the selection of 
 the best meter a difficult task, owing to the various classes 
 of service to be provided for. In commercial distributions 
 of current we have incandescent lamps, arc lamps and 
 motors to meter beside other forms of energy consumers, 
 such as electric heaters, etc. Frequently all of these forms 
 of translating devices are found in one installation, and it 
 is desirable to meter them through one meter. The current 
 may be direct or alternating, or both, and the various 
 conditions that the meter has to meet will be treated 
 separately. 
 
 First. Varying Voltage and Load. 
 
 If a constant load at a constant voltage always passed 
 through the meter, its calibration would be a simple matter, 
 but, in the majority of installations, the variation is between 
 the limits of an eight c. p. lamp and the full capacity of 
 the meter. The meter must record the true watts passing 
 at all loads, and its accuracy on light loads is extremely 
 important. It is seldom that a good meter cannot be 
 relied upon to record its full load within one or two per 
 cent, of its correct value, but this same meter may be 50 
 per cent, slow on light loads and fail to record at all on an 
 eight c. p. lamp. In large systems of distribution the 
 voltage remains very constant, but in outlying districts, 
 
 34 
 
METER SELECTION. 35 
 
 and in small plants with insufficient copper in the lines, the 
 voltage often varies ten per cent, either way from its 
 normal value. The meter must record accurately under 
 these variations of voltage. Light load accuracy has a far 
 reaching effect on the earning capacity of the plant, and 
 is so important that it is taken up more fully in Chapter 
 VIII. 
 
 Second. Varying Frequency. 
 
 Only in a few of the larger plants are the alternating 
 generators operated in parallel, and in many of the smaller 
 plants it is often the case that two circuits from different 
 machines may vary ten per cent, in frequency, owing to 
 different speeds of the generating units. Meters in service 
 are often called upon to register at slightly different 
 frequencies from the one they are designed for. This 
 irregularity may also be caused by the varying frequency 
 of a circuit caused by change of speed of the generating unit. 
 The meter must record accurately under these conditions 
 the total watts passing in the circuit, or fail to meet the 
 requirements of commercial service. 
 
 Third. Power Factor. 
 
 When a circuit is inductive it is said to have a power 
 factor or a ratio of the true to the apparent watts passing 
 in the circuit. Many circuits are purely inductive, such as 
 those feeding induction motors, but a single installation 
 may have connected incandescent lamps, or non-inductive 
 load, as well as arc lamps and motors. The power factor 
 under these conditions can vary from .55 to unity. The 
 meter is called upon to register accurately throughout this . 
 range. These conditions are not at all unusual, but are.- 
 met in practice every day. 
 
36 AMERICAN METER PRACTICE. 
 
 Fourth. Sine Wave Form. 
 
 The changes of direction of alterations of a current vary 
 in intensity from positive to negative according to the sine 
 law. The variation approaches the true sine wave, but 
 many wave forms met in practice are jagged, peaked or 
 flat. A meter for commercial use must fit any wave form; 
 that is, record as accurately on one form as another. 
 
 Fifth. Short Circuits. 
 
 The effect of a short circuit on a meter is one that should 
 be carefully determined. The instantaneous rush of current 
 is great and creates a strong magnetic field which extends 
 to the permanent magnets and tends to demagnetize them 
 unless a shield is interposed between them. This shield 
 may be of iron, but a closed secondary band is sometimes 
 placed around the field coil which sets up a counter magnetic 
 iield when a short circuit takes place. Another effect of 
 :short circuit on meters where a disk or drum armature is 
 msed is to distort or bend the disk by the powerful moment- 
 ary thrust exerted which acts exactly like a blow. This 
 is prevented in some meters by evenly distributing the 
 field coils in such a way as to balance this thrust. 
 
 Sixth. Permanency of Magnetic Drag. 
 
 If the meter stand successfully the short circuit test 
 without affecting the permanence of the magnets, it is 
 proof against immediate deterioration, but the action of 
 time may still be detrimental. Some essentials, however, 
 of correct form and mode of manufacture will be of service 
 in judging the magnets. A very hard steel, usually an 
 alloy of tungsten, is highly tempered and magnetized by 
 contact with powerful electro magnets and partially de- 
 magnetized by weak alternating current magnetizations. 
 
METER SELECTION. 37 
 
 This process is repeated until the steel is brought to a state 
 of permanent magnetization. This state of permanent 
 magnetization has been determined by experiment, and 
 each manufacturer knows from the quality of steel used 
 just the balancing line where the magnets will not lose or 
 gain in magnetism. Conditions of length and area and 
 length of air gap affect permanence. The magnet 
 should be long and the pole pieces of sufficient area to allow 
 of a good distribution of the magnetic flux. The length 
 of the air gap should be small, otherwise the magnetic flow 
 is retarded and the magnet is unable to maintain perma- 
 nently a flow across the gap. 
 
 Where the magnets have been forged, the manufacturer 
 usually lets them set to assume their final shape. Even 
 then the distance between the poles may change slightly 
 and the accuracy of the meter be destroyed. Any meter, 
 therefore, which combines long magnets, short air gap and 
 permanently fixed pole pieces has an advantage over one 
 lacking these points. 
 
 Seventh. Torque. 
 
 The torque of a meter is directly proportional to its 
 efficiency as a motor. In other words, suppose between the 
 shunt and field losses it takes ten watts to operate the 
 meter. These ten watts are the input into the motor meter, 
 and the efficiency of this motor meter will give the ratio 
 of the torque exerted on the armature. A good meter, then, 
 should be as highly efficient a motor as possible to enable 
 the losses to be made as small as is compatible with 
 efficiency. 
 
 When we consider the counter frictional torque of a 
 meter, the value of having the current torque as large as 
 possible is immediately apparent. The torque of a meter 
 
38 AMERICAN METER PRACTICE. 
 
 is also its measure of ability to overcome friction, namely, 
 if the ratio of frictional and current torque be i to 100, the 
 doubling of the friction would only cause a loss of one per 
 cent, additional, whereas, if the ratio were i to 10, an ad- 
 ditional loss of ten per cent, would be experienced when 
 the friction is doubled. Hence, the securing of as large a 
 torque as possible, without too great energy losses, is a 
 matter of careful design and of extreme importance. 
 
 Eighth. Friction and Friction Balancing. 
 
 The support of the armature and the dial train are the 
 points of friction in a meter, and in many meters are 
 balanced by an initial turning impulse imported to the 
 armature. The turning impulse remains constant, but the 
 frictional equivalent is continually varying, not only in one 
 location, but for various locations of the meter. As this 
 friction changes with wear, we find the meters creeping in 
 some localities and very slow in others, owing to the vary- 
 ing degrees of vibration present. Many forms of friction 
 balance are used, but none are perfect, and the true solu- 
 tion of the problem lies in reducing the counter frictional 
 torque to a minimum and raising the current torque to a 
 maximum, so that the error due to friction is inappreciable. 
 In selecting a meter, therefore, a careful determination of 
 this ratio is all important. 
 
 Ninth. Energy Losses. 
 
 Work cannot be done without consumption of energy, 
 and a balancing ratio between the energy needed and the 
 torque desired must be reached in the determination of the 
 allowable losses in the meter. 
 
 The loss in the potential circuit ranges from i J to 15 
 watts in different makes of meters. Only a part of the 
 
METER SELECTION. 3D 
 
 energy is available for useful work ; the remainder is given 
 off in heat. The field losses vary with the load, but should 
 never exceed one per cent, of the rated capacity of the 
 meter on full load. 
 
 Large losses in the shunt and field circuits are, in poorly 
 designed meters, not compensated for by a resulting large 
 torque, or in other words, the efficiency of the motor meter 
 is very low. The determination of this efficiency is easily 
 made by measuring the pull of the disk by means of a 
 sensitive spring balance when a given load is on the meter. 
 The reduction of the watt losses of the meter to their 
 mechanical equivalent at the point of attachment of the 
 spring balance will give the ratio between the actual work 
 done and the energy consumed. The meter possessing the 
 highest efficiency and largest torque has a much better chance 
 of remaining permanently accurate, provided its counter 
 frictional torque is lower than in one that is less powerful. 
 
 Tenth. Armature Support and Vibration. 
 
 The critical point in the permanence of the constants 
 of a meter is the insurance of a permanent non-frictional 
 support for the armature. Much time and money have 
 been expended in this development, and the two outcomes 
 are a highly polished jewel, on which runs the hardened 
 steel point of the armature spindle, and magnetic flotation 
 of the armature. 
 
 The jewel and hardened steel point can hardly be called 
 a permanent solution of the problem, as, under certain con- 
 ditions, the jewel wears and breaks down very rapidly, 
 destroying at the same time the smoothness of the steel 
 shaft. 
 
 The causes of jewels breaking down are vibration and 
 dust. It is almost impossible to eliminate vibration of the 
 
40 AMERICAN METER PRACTICE. 
 
 meter support in some locations. The passing of cars and 
 wagons, and heavy jars given to the building, jolt the meter, 
 causing the armature shaft to strike a blow on the jewel 
 which frequently cracks and very quickly destroys its 
 polished surface. On alternating current circuits the 
 process is even more disastrous in its final results. The 
 slight vibration or tremor in the armature drum or disk 
 imparted to it by the shunt winding, which is always in 
 circuit, hammers the jewel at the rate of the alternations 
 of the service circuit. The millions of minute blows gradu- 
 ally destroy the smoothness of the jewel and it finally 
 breaks down. Magnetic flotation of the armature not 
 only cures this evil, but reduces the initial friction of the 
 meter to a large extent. The armature is floated in space 
 by means of a permanent magnetic system, and any vibra- 
 tion which produces a lateral thrust has no effect on the 
 meter. A meter possessing magnetic flotation of the 
 armature and fulfilling the previous conditions, if it meet 
 the requirements of the service, comes nearer the ideal 
 conditions than any other. 
 
 The presence of dust or grit in the meter grinds the 
 polish off the jewel and quickly wears it out. In magnet- 
 ically floated armature guides this dust would be even more 
 disastrous. 
 
 Eleventh. Air Tight. 
 
 To keep out dust, insects, acid fumes and all foreign 
 matter the meter must be air-tight. The carelessness of 
 meter manufacturers in this respect has been appalling, 
 but the responsibility for this defect rests ultimately upon 
 the meter buyer, as he should insist on this point. 
 
 In order that the meter may retain its calibration it must 
 have the same conditions present in service that it had when 
 
METER SELECTION. 41 
 
 installed, and the exclusion of all matter which can in any 
 way change the relation of the different elements to each 
 other must be insured. 
 
 Twelfth. Temperature Co-Efficient. 
 
 The temperature co-efficient should be unity through a 
 wide range. The range of temperature over which meters 
 work during a year's service is as high as 100. Meters 
 tested hot and cold often show as much as five per cent, 
 difference in their reading, owing to the different resistance 
 of the windings at different temperatures. 
 
 Thirteenth. Insulation. 
 
 The failure of many meters is caused by defective insula- 
 tion. A meter is easily tested for insulation, and should 
 be rejected unless it can show several megohms between the 
 windings and meter frames. 
 
 Fourteenth. Mechanical Features. 
 
 An eminent engineer once said that if a machine look 
 right it is almost sure to be right in design, probably draw- 
 ing this conclusion from a sense of the fitness of things. All 
 the electrical conditions maybe successfully met by a meter, 
 but its mechanical design may be so poor as to make it 
 worthless. Besides the elimination of friction of the mov- 
 ing parts, it is an important consideration that the general 
 get-up of the meter should be pleasing and carefully worked 
 out. Meters receive a great deal of handling, and fre- 
 quently rough handling. The mechanical details need not 
 be elaborated, but the proper construction of the meter 
 frame and parts should be worked out along the lines of 
 strength and rigidity rather than lightness. 
 
42 AMERICAN METER PRACTICE. 
 
 The commercial problems involved in meter selection are 
 complicated by reason of the great variety of service fur- 
 nished. Following close upon the determination of a meter's 
 electrical and mechanical excellence come the questions of 
 first cost and operation, which vary with the kind of service 
 to be metered. The different classes of service which may 
 be met in practice, generated and sold by one company, are 
 as follows : 
 
 1. Plain direct current two-wire systems of 115, 220 or 
 500 volts. 
 
 2. Three-wire direct current systems of 115-230 and 
 220440 volts. 
 
 3. A combination of either one or two, with a single- 
 phase two-wiring alternating current system. 
 
 4. A combination of either one or two, with a polyphase 
 system. 
 
 5. Plain single or polyphase distributions for light and 
 power. 
 
 6. Series arc systems in conjunction with any of the 
 above. 
 
 Many more combinations might be given, but the above 
 are the most common and will serve as illustrations. To 
 meet the first case a two-wire direct current meter of 
 either the chemical or motor type is what is needed. The 
 most extensively used meters for this service in America 
 have been the Edison chemical meter and the Thomson 
 recording wattmeter. The chemical meter is rapidly 
 passing into disuse. In principle it is an ampere hour- 
 meter possessing many points of value, but, owing to the 
 immense amount of work connected with its operation, it 
 has been steadily losing ground to the Thomson or corn- 
 mutated type of meter. For many years the Edison 
 chemical meter has been the mainstay of direct current 
 
METER SELECTION. 
 
 stations, and, when carefully operated, furnished a very ac- 
 curate register; in fact, in many instances a far more 
 accurate record can be obtained with it than with any 
 mechanical meter. An incident is recalled within the 
 writer's experience where the vibration was so great that 
 special screw clips were designed to hold the bottles in 
 circuit, the conditions being such that any mechanical 
 meter would have been ruined in less than half a day. 
 
 The Thomson and Duncan commutated meters are 
 practically, at the present writing, the only direct current 
 meters in use in America, although in the near future there 
 promise to be other forms. At the present there are only 
 two selections to make, either the chemical or commutated 
 types. Modern practice has decided in favor of the com- 
 mutated type. 
 
 The second case falls under the same limitations as the 
 first and may be dismissed with the same considerations. 
 
 For the sake of simplification we will confine ourselves 
 in the selection of the best meter to fulfill the conditions of 
 the third class to motor types of meters. 
 
 Here we have a varied class of service, direct and alter- 
 nating current, and we must bear in mind that it is easier 
 and better to have but one type of meter in service. 
 
 Let us assume, as one phase of case three, a low tension 
 Edison net work with an outlying district fed by alternat- 
 ing current. Then, on all sides where the two districts meet 
 there will be lapping over of the direct and alternating 
 service. If two types of meters be used, which are not 
 interchangeable, the meter service is not as flexible as it 
 would otherwise be. It would also necessitate the carry- 
 ing of a much larger stock of meters to meet all demands. 
 Again, the service to any installation could not be changed 
 from direct to alternating or vice versa without changing 
 
.44 AMERICAN METER PRACTICE. 
 
 the meter as well. The advantages in using a meter com- 
 mon to both systems may be summed up as follows: 
 
 1 . Flexibility of meter service. 
 
 2. Carrying of a smaller stock. 
 
 3. Simplification in the repair and testing departments. 
 
 4. The same quality of meter service to all customers. 
 Any form of meter, therefore, which will operate with 
 
 equal accuracy on direct and alternating current without 
 any change in calibration would, other considerations being 
 equal, be the proper one to use. 
 
 When the problem is further complicated by having a 
 polyphase system to meter in combination with a direct 
 current system, conditions arise which have to be carefully 
 considered. If a meter having an armature and commu- 
 tator is used for the direct current it will take two or more 
 meters per installation to meter an unbalanced polyphase 
 service, so that the first cost of meters must be taken into- 
 consideration, that is, whether it is cheaper to install 
 two meters or one polyphase induction meter. Other con- 
 siderations also enter. The consumer in general does not 
 understand why two meters should be put in to register a 
 single service, and no amount of explanation will disabuse 
 his mind of the idea that the company is getting the better 
 of him. Again, the maintenance and repairs on two meters 
 are double that of one and will add to the cost of operating 
 the meter department. These considerations outweigh 
 those advanced for the use of one type of meter, and it is 
 preferable, in this case, to use two types of meters. In case 
 five, where only one class of service is given, namely, a single 
 or polyphase distribution, the selection of a meter is largely 
 controlled by local preference. There are a number of 
 meters which can be used with equal success, questions of 
 first cost, durability and small maintenance are deciding; 
 
METER SELECTION. 
 
 factors, and, as the service is not mixed, it is best 
 
 use a single meter per installation. It is frequently the 
 
 case that a central station has no choice in the selection of 
 
 its meters ; under such conditions it is hard to work out a 
 
 satisfactory system unless the meter happens to be a good 
 
 one. 
 
 If the meter selected comply with both the electrical 
 and commercial requirements its successful operation should 
 be assured with ordinary care. 
 
CHAPTER IV. 
 Torque and Friction. 
 
 The designing of a meter for permanent commercial 
 accuracy involves a great many serious problems. In 
 Chapter III we have alluded to the principal qualifications 
 to be fulfilled by a good meter. The subject of torque was 
 treated briefly, but its true significance in meter design 
 needs fuller elaboration. 
 
 The word "torque" is used to express the force with 
 which a rotating body tends to revolve, and is usually 
 expressed in foot pounds. Counter torque is a force acting 
 in opposition to torque; thus, the friction of a moving body 
 may be expressed as counter torque, the overcoming of 
 which consumes a certain proportion of the torque of the 
 rotating body. In a meter the torque is a minute fraction 
 of the foot pound unit, and the counter torque or friction a 
 smaller quantity still. If a half dozen different makes 
 of recording wattmeters be set up and tested when new, 
 a very slight difference may be noticed in their readings on 
 full and medium loads, but very decided differences are 
 recorded on light loads from five per cent, and under of the 
 full load capacity. In other words, each meter has a differ- 
 ent ratio between the force impelling it forward and the 
 friction of the moving parts holding it back. The friction 
 in each meter may vary widely, and the impelling force or 
 torque may do the same. Consequently, there is a differ- 
 ent ratio between friction and torque for each individual 
 meter. 
 
 46 
 
TORQUE AND FRICTION. 47 
 
 The continued accuracy of any meter depends primarily 
 on preserving permanently the ratio between torque and 
 counter-frictional torque. The variable quantity in this 
 ratio is the counter-frictional torque; hence, the torque 
 should be so large in proportion that a doubling or trebling 
 of the frictional torque would have no appreciable effect 
 on the accuracy of the meter. For example, suppose the 
 ratio existing in a given meter on a given load be i to 100, 
 then, by doubling the friction, the ratio becomes 2 to 100. 
 In the first ratio the error would be one per cent., in the 
 second, two per cent. On the contrary, if the first ratio 
 on the same load were i to 10, the doubling of the friction 
 would make an error of 20 per cent, in the reading of the 
 meter instead of two per cent, as in the second ratio. 
 Carrying on this process, multiplying each friction by ten 
 in both ratios gives a ten per cent, error in the first one and 
 a complete stoppage of the meter in the second. If the 
 frictional counter torque were not of a negative character 
 the increasing of this friction in the second ratio above ten 
 would result in a reversal of the meter. In speaking of the 
 frictional torque, therefore, we have to think of it as a neg- 
 ative torque, which is to be subtracted from the current 
 torque, but cannot exert its influence to revolve the meter 
 in an opposite direction. 
 
 The current torque remains constant for a given load on 
 the meter, but necessarily varies with the load; hence, if 
 the ratio be i to 10 on ten per cent, of the capacity of the 
 meter, it becomes i to 100 on full load. Therefore, the 
 percentage by which the meter runs slow on full load may 
 be read as a multiplier of the frictional torque; for instance, 
 eight per cent, slow means an increase of friction of eight 
 times the original amount. It is assumed, in this state- 
 ment, that the meter ran properly originally and became 
 
48 AMERICAN METER PRACTICE. 
 
 slow through increase in friction. Meters can run slow 
 from other causes, but these causes are not at present under 
 discussion. 
 
 In order to present more clearly the ratio of current 
 torque and counter-frictional torque, a meter of five amperes 
 capacity is selected for illustration. 
 
 For the sake of analysis the friction-torque ratio of the 
 meter is assumed to be as stated in the second column : 
 
 Friction-Torque Per Cent. 
 
 Load. Ratio. Slow. 
 
 10 lamps =100% 1-50 2% 
 
 5 " = 50% 1-25 4% 
 
 2* " = 25% 1-12* 8% 
 
 2 " = 20% I-IO 10% 
 
 i " = 10% i- 5 20% 
 
 8c.p.lamp= 5% 1-2^ 40% 
 
 Stopped = 2% i- i 100% 
 
 In order to avoid such a poor showing as the above, a 
 friction compensator is introduced which, theoretically, com- 
 pensates for the friction and enables the meter to run 
 correctly under all loads. Since this friction compensator 
 is not automatic in its operation, that is, does not increase 
 as the friction increases, it affords only imperfect relief, 
 as the increasing of the frictional torque by ten would stop 
 the meter on 20 per cent, of its full load. 
 
 The remedy lies in increasing the ratio. If the friction 
 remain the same, the doubling of the torque in the fore- 
 going table would result in the following: 
 
TORQUE AND FRICTION. 
 
 Load. 
 
 Full 
 
 Half . . . 
 Quarter. 
 
 20%.... 
 10% 
 
 2% 
 
 Friction-Torque. 
 I-IOO 
 I- 50 
 I- 25 
 I- 20 
 I- 10 
 
 J - 5 
 
 I- 2 
 
 Increasing the friction by ten in this instance would stop 
 the meter on ten per cent, of its load, and make it ten per 
 cent, slow on full load instead of 20 per cent, slow on full 
 load as under the conditions assumed for the previous 
 table. 
 
 Carrying out this principle one could get a meter that 
 for all commercial considerations would be correct through 
 all ranges of load. Two things must be done to accomplish 
 this result, namely, increase the torque and reduce the 
 friction. The most economic course is to eliminate friction, 
 as this keeps down the power consumption of the meter. 
 When the frictional equivalent is reduced to its minimum 
 for a given form of construction, the friction -torque ratio 
 at full load for any meter should not be less than 1 1400. 
 This ratio with no friction compensator gives a two and 
 one-half per cent, error on ten per cent, of full load, which 
 is close enough for commercial use; it would mean an 
 error of 1.2 watts with one light burning on a five ampere 
 meter. 
 
 It is evident from the foregoing that there is an economic 
 limit to the increase of the torque of a meter to obtain 
 greater accuracy. When this increase in torque is directly 
 proportional to the increase in current necessary to produce 
 it, it is possible to consume more power in obtaining 
 
50 AMERICAN METER PRACTICE. 
 
 extreme accuracy than the amount of power lost through 
 inaccuracy. If the ratio 1 1400 were doubled, making it 
 i :8oo, and this entailed increasing the power consumption 
 of the meter by three watts, nothing would be gained; on 
 the contrary, a loss would be sustained of 1.8 watts on any 
 load the meter registered. 
 
 By assuming a fixed frictional equivalent, the economical 
 torque of any meter can be worked out. In all forms of 
 motor meters the design should be such that the maximum 
 torque should be obtained for a given input in current. 
 If all meter losses in a given case amount to five watts^. 
 the meter should have a resultant torque approaching as 
 closely as possible its mechanical equivalent, or about t | 
 horse-power. As a rule meters are very inefficient, con- 
 sidered as motors, and cannot be otherwise from their con- 
 struction. 
 
 In a 5oo-wattmeter, assume that the friction has been 
 reduced to its lowest possible limit, and that the meter 
 consumes at full load six watts in its shunt and field coils. 
 With this consumption of energy, also assume the friction- 
 torque ratio to be 1:100. The meter, if uncompensated, 
 would lose one per cent., or five watts, at full load from 
 friction. If eight watts were consumed by the meter and 
 the torque thereby increased by 33^- per cent., the friction- 
 torque ratio would be 1:133^, and the error at full load 
 three and three-fourths watts. If nine watts were con- 
 sumed by the meter and the torque thereby increased by 
 50 per cent., the friction-torque ratio would be 1:150, and 
 the error at full load would be three and one-third watts. 
 The net gain, therefore, would be only one-third watt, so 
 this ratio of 1 1150 could be assumed to be about the point 
 of economical balance, beyond which there would be no 
 gain in further increasing the torque at the expense of 
 
TORQUE AND FRICTION. 51 
 
 further loss of power. Where the friction is an absolute 
 fixed quantity, the friction-torque ratio of any given meter 
 may be worked out in the same way. 
 
 A graphic method of determining the friction-torque 
 ratio of a given meter is as follows: Adjust the friction 
 compensator on no load so that the meter is just balanced ;; 
 that is, so that it will creep under the slightest vibra- 
 tion. When this is done, turn on the full capacity of 
 the meter and take its speed under full load; then 
 remove the friction compensator and note the difference 
 in the speed of the meter under the same load. The per- 
 centage of slowness will be inversely proportional to the 
 friction-torque ratio; thus, two per cent, would mean 
 1 150 ; one per cent., i :ioo; and so on. When the friction- 
 torque ratio is determined, it is very easy to figure out the 
 gain or loss that the consumption of additional power to 
 increase the torque would entail. 
 
 Meters with ratios under 1 1400 are the rule rather than 
 the exception, but their frictional equivalent is such that 
 it is impossible to better them. ' Many meters fail in com- 
 mercial service by having an unstable frictional equivalent, 
 through faulty mechanical design or improper protection 
 of the revolving parts from foreign matter. 
 
 The correct end of all meter design is towards an un- 
 changing frictional equivalent of infinitesimal amount and 
 the largest possible economic torque determined by the 
 rules laid down above. Friction compensators should be 
 discarded where possible. 
 
 Meters vary in the amount of power required in their 
 potential circuits, and, also, in the amount used in the field 
 windings. The values of these combined amounts should 
 be taken at full load of the meter and a tabulated set of 
 friction-torque ratios put down. The ratio for a given loss 
 
52 AMERICAN METER PRACTICE. 
 
 of power can be found by ascertaining the loss in registra- 
 tion at full load of the meter caused by the friction, and 
 compensating for this by giving the meter more torque at 
 an increased expenditure of energy. From the tabulated 
 values it is a simple matter to select a friction-torque ratio 
 that cannot be increased unless more energy is expended 
 than will be saved by the increased accuracy obtained. 
 
 In the following tables three different amounts of power 
 loss are assumed in relation with the same friction-torque 
 ratios; the balancing ratios obtained represent the correct 
 design of the meter to obtain the highest economy under 
 any one of the given conditions. 
 
 TABLE I 
 
 Consumption in Watts Correct Watts True 
 
 Ratio in Meter at Full Load. for Balancing Ratio. Balancing Ratio. 
 
 1-50 2 5 - 125 app. 
 
 i- 100 2 4 - 200 
 
 I- 200 2 3H - 3 2 5 
 
 I- 400 2 2% - 550 
 
 I- 800 2 2^ -1000 
 
 i-iooo The highest economic balance at 2^ watts consumption of 
 current. 
 
 TABLE II 
 
 - 50 4 8 i-ioo app. 
 
 -zoo 4 6 1-150 
 
 -200 4 5^ 1-275 
 
 -400 4 5 1-500 
 
 -500 The highest economic balance at 5 watts consumption of 
 current. 
 
 TABLE III 
 
 1-50 6 ii i- 90 app. 
 
 i-ioo 6 9 1-150 
 
 1-200 6 7% 1-250 
 
 1-250 Approximate highest economic balance at 7% watts con- 
 sumption of current. 
 
TORQUE AND FRICTION. 53 
 
 These tables show that for different values of the fric- 
 tional equivalent the amount of energy necessary to pro- 
 duce the best results varies accordingly. For example, 
 in Table I the ratio 1-50 with 2 watts consumption of 
 current is an inefficient ratio; the true ratio should be 
 1:125 with 5 watts consumption of current. 
 
 In Table II the ratio 1-50 shows that it should be i-ioo 
 with 8 watts consumed instead of 4. As the energy torque 
 remains the same and the friction decreases, a ratio is 
 reached wherein no change either way can be made to 
 make the meter more efficient. In Table I this ratio is 
 i- 1 ooo with i\ watts consumption of energy. 
 
 As the consumption of energy increases, as shown in 
 Tables II and III, the ratios diminish in direct proportion. 
 If a ratio of 1-50 consume 10 watts, it is balanced, and the 
 meter cannot be improved by increasing the watt con- 
 sumption. 
 
 Tables can be made of various watt losses of different 
 meters, and economic balance of any meter obtained after 
 its ratio is found. 
 
 The foregoing tables will be useful for designers and also 
 to the central station in the selection of meters. It is 
 assumed in the foregoing that the frictional equivalent is 
 uncompensated, but, as the majority of manufacturers try 
 to compensate the frictional losses of their meters, the ratio 
 of almost all meters on the market can be obtained as out- 
 lined without much difficulty. 
 
 The Stanley meter is uncompensated, the frictional equiv- 
 alent being so slight that it has a very large ratio. The 
 ratio can be determined on this meter by noting the num- 
 ber of watts on which it starts. This per cent, of the full 
 load capacity would determine the ratio of the meter with 
 reasonable accuracy. 
 
54 AMERICAN METER PRACTICE. 
 
 Too much stress connot be laid on low and permanent 
 frictional equivalent. If these elements can be fixed it is an 
 easy matter to build an accurate meter for all loads. 
 
 It is customary among meter manufacturers to maintain 
 the torque for the same percentage of load constant through 
 various sizes of meters. Thus a five ampere meter may run 
 at the rate of one revolution in four seconds on full load, 
 and a 100 ampere meter run at the same rate on full load. 
 
 The registration on the dials is corrected either by multi- 
 plying by a constant or changing the ratio between the 
 revolving armature and the registering train. 
 
 A 100 ampere meter, therefore, with a ratio of 150 
 would have a ratio of i to 5 on 20 lights or 20 per cent, 
 .-slow on this load on two per cent, of its capacity it would 
 stop, that is, on four lights. Of course, the friction is 
 compensated for as far as possible, but any increase in 
 friction makes the above conditions more aggravated. 
 
 Any meter therefore with such a ratio is absolutely un- 
 reliable and can never be made to run correctly. The 
 central station is the loser. 
 
 A different set of tables must, therefore, be made up for 
 each capacity of meter, and its balancing ratio obtained to 
 determine just where it stands with regard to accuracy. 
 
 If a 100 ampere meter with a ratio of 1-50 be installed 
 and the friction compensated for, it may run approxi- 
 mately correct for a short period, but the friction is liable 
 ,to treble or quadruple itself at any time, and the com- 
 jpensator will only take care of the original friction value. 
 
 A curve of the meter's running can be plotted showing 
 the per cent, slow at various increases in friction. 
 
 If we take a meter of 500 watts capacity as a base from 
 which to deduct different values for friction and torque, 
 we .arrive at ratios given in the following tables. 
 
TORQUE AND FRICTION. 
 
 We will suppose that, if a certain torque be obtained b 
 given input in watts in the meter, a larger or smaller 
 input will increase or decrease the torque in direct pro- 
 portion. 
 
 If the meter have no compensation for its friction, it will 
 run at various percentages slow according to the load; 
 thus a meter two per cent, slow on full load will be four 
 per cent, on half load and so on. 
 
 TABLE I 
 
 Watt Loss 
 in 
 
 Watt Loss Total Losses 
 due in Meter and 
 
 Friction. 
 
 Torque . 
 
 Meter. 
 
 Ratio. 
 
 to Friction. 
 
 by Friction 
 
 % Slow. 
 
 05 
 
 50 
 
 IO 
 
 I IOOO 
 
 1 A 
 
 10^ 
 
 I-IO Of I< 
 
 . I 
 
 5 
 
 IO 
 
 I- 500 
 
 i 
 
 II 
 
 i/5 of i < 
 
 . 2 
 
 5 
 
 IO 
 
 I 250 
 
 2 
 
 12 
 
 2/5 of i< 
 
 -4 
 
 50 
 
 IO 
 
 I- 125 
 
 4 
 
 14 
 
 4/5 of i < 
 
 i 
 
 5 
 
 10 
 
 I- 50 
 
 IO 
 
 2O 
 
 2% 
 
 2 
 
 5 
 
 10 
 
 I- 2 5 
 
 20 
 
 30 
 
 4% 
 
 4 
 
 5 
 
 10 
 
 i- 12^; 
 
 2 40 
 
 5 
 
 8% 
 
 5 
 
 5 
 
 10 
 
 I- 10 
 
 50 
 
 60 
 
 10% 
 
 10 
 
 5 
 
 10 
 
 I- 5 
 
 IOO 
 
 no 
 
 20% 
 
 25 
 
 5 
 
 10 
 
 I- 2 
 
 250 
 
 260 
 
 50% 
 
 .50 
 
 5 
 
 IO 
 
 I- I 
 
 500 
 
 5io 
 
 stopped. 
 
 TABLE II (Doubling the Torque by Doubling Input.) 
 
 2O*4 I/2O Of I ' 
 
 i/io of i; 
 i/5 of i; 
 2/5 ofi; 
 
 2% 
 
 4% 
 
 .05 
 
 IOO 
 
 20 
 
 I-20OO 
 
 M 
 
 20 
 
 . I 
 
 IOO 
 
 20 
 
 I IOOO 
 
 Vz 
 
 20 
 
 .2 
 
 IOO 
 
 20 
 
 I- 5OO 
 
 I 
 
 21 
 
 4 
 
 IOO 
 
 20 
 
 I- 250 
 
 2 
 
 22 
 
 J 
 
 IOO 
 
 2O 
 
 I- IOO 
 
 5 
 
 25 
 
 2 
 
 IOO 
 
 2O 
 
 I- 50 
 
 IO 
 
 3 
 
 4 
 
 IOO 
 
 20 
 
 I- 25 
 
 20 
 
 40 
 
 5 
 
 IOO 
 
 2O 
 
 I- 20 
 
 25 
 
 45 
 
 IO 
 
 IOO 
 
 2O 
 
 I- 10 
 
 50 
 
 70 
 
 25 
 
 IOO 
 
 20 
 
 I- 4 
 
 I2 5 
 
 145 
 
 5 
 
 IOO 
 
 20 
 
 I 2 
 
 250 
 
 270 
 
 50% 
 
56 
 
 AMERICAN METER PRACTICE. 
 
 TABLE III 
 
 Watt Loss 
 in 
 
 Watt Loss 
 
 due Total Losses 
 
 Friction. 
 
 Torque. 
 
 Meter. 
 
 Ratio. 
 
 to Friction. 
 
 in Meter 
 
 
 .001 
 
 10 
 
 2 
 
 -400 
 
 J M 
 
 3/4 
 
 
 05 
 
 IO 
 
 2 
 
 2OO 
 
 2^ 
 
 4^ 
 
 
 .1 
 
 10 
 
 2 
 
 -IOO 
 
 5 
 
 7 
 
 I 
 
 .2 
 
 IO 
 
 2 
 
 - 5 
 
 IO 
 
 12 
 
 2 
 
 4 
 
 10 
 
 2 
 
 ~ 25 
 
 20 
 
 22 
 
 4 
 
 i 
 
 IO 
 
 2 
 
 - 10 
 
 50 
 
 5 2 
 
 10 
 
 2 
 
 10 
 
 2 
 
 - 5 
 
 IOO 
 
 IO2 
 
 20 
 
 4 
 
 10 
 
 2 
 
 - 2^ 
 
 200 
 
 2O2 
 
 40 
 
 5 
 
 IO 
 
 2 
 
 2 
 
 250 
 
 252 
 
 50 
 
 10 
 
 10 
 
 2 
 
 - I 
 
 5 00 
 
 5 02 
 
 stc 
 
 % Slov 
 
 TABLE IV 
 
 .001 
 
 20 
 
 4 
 
 [-800 
 
 . 62 J 
 
 ^ 4.62 
 
 H Hof 
 
 i% 
 
 .05 
 
 2O 
 
 4 
 
 400 
 
 I ^^ 
 
 sM 
 
 &of 
 
 i% 
 
 .1 
 
 20 
 
 4 
 
 200 
 
 2^ 
 
 6^ 
 
 Hof 
 
 i% 
 
 . 2 
 
 2O 
 
 4 
 
 -IOO 
 
 5 
 
 9 
 
 1% 
 
 
 4 
 
 20 
 
 4 
 
 - 50 
 
 IO 
 
 14 
 
 2% 
 
 
 i 
 
 20 
 
 4 
 
 - 20 
 
 25 
 
 29 
 
 5% 
 
 
 2 
 
 20 
 
 4 
 
 IO 
 
 50 
 
 54 
 
 io9o 
 
 
 4 
 
 20 
 
 4' 
 
 - 5 
 
 IOO 
 
 104 
 
 20% 
 
 
 5 
 
 20 
 
 4 
 
 - 4 
 
 125 
 
 129 
 
 25% 
 
 
 IO 
 
 20 
 
 4 
 
 2 
 
 250 
 
 254 
 
 50% 
 
 
 20 
 
 2O 
 
 4 3 
 
 I 
 
 500 
 
 54 
 
 stopped 
 
 
 In the Tables I, II, III and IV we have meters con- 
 suming 2, 4, 10, 20 watts respectively. These meters are 
 supposed to operate at the same efficiency per watt, that 
 is, have the same torque per watt. 
 
 Various friction values are assumed, the unit of friction 
 being taken as some proportion of the torque unit. 
 
 In Table I a meter having a friction value of .05 and 
 torque value of 50 with a watt consumption of 10 has a 
 ratio of i-iooo, and the total meter losses are loj watts. 
 
TORQUE AND FRICTION. 57 
 
 The same meter, Table II, with double the torque and 
 double the watt input has a ratio of 1-2000 and total 
 meter losses on full load of 20 J watts. 
 
 It is seen at a glance that the lower torque meter Table 
 i is the more efficient. 
 
 In Table III a low torque meter having the same frictional 
 value with 2 watts input has total watt losses on full load 
 of 4*. 
 
 These tables enable one to ascertain the correct watt 
 input of a meter having a given frictional equivalent for 
 the most economical running. 
 
 In these tables the watt losses at full load are given for 
 various friction values, *and it is seen that the low torque 
 meters are the most economical on low friction values and 
 the high torque meters on large friction values. In other 
 words, every meter having a given frictional equivalent 
 has a balancing ratio at which it will operate most 
 economically and correctly. 
 
 Table I from .05 to .2 frictional equivalents, we find by 
 comparison that Table III is far more economical, while for 
 values from .2 to 10, Table I is the best. 
 
 Like comparison can be made for various values in the 
 different tables. 
 
 These tables are calculated for the purpose of showing 
 the losses due to uncompensated friction. As nearly all 
 meters are compensated, these tables do not apply when 
 the meter is correct, but should the friction equivalent 
 become increased while in service, the losses set forth 
 would occur. 
 
 The lines to be pursued in obtaining a correct meter are 
 the reducing of the frictional equivalent to its lowest 
 permanent value and the designing of the meter electrically 
 for its greatest efficiency per watt input. 
 
58 AMERICAN METER PRACTICE. 
 
 In other words, if a meter consume four watts on full 
 load and have a torque value of 20, the frictional equivalent 
 must be at least as low as .4 of this amount, or it becomes 
 more economical to consume more than 4 watts to obtain 
 a larger torque. The ratio must not fall below 1-50. 
 If a ratio between torque and friction of i-ioo be 
 obtained, the wattage can be reduced to 2 with more 
 economical results as shown (Table III). 
 
 With high friction values we get better results by using 
 more watt input. For instance, with a ratio i-io (Table 
 III) we get 52 watts loss; the same friction value (Table IV) 
 gives us a ratio 1-20, watt loss 29, while in Table i we 
 have a ratio 1-50, watt loss 20. The same friction value, 
 Table II, gives a ratio i-ioo, watt loss 25, showing that 
 ratio 1-50, watt loss 20, is the balancing ratio. 
 
 In the foregoing it is assumed that the central station, 
 not the consumer, pays for the loss in the meters. 
 
CHAPTER V. 
 The Edison Chemical Meter. 
 
 Almost from the inception of electric lighting the Edison 
 chemical meter has been used for the metering of direct 
 currents. In theory it is an ampere hour meter, and with 
 proper manipulation and care it proves a wonderfully 
 accurate meter in service. 
 
 If two plates of zinc be immersed in a solution of sul- 
 phate of zinc and a current passed from one plate to the 
 other through the solution, zinc is carried over from the 
 
 FIG. 18 
 
 positive plate and deposited on the negative in direct pro- 
 portion to the amount of current flowing. This principle is 
 utilized in the Edison chemical meter. 
 
 In a solution of zinc sulphate having a density i.n, one 
 ampere of current will deposit 1.224 grams of zinc per hour. 
 This rate of deposit would necessitate liaving plates weigh- 
 ing a ton to register large amounts of current. This diffi- 
 culty is surmounted by using a shunt of low resistance and 
 allowing only a small fraction of the current to be used in 
 registering, the amperes flowing in the circuit. Fig. 18 
 
 59 
 
60 
 
 AMERICAN METER PRACTICE. 
 
 represents, diagrammatically, the arrangement of the meter 
 for the two-wire system. It is observed that there are two 
 bottles each containing two plates held at a fixed distance 
 from each other by rubber bolts and spacers. Two bottles 
 instead ot one are used to serve as a check on each other. 
 A is the low resistance German silver shunt, B and 
 C are spools on which is wound copper wire of com- 
 paratively high resistance. In the 12 light meter the shunt 
 has approximately .02 ohm resistance and the bottle cir- 
 cuit 19.50 ohms, of which the spool has about 16.50 ohms 
 and the bottle 3 ohms, making a relation of i to 975 be- 
 
 fWWVWWWWW* 
 
 FIG. 19. 
 
 tween the shunt and bottle. From this relation a constant 
 is worked out for the meter to determine the amount of 
 money one grain or fraction thereof represents. This con- 
 stant varies with the voltage of the circuit employed if the 
 service be charged for by watt-hours. 
 
 In the three-wire meter, Fig. 19, a shunt is inserted in 
 each outside leg, making simply two two-wire meters under 
 the same cover. The meter is made in a number of com- 
 mercial sizes, and these can be made of longer recording 
 capacity by increasing the resistance of the bottle circuit. 
 
EDISON CHEMICAL METER. 61 
 
 The zinc plates are mounted on hard copper rods threaded 
 into the top end. The glass bottles resemble fruit jars 
 with two holes in the cover for the extended rods. Of late 
 years a large cork was used instead of the glass top and 
 considerable time saved in assembling the cells. The 
 whole meter was mounted in a well seasoned wooden box 
 having a sheet iron door with a shelf provided midway in 
 the meter for the reception of the bottles. The low re- 
 sistance shunts were contained in the lower part of the 
 meter. The hard copper rods were pushed into hard 
 copper spring contact clips for connecting them in circuit. 
 In cold climates a lamp in series with a thermostat is 
 placed in the meter compartment for furnishing heat if the 
 temperature falls near freezing point. The temperature 
 coefficient of the meter is preserved at unity by winding 
 the resistance spools with copper wire, as copper increases 
 in resistance and German silver decreases as they grow 
 warmer. The solution decreases in resistance as it grows 
 warmer, so the copper spools become a balancing factor 
 between the German silver shunt and the solution. The 
 fall of potential across the shunts is so slight as to make the 
 energy losses of the meter inappreciable ; the nature of the 
 loss being the same as a line loss, and not a consumption of 
 additional energy. 
 
 By following the meter through its practical operation 
 in the meter department its different features will be 
 clearly brought out. The zinc plates must be chemically 
 pure, or, otherwise, battery action will be set up in the cell 
 and a counter electromotive force generated which will 
 destroy the accuracy of the register. 
 
 When the plates are received from the makers they are 
 rough stampings with the hard copper rods screwed into 
 .their ends. The first operation is to polish them smoothly 
 
62 AMERICAN METER PRACTICE. 
 
 on a sand wheel and then amalgamate them in mercury. 
 The surplus mercury is removed by a revolving brush. The 
 finished plate is smooth and silvery in appearance. The 
 mercury has the effect of rendering the plate active and 
 eats its way inward as the zinc is deposited from one plate 
 to the other. The bare zinc plate unmalgamated seems 
 to possess a sort of resisting property to free action which 
 hinders it from being deposited at a uniform rate. The 
 joint between the zinc plate and copper rod should be 
 painted with black asphaltum paint to keep the mercury 
 from eating away the copper and allowing the rod to be- 
 come loose in the joint. This paint is put on before the 
 plate is amalgamated. When ready to be ^eighed it is tested 
 across the joint with a galvanometer to insure the presence 
 of a perfect electrical joint. It is found that, after the 
 plates have been in service for some time, this joint fre- 
 quently becomes defective, and a resistance is thereby added 
 to the bottle circuit which makes the meter register less 
 than it should. Analytical balances of the greatest del- 
 icacy are used in weighing the plates and the plates are 
 weighed to within 5 milligrams. Two methods of weigh- 
 ing are in use; the single and double system. As only 
 one plate loses weight, the positive one, the single method 
 of weighing records its weight on suitable slips, and when 
 the plate comes back at the end of the month its lost weight 
 is proportional to the amount of current used. If the plate 
 be reversed in the meter, the weighed plate gains in weight 
 instead of losing. The gain or loss represents the amount 
 of current used. The double weighing system weighs both 
 positive and negative plates on a double rider balance, the 
 net difference in weight being recorded. When the plates 
 are returned the net difference is again noted and the loss 
 recorded on the slip. 
 
EDISON CHEMICAL METER. 63 
 
 After the plates are weighed by one operator they should 
 be check weighed to eliminate errors by another operator. 
 Each plate is given a number and letter written on the back 
 of the plate itself, so as to identify it on its return to the 
 meter department. From the weighing bench the plates 
 are taken to be assembled, which consists of bolting 
 positive and negative together, and separating them 
 by insulating spacers of a fixed width. The plates 
 are reground and used over and over again until they 
 wear out. The preparation of the solution must be care- 
 fully done, and all the ingredients must be chemically pure. 
 Distilled water and sulphate of zinc are the two ingredients, 
 and these are mixed in solution to a density of i.n actual. 
 After mixing, the solution is filtered to remove any foreign 
 particles that may have fallen into it. The bottles are 
 washed clean, and the assembled plates put in with the 
 solution covering them for about half an inch. When a 
 large number of bottles are handled daily, special tanks of 
 porcelain or glass should be arranged to hold the solution, 
 and the bottles filled from them by a flexible hose with 
 faucet attached. 
 
 A given number of meters should be read each day, and 
 they should be listed according to location and a route 
 list given to the man changing the bottles. 
 
 The great secret of success in the metering of current in 
 this way lies in the exercise of extreme care in each 
 operation. Carelessness in any one part of the work 
 affects the whole system. In the installation of the bottles 
 in the meter care must be taken to place the positive plate 
 to the positive clip of the meter, otherwise reversed plates 
 will be the result. It is the practice in some plants to put 
 one bottle in reversed and one bottle the right way and take 
 the average of the sum of the readings, but this method is 
 
64 AMERICAN METER PRACTICE. 
 
 open to objection if the bottles become "smutted" from 
 excessive load. 
 
 The "smutting" of the plates is caused by the too rapid 
 deposition of the zinc, and, if carried on too long, the plates 
 are bridged across with black bridges resembling coke. 
 This short circuiting of the plates shunts them, and makes 
 them register less than they should. When a meter is 
 found to smut badly, even if it do not short circuit, the 
 bottles should be renewed oftener or a larger meter put 
 in. It has been found, by a careful series of tests by the 
 author, that a heavy smutting increases the registration 
 by about 10 per cent., and a short circuit decreases the 
 registration by a varying quantity dependent on how many 
 bridges have been built up between the two plates. It 
 was found that, where there was excessive vibration, 
 the bottles gradually crept out of the clips and ceased to 
 register. Again, it was found that verdigris and dirt fre- 
 quently introduced a resistance in circuit which kept the 
 meters from registering all that they should. Consider- 
 ations of this nature led to the adoption of a cast brass 
 clip in which holes were provided for the reception of the 
 copper rod terminals of the cells. A set screw was pro- 
 vided for setting the rods up hard in the hole, and it was 
 impossible to get them loose. The grinding movement of 
 the screw on the rod insured a good contact. The great 
 improvement in the registration of the meters led to their 
 adoption for the entire system of a large plant. It was the 
 watching of every minute detail and the careful handling 
 of this meter that made it a commercial possibility for 
 20 years. 
 
 The great drawback to the meter is that it cannot be 
 read by the consumer. The consumer was, so he thought, 
 absolutely at the mercy of the company furnishing current, 
 
^ OFTHE 
 
 , UNIVERSITY ) 
 
 Of 
 
 EDISON CHEMICAL METER. 65 
 
 and he could seldom follow the process of measuring the 
 current intelligently when the method was explained to 
 him. Another great drawback to the meter is that it has 
 no rate of measurement from which to determine its ac- 
 curacy. In other words, if a reading failed for one month 
 owing to some accident, the only recourse was to take a 
 check reading and estimate the bill. In a mechanical 
 meter, if the reading be disputed, a test can be made of 
 the meter and a determination made of its accuracy, from 
 which a bill can be checked up with a fair amount of pre- 
 cision. On the other hand, the chemical meter could not 
 "run slow" for more than a month, as it was practically a 
 new meter after each renewal. 
 
 After the bottles have stood for a month, even if no cur- 
 rent be used, a slight oxidation takes place which increases 
 the weight of the plate slightly. 
 
 When the bottles are taken apart, if the single plate 
 method of weighing be used, the plates are dipped for a 
 moment in dilute sulphuric acid and quickly rinsed. This 
 removes the slight over-weight caused by oxidation. When 
 the double plate method of weighing is used, the oxidation 
 on the negative and positive plates cancel each other, so no 
 dipping is required. 
 
 The relation between the shunt and bottle circuits 
 rarely changes, and when it does it is nearly always 
 due to an open circuit in the spool. The infinitesimal 
 difference of potential between the adjacent coils in 
 the spools renders a short circuit in them practically out 
 of the question. However, the precaution is taken to boil 
 the spools in paraffine. 
 
 The labor of maintaining the meters is excessive, and 
 the details, when a large number of meters are operated, 
 cumbersome. 
 
66 AMERICAN METER PRACTICE. 
 
 The chemical meter is passing out of use. Its com- 
 mercial accuracy is equal to many mechanical meters, but 
 there are advantages in the mechanical type which offset 
 some of the advantages which the chemical meters possess 
 over them. 
 
 If this description of the chemical meter were likely to 
 prove other than of historical interest, many interesting 
 details of operation and arrangement of the meter depart- 
 ment would be gone into, in the belief that they might prove 
 of service. The meter has been barely sketched for the 
 purpose of illustrating a type which so long held first place 
 in the meter world, but which has passed its day of useful- 
 ness. Some of the requisites of a good meter are fulfilled 
 in this meter. Its light load accuracy is very good on 
 small sizes of meters, but not on large sizes, owing to the 
 modifying effects of increased oxidation due to large plates. 
 
 Vibration does not alter the accuracy of the meter if the 
 copper rods be held firmly in contact by set screws. As 
 the meter's accuracy is altogether dependent on the con- 
 tinued care and accuracy of many operations, individual 
 inaccuracies are not as likely to occur as for a whole system 
 to become deranged by some chemical impurity in the 
 solution or some such similar cause. 
 
 The meter is not adapted to the measurement of alter- 
 nating current, but can measure in ampere hours any con- 
 stant voltage direct current by obtaining the proper 
 constant for reduction to watt hours. 
 
CHAPTER VI. 
 The Thomson Recording Wattmeter. 
 
 The Thomson recording wattmeter is one of the most 
 widely used meters in America, up to the present time. Its 
 extended use has been in great measure due to its being 
 practically the only meter on the market for direct current, 
 and to its further quality of adaptability to all forms of 
 service. 
 
 Its ability to register either alternating or direct current, 
 whether the former be inductive or non-inductive, without 
 change of calibration makes it especially suitable to the 
 needs of central stations furnishing different kinds of 
 current. 
 
 In principle, the meter is a modification of a Siemens' 
 dynamometer, the stationary coils of which are energized 
 by the current flowing, and the movable coils proportion- 
 ately to the potential of the circuit in which the meter is 
 inserted. There is no iron employed in the shunt or series 
 field coils, the shunt coil or armature being wound drum 
 fashion over a non-magnetic skeleton support and provided 
 with commutator and brushes. 
 
 The operation of the dynamometer then becomes that of 
 a simple motor. The current flowing energizes the fields 
 and the potential of the circuit, the shunt field coil and 
 armature. 
 
 It is evident that the torque of the armature will be pro>- 
 portional to the product of the field current by the E. M. F. 
 of the circuit or the watts passing through the meter. If 
 
 67 
 
68 AMERICAN METER PRACTICE. 
 
 the potential be constant, the meter may be used as a re- 
 cording ammeter, as the torque of the armature then varies 
 as the amperes. If the current be constant and the voltage 
 be variable, the meter maybe used as a recording voltmeter. 
 
 The construction of the meter has been carefully worked 
 out mechanically. The armature spindle is supported on a 
 spring seated removable jewel and the shaft end provided 
 with a hardened steel removable point, which may be re- 
 newed when the jewel becomes defective and needs re- 
 placing. Secured to the shaft just above the jewel bear- 
 ing is a copper disk revolving between the poles of perma- 
 nent magnets, which act in the usual manner to furnish a 
 brake proportional to the torque on the armature. 
 
 The commutator is made of silver, highly polished, and 
 the flexible copper brushes bearing on it are silver tipped. 
 Silver was chosen for this purpose after a long series of 
 experiments with other metals, owing to its not forming a 
 high resistance oxidized film over its surface when exposed 
 to the air for extended periods. The usual worm gear 
 actuates the dial train which records in watt hours, modified 
 by whatever constant the meter carries. At constant i 
 the meter records a watt hour for every revolution of the 
 armature. 
 
 In two-wire meters the field coils are placed one on each 
 side of the armature and connected in series ; in three-wire 
 meters the same position is occupied, but one coil is con- 
 nected in each side of the circuit, the armature acting in 
 common for both fields. Either field and the armature 
 constitute individual recording elements which will act 
 independently of the other field. Hence, any unbalancing 
 of the regular three-wire system does not effect the ac- 
 curacy of the meter. If the voltage on both sides of the 
 system be the same, the register will be true watts; if 
 
THOMSON RECORDING WATTMETER. 69 
 
 different, the error will be proportional to the difference 
 existing. As the same error exists if the armature be fed 
 from across the two outside legs, it is manifestly no gain to 
 feed the meter in this way. 
 
 In series with the armature is a resistance coil to regu- 
 late the flow of current in the potential circuit, and also to 
 reduce the fall of potential between the armature coils and 
 commutator segments. The initial friction of the meter is 
 compensated by a coil in series with the armature circuit; 
 this coil is wound parallel to the field windings and placed 
 inside of one of them, it is immaterial which. The coil is 
 so proportioned as to overcome in great measure the friction 
 of the moving parts, enabling the meter "to register more 
 accurately on light loads. The torque exerted by this 
 auxiliary coil is about two per cent, of the full load torque of 
 the meter. Care must be taken in installing the meter to 
 secure it as far as possible from vibration, otherwise the 
 meter will creep, unless the starting coil be adjusted to suit 
 the local conditions. 
 
 Later types have an adjustable starting coil. This ad- 
 justable starting coil is wound with a great many more 
 turns than the permanent coil used in older types. It is 
 carried in a little frame which is movable in and out of the 
 field coil by means of a screw. When the meter is installed, 
 the local conditions of vibration are noted and the auxiliary 
 starting coil so adjusted as just to balance the frictional 
 torque of the meter in its present position without allowing 
 it to "creep." Should the jewel become slightly defective 
 after several months' use, the starting coil can be reset to 
 overcome this additional friction of the meter. This 
 arrangement is an improvement over the old method of 
 having a permanent additional torque imparted to the 
 meter no matter what the conditions incident to its location. 
 
70 
 
 AMERICAN METER PRACTICE. 
 
 When the meter is used on alternating current circuits, we 
 may think of its action by considering the circuit for an 
 instant of time. This was gone into fully in Chapter II, 
 and need not be further elaborated, but, as the true power 
 
 FIG. 21. 
 
 in an alternating circuit for any instant of time is the 
 product of the instantaneous values of current and E. M. P., 
 the meter registers as accurately on inductive as non- 
 inductive load. 
 
THOMSON RECORDING WATTMETER. 
 
 71 
 
 To fulfil the demands of service the meter is manu- 
 factured in a great variety of sizes and types of ' ' low effi- 
 ciency" and "high efficiency," the latter type succeeding 
 the former in general commercial use, owing to its greater 
 accuracy on light loads and lower watt consumption in the 
 potential circuit. The potential circuit of the high effici- 
 ency type in the 115-230 volt commercial sizes for light 
 consumes about 4 watts of energy, the 230-500 volt power 
 meters 10 and 20 watts respectively. 
 
 FIG. 22. 
 
 We now pass to the general consideration of the meter 
 and the means of connecting it for registering various 
 forms of commercial service. 
 
 Fig. 21 may be taken to represent the general interior 
 appearance of the various forms. 
 
 In the diagrammatic representations, Figs. 22 and 23 are 
 shown the methods of connecting all meters of the two- 
 wire type, whether for light or power purposes, direct or 
 
72 
 
 AMERICAN METER PRACTICE. 
 
 alternating current. The service feeds on the left, and the 
 disk rotates counter-clockwise as seen from the top. 
 The high efficiency three-wire type is connected as shown 
 in Fig. 24, in sizes up to 150 amperes. Over this capacity 
 the circuit is metered by two meters of the two-wire type. 
 In the metering of polyphase systems by the Thomson 
 meter a variety of connections are used. The characteristics 
 of the circuit govern the method pursued. 
 
 FIG. 23. 
 
 Before passing to the measurement of Alternating Poly- 
 phase currents we wish to call attention to a grave defect 
 in the three-wire meter, Fig. 24, which may lead to the loss 
 of revenue if the consumer be aware of the conditions. It 
 is seen that the armature circuit is fed between the neu- 
 tral and one outside leg, and if for any cause a fuse blow 
 on this side of the system before the meter, the remaining 
 lamps on the opposite side of the system can be burned 
 without causing the meter to register. Nor can this defect 
 
THOMSON RECORDING WATTMETER. 73 
 
 be remedied by connecting the armature across the two 
 outside legs, as, then, if either fuse blow, the meter stops 
 unless some lamps are in circuit on that side of the system. 
 In that event the armature would receive less than one- 
 half its voltage and the meter register less than one-half of 
 what it should. Therefore, the removal of a fuse, or the 
 splitting up of the main switch before the meter into single 
 pole switches, could be used by designing and unscrupulous 
 persons to rob the company furnishing current of a large 
 
 FIG. 24. 
 
 amount of revenue. No doubt the manufacturers of this 
 meter are fully aware of this defect, but the central station 
 is certainly taking chances when installing it of being 
 deprived either through accident or by intent of a portion 
 of its just revenue. There are a number of simple rem- 
 edies that can be applied in the way of small auxiliary 
 devices that will automatically cut the armature over to 
 the live side of the circuit and preserve the proper polarity 
 of the leads so as not to reverse the armature. Some such 
 
74 
 
 AMERICAN METER PRACTICE. 
 
 provisions should be made to make this meter a safe one, 
 as no one can take action against the consumer for 
 removing a plug and burning half of his lights, when he in 
 no way touches or molests the meter. 
 
 The temptation for a certain class of customers, when 
 they understand the meter, is too great for them not to 
 
 Load 
 
 FlG. 25. 
 
 take advantage of what is obviously the fault of the com- 
 pany in providing such a simple and safe meter to "beat." 
 The theory of the measurement of polyphase systems 
 by means of non-inductive meters has been set forth in 
 Chapter II; hence, simply the application of the principle 
 therein outlined will be given in the diagrams for metering 
 the different systems. 
 
THOMSON RECORDING WATTMETER. 
 
 75 
 
 In balanced three-phase circuits, one two-wire meter 
 connected in either one of the three legs with the armature, 
 fed by the energy voltage of the circuit, will register one- 
 third of the total power passing; hence, multiplying by 3, 
 we obtain the total energy passing in the three legs of the 
 circuit. Such a connection is illustrated diagrammatically 
 in Fig. 25, where the armature forms part of one of the legs 
 
 Shunf 
 
 FIG. 26. 
 
 of the set of Y connected resistances. The voltage from 
 the common center to either point of the Y is energy 
 voltage of the system. This connection may be used for 
 either light or power, but must always be used on a bal- 
 anced circuit and one that has the same power factor for 
 -each phase. 
 
 Unbalanced three-phase circuits are metered by placing 
 two two- wire meters in two legs of the circuit and connect- 
 ing one end of the armature circuit of each meter to the 
 
76 
 
 AMERICAN METER PRACTICE. 
 
 feeding leg, and the other end to the unmetered leg. This 
 connection is illustrated in Fig. 26. 
 
 In an unbalanced three-phase system employing four 
 wires in its distribution, the power flowing in any of its 
 three phases is found by connecting a two-wire meter in 
 that phase, and feeding the armature between the leg 
 metered and the fourth wire or common neutral. 
 
 This is practically what is done in the balanced three- 
 wire system, the difference lying in the extension of the 
 
 FIG. 27. 
 
 neutral point of the three phases to the distribution. To 
 meter the total power flowing in the four-wire unbalanced 
 three-phase system, the three phases must be metered as 
 shown in Fig. 27, and the results of all the readings added. 
 In the two-phase systems employing four wires in the 
 distribution, two meters are used if the two phases be 
 unbalanced as indicated in Fig. 28. If the phases be 
 exactly balanced, one meter connected in one phase is 
 sufficient, the total energy passing being double its indica- 
 tion. The connection for the meter is shown in Fig. 29. 
 
THOMSON RECORDING WATTMETER. 
 
 77 
 
 Two-phase systems employing a common return, and 
 distributed by means of three wires, are metered in the 
 same manner as three-phase systems. 
 
 The measurement of monocyclic circuits by means of 
 Thomson meters has been, in many cases, erroneously con- 
 summated. In Fig. 30 one meter is shown connected into a 
 monocyclic secondary circuit, the meter being of the three- 
 wire type with a common armature for both fields. Such a 
 
 FIG. 28. 
 
 connection will not give the correct register of the energy 
 passing in the circuit. This connection has been in use 
 upwards of five years, but, from carefully conducted tests 
 by the author, it has proven beyond a question that about 
 33 YI P er cent - of the current remains unregistered. The 
 manufacturers have now ceased to put out this meter as 
 
78 
 
 AMERICAN METER PRACTICE. 
 
 correct, and meter monocyclic current by a regular three- 
 phase induction meter, or by means of two Thomson meters 
 connected in the same manner as shown in Fig. 26, for un- 
 balanced three-phase circuits. It is readily seen, where 
 two legs of an unbalanced three-phase system are placed 
 in relation with the potential circuit of only one of them, 
 that a correct registration cannot be obtained. In Fig. 
 30, the current in -the left hand field for a portion of each 
 
 TAAAAA! 
 
 WrWl 
 
 
 
 
 vwwvw i 
 
 Load 
 
 
 Load 
 
 FIG. 29. 
 
 period opposes the current in the right hand field, as they 
 differ in phase by 120 degrees. The torque exerted on the 
 armature is the product of the algebraic sum of the two- 
 field currents and potential circuits, instead of the arith- 
 metical sum of the two-field currents and potential circuit. 
 Another form of monocyclic secondary circuit is that 
 shown in Fig. 31, wherein lights and motors are fed from 
 the same circuit, the teaser current being measured in- 
 dependently of the light circuit. As is shown, two meters 
 
THOMSON RECORDING WATTMETER. 
 
 79 
 
 are used to meter the total energy passing. Various 
 adaptations of this meter have been made to fulfil special 
 conditions, such as the measurements of arc circuits and 
 storage battery circuits where the one meter measures the 
 output as well as input of the battery. 
 
 The Thomson meter, judged in the light of the requisites 
 of a good meter, fails in several qualifications, and has 
 several faults peculiar to itself. 
 
 FIG. 30. 
 
 The torque of the meter is small with regard to its 
 frictional counter torque, having a ratio of 50:1 in a new 
 meter, but more like 20:1 after it has been in service a few 
 months. The effects of this are slowness on light loads and 
 consequent loss of revenue. 
 
 In the matter of friction balancing, we have seen that a 
 coil is provided which produces a turning movement on the 
 
80 
 
 AMERICAN METER PRACTICE. 
 
 armature supposedly equal to friction of the moving parts. 
 As this friction is an ever varying and often ever increasing 
 quantity, the friction balancing produced by the coil is un- 
 satisfactory. When the friction is diminished, by what- 
 ever cause, the meter will "creep" on no load, and lead to 
 complaint and distrust on the part of the consumer. 
 
 FIG. 31. 
 
 The supporting of the armature, which weighs, complete, 
 over half a pound, on a hardened steel point resting on a 
 highly polished jewel surface, furnishes a very small amount 
 of friction while the jewel and shaft end retain their integ- 
 rity, but causes no end of trouble the minute either of them 
 becomes rough. 
 
- THE T>* 
 
 UNIVERSITY 
 
 THOMSON RECORDING WA TTMETE^S&^l^^ 
 
 The armature is subjected to vibration, and rapidly 
 pounds the jewel to pieces. This causes the friction of the 
 meter to increase enormously and frequently stops it. . 
 
 The tropical type of meter is supposed to be air-tight, but 
 it is not sufficiently so to exclude acid fumes, insects and 
 dust, and all of these exert a deleterious effect on the 
 accuracy of the meter. 
 
 An apparently trivial defect is the flaking off of the 
 paint used on the permanent magnets. This paint is 
 magnetic, and will stick out like stiff bristles from the poles 
 of the magnet and drag on the disk, retarding the meter 
 appreciably and sometimes stopping it. 
 
 The remedy for this trouble lies in tinning or copper 
 plating the magnets instead of painting them. 
 
 From a long series of records kept very accurately, it was 
 found that this defect was a very serious one and led to the 
 loss of much revenue. 
 
 Stray iron filings produce a like effect, and, in the small 
 meters having iron set screws in the binding post, it has been 
 found that, in setting up the screws with a screw-driver, 
 small clippings of iron have been dropped on the disk with 
 the above result of slowing or stopping the meter. 
 
 Where vibration is present, the commutator becomes 
 rpugh from the sparkling of the brushes and creates friction 
 which greatly interferes with the meter's accuracy. A 
 piece of hard linen tape passed around the commutator 
 and under the brushes will polish it nicely when pulled 
 briskly to and fro. Only in extreme cases should crocus 
 cloth be used, as it lodges small silver particles in between 
 the commutator bars, tending to short circuit them unless 
 a strong air blast is used to dislodge the silver dust. 
 
 The care and maintenance of the meter are outlined more 
 fully in later chapters. 
 
CHAPTER VII. 
 
 The Duncan Recording Wattmeter, for Alternating Current. 
 
 The Duncan intergrating wattmeter for alternating cur- 
 rent is, in its present state, the work of many years of 
 development. Attention has been paid to the mechanical 
 requirements of a good meter with the pleasing result 
 
 shown in the accompanying cut, 
 Fig. 3 2 , which shows the exterior 
 appearance of the meter. Fig. 
 33 gives a clear idea of the 
 interior construction, the me- 
 chanical and electrical details of 
 which will be briefly reviewed. 
 The meter is essentially a 
 rotating field alternating current 
 single-phase motor with an in- 
 verted aluminum cup for an 
 armature, and its fields consist 
 of series and shunt coils placed 
 at angles of 90 degrees. 
 
 The series field coil of an- 
 nealed sheet steel laminations 
 carries the form wound series 
 coils on inwardly projecting 
 
 pole pieces, while the shunt field is placed inside of the 
 armature drum with its axis at right angles to the axis of the 
 series fields. The shunt field is also wound on a laminated 
 core v/hich has a central hole to allow the shaft of the 
 
 FIG. 32. 
 
DUNCAN RECORDING WATTMETER. 
 
 83 
 
 revolving armature to pass through it. This rotating shaft 
 carries the usual gear for operating the dial train, and is 
 supported at its lower extremity on a spring seated sapphire 
 jewel bearing. The pivoted point of the shaft is removable 
 
 FIG. 33. 
 
 in event of its becoming injured. Just above the jewel 
 bearing, the shaft carries the retarding disk, which revolves 
 between the poles of two permanent magnets, and acts as a 
 magnetic brake, the power of which is proportional to the 
 speed. 
 
84 
 
 AMERICAN METER PRACTICE. 
 
 The operation of the meter is as follows : 
 
 The current flowing in the series coils sets up a plane of 
 magnetization at right angles to that set up in the shunt 
 field coils, but, owing to the self-induction of the latter, and 
 the impedance in series with it, its magnetization lags be- 
 hind that of the series coil, creating a shifting or rotating 
 magnetic field which actuates the closed secondary or 
 armature in a well-known manner. The torque produced 
 varies with the energy flowing. If the load on the meter be 
 non-inductive, the magnetism of the shunt field must lag 
 exactly 90 degrees behind the E. M. F. of the circuit. 
 
 FIG, 34. 
 
 Referring to the diagrammatic view of the relation of the 
 armature with the series and shunt coils, Fig. 34, we find at 
 5, a secondary coil closed through the resistance 8. As the 
 current in the shunt field coil lags less than 90 degrees be- 
 hind the E. M. F., the resistance 8 is so adjusted that the 
 induced current flowing in the secondary 5, acting in con- 
 junction with that flowing in the shunt field coil, forms a 
 resultant field in quadrature with the E. M. F. 
 
DUNCAN RECORDING WATTMETER. 85 
 
 The initial friction of the moving parts is compensated 
 for by the action of a small movable disk of iron, 
 surrounded by a band of copper shown at 6, Fig. 34. This 
 compensator is carried on an arm whose axis of rotation is 
 co-incident with that of the armature shaft, and is movable 
 concentrically to the armature over an angle of about 1 20 
 degrees. The purpose of the compensator is to give an 
 auxiliary torque to the armature independently of the 
 series fields, this torque being so adjusted as to just counter- 
 balance the friction of the moving parts. When the axis 
 of the compensator makes an angle with the axis of the 
 shunt field coil, its magnetism is distorted owing to the iron 
 in the compensator. 
 
 Eddy currents, set up in the armature by these lines of 
 force cutting it, lag behind the current in the shunt field 
 coil. The consequent magnetic pole of the inner surface 
 of the armature, being of like polarity, repels, while the 
 outer pole is attracted to the unlike pole set up by the eddy 
 currents generated in the copper band of the compensator. 
 Hence, we have a torque proportional to the angle which 
 the axis of the compensator makes with the axis of the 
 shunt field coil. If the compensator is moved backwards 
 or clockwise from the axis of the shunt field, the armature 
 receives a corresponding torque of reverse direction. 
 
 The very easy manner in which this compensator can be 
 shifted enables the meter to be adjusted to suit a local con- 
 dition; that is, if vibration be present, the meter may creep 
 under the adjustment suitable for no vibration, but can be 
 adjusted to just balance the friction on the customer's 
 premises without altering the adjustment on full load. 
 
 This compensator answers the same purpose that the 
 auxiliary coils in other meters do with the advantage, in its 
 form, of being more readily adjusted. 
 
86 AMERICAN METER PRACTICE. 
 
 The small cross section of the series pole pieces allows 
 the coils to be of small diameter, thus cutting down the 
 length of wire for a given number of ampere turns and re- 
 ducing the PR losses of the fields. 
 
 The impedance coils which, in a single-phase meter of this 
 character, have more inductance than resistance, are in 
 series with the shunt field circuit, and serve to displace the 
 phase between the series and shunt fields, giving the ro- 
 tating magnetic field necessary for the rotating of the 
 armature. The high inductance of these coils reduces the 
 wattage in the shunt coil to a minimum in this particular 
 meter down to about 2 watts. We have already pointed 
 out the manner in which the phase of the potential circuit 
 is made to lag exactly 90 degrees behind the current in 
 the series fields. 
 
 The case of the meter consists of a heavy cast iron back 
 with three supporting lugs. The cover is of sheet metal, 
 hinging to the case by means of a slot and tongue, and 
 fastened at the top by a screw over the head of which passes 
 the seal wire. 
 
 The cover fits into a groove in the back lined with soft 
 rubber, keeping out insects, dust and moisture, and is also 
 provided with the usual window for reading the dials and 
 auxiliary window for counting the revolutions of the ar- 
 mature. The wires do not enter into the main body of the 
 meter, but are fastened into a binding block at the top of 
 the case. In the two-wire meters, one leg only passes into 
 the meter, the other wire of the circuit feeding the shunt 
 field coils by a half tap. The three-wire meters have the 
 two outside legs taken into the meter either with or without 
 the half tap from the neutral. 
 
 The meters are designed to register accurately on indue- 
 tive and non-inductive loads. 
 
DUNCAN RECORDING WATTMETER. 
 
 87 
 
 The following diagrams are explanatory of the different 
 ways of connecting the meters for various classes of service. 
 Other ways of connecting these meters for various kinds of 
 service will suggest themselves from the explanations, given 
 in Chapter II, on the measurement of power by inductive 
 meters wherein the various connections for two and three- 
 phase systems were given. 
 
 Fig. 35, shows a simple two-wire meter service in which 
 only one main leg is taken into the meter, the potential 
 circuit being fed by a half tap from the other leg. 
 
 FIG. 35. 
 
 Fig. 36, is the same as Fig. 35, except for the larger lugs 
 used for meters over 100 ampere capacity. 
 
 The three-wire system is metered by taking in the two 
 outer legs. 
 
 Fig- 37, the potential circuit, is fed by the maximum 
 voltage of the system. This connection is used for meters 
 of low efficiency type; the regular high efficiency meter is 
 connected as shown in Fig. 38. 
 
 Regular two-wire meters are used for balanced two and 
 three-phase systems and for balanced two-phase systems 
 
88 
 
 AMERICAN METER PRACTICE. 
 
 employing four wires, as shown in Figs. 39 and 40 respect- 
 ively. Of course, the total amount of power consumed in 
 
 PROM 
 
 TRANSFORMER 
 
 TO 
 
 TRANSLATING DEVICES. 
 
 FIG. 36. 
 
 the circuit is found by multiplying the register of the 
 meter by the number of phases. 
 
 FROM J 
 
 TO 
 
 TRANSFORMER. J * 
 
 TRANSLATING DEVICES. 
 
 JLf V 
 
 
 FIG. 37. 
 
 This meter may be taken as a type representative of 
 quite a number of meters such as the Shallenberger, 
 Shaeffer, Guttman, Packard, all of which employ similar 
 
DUNCAN RECORDING WATTMETER. 
 
 89 
 
 principles, but of course differ widely in mechanical details. 
 All of the above, however, have a spring mounted jewel 
 for the armature support in common. 
 
 FROM TRANSFORMER 
 
 _J 
 
 TO TRANSLATING DEVICES. 
 
 
 
 \ 
 
 
 
 CD 
 
 
 FIG. 38. 
 
 In this one point all these types of meters fail to fulfill ideal 
 conditions, by having a variable frictional support for the 
 
 gn T0 
 
 
 MOTOR. 
 
 armature which changes rapidly for the worse with ill 
 usage. 
 
90 AMERICAN METER PRACTICE. 
 
 The torque in this type of meter is usually small, but the 
 ratio between torque and frictional equivalent may be 
 large when the meter is new, and a very accurate register of 
 load through wide ranges may be obtained. The meter is 
 
 FIG. 40. 
 
 TO MOTOR 
 
 not absolutely air-tight, and is subject to all the objections 
 which this implies. 
 
 What may be said of this meter will apply generally to all 
 meters of this type. 
 
r^r OF THE 
 
 UNIVERSITY 
 
 CHAPTER VIII. 
 
 The Duncan Wattmeter Commutated Type, for Direct 
 
 Current. 
 
 Meters of the commutated type for use on direct and 
 alternating current systems have been very few in number; 
 probably the best known is the Thomson recording watt- 
 meter. 
 
 The Duncan meter has recently been placed on the 
 market, and its electrical and mechanical design insure for 
 it a lasting place in the commercial meter field. 
 
 The meter is of the non-inductive type and can be used to 
 register direct and alternating currents with equal accuracy. 
 
 Alternating current, whether inductive or non-inductive, 
 is registered without a change in calibration; but com- 
 mutated types of meters are not used for alternating cur- 
 rent as much as inductive type meters, owing to the cheap- 
 ness of the latter. 
 
 In principle the meter is a Siemens dynamometer, the 
 stationary coils of which are energized by the current 
 flowing in the circuit, and the movable coils or armature 
 proportionally to the potential of the circuit in which the 
 meter is inserted. There is no iron employed in the series 
 or shunt fields, the armature is wound drum fashion over 
 a non-magnetic skeleton support and provided with com- 
 mutator and brushes. The field coils are placed in mag- 
 netic relation to the armature, and the whole operates as a 
 simple motor, the current flowing energizing the fields 
 the potential of the circuit, the shunt field coil or 
 
 91 
 
92 
 
 AMERICAN METER PRACTICE. 
 
 armature. The difference in operation from a shunt 
 motor is that the shunt field circuit revolves and the 
 series circuit remains stationary. 
 
 The details of electrical and mechanical design will prove 
 more interesting than a repetition of the theory of opera- 
 
 
 FIG. 41. 
 
 tion of meters of this type contained in the Chapter VI, 
 on Thomson wattmeter, with which we are already 
 familiar. 
 
 Fig. 41, shows meter with case on and it is noticed that 
 
DUNCAN METER. 
 
 93 
 
 the meter hangs from the top lug of the frame and is lev- 
 elled horizontally by the side lugs. This is a very con- 
 venient feature in installing, as the workman can insert 
 his top screw and then hang the meter on it. The capacity 
 and voltage of the meter are marked in plain figures on the 
 
 
 FIG. 42. 
 
 case. The case is made air-tight and is hinged at its lower 
 extremity to the meter frame as shown in Fig. 45. When 
 the case is lowered the hinge holds it in the position shown, 
 and when it is lifted a little it comes off. This is a very 
 convenient arrangement for inspection and testing. 
 
94 
 
 AMERICAN METER PRACTICE. 
 
 The frame is made of a single aluminum casting, very 
 light and rigid, and handsome in appearance. Around the 
 front edge of the frame is a groove lined with felt into which 
 the cover is securely clamped, rendering the meter dust and 
 insect proof when closed. A seal is inserted over the screw 
 
 
 FIG. 43. 
 
 holding the cover closed, rendering access to the meter by 
 unauthorized persons impossible, except by breaking the 
 seal. 
 
 Fig. 42, is a front view of the interior of the meter with 
 the case removed. In the foreground are the auxiliary, 
 
DUNCAN METER. 
 
 95 
 
 or starting coil, and field coils, showing their relation to the 
 armature coils. The aluminum brake disk is shown below 
 revolving between the poles of two permanent magnets. 
 In Fig. 43, the starting coil and one field coil are removed, 
 
 FIG. 44. 
 
 showing the accessibility of the armature, and a still 
 further illustration of this feature is seen in Fig. 44, where 
 the essential elements of the meter are taken apart without 
 disturbing the other parts. The great accessibility to 
 all parts of the meter, and the ability to take it 
 
96 AMERICAN METER PRACTICE. 
 
 apart by the removal of a few screws, are valuable features 
 when repairs or inspection are necessary. 
 
 The perforated cage at the back of the meter encloses 
 from mechanical injury the resistance coil in series with the 
 armature circuit. This coil is wound of fine resistance 
 wire, securely held in place and thoroughly insulated from 
 the meter frame. The round hole near the top of the 
 meter, Fig. 46, is for the entrance of the feeding wire for 
 
 FIG. 45. 
 
 the series coils, and the half tap for the potential circuit is 
 led in from the top. The terminals for the leading-in wires 
 are mounted on a wooden block just inside the frame, and 
 are inaccessible when the meter is closed. 
 
 Fig. 47 shows the case removed from the hinge, but 
 otherwise is a repetition of Fig. 46. 
 
 Figs. 48 and 49 are views of the revolving element and 
 brushes respectively. The worm gear at the top of the 
 
DUNCAN METER. 
 
 97 
 
98 
 
 AMERICAN METER PRACTICE. 
 
DUNCAN METER. 
 
 09 
 
 revolving element meshes into the actuating gear of the 
 dial train in the usual manner, the relation of the gears in 
 the dial train being so proportioned that all sizes of meters 
 register directly in watt hours no matter what the speed of 
 the revolving element may be. This feature eliminates 
 errors due to the faulty recording of the multiplier by the 
 meter reader or office clerk. 
 
 FIG. 48. 
 
 The commutator is built up of eight segments with air 
 insulation between the bars, and is composed of a non-oxi- 
 dizing metal. The brushes bearing on this commutator 
 are tipped with the same metal. 
 
 The armature is wound with very fine silk covered wire, 
 with a large number of turns in ?ach section, and is 
 
100 
 
 AMERICAN METER PRACTICE. 
 
 thoroughly insulated from the shaft. The aluminum brake 
 disk is carried just above the hardened steel removable 
 pivot which bears on the spring mounted sapphire bearing. 
 The spring supporting this sapphire is so proportioned as to 
 cushion the blows due to vibration by the revolving el- 
 ement so as to prolong the life of the jewel. The exact 
 strength of this spring is a very important feature. The 
 jewel bearing is contained in an ordinary filister head screw 
 which can be easily removed for inspection or repairs. 
 
 BHBBSBEBHB9BI 
 
 
 FIG. 49. 
 
 The field coils are supported on studs mounted in the 
 frame of the meter, and can be removed as shown in Fig. 
 43. The whole mechanical design of the meter is good, and 
 is pleasing to the eye, embodying at the same time neatness 
 and rigidity. 
 
 The dial train records in K. W. hours, and has five re- 
 cording circles in a straight line. The figures are plain and 
 
DUNCAN METER. 101 
 
 the elimination of the constant or multiplier is a good 
 feature. 
 
 The meters are manufactured for two and three-wire cir- 
 cuits; diagrams of the various connections need not be 
 given as they are the same as those shown in Chapter II, 
 for the various classes of service. 
 
 The meter will be considered in connection with the 
 various requisites of a good meter given in Chapter III. 
 
 First: Varying Voltage and Load. 
 
 On loads from 5 to 100 per cent, the meter has a very good 
 accuracy curve, but, as the friction is equal to about 2 per 
 cent, of full load torque value, any increase in friction has a 
 deleterious action on accuracy. The mechanical design 
 of the meter is such, however, as to reduce this loss to the 
 lowest limit for meters of this class. The voltage is reg- 
 istered correctly through a wide variation, no appreciable 
 error arising from this source. 
 
 Second: Varying Frequency. 
 
 As the meter has no iron in its magnetic circuits, the 
 effect of varying frequency is eliminated. The meter needs 
 no change of calibration for circuits of 7,200 or 16,000 alter- 
 nations. 
 
 Third: Power Factor. 
 
 Variations of power factor are met accurately, a full 
 description of power factor variations for non-inductive 
 meters being given in Chapter II. 
 
 Fourth: Wave Forms. 
 
 Owing to the absence of iron in the magnetic circuits 
 various wave forms have no effect on the accuracy of the 
 meter. 
 
102 AMERICAN METER PRACTICE. 
 
 Fifth: Short Circuit. 
 
 A short circuit on the lines could cause a rush of current 
 which would largely increase the magnetic field surround- 
 ing the field coils and, if heavy enough, cause an alteration 
 of the magnetic strength of the permanent magnets. 
 There is no provision made for shielding against such an 
 occurrence. 
 
 Sixth: Permanence of Magnetic Drag. 
 
 The manufacturers claim a very high degree of perma- 
 nence for the magnets owing to a special process which is 
 the outcome of long experience. The meter has not been 
 in commercial use long enough to verify or disprove this 
 statement. 
 
 Seventh: Torque. 
 
 The torque of the meter is as large as it is possible to 
 obtain from the energy expended and the design of the 
 meter. The torque is about fifty times as great as the 
 counter frictional torque, and the ratio of the meter is 
 therefore one to fifty. From Chapter IV, we find that it 
 would not be good design to further increase the torque at 
 the expense of further energy losses. Hence, the meter 
 fulfills the requirements of torque for the given frictional 
 load. 
 
 Eighth: Friction and Friction Balancing. 
 
 Every device known to the art has been used to reduce 
 friction from the moving parts, but as all commutated 
 meters have a brush friction which it is impossible to 
 eliminate,, besides a dial friction and a friction due to the 
 weight oflhe revolving member on the lower bearing, it is 
 not entirely frictionless. A compensating coil is provided 
 
OF THE 
 
 UNIVERSITY 
 
 DUNCAN. METER. 
 
 to overcome the initial friction of the meter, which is found 
 to be about 2 per cent, of full load torque value. This 
 friction compensator is not adjustable. 
 
 Ninth: Energy Losses. 
 
 The losses in the potential circuit are, for the no volt 
 series, about four watts, and vary in the field circuits 
 according to the size of the meter, averaging about i per 
 cent, of the capacity of the meter at full load. 
 
 Tenth: Armature Support. 
 
 The armature is supported on a highly polished sapphire 
 bearing, which is mounted on a spring of such tension as to 
 produce a cushion for the armature to all jars or vibration 
 to which the meter may be subjected. This type of bearing 
 is the best known to-day, where end bearings are used at 
 all. The hardened steel pivot resting on the sapphire is 
 ground to a ball point and highly polished. The aluminum 
 brake disk lightens the revolving element and reduces the 
 friction due to weight. 
 
 Eleventh: Air Tight. 
 
 The construction of the cover and its seat in the felt- 
 lined groove render the meter practically air-tight. It is 
 proof against acid fumes, dust, insects, etc., which are all 
 so injurious to meter accuracy. 
 
 Twelfth: Temperature Co-efficient. 
 
 The temperature co-efficient has been carefully worked 
 out and is, for all practical purposes, unity. 
 
 Thirteenth : Insulation . 
 
 The insulation between wires and frame exceeds one 
 megohm. 
 
104 AMERICAN METER PRACTICE. 
 
 Fourteenth: Mechanical Features. 
 
 In the various figures illustrative of the meter, the 
 mechanical features were dwelt upon. The main points 
 are rigidity and lightness, and this meter combines both in a 
 remarkable degree. 
 
 It is noted from the foregoing that the requisites of a 
 good meter are fulfilled in a satisfactory manner by the 
 Duncan commutated meter, and it is a worthy example 
 of careful design and well-thought-out details. 
 
CHAPTER IX. 
 The Stanley Recording Wattmeter. 
 
 The meters heretofore described have one essential 
 characteristic feature in common, a jewel bearing for sup- 
 porting the armature shaft. It is well known that in 
 alternating current meters of the induction type the alter- 
 nations of the current impart a vibration to the armature 
 shaft, which in time roughens and destroys the smoothness 
 of the supporting jewel surface, thereby introducing friction 
 and impairing the accuracy of the record. 
 
 In the Stanley meter the lower bearing or jewel support 
 is entirely dispensed with, the armature being magnetically 
 floated. The flat aluminum disk armature is mounted in a 
 soft steel core, which is sucked up, so to speak, to a position 
 of flotation by a magnetic system composed of permanent 
 magnets. 
 
 The shaft is aligned in a vertical position in this 
 system by hardened steel wire passing through phosphor 
 bronze guide rings in center of armature core. 
 
 These guide rings are very accurately placed, so that the 
 axis of rotation corresponds as closely as possible with the 
 axis of the magnetic system. Upon the correctness of this 
 adjustment depend in great measure the frictionless qual- 
 ities of the support. 
 
 Fig- 50 presents a view of the exterior of the meter. Fig. 
 51 shows the front part removed, disclosing the front 
 compartment containing the permanent magnets, etc. 
 
 In the foreground, Fig. 5 1 , are seen the permanent mag- 
 nets with their poles in contact with pole pieces pierced 
 
 105 
 
106 AMERICAN METER PRACTICE. 
 
 with holes, with projecting rings corresponding to like 
 rings in the armature core. This enables the core to assume 
 a definite position in the magnetic system. A clear 
 understanding may be had of this magnetic relation 
 if we liken it to that of a solenoid, the position assumed 
 
 FIG. 50. 
 
 MODEL G METAL ENCLOSED. 
 
 by an iron core in the solenoid corresponding to the mag- 
 netic flotation of the armature shaft in the flux established 
 by the permanent magnets. 
 
 Adjustable pole pieces are provided which, in con- 
 junction with the armature disk revolving between 
 
STANLEY RECORDING WATTMETER. 107 
 
 them, furnish a magnetic brake of the usual character 
 for the meter. These pole pieces are adjusted vertically, 
 thereby varying the reluctance of the air gap, and, the 
 magnetic flux passing through it, resulting in a like 
 variation of the power of the braking action. 
 
 FIG. 51. 
 
 FRONT VIEW OP SUPPORTING FRAME, SHOWING BRAKE 
 MAGNETS, DIAL, ETC. 
 
 Fig. 52 shows in /more detail the magnetic suspension 
 elements of the^recent model G meter, and the method 
 of suspension of tffe ^magnetically floated armature. 
 
 The magnetic / suspension elements consist of the 
 
108 AMERICAN METER PRACTICE. 
 
 1 
 
 permanent magnet Y, and the steel plugs or pole 
 bushings, A and B. 
 
 The rotating parts floated in air are the aluminum disk 
 and the soft steel vertical shaft, called suspension core, 
 to which the disk is rigidly secured. 
 
 The lower end of the suspension core is flanged larger, 
 while the upper end is turned smaller in diameter than 
 the body of the core. 
 
 FIG. 
 
 The flanged lower end of the suspension core enters a 
 cup formed in the lower pole bushing B\ its upper end, 
 when in magnetic suspension, is just inside of a recess 
 in the upper pole bushing A . 
 
 The difference in diameters between the flange of the 
 suspension core and the cup in the lower pole bushing B, 
 as also between the upper end of the suspension core and 
 
STANLEY RECORDING WATTMETER. 109 
 
 the recess in the upper pole bushing A , is such that there 
 is a predetermined space all around the flange in pole 
 flushing B, and all around the upper end of the core in 
 pole bushing A . 
 
 When the suspension magnet is not in place the 
 flanged end of the suspension core, due to gravitation, 
 will naturally rest on the surface of the cup in pole 
 bushing B\ but, due to difference in diameters, it will 
 not touch the circumference of the cup at any point, 
 there being an air space between the periphery of the 
 flange and the circumference of cup. 
 
 When the suspension magnet is not in place, and when 
 the lower end of suspension core rests on the surface of 
 the cup in pole bushing B, the upper end of the core 
 will be slightly below the lower edge of the upper pole 
 bushing A. 
 
 Magnetism passes between the ends of the magnetized 
 plugs (pole bushings A and B), attracts the steel suspen- 
 sion core, with the disk attached, in an upward direction, 
 lifts the flanged lower end of the suspension core from 
 contact with the surface of the cup in the lower pole 
 bushing B, and carries the upper end of the suspension 
 core into the recess in the upper pole bushing A . 
 
 By the laws of magnetism, the magnetic field, into 
 which the suspension core with its attached disk is thus 
 attracted, acts uniformly upon the suspension core, hold- 
 ing the rotating parts in a predetermined definite position in 
 space, free from mechanical support or contact of any kind. 
 
 The actuating part of the meter is contained in a closed 
 iron compartment which shields it magnetically from out- 
 side influence^. A slot in the case permits the arma- 
 ture to revolve between the poles established by the 
 
 series and shunt field coils in circuit with the energy to be 
 / 
 
110 AMERICAN METER PRACTICE. 
 
 measured. The field coils are ribbon-wound and mounted 
 one below and one above the disk. The shunt field coils , four 
 in number, are carried on a laminated yoke, having pro- 
 jecting pole pieces straddling the ribbon-wound series coils. 
 
 The method of obtaining the "lag compensation" for 
 the potential circuit differs in this meter from the usual 
 form of impedance coil with short-circuited secondary. 
 
 The potential circuit is wound around a laminated 
 magnetic yoke having two air gaps in its magnetic circuit. 
 The reluctance of these air gaps is so proportioned that 
 an initial lag angle of about eighty degrees is obtained. 
 The introduction of the armature disk into these air gaps 
 is so designed that the resultant eddy currents set up 
 therein react upon the magnetic flux in such a manner 
 as to displace the effective flux by ninety degrees from 
 the impressed voltage of the system. 
 
 This phase-angle depends upon the conductivity of the 
 inserted closed-circuited disk, and is the complement of 
 the angle of lag of the energizing coil. 
 
 The revolving magnetic field set up between the 
 potential and current fields, cutting across this closed 
 secondary or armature, induces eddy currents therein. The 
 reaction of these currents on the primary field produces 
 rotation of the armature at a speed proportional to the 
 energy passing. 
 
 The service feeds the meter on the left side, the disk 
 appearing, when observed from above, to revolve in the 
 same direction as the hands of a clock. The meter is sealed 
 so as to be air-tight, and is guaranteed by the makers for a 
 period of three years. An especially good feature about 
 the design of this meter is the absolute exclusion of dust, 
 insects, etc., and it is, in this respect, ahead of the meters 
 now on the market. 
 
UNIVERSITY J 
 
 STANLEY RECORDING WATTMETER. Ill 
 
 The energy consumed in the potential circuit is about 
 2 J watts; the PR losses in the fields are small, owing to the 
 small radius of the winding and consequent short length 
 of conductor. Great accuracy is claimed for this meter. 
 As the meter is suitable only for alternating current, 
 the usual methods of connecting it in circuit are the 
 same as apply to all induction meters. The absence of a 
 compensating device for friction shows how much the ratio 
 of frictional equivalent to torque is reduced by this method 
 of supporting the armature spindle. A slight oscillation is. 
 noticed in the armature disk when no load is on which is 
 caused by the slight difference in radical conductivity of 
 the disk in which eddy currents are generated. A zero 
 point is soon reached and the oscillation stopped. It 
 will prove of great interest to watch the commercial life 
 of this meter which in so many respects is a radical de- 
 parture from the ordinary types. Its history is now in its 
 infancy, and no trustworthy conclusions can be drawn as to 
 its life without more extended service. The meter is man- 
 ufactured in the usual commercial sizes and for various 
 '-equencies. This meter promises to fill more fully the 
 requisite of a good meter than any other type, owing to the 
 permanence of its fractional equivalent. The effects of short 
 circuits on the permanent magnets are entirely eliminated, 
 as the field coils are mounted behind a steel shield which 
 effectually keeps any magnetism from straying beyond it. 
 The permanent magnets are long and the pole pieces are 
 held in an absolute relation by being mounted in a brass 
 yoke. This prevents any widening of the air gap and 
 consequent change in meter speed. While the torque of 
 this meter is not any larger than many others, the fric- 
 tional equivalent is so exceedingly small and so constant 
 that a very high degree of permanent accuracy is obtained 
 
112 AMERICAN METER PRACTICE. 
 
 on very light loads. The supporting of the shaft by a 
 magnetic flotation removes all vertical friction, and the 
 horizontal friction of the guides is practically eliminated 
 by placing the armature core in the exact center of mag- 
 netic stress, so that the pull in every side is balanced. 
 The friction of the dial train is probably greater than the 
 other friction, but is inappreciable; hence, it is unnecessary 
 to have a frictional balance. 
 
 The absolute permanence of the armature support 
 renders the meter impervious to vibration, and as 
 it is hermetically sealed, all acid fumes, dust, etc., are 
 excluded, and the elements of the meter retain their same 
 relation towards each other for a long period of time. 
 
CHAPTER X. 
 The Guttman Wattmeter. 
 
 Induction wattmeters of different makes have their main 
 characteristics in common, viz., a series field in quadrature 
 with a potential circuit and some form of friction com- 
 pensator. 
 
 The Guttman wattmeter in its present form possesses 
 
 FIG. 53. 
 
 considerable merit, and has a very pleasing mechanical 
 appearance. Fig. 53 shows the meter with case on, and 
 Fig. 54 with case removed. 
 
 113 
 
114 
 
 AMERICAN METER PRACTICE. 
 
 A better idea of the rotating mechanism of the meter 
 is obtained in Fig. 55, wherein the cover, dial and perma- 
 nent magnet are removed. 
 
 Like the Stanley meter, the armature and retarding disk 
 are combined in one to form the rotating armature. 
 
 In the old style Guttman meter, the armature took the 
 form of a long spirally slotted cylinder, and a separate disk 
 was acted upon by a permanent, magnet to furnish the 
 brake. 
 
 FIG. 54. 
 
 The same principle of operation is used in the new meter 
 as in the old, except that the spirally slqtted cylinder and 
 disk are combined in one. The torque of the meter is 
 obtained by the interaction of he slotted armature disk 
 with the magnetic fluxes set up by a laminated shunt 
 magnet and a small set of series fields. The armature is 
 carried on a short shaft having a ball shaped pivot to fit 
 
GUTTMAN WATTMETER. 
 
 115 
 
 the supporting jewel, and is geared by means of a worm at 
 its upper end to the usual dial train. The whole armature 
 weighs only about five-eighths of an ounce, so that the wear 
 on the jewel bearing is very slight. The shunt and series 
 
 FIG. 55. 
 
 magnetic circuits are composed of two parts ; a laminated 
 steel magnet body carrying the shunt field coil, and the 
 "bridge" carrying the series coils. 
 
116 AMERICAN METER PRACTICE. 
 
 On the inner side of the shunt field magnet is held the 
 compensating device, which is adjustable by means of a 
 screw. A good view of the shunt and series magnetic 
 fields is obtained in Fig. 56, which shows their mechanical 
 
 FIG. 56. 
 
 relation to each other plainer than any amount of de- 
 scription could. 
 
 In the series field magnetic circuit we have two air gaps 
 which add slightly to the reluctance of the entire magnetic 
 
GUTTMAN WATTMETER. 117 
 
 circuit, but this addition is so slight as to scarcely affect the 
 self-induction of the shunt field coil. 
 
 The compensation of the meter for inductive loads is 
 effected by means of a heavy copper band with a small gap 
 in it, attached to the bridge on the lower side. This gap 
 is closed by means of a resistance wire of such size as to 
 permit enough induced current to flow to effect the desired 
 compensating effect. 
 
 The secondary current induced in the band, when of the 
 correct amount, reacts upon the field just enough to cause 
 it to be in apparent quadrature with the impressed electro- 
 motive force. The secondary induced currents in the 
 armature will then be of the right phase to measure 
 the true energy passing in the circuit when acted upon 
 by the flux from the series coils. For low frequencies 
 the resistance of this connection across the gap in the 
 band is decreased, and it is made adjustable for different 
 frequencies. 
 
 The series coils are held by means of aluminum clamps, 
 and are placed in such position as to obtain the maximum 
 rotating effect on the disk. 
 
 Each coil is 1.4 inches long and 1.25 inches in diameter; 
 being small, they have very little resistance or self- 
 inductance. 
 
 The permanent magnet for the retarding effect is, in the 
 main, of the usual form, but differs slightly in its shape 
 from either the Thomson or Duncan. 
 
 The meters are sent out sealed from the factory, and the 
 connections are made to the service wires without getting 
 into the interior of the meter. 
 
 The meters are made in two and three- wire, and are con- 
 nected in circuit in the same manner that any induction 
 meter would be. 
 
118 AMERICAN METER PRACTICE. 
 
 For all meters above fifty amperes a series transformer 
 is used, and a potential transformer is employed where 
 more than 250 volts are required. 
 
 The mechanical features of the meter are excellent; it is 
 light, compact, well made, and is dust and insect proof. 
 There has hardly been time as yet to determine its dura- 
 bility. The dial reads directly in watt hours, and the 
 ratios of the armature shaft and dial train are changed to 
 suit the different speeds in large and small meters. 
 
 The meter has a good temperature co-efficient and low 
 watt losses in the shunt and field circuits. The loss in the 
 potential circuit varies with the voltage and frequency, 
 but ranges from J to ij watts. The field losses average 
 two watts at full load. Owing to the lightness of the 
 armature, the fractional torque is low, and accuracy is 
 claimed on two per cent, of the rated capacity. 
 
CHAPTER XI. 
 The Westinghouse Induction Meter. 
 
 Probably one of the earliest forms of induction meter 
 used for commercial purposes was the Shallenberger ampere- 
 hour meter. This meter has become obsolete in modern 
 practice, although it may still be found connected in a 
 number of instances. 
 
 The Westinghouse company, during the past few years, 
 has been manufacturing a meter that registers watt-hours 
 instead of ampere-hours. It is made single and multi- 
 phase in all standard frequencies. 
 
 Fig. 57 shows the exterior of a single-phase meter, 
 and gives a good idea of the neatness and compactness 
 of its mechanical features. 
 
 In theory the meter is similar to all induction meters in 
 that the field maintained by the potential circuit is in 
 quadrature to the series field when registering non-inductive 
 loads. 
 
 The laminated sheet steel yoke carrying the shunt and 
 series windings is clearly shown in Fig. 58. It will be 
 noticed that the shunt coils are two in number, and are 
 carried on the half of the yoke from which projects two 
 annular shaped pole pieces. The series field consists of a 
 few turns of wire wound round an upwardly projecting 
 pole piece. 
 
 In series with the potential winding is an impedance 
 coil which approximately places the shunt field coil in 
 quadrature with the series field. A short-circuited 
 secondary winding wound over the potential coils is 
 
 119 
 
120 
 
 AMERICAN METER PRACTICE. 
 
 adjusted through the resistance coil shown in Fig. 58, in 
 such a manner as to place the fields in exact quadrature. 
 
 In this secondary winding is placed a resistance bar on 
 which slides the connector A, Fig. 58, which moves to right 
 
 FIG. 57. 
 
 or left to furnish a balancing friction compensator. Before 
 adjusting the meter to run under load, an adjustment for 
 frictional load should be made. Move the connector A t 
 Fig. 58, to the right until the meter runs with a positive 
 movement without any load. The steadiness of this 
 
WESTINGHOUSE INDUCTION METER. 
 
 121 
 
 movement of the meter disk will determine whether any 
 stickiness or intermittent friction exists. When this is 
 determined the sliding contact should be moved to the left 
 until the meter stops. 
 
 The meter may then be adjusted on full load, and 
 for all loads down to about 2 per cent, of the full load. 
 Coils may be cut in or added to the closed secondary 
 surrounding the shunt field coils in order to secure the 
 proper position of the sliding contact A, Fig. 58, near the 
 
 FIG. 58. 
 
 middle of the resistance bar on which it slides. The 
 closed secondary coils are wound in such a manner as to 
 allow of their being easily connected to make the meter 
 correct on frequencies varying over a wide range as from 
 7,200 to 16,000. 
 
 The revolving element consists of a short shaft on which 
 is mounted an aluminum disk of small diameter, this 
 aluminum disk revolves between the poles projecting in- 
 ternally from the magnetic yoke. The magnetization set 
 
122 AMERICAN METER PRACTICE. 
 
 Tip by the current flowing in the shunt and series fields 
 acts upon the aluminum disk, creating therein induced 
 currents which react on the rotating field in a manner to 
 revolve the disk with a torque varying according to the 
 energy flowing. 
 
 When the load on the meter is non-inductive, the current 
 in the shunt field lags 90 degrees behind the E.M.F. of the 
 circuit. As the power factor decreases, the shunt field 
 becomes more in phase with the E.M.F. of the circuit. 
 Hence, the meter registers accurately inductive or non- 
 inductive loads. 
 
 The latest form of meter has a ball bearing support. 
 The end of the armature shaft contains a cup-shaped 
 jewel which rests on a small, highly polished steel ball, 
 which is, in turn, supported in a cup-shaped jewel bearing. 
 The steel ball constantly turns as the meter revolves, and 
 furnishes an excellent support. In fact, so accurate are 
 these meters that a straight line accuracy curve is ob- 
 tained from 2 per cent, to 100 per cent, load, a slight falling 
 off is observed after full load, and on 50 per cent, over- 
 load the meter is from J to i per cent. slow. 
 
 The three-wire meters are provided with two field coils 
 wound over the projecting pole piece, one in series with 
 each outside leg of the three-wire system. Otherwise, the 
 meters are identical with the two-wire type. The shunt 
 field coil is connected between the neutral and an outside 
 leg. Polyphase meters have two distinct single phase 
 units, placed one above the other, and acting on a single 
 armature having two disks. These two single phase units 
 are connected up exactly as though they were separate 
 meters under separate covers. The meter case is 
 necessarily longer than the single phase units, but other- 
 wise the meters are the same in their make-up. 
 
WESTINGHOUSE INDUCTION ME 
 
 The regulating device consists of the usual permanent 
 magnet, which acts as a brake on the disk which revolves 
 between its poles. The meter is made fast or slow by 
 moving it in or out. These magnets are given a special 
 treatment which ensures their remaining permanent for 
 a number of years. The dial face is provided with five 
 dials which read from right to left. A complete revolution 
 of the first right hand dial indicates one kilowatt hour, and 
 the succeeding dials increase in multiples of ten. No 
 multiplier is used, and the meters read directly in kilowatt 
 hours. The meters are connected for various circuits as 
 outlined in Chapter III. The current consumed in the 
 potential circuit varies from i watts to 2 watts, and the 
 loss in the fields at full load is negligible owing to the 
 small number of turns and the small radius of the coils. 
 
 The torque of the meter is small, but large in proportion 
 to the weight of the revolving element which is unusually 
 light. The ratio between torque and friction of the 
 meter is large, hence the great degree of accuracy 
 obtained over a wide range of loads and on extremely light 
 loads. 
 
 The meter box is sealed so that it is air-tight and dust 
 and insect proof. The inleading wires do not enter the 
 body, but are designed to connect under binding posts 
 contained at the top of the meter. 
 
 The cast iron body containing the meter forms a thorough 
 shield from outside magnetic influences. The presence 
 of strong external magnetic fields, has no influence on the 
 accurate registration of the meter. 
 
 It will be of interest to watch the life of the ball bearing 
 armature support ; it has every indication of being a vast 
 improvement over the pointed steel shaft both in point of 
 life and of lower f fictional value. 
 
CHAPTER XII. 
 
 General Management of the Meter Department Records 
 Testing General Policy. 
 
 The general operation of a meter department divides 
 itself primarily into two distinct branches; the keeping 
 of records, and the technical branch which includes testing, 
 repairs, etc. In large central stations the meter depart- 
 ment employs a large number of men, who are- organized 
 with a clerical and technical force. In some stations the 
 record part is kept distinct from the technical part, the 
 records being a branch of the auditing department. It is 
 general practice, however, for the meter department 
 to keep all its own records and be responsible as a depart- 
 ment for everything which eminates from it. There is 
 necessarily a close bond between this department and the 
 auditing department, and the place where the work of the 
 one is turned over to the other varies in different central 
 stations. The keeping of any kind of records has been 
 worked out along several well defined lines, the alpha- 
 betical system being most generally adopted, as fur- 
 nishing the most convenient natural divisions. The 
 treachery of the memory is proverbial, and it is well to 
 trust nothing of the least moment to verbal communi- 
 cation; hence, the keeping of records not only involves 
 the keeping of ledger accounts, but all of the minute 
 details of the business. If the business be small, it is 
 possible to carry it on with less "red tape" than would 
 otherwise be necessary, but as the number of details 
 
 124 
 
RECORDS. 125 
 
 increase, it is essential to have some well defined system 
 along which to work out and record them. Each central 
 station has its own peculiar forms and systems; therefore, 
 what is here written must be considered as illustrative of 
 simply one system which has been found to fulfill the 
 demands of the service. 
 
 Let us assume that the central station business is 
 divided into three main departments, the engineering, 
 contracting and auditing. The meter department lies 
 under the head of the engineering department, but is 
 in close touch with the other two. The contracting 
 department, we will assume, has charge of soliciting 
 new business, rating, etc., and is a department which 
 represents the company's policy and interests to its 
 customers. 
 
 All communications touching new installations, cut- 
 offs, extensions of service, etc., are received from the con- 
 tracting department to be acted upon by the engineering 
 department. It is good practice to have all such orders go 
 to a special clerk (or clerks) in the engineering department 
 who disposes of them according to their character. Orders 
 issued to the meter department, when fulfilled, are returned 
 O. K. to the clerk, and checked up against the original 
 order from the contracting department. Any discrep- 
 ancy is at once noted and the proper steps taken to remedy 
 the error, or orders may be issued directly to the meter 
 department from the contracting department. In either 
 event the order is re-issued as a record in that department. 
 After an order for an installation has been received, there 
 are various orders for lamps, meters, etc., to be issued by 
 the meter department to the stock-keeper. All of these 
 minor orders are returned O. K. when fulfilled, and a record 
 is kept for reference. 
 
126 AMERICAN METER PRACTICE. 
 
 To avoid the issuance of a great number of lamp orders 
 during the day, each installation man is issued an amount of 
 lamp stock sufficient for one day's needs. On his daily 
 report the name, address and number of lamp used for 
 installation are recorded, and the total must balance with 
 the number of lamps returned in the original order. This 
 system has been found to be exceedingly satisfactory and 
 labor saving, and any shortages are easily checked at the end 
 of the day. The meter orders are returned by the installa- 
 tion man, the number, size, etc., of the meter being checked 
 and 0. K'd by the store-keeper on the issuance of the meter 
 from his stock. In this way a double check is placed on the 
 order and the assurance of obtaining the meter ordered. 
 The daily report, therefore, of the installation man is a 
 check and record of the fulfillment of his orders for that 
 day. It is general practice to keep account of the meter 
 as an individual meter on the card system, the card telling 
 when the meter was received, when first installed, and its 
 general history thereafter. This may prove interesting to 
 one who is collecting data on repair and maintenance of 
 meters, but otherwise is not essential. The meter should 
 be treated as an arc lamp, or any other piece of apparatus. 
 Any peculiarity in running is readily traceable to some 
 source. There should be no such thing as an erratic 
 meter; one that retains peculiar characteristics which 
 would form interesting history. 
 
 If a new meter is sent into service, and, after being 
 installed several months, has to be brought in for repairs, 
 the fields, armature, or any part of it, may be changed and 
 the meter thus loses its original individuality. When re- 
 tested and sent out it should conform to the known effi- 
 ciency curve obtainable from the type of meter. If it does 
 not give this efficiency curve, the trouble which exists- 
 
RECORDS. 127 
 
 somewhere must be found and remedied. The life of the 
 meter can be prolonged indefinitely by the renewal of parts. 
 In five years it is possible for only the frame to be in the 
 same condition as when first installed. There should be no- 
 such thing as the history of a "cranky" meter, as that 
 expression merely means insufficient knowledge or care in 
 locating its troubles. 
 
 An indexed list of tested stock, giving the sizes, numbers, 
 voltage, etc., of the meters ready for insulation is all that 
 is necessary, the meter when ordered out being scratched 
 off the list. The record of the meter after it is installed is 
 preserved on the meter slip as well as in the order book. 
 The meter number is not only a guide to the meter reader 
 and bill clerk, but also enables the consumer to check from 
 his bill against the meter number in his premises. From the 
 foregoing it is clear that the keeping of an individual record 
 of each meter as a meter is superfluous and involves labor 
 which can easily be avoided. The systematic reading 
 and recording of the registers of the consumers is carried 
 on by means of meter slips, giving the name, address, meter 
 number, size, etc., of each meter on the system. The 
 sample meter slip, Fig. 59, is one which was in use several 
 years in a large central station, and was found satis- 
 factory. If the reader's slip be the only record of the read- 
 ing, some form of ledger is usually kept to preserve the 
 reading in case of accident to the slip. It is much more 
 convenient, however, to run a duplicate set of slips which 
 will be exact copies of the reader slip. The advantages of 
 such a duplicate record are many. There is always in 
 the office a duplicate set of the readings of a customer for 
 reference, the slips serve to replace each other in the 
 event of either being lost or misplaced, and further, the 
 slips are more flexible than the ledger. It is a recognized 
 
128 
 
 AMERICAN METER PRACTICE. 
 
 practice in large stations to read approximately the 
 same number of meters each day and have them listed ac- 
 cording to their situation so as to avoid as much walking as 
 
 FIG. 59. 
 
 possible. These meters are indexed and arranged alpha- 
 betically while they are in their cases, and also numbered 
 according to location, so the reader has no difficulty in 
 
OF THE 
 
 UNIVERSITY 
 
 *fUFOR^ 
 
 RECORDS. 
 
 making up his route. Any new customer in the route 
 is given the decimal part of a number in between; for 
 example, 36.1 would be the number given a customer 
 coming between 36 and 37 in a given route and 36.2 
 an additional customer between 36 and 37. After the 
 route is read the readings are copied in the office on the 
 duplicate slips and rendered from them into a ledger file 
 which serves as a journal, from which to post the net bills 
 into the consumer's ledger. The renderings in this file are 
 checked from the reader's slips, so that any errors in copy- 
 ing the readings to the duplicate slips and errors in render- 
 ing are checked at the same operation. This system has 
 been found to give eminent satisfaction, and furnishes a 
 check on each stage of the work without checking each 
 individual operation. The sample ledger file sheet, Fig. 60, 
 is shown, from which a clearer idea can be obtained of the 
 manner of rendering the bill. The dividing line between 
 this clerical work of auditing and meter departments is 
 usually drawn at the K. W. column on the rendering sheet, 
 the making out of the bills, posting, etc., being done in the 
 auditing department. Of course, this is a mere arbitrary 
 practice and is not followed in many stations. Each bill 
 should receive the scrutiny and judgment of the head of 
 the meter department, or his assistant. Any abnormal 
 increase or decrease in a customer's bill should be investi- 
 gated and the reason for the same found out if possible, 
 so as to avoid clerical errors. The meters in circuit 
 should be tested periodically and a record kept of the test, 
 the card system being the usual form of record. Various 
 forms of cards are used, but the one shown in Fig. 61 is 
 both useful and convenient. Besides keeping the record of 
 the test for reference, it is interesting and instructive to 
 record in a day-book the gross amount saved by these 
 
REMARKS 
 
 
 
 
 
 
 
 
 
 
 Net Bill 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 1 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ! 
 
 O 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 j 
 
 
 y 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 snoQ 
 
 
 
 
 
 
 
 
 
 
 Consumption 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 S 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 1 
 
 
 
 
 
 
 
 
 
 
 a 
 
 
 
 
 
 
 
 
 
 
 
 t 
 
 
 I 
 
 1 
 
 j 
 
 1 
 
 1 
 
 | 
 
 S 
 
 NAMES AND ADDRESS 
 
 
 
 
 
 
 
 
 
 
 it 
 
 
 
 
 
 
 
 
 a 
 
 i 
 
 o 
 
 
 
 1 
 
 5 
 
 B z 
 
 FIG. 60. 
 
RECORDS. 
 
 131 
 
 tests. Simply the customer's name is written, and in 
 separate columns the average per cent, fast or slow of his 
 meter with what this means in dollars and cents taken on 
 the basis of his last bill is noted. 
 
 At the end of each month the amount saved by these 
 tests for one month is found by subtracting columns fast 
 from slow. A record of this sort demonstrates clearly the 
 saving that is effected by testing the meters, and from it an 
 idea is obtained of how often it pays to test the meters in a 
 
 METER 
 RECORD 
 
 FIG. 61. 
 
 consumer's premises. Further discussion will be found 
 under head of "Testing." 
 
 The filing of old slips and cut-offs must be done in such 
 a way as to be readily accessible when needed. The alpha- 
 betical system furnishes its own index and is the best to 
 follow. Any form of suitable filing case will do, and it is the 
 practice in many stations to file all documents pertaining 
 to a customer in one envelope, in this way reducing the en- 
 
132 AMERICAN METER PRACTICE. 
 
 tire filing system to one cabinet. Under no circumstances 
 should such records be destroyed. They become invalu- 
 able in the event of a law suit. There are many minor 
 records, little details which vary with local conditions, but 
 the value of keeping in writing, easily accessible, every com- 
 munication pertaining in any manner to the work cannot be 
 too fully emphasized. We pass now to the more impor- 
 tant part of the work, where we encounter the various 
 problems connected with the correct commercial metering 
 of the station's output. 
 
 This part of the work is the golden key which controls 
 the earning power of the plant. It may be divided into 
 four branches: Repairs, Testing in the meter department, 
 Testing in the consumer's premises, General Policy. 
 
 The repair department should be so thoroughly fitted up 
 as to render unnecessary the sending back of the meter to 
 the manufacturer for repairs. A thoroughly competent 
 electrical mechanic must be secured who is capable of 
 making all repairs of every character. Besides the com- 
 plement of small tools used, the workshop needs one 
 small drill press, one grinder and buffer, one lathe head- 
 stock. The electrical instruments needed are voltmeters 
 according to the variety and kind of service used ; a portable 
 testing set such as the "Queen Acme."* 
 
 Meter repairs are distinct enough in character for us to 
 retain the division between the inductive and non-inductive 
 types. The repair of induction meters, having a closed 
 metallic secondary for an armature, is simple. Beginning 
 with the armature and its' shaft, generally the only repairs 
 needed will be a new shaft end, if the shaft end be 
 
 *3ee Kempster B. Miller's "American Telephone Practice" for 
 full description. 
 
OF THE 
 
 UNIVERSITY 
 
 RECORDS - 
 
 removable. Except in case of mechanical violence metallic 
 armatures should never need renewing. The potential wind- 
 ings of an induction meter are liable to burn out as are also 
 the usual impedance coils or coil in series therewith. These 
 coils are usually form wound, and can be obtained from the 
 manufacturers at possibly lower cost than they can be 
 wound in the repair shop. Charts of the proper impedance 
 of the different sections of the potentional windings can 
 be obtained from the manufacturer. The known resistances 
 of the different parts of the potential circuit are more service- 
 able than their impedances, as then the wheatstone bridge 
 can be used to verify this integrity from shunt or open cir- 
 cuit. When the resistance is known, a good method is to 
 connect up in series with the coil to be tested a known 
 resistance. A direct current passed through this circuit 
 should give falls of potential across each coil relative to the 
 resistance of the coils, and deviation from the determined 
 fall of potential would indicate a fault. For example, we 
 wish to test an impedance coil which is supposed to have 
 400 ohms resistance in series with this coil. We place a 
 known resistance of 600 ohms and connect up to a 100 v. 
 direct current circuit. If the 400 ohm coil be all right the 
 fall of potential across it, measured by a volt meter, is 40 
 volts; if partially or wholly short-circuited some value 
 under 40. 
 
 In replacing defective impedance coils, the resistance is 
 relatively of least importance. The exact number of turns 
 of the replaced coil must be present in the new one. 
 
 The series field coils of a meter are frequently burnt out 
 from a heavy overload. These coils can be wound in the 
 repair shop or bought from the manufacturer. As a gen- 
 eral rule it is found preferable to buy all spare parts from 
 the manufacturer and simply assemble them in the repair 
 
134 AMERICAN METER PRACTICE. 
 
 shop. The greater facilities possessed by the manufac- 
 turer for turning out the various parts enable him to 
 sell at a profit for a lower price than it costs to make them 
 specially. 
 
 It frequently happens that the magnetic drag on alter- 
 nating current meters loses some of its magnetism. This is 
 due, no doubt, to the almost imperceptible vibration of the 
 meter from the alternations of the current. New magnets 
 of the requisite strength should replace magnets that have 
 lost magnetism, as they are unreliable. 
 
 The three main matters which fill almost the entire 
 field of repairs on induction meters are the jewel and shaft 
 end, potential circuit and field coils. 
 
 Let us follow a meter through its different stages in the 
 repair shop. It has been brought in from the circuit 
 for repairs. Before examining, its reading, number, 
 size, etc., are taken and recorded in a day book which 
 serves as a running record of the daily receipt of meters 
 by the repair shop. The cover of the meter is removed, 
 thoroughly cleaned and repainted. The meter is then 
 connected in circuit and a load put on it to see whether it 
 is electrically or mechanically imperfect. A rough test of 
 this character demonstrates whether the potential circuit 
 is intact, and a glance at the field coils will tell whether 
 they need renewing. If these parts be all right and the 
 meter still slow, the jewel is examined, and, if cracked 
 or broken, the shaft end as well as the jewel is renewed. 
 If the meter has no jewel or shaft end, the friction must be 
 looked for in the guide journal of the armature shaft. 
 After the rough test, the meter is taken apart and cleaned 
 thoroughly, particular attention being paid to the moving 
 parts. After re-assembling, it is tested roughly for accuracy 
 and grounded frame. It is then ready for the final test 
 
RECORDS. 135 
 
 in the calibrating room, the features of which will be gone 
 into later. This general plan is followed with each meter, 
 no matter whether cut off for repairs or simply discon- 
 nected from service. The mechanical defects in the running 
 of a meter are easily traced. They necessarily lie in either 
 or both bearings of the armature shaft, in the dial train or the 
 clearance either of the armature or the retarding disk. 
 The remedy in any case is easily applied, and presents no 
 difficulty. Taken for granted that the meter is theoretically 
 designed properly, the electrical defects have mainly to do 
 with the continuity or insulation of the different circuits. 
 If both be intact, no repairs are necessary in this 
 quarter. 
 
 In considering the repairs of non-inductive meters, we 
 will confine ourselves to the motor type and take as an 
 example the Thomson wattmeter. 
 
 The repairs to this meter are varied in character owing 
 to its having an armature composed of wire wound in coils 
 and a commutator and brushes. The same general course, 
 however, is pursued that has already been outlined. 
 Mechanical defects which may influence the accuracy 
 of the meter are the same as in the induction meter, and 
 need not be further enumerated. 
 
 After the meter is connected in circuit, if it runs at all, a 
 voltmeter is connected across the terminals of the brush 
 holders, and the fall of potential across the armature coils 
 obtained. This fall of potential varies with different 
 types of meters, but should remain constant through one 
 revolution or armature. If this fall of potential vary 
 more than two volts in the different sections, it indicates a 
 defect in the armature, and if as high as five volts, the 
 meter can on light loads be seen to slow up when the 
 defective section is reached. Armatures which are found 
 
136 AMERICAN METER PRACTICE. 
 
 defective must be replaced by new ones, as it does not pay 
 to find the defective section and re-wind it. To replace 
 the armature the leads are unsoldered from the commu- 
 tator, the set-screws holding the armature body to the shaft 
 loosened, and the old armature slipped off. The new 
 armature is put on, its leads soldered and the commutator 
 given a lead of about 90 degrees from armature coils. Be- 
 fore placing the new armature in the meter it is well to test 
 it in an improvised circuit, corresponding to the potential 
 circuit of the meter. A set of brushes is so arranged that 
 the commutator is revolved under them in the same 
 manner as in the meter. A voltmeter connected as 
 before indicates whether the new armature is all right. 
 The question of re-winding armatures in the repair 
 shop is worth considering where a large number of 
 Thomson meters are used, and the armature renewals, 
 heavy. A regular armature winding machine, such as 
 is used in the factory, can be purchased, or one may be 
 improvised to answer the purpose. The winding is simple,, 
 the sections following each other in regular rotation and 
 the loops brought out for each coil. It is nice, delicate 
 work and can be done advantageously by girls. As many as, 
 eight to ten armatures can be turned out by one girl in a 
 day, so that the cost of labor is very small. In small 
 departments it would not pay to go into this kind of work,, 
 and it would be found cheaper to throw the old armature 
 away without re-winding it. The commutator, if scarred 
 or very dirty, should be polished with a fine piece of crocus 
 cloth and then blown out with a strong blast from a bellows 
 to remove all silver particles that may have lodged be- 
 tween the commutator bars. An extra polish is put on by 
 finishing with a piece of legal tape, and in almost all cases a 
 piece of legal tape is all that is needed. The brushes are 
 
RECORDS. . 137 
 
 cleaned by removing the brush holder and resting the 
 brush against a firm surface and then polishing with a 
 piece of crocus cloth pasted on a stick. After the brush- 
 holder is replaced, if the commutator have a tendency 
 to spark, hold the brushes tightly against the com- 
 mutator with two fingers and revolve the shaft with 
 the other hand, the result is the freeing of both commutator 
 and brushes of any roughness or foreign particles. The 
 brush tension of the meter is of much importance. It 
 should be strong enough to hold the brush on sufficiently 
 hard to make good contact and not to fly off under slight 
 vibration, causing arcing. Anything which causes arcing 
 at the brushes is disastrous to the accuracy of the meter. 
 If the tension is made too strong it is difficult to make the 
 meter run correctly on light loads without increasing the 
 starting coil. There is found by experiment just the 
 right tension for the best results and the tension is ascer- 
 tained by the "feel" of the brush when drawn away from 
 the commutator. The resistance of the coil at the back 
 of the meter in series with the armature is usually marked 
 on it, and if this coil is burnt out it is replaced with one of 
 the same number of ohms. The starting coil may be also 
 frurnt out and both starting coil and resistance may be 
 wound to advantage in the repair shop. In rewinding 
 the starting coils the same number of turns must be put on. 
 When the integrity of the potential circuit is established in 
 every part, the jewel and shaft end being in perfect con- 
 dition, and the copper disk and armature having the proper 
 clearance , the meter should be approximately correct. After 
 the test in the repair room, it is turned over to the testing 
 room for calibration. In changing a meter from a low 
 efficiency to the high efficiency type, the alteration is con- 
 fined to the potential circuit. In fact, the entire potential 
 
138 AMERICAN METER PRACTICE. 
 
 circuit is changed, a new armature, resistance and start- 
 ing coil being put in. The fields, frame and magnetic 
 drag remain the same. 
 
 The change results in an increased efficiency of the 
 meter on light loads with a decreased consumption of 
 energy in the potential circuit. The advantages of this 
 change make it highly desirable and all the low efficiency 
 meters can be gradually changed to high efficiency with- 
 out prohibitive cost. In changing over meters from 50 
 volts to 100 volts, two courses are open, the one to wind an 
 increased resistance until the ampere flow in the potential 
 circuit is the equivalent of what it was at 50 volts. This 
 makes a meter of low efficiency with double the capacity 
 of lights at 100 volts. The other course is to change the 
 armature resistance and starting coil to similar high 
 efficiency 100 volt parts. In either case the capacity of 
 the meter is double that of 50 volts, and at the same con- 
 stant as the regular type 100 voltmeter of the same size. 
 An exception, however, is the ten ampere 50 voltmeter, 
 which also has its magnetic drag halved to bring it to 
 constant one-half. The same general policy is pursued in 
 changing to any voltage. Meters are tested to ascertain 
 whether they will register a given load for a given time 
 accurately. In other words, the energy passing through 
 the meter must make the meter revolve in exact pro- 
 portion to its amounts. Nearly all mechanical meters 
 make one revolution of the armature for one watt hour, 
 hence we derive the formula: 
 
 No. revolutions X constant X 3600 
 
 = seconds in which 
 
 watts 
 
 the meter should make the given number of revolutions. 
 To count the number of revolutions in a given time accu- 
 rately a stop watch is used. A mark is made on the 
 
RECORDS. 
 
 139 
 
 revolving disk, and a meter under test is given a run of say 
 20 revolutions. The exact elapsed time is caught by the 
 stop watch, and the rate at which energy was flowing 
 through the meter is obtained from the instruments in 
 circuit. The application of the above formula will de- 
 termine whether the meter is fast or slow. 
 
 FIG. 62. 
 
 The indicating instruments to determine the amount of 
 energy passing in the circuit are connected in a number of 
 ways. In Fig. 62 the ammeter and voltmeter are connected 
 before the meter, hence the amount of current used by the 
 potential circuit is registered on the ammeter and fall of 
 potential across the fields of the meter by the voltmeter. 
 
 FIG. 63. 
 
 If the meter is calibrated on this basis, the consumer pays 
 for the current used in the potential circuit of the meter as 
 well as the loss in the fields. The usual combined losses in 
 the meter are about 2 per cent, of full load rating, hence in 
 this connection the consumer would pay for 2 per cent, 
 more energy than was actually delivered to him. In 
 Fig. 63, where the ammeter is placed after the meter, the 
 
140 
 
 AMERICAN METER PRACTICE. 
 
 loss in the potential circuit is sustained by the supplier 
 of current, and the loss in the fields by the consumer. 
 This is generally the accepted plan for dividing the losses, 
 and is fairer to the consumer than the other method. 
 In Fig. 56 the ammeter and voltmeter are both connected 
 after the meter, and the central station bears both field 
 and potential circuit losses. We have pointed out in 
 previous chapters that the average loss in meters was 15 
 per cent, of the total amount of current supplied. This 
 being the case, the connection for calibration shown in 
 Fig. 64 will help to eliminate this loss by 2 per cent., and 
 should always be used for this reason in calibrating meters 
 for service. Care should be exercised in connecting the 
 
 FIG. 64. 
 
 voltmeter that its current is not registered by the am- 
 meter. Connections for three-wire meters are modifica- 
 tions of the above. It is assumed that the current to be 
 measured is non-inductive. If an indicating wattmeter 
 replace the volt and ammeter, the series coils of the watt- 
 meter take the place of the ammeter and the potential coils 
 that of the voltmeter. 
 
 TESTING IN METER DEPARTMENT. 
 
 The outfit for calibrating meters in different stations 
 varies with the degree of importance attached to this 
 branch of the work. It is needless to emphasize the fact 
 that a correct calibration of the meters before they are 
 installed is of the utmost importance. 
 
TESTING. 141 
 
 The instruments necessary for the work vary with the 
 class and size of the meters to be tested and 'naturally 
 divide themselves into two classes, direct and alternating. 
 Direct current testing will be considered first. The usual 
 commercial circuits are distributed at no, 220 and 500 
 volts potential. 
 
 To measure the energy passing in a direct current circuit 
 by means of an indicating wattmeter it is necessary to take 
 reversed readings, the mean of which is the true reading. 
 It is found quicker to use a volt and ammeter and obtain 
 the product of the readings. If a station is distributing 
 three direct potentials no, 220 and 500, two standard volt- 
 meters are needed, one with double scale reading from o to 
 150 and o to 300, the other for 500 voltsreading from 250 
 to 600 volts. The number of ammeters can be limited to 
 two, one double scale instrument reading from o to 10 and 
 oto5oini-io amperes and J amperes divided respectively, 
 and another instrument reading from o to 500 amperes for 
 heavy work. 
 
 These instruments should be of standard dead beat type 
 and extremely accurate. The accuracy of all the instru- 
 ments used in testing should be checked once a week, and 
 oftener if inaccuracy is suspected. 
 
 The design and requirements of the testing board vary 
 with the number of meters tested per day, but for any 
 capacity the board should be laid out in such a way as to 
 enable testing to be done as quickly as possible. Before going 
 into the arrangements of switches and instruments a brief 
 discussion of the kind of current to be used in testing is in 
 order. If the station is provided with a storage battery, 
 leads direct from the battery should supply the testing 
 fcoard, so as to avoid the fluctuation of the voltage on the 
 lines. It is customary to charge station batteries at night 
 
142 
 
 AMERICAN METER PRACTICE. 
 
 from about 12 P. M. to 7 A. M. and let them " float " on the 
 lines in the day time. The leads to the testing board 
 would be at practically the same voltage all day. In 
 event of no battery being installed, leads direct from the 
 bus bars of the plant would furnish a steadier current than 
 the lines, and the voltage could be reduced by inserting a 
 resistance in series with the circuit. Leads from the service 
 mains are the least to be desired, owing to the fluctuation 
 
 Gen. 3 volts 
 
 Motor 120 volts 
 FIG. 65. 
 
 Gen. 50 voll 
 
 caused by varying motor loads. In a well-laid-out system 
 the fluctuation is small, but can easily be 3 per cent, 
 either way and is a source of error because it necessitates 
 guessing at an average value. 
 
 A testing system which is by far superior to either of 
 these three is obtained by installing a small storage battery 
 of sufficient number of cells to furnish a three-wire service 
 of 100 to 220 volts. These cells can be quite small, 
 
TESTING. 
 
 either of 7 or 10 ampere-hour capacity, as they furnish 
 current only for the potential circuits of the meter. The 
 use of this battery in photometer work is described in 
 Chapter XVI. 
 
 The load on the series coils of the meter is furnished by 
 either one or two large cells in series of about 600 ampere- 
 hour capacity, and a normal discharge rate of 75 amperes. 
 The short circuit value of two of the cells in series would be 
 about 300 to 400 amperes. These batteries are connected 
 in series with the field coils of the meter through a variable 
 resistance. The connections for the charging set, switch- 
 board and batteries are shown in Fig. 65. 
 
 For storing purposes, this small motor, driving the low 
 voltage generator for the large cells and the booster in 
 series with the line current for the small cells, gives a very 
 neat and effective outfit. 
 
 The motor is mounted in between the booster and gen- 
 erator on one bed plate and is directly coupled to each. 
 
 The meter testing board if laid out for use with line cur- 
 rent must employ either lamp banks or rheostats capable 
 of dissipating large amounts of energy. The system just 
 described uses only about one-fiftieth of the energy con- 
 sumed by the other method. The variable resistance 
 placed in circuit with the field coils of the meter gives any 
 desired load without in any way affecting the voltage of the 
 potential circuit. 
 
 Five-hundred-volt meters can be tested on this battery 
 system by obtaining the fall of potential across the arma- 
 ture and starting coil and applying this voltage from the 
 batteries across this circuit, the field, as in the other low 
 voltage meters, being in series with the low voltage circuit 
 from the large batteries. In effect this test gives the 
 same result as though 500 volts were used. 
 
144 
 
 AMERICAN METER PRACTICE. 
 
 Before laying out the testing board, it must be decided by 
 what current the meters are to be tested , as great simplifi- 
 cation results in using the battery system. 
 
 For small stations the expense of a battery outfit would 
 hardly be justified, and, for this reason, a complete testing 
 
 OuQuudS 
 
 FIG. 66. B 
 
 board employing one lamp bank for the three voltages will 
 be given as well as a diagram of connections for battery 
 system. 
 
 In Fig. 66, we have a three-wire no to 220 volt service 
 and a 500 volt, no volt A. C. and D. C. two-wire 
 
TESTING. 145 
 
 service feeding a common three-wire bus C. From C are 
 let three sets of three-wire leads to a triple row of binding 
 posts at D, E and F. The upper rows of these binding 
 posts feed into bus bar G. Single pole switches are in- 
 serted in leads from C to lower row of binding posts. 
 The bus bar G feeds ten rows of lamps which are con- 
 nected up as shown in diagram, five on each side of the 
 system. The meter to be tested is fed from lower row 
 of binding posts at either D, E or F, and the return 
 leads connected to top row feeding lamp banks. Each 
 row of lamps is provided with single pole switches on each 
 leg, and a tie over switch at H enables both sides of the 
 lamp bank to be used on either side of the system. The 
 rows of lamps are fed in such manner that the wires of 
 adjacent rows are of the same potential and are perma- 
 nently tapped together. The right hand part of the bank is 
 subdivided by breaking the rows of lamps above the 
 second lamp by means of a row of single pole switches. 
 
 The great flexibility of this arrangement is apparent, as 
 a two or three-wire meter can be tested from a load of one 
 lamp up to entire load of lamp bank. 
 
 In testing on 500 volts, all switches are opened except 
 those at the ends of each five rows. The current then 
 flows in series with five lamps in multiple and up to 
 the full capacity of the bank. This form of lamp 
 bank is very convenient in operation and is cheaper 
 in first cost than two distinct banks for low and high 
 voltage. 
 
 The lamps are arranged on the back of the board so as 
 to shield the operator's eyes from the glare and heat. The 
 most convenient arrangement is to have the lamps, switches, 
 etc., mounted on a slate or marble panel, set at the rear 
 dge of the meter testing table. 
 
146 AMERICAN METER PRACTICE. 
 
 Fig. 67 gives diagrammatically the connections used with 
 the large and small storage batteries. A is field battery, 
 B potential batteries, C variable resistance in series with 
 field battery, D regular type Thomson meter. 
 
 As shown the connections are very simple, and need only 
 be modified for testing three-wire meters by placing both 
 fields of the meter in series. It is customary, where a 
 large number of meters are tested daily, to connect up a 
 number of them in series with each other and a carefully 
 calibrated standard meter of the same type. This method 
 is possibly quicker than testing each meter individually, but 
 is liable to several errors which may be overlooked by some. 
 
 FIG. C7. 
 
 When the meters are connected in series, the fall of 
 potential across the fields of the meter diminishes the 
 potential delivered to the next meter by that amount, 
 hence the potential circuit of the succeeding meter operates 
 at a lower voltage than its predecessor; in fact, it will regis- 
 ter less by the exact amount of PR losses in the fields of the 
 preceding meter. The current consumed in the potential 
 circuit of the succeeding meter is registered on the pre- 
 ceding one, so, as we descend down the line, each meter 
 registers less than its predecessor by the watt losses of the 
 former. If the watt losses at full load be 10 watts, and 
 
TESTING. 147 
 
 10 meters were in series, the first meter would register 100 
 watts more than the last one. As the load is varied this 
 percentage varies as the I*R losses in the fields vary, so 
 that, in order to get a correct register, the potential cir- 
 cuits of each meter must be fed separately from the field 
 circuit. This involves disconnecting the potential circuits 
 and feeding them separately, which is troublesome. In 
 the battery systems, where the potential and field cir- 
 cuits are independent in origin, no such trouble exists. 
 The potential circuits need not be disconnected, as they are 
 fed independently by the high voltage battery circuit, which 
 has its positive voltage connected to positive of low voltage 
 circuit in the meter. The individual adjustment of a meter 
 takes as long, whether it be connected in series with many 
 others or by itself; and, as each meter has to receive its in- 
 dividual adjustment, it is no gain in this respect to connect 
 it in series. Meters require a heat-run of from 20 minutes 
 to a half hour to allow the resistances of the potential and 
 field circuit to assume their normal conditions. For this 
 reason, it is a gain to have a number of meters connected in 
 circuit ready to receive their individual calibration. The 
 use of the battery system enables meters to be adjusted 
 with great speed and accuracy. The variable resistance in 
 series with the fields is adjusted to allow a fixed number of 
 amperes to flow. The voltage of the potential circuit is 
 constant and is not subject to varying field strength, as the 
 circuits are independent in origin. For any given ampere 
 value the watts are known and will always be exactly the 
 same for this value; hence, to test the meter, a series of 
 ampere values are taken and the meter speed noted and 
 adjusted. It does not become necessary to make frequent 
 calculations of energy value because it is always constant, 
 the only variable being the load, and that is made whatever 
 
148 AMERICAN METER PRACTICE. 
 
 one chooses. In employing only one source of energy, a 
 varying field strength gives a varying voltage owing to the 
 drop in the leads and field coils, and a new calculation has 
 to be made for each change of field strength. 
 
 To sum up, it appears that the battery system has a 
 number of advantages over any other, whether the meters 
 are tested individually or in series. 
 
 Alternating current meter testing, where a central 
 station employs Thomson meters for both direct and alter- 
 nating current, may be carried out altogether on the direct 
 current testing board, as the calibration is practically the 
 same for either alternating or direct current. 
 
 For induction meters we employ lamp banks, rheostats or 
 motors to furnish a load for the meter. If the load is 
 purely non-inductive, an ampere and voltmeter can be 
 used to measure the energy input, but, as the load may be 
 frequently inductive, it is best to fit up the testing board 
 with indicating wattmeters. In connecting up the meters 
 for various kinds of service, the instructions given in 
 Chapter II must be followed in determining the true watt 
 input flowing through the meter. In other respects the 
 same general rules are followed as given for direct current 
 meter testing. 
 
 TESTING IN CONSUMER'S PREMISES. 
 
 The test of a meter in the consumer's premises is made 
 for either of two reasons : a complaint by the consumer of 
 high bills, or to check the accuracy of the meter for the 
 benefit of the central station. For whatever cause, this 
 test is made in several ways. The merits of all will be con- 
 sidered. The meter may be tested by volt and ammeter, 
 indicating wattmeter, calibrated lamp bank or standard 
 recording meter in series with the meter to be tested. In 
 
TESTING. 149 
 
 testing with volt and ammeter an approximation may be 
 reached, but the method is open to various objections. 
 The fluctuation of voltage on the lines, and the fluctuation of 
 the load in the premises, make it impossible in many in- 
 stances to obtain more than an approximate test. Es- 
 pecially is this the case where the load is a motor running 
 a shop or direct connected elevator. The man making the 
 test has four moving objects to attend to at one time, am- 
 meter, voltmeter, stop watch and moving meter disk, and 
 the consequence is that he can never be sure that the test 
 is a correct one. On alternating current lines, where the 
 load is inductive, it is not possible to get a correct test by 
 this method unless the exact power factor of the circuit 
 passing through the meter is determined. In practice this 
 is not feasible so the above method is never used. An in- 
 dicating wattmeter may be used instead of the volt and 
 ammeter, and has the advantage of reducing the number 
 of objects to be watched by one. On alternating circuits, 
 where the load is inductive, the indicating wattmeter is 
 very much used and is the proper thing in testing where the 
 load is not subject to variation. 
 
 The third method, a calibrated lamp bank, necessitates 
 the shunting of the service and disconnecting the meter 
 from its load. The lamp bank is composed of lamps of 
 known wattage at known voltages, and the only instrument 
 needed is a voltmeter. The lamp bank is connected to the 
 meter, and furnishes a load with which to test it. If the 
 voltage on the lines is variable, as is likely to be the case 
 where a heavy day motor load is connected to the system, 
 the wattage of the calibrated lamp bank has to be averaged 
 in the same manner that the indicated watt readings were ; 
 there is, however, this advantage, that the load is steady. 
 In large installations, where the meter would need at least 
 
150 
 
 AMERICAN METER PRACTICE. 
 
 the equivalent of 100 i6's connected for a test, the lamp 
 bank method is too cumbersome to lend itself readily to 
 economical transportation from house to house. One of 
 the gravest objections to this method of testing is the time 
 and trouble taken in shunting the wiring of a large installa- 
 
 FIG. 68. 
 
 tion, otherwise the consumer is cut out of light for a half 
 hour or more. 
 
 The fourth method, the placing of a standard recording 
 wattmeter in series with the meter in the premises, permits 
 of an accurate test to be made under any variation of load 
 
TESTING. 151 
 
 and voltage. The standard portable meter is easily rigged 
 up from the usual stock type by attaching chains with 
 turnbuckles to the frame of the meter, in such a way as to 
 allow the standard meter to be suspended from the one to be 
 tested by means of hooks. To properly hang the meter 
 four points are selected on the frame, and chains connected. 
 These lead to a hook on each side of the meter which is used 
 to hang the meter from the one to be tested. A diagram- 
 matic representation is shown. Fig. 68 shows connection of 
 chain and a view of the portable meter as it would appear 
 when suspended from the one to be tested. The meters are 
 then placed in series, and the load in the premises applied. 
 The meter disks are held in the same relative position with 
 reference to a mark on them and allowed to start at the 
 same time. After one or two revolutions it becomes 
 apparent whether the meter to be tested is fast or slow by 
 the relative position of the marks on the disks. If slow, 
 the meter being tested is cleaned, the jewel examined and, 
 if defective, is replaced by new jewel and shaft end, if the 
 parts are accessible. Another trial will show the effect 
 of these operations, and, if still slow, an adjustment of the 
 magnets acting as a drag can be made until the meters run 
 in unison. While this test is being made the operation is 
 entirely independent of any variation in voltage or 
 load as both meters are affected alike by any change in 
 either. 
 
 On direct connected elevators and motors of all sorts, 
 this method is the only one which will allow of a correct test 
 being made on the meter with the load which it has to 
 carry. 
 
 The ratio of the energy passing through the standard 
 and tested meters varies with the position of the standard 
 meter; that is, whether it is placed before or after the one 
 
152 AMERICAN METER PRACTICE. 
 
 to be tested. If before, it receives more energy by the 
 watt losses of the tested meter; if after, less energy by the 
 watt losses of the tested meter, that is, assuming both meters 
 to be of the same capacity and efficiency. 
 
 In making a test it is immaterial whether the meters 
 are connected before or after, as long as the watt losses are 
 added or subtracted from the standard. In many makes 
 of meters the watt losses, for practical purposes, may be 
 neglected in making a test. 
 
 To make the matter clearer, suppose the consumer has a 
 T. H. meter of five ampere capacity, and the standard 
 portable meter is of five ampere capacity. If the standard 
 meter is connected in front of the meter to be tested, and 
 the watt losses of the two meters on the same load are 
 equal, then, when 500 watts are delivered to the consumer 
 with a 10 watt loss in each meter, we have the following: 
 
 520 watts delivered to service side of standard meter. 
 4 " loss in potential circuit " 
 6 " " " field " " " =I 2 R. 
 
 510 delivered to meter to be tested. 
 
 Assuming the internal losses of the meters to be unregis- 
 tered, we have the standard meter recording 510 watts of 
 energy and the meter to be tested 500 watts. In other 
 words, on full load the standard would run one-half of i 
 per cent, faster than the meter to be tested. In com- 
 mercial tests a percentage of this magnitude may be 
 neglected, but, as the load decreases, the percentage of 
 error becomes greater as the PR losses in the fields become 
 less; that is, on a load of 50 watts the standard meter would 
 receive approximately 4.6 watts more energy than the 
 meter to be tested, or 9 per cent. In adjusting a meter, it is 
 always made correct on full load first and then its efficiency 
 
TESTING. 153 
 
 on one light load ascertained. If the standard meter is con- 
 nected behind or after the meter to be tested, it runs slower 
 than the one to be tested, and this is bad practice, for this 
 reason. If the consumer is watching the test and sees his 
 meter slower than the standard, he is satisfied that the 
 meter is performing properly ; if his meter, on the contrary, 
 runs faster than the standard, no amount of explanation 
 will satisfy him. Therefore, it is good practice to place the 
 standard meter always in front of the meter to be tested. 
 
 If the test is being conducted for the consumer, he or his 
 representative usually watches the operation and, if they 
 possess no technical knowledge on the subject, the graphic 
 method of placing two meters in series appeals to them 
 more strongly than when conducted by indicating instru- 
 ments of which they know nothing. 
 
 Again, when instruments are used a stop watch is nec- 
 essary to catch the number of revolutions accurately, so 
 that we have to contend with the personal error of the 
 observer in reading the indicating meters and catching the 
 time. It is impossible for one man to do this correctly, he 
 cannot watch his instruments and the meter at the same 
 time. Two men are necessary, and the expense of testing 
 is doubled. The series method needs only one man, no 
 reading or calculations are necessary, a variable load or 
 voltage affects both meters alike, the per cent, fast or slow 
 may be ascertained to any degree of accuracy by taking a 
 large number of revolutions for comparison. All of these 
 advantages, and the saving in expense, place this method far 
 ahead of the others. 
 
 Meters of 100 ampere capacity and over, owing to the 
 large wires to which they are connected, are troublesome to 
 test and make the necessary connections. On any system 
 the number of such meters on the circuit is a very small 
 
154 AMERICAN METER PRACTICE. 
 
 percentage of the whole, and it has been found by expe- 
 rience that the quickest and cheapest way to test these 
 meters is to remove them altogether and replace them by 
 new ones. Suppose in a system of 5,000 meters, 50 of 
 them were 100 ampere capacity and over. To test these 
 meters once a year it would be necessary to change 
 four of them each month; that is, a reserve of four or 
 five meters of different sizes would enable all these meters 
 to be brought back, tested and sent out again. 
 
 The sizes of portable meters, then, to be carried in stock 
 are few in number, say five or six. These meters are care- 
 fully calibrated before being put in service and are checked 
 before and after each day's work. Their calibration, with 
 proper handling, remaining remarkably constant. The 
 jewel screw is always kept lowered and the disk firmly 
 wedged when not in use. 
 
 On a large system of meters, where two or more meter 
 testers are employed, the work is so arranged that each man 
 tests only one size of meter during a day. The tester 
 follows the regular meter routes, and on one day tests one 
 size of meter and the next another and so on, so that the 
 number of portable meters even for a large system may be 
 kept down to one meter for each size. 
 
 The desirability of regular meter testing has been proved 
 so conclusively that it need not be gone into, but the 
 question of how often to test is one which varies with local 
 conditions and the kind of meter used. As a general con- 
 dition twice a year is about right, but in localities where the 
 meters are subject to vibration from heavy traffic more fre- 
 quent tests result in very material saving in the meter bills. 
 
 If a list of the tests made and the amounts saved be 
 kept, such a record enables the question of how often it 
 pays to test to be settled beyond dispute. 
 
GENERAL POLICY. 155 
 
 The General Policy of the meter department should have 
 two ends in view, the giving of as perfect meter service as 
 it is possible to establish, and the maintaining of harmo- 
 nious relations with the consumer. Some natural divisions 
 suggest themselves, such as: 
 
 1. Should central stations make meter tests without 
 charge ? 
 
 2. The education of the Consumer. 
 
 3. Duties of the Meter Readers. 
 
 4. Cleaning meters. 
 
 5. Jewel renewals. 
 
 1 . It has been found by experience that bills complained 
 of are much more easily settled after a test of the meter has 
 been made. This test furnished to the adjuster enables 
 him to know definitely whether the meter is fast or slow and 
 facilitates a settlement in that way. From the viewpoint 
 of cost to the central station it averages about 50 cents to 
 test a meter in the premises, and, as the meters as a rule are 
 slow, the adjusting of the meter to run correctly more than 
 compensates for the cost of the test. As an economic 
 practice, the testing of meters for consumers is desirable, 
 and under no circumstances should they be discouraged 
 by charging for the test. The writer has had occasion to 
 change his views on this point within the past two years 
 since collecting data on the. subject, which data showed 
 conclusively that the central station saves many times 
 over the cost of testing by adjusting the meters to 
 run correctly. The manner in which the record is 
 kept is detailed more fully under "Records" in this 
 chapter. 
 
 2. The widely extended use to which electricity is put 
 in every branch of life is gradually lessening the general 
 
156 AMERICAN METER PRACTICE. 
 
 ignorance of the public on electrical sub j ect s . The ma j orit 7 
 are still wrapped in profound ignorance of the simplest 
 fundamentals, and the meter to such people is a wholly 
 mysterious, eccentric and untrustworthy piece of apparatus. 
 
 It would be a useless effort to try to educate the con- 
 sumer in the principles upon which his meter operates, 
 but he should have a knowledge of the manner in which his 
 bill is determined. Many of them can read the dials cor- 
 rectly, but cannot figure out the bill from the readings. 
 An excellent method of placing in each consumer's hands 
 the knowledge of how to arrive at his bills is to get up an 
 instruction card, showing a dial with a sample reading 
 thereon and go through the process of getting the bill. 
 This card can be mailed to each consumer or left by the 
 meter readers, and, as a result, the consumer's curiosity is 
 excited to know the amount of his bill, and he tests his 
 meter by the dial to find out whether it is correct. The 
 better he comes to know the meter, the more respect he 
 has for its correctness, and fewer complaints are turned in 
 of "high charges." If we could conceive of the condition 
 of every consumer being able to read and test his meter, 
 the complaint adjuster would have very little to do, as then 
 only the real errors would be brought in. 
 
 The results from distributing the instruction cards have 
 been found very beneficial in practice, both in the 
 education of the consumer and in establishing better re- 
 lations. 
 
 3. The duties of the meter readers are not confined to 
 simply taking off the position of the dial hands. It 
 is the commonly accepted idea that cheap labor should be 
 employed on this work, and its duty confined to the mere 
 taking of the reading. If this practice be followed, it be- 
 comes necessary to employ inspectors to examine and 
 
GENERAL POLICY. 157 
 
 clean the meters at regular intervals if an efficient meter 
 service is the end sought. The cost of meter readers and 
 inspectors to the central station is greater than if the offices 
 of the two are combined by making the meter reader 
 an inspector as well. 
 
 4. The practice of giving the meter a thorough cleaning 
 once every two months cannot be too highly commended, 
 and if the meter has commutator and brushes, the passing 
 of a piece of linen tape between them and pulling briskly 
 back and forth gives a high polish which reduces friction. 
 At the same time that the meter is cleaned the jewel is 
 examined and, if defective, renewed as well as the removable 
 shaft end. 
 
 5. The life of a jewel varies with the treatment it 
 receives. Vibration and dust tend to shorten this life, and 
 if the vibration be excessive it becomes extremely short. 
 However, outside of the effects of vibration and dust, the 
 life of a jewel has a somewhat definite period, after which it 
 rapidly deteriorates. In order to avoid the evil effects of 
 letting defective jewels remain in the meter, introducing a 
 friction which causes a direct loss of revenue, it is evident 
 that some definite system of jewel renewals must be in- 
 stituted. If a meter contain a defective jewel there is no 
 economy in letting it remain in the meter, and the con- 
 dition of meter service in which no defective jewels are 
 allowed to remain in would certainly be far superior to any 
 other. Therefore, when the meter reader cleans the meter 
 at periods of two months apart, if he renews the jewel at 
 the same time, he eliminates the loss due to defective 
 jewels. He need not test the jewel to see if it needs re- 
 newing, after the system is instituted the old jewel is taken 
 out and the new one put in, the old jewel being tested after 
 it is brought back to the meter department. The good 
 
15.8 AMERICAN METER PRACTICE. 
 
 jewels are sent out again ; the defective ones discarded. In 
 this way nothing is wasted, and the bad jewels are effectively 
 weeded out of the system. The meter reader's duties then 
 are to read every meter on his route, to inspect, clean and 
 renew jewels in every other meter. The greater part of the 
 time taken in reading is consumed in getting from meter to 
 meter, the actual time of taking the reading is small; hence, 
 the cutting down of the route owing to the extra time con- 
 sumed in cleaning every other meter is not as great as 
 appears at first sight. If a reader on a given route can 
 read 100 meters per day, on the same route he can read 75 
 meters and clean every other one, the reduced distance by 
 leaving off 2 5 meters more than compensates for time taken 
 in cleaning 37 meters. 
 
 As a concrete example, suppose a central station em- 
 ploys four meter readers and three inspectors, the inspector 
 cleaning and renewing jewels once every two months, the 
 addition of two more inspectors would take the place of four 
 meter readers and a much more efficient service secured. 
 In other words you eliminate the going over of the same 
 ground by different men for different purposes, and secure 
 better service at less expense by combining their offices. 
 While cleaning the meter a rough test can be made of its 
 correctness on light loads; in this way many defective 
 meters are replaced to advantage by new ones. 
 
 The examination of the meter loop to see that the cus- 
 tomer's installation is all recorded, and the lookout at all 
 times for ways and means of beating the meters, are among 
 the many if lesser duties of the readers. A good meter reader 
 must be alert, intelligent, courteous. It pays to employ 
 good men for any service, it pays particularly well here. 
 
CHAPTER XIII. 
 Reading Meters. 
 
 Meter reading is a simple operation, yet many fail to grasp 
 the relations of the hands of the dials to one another, in 
 such a way as to avoid making mistakes in taking a reading. 
 
 The majority of meters have five circles composing the 
 dial face, and each circle is divided into ten divisions. 
 Under or above each circle is marked the amount of 
 energy each one indicates in a complete revolution. These 
 units are usually expressed in watt hours, and the lower 
 right hand circle is usually marked "1,000 watt hrs.," 
 indicating that one revolution of this circle registers 1,000 
 watt hours, and each division in it 100 watt hours. The 
 next circle, reading from right to left, will be marked 10,000 
 watt hours, and so on in multiples of ten to the last circle 
 at the left hand side of the dial face. Therefore, reading 
 from right to left, each circle indicates ten times the 
 amount of its predecessor, hence, each division on a circle 
 indicates an entire revolution of the preceding one. If 
 this relation be borne in mind it is almost impossible to 
 make a mistake in reading the meter, even if the hands be 
 slightly misplaced. The hands on the first, third and 
 fifth circles turn clockwise, and on the other two counter- 
 clockwise, which leads to confusion with a beginner when 
 first attempting to read a meter. 
 
 The reading is started with the dial indicating 1,000 
 watt hours and the indication of the hand put down at the 
 nearest 100 watt hours. The second circle is likewise 
 marked down, the number being placed to the left of the 
 
 159 
 
160 
 
 AMERICAN METER PRACTICE. 
 
 first circle's indication. The third, fourth and fifth dials 
 are likewise recorded. 
 
 The relation of the dials to each other must be borne in 
 mind when reading, and it must be remembered that, when 
 a hand stands between two numbers, the one last passed is 
 the one to record. 
 
 The subject can be better explained by means of ex- 
 amples, a few of which are herewith given. 
 
 In Fig. 69, we have 900 watt hours registered as the first 
 circle and the hand of the second dial apparently on o. It 
 is evident that this cannot be on zero, as the space from 9 
 
 10,000,000 
 
 1,000,000 
 
 I00,0u0 
 
 10,000 
 
 1,000 
 
 FIGS. 69 (TOP ROW OF DIALS), 70 (MIDDLE ROW) AND 71 (BOTTOM ROW). 
 
 to o represents one whole revolution, circle i ; hence, we 
 read this 9,900. In dial 3, the hand is again on zero, but 
 it cannot be actually zero as the previous dial did not 
 complete a whole revolution, therefore we read it 9, and 
 the reading stands 99,900. Circle 4, is read 9 in the same 
 manner and the total reading is 999,900. On the fifth 
 circle the pointer is on i, but, as dial 4 did not complete its 
 revolution, we read this zero. It is very easy for any 
 one not bearing in mind the relation of the different circles 
 
READING METERS. 161 
 
 to each other to make this dial read 1,000,900 watt 
 hours. 
 
 In Fig. 70, we have a very easy reading 253,400 watt 
 hours, and in Fig. 71, a succession of 5*5, making a reading 
 5,555,500 watt hours. Asi,ooowatt hours is one kilowatt 
 hour, the total indications of the dial face with multiplier 
 i is 10,000 kilowatt hours. 
 
 After one becomes familiar by practice with reading 
 meters, the dials may be read from left to right with facility, 
 but the safer way is from right to left. Nearly all meters 
 have multipliers in the larger sizes by which the indication 
 on the dial face is multiplied to obtain the number of 
 watt hours that has been used. In some of the small sizes, 
 the indication of the dial face is multiplied by a fraction to 
 obtain the true watt hours. 
 
 Some manufacturers, instead of using a multiplier for 
 different sized meters, change the relation of the worm gear 
 reduction between the dial train and the revolving arma- 
 ture shaft in order to make the dial face always read 
 directly in watt hours. This practice is to be commended, 
 as it tends to simplify the records and eliminate errors by 
 getting the wrong multipliers recorded through oversight. 
 
 Ampere hour meters have dial faces which indicate 
 ampere hours instead of watt hours. The readings are 
 taken in the same manner as described for watt hour meters. 
 
 The majority of consumers of electric current, particu- 
 larly store keepers, mill owners and factories, would like 
 to know how to read their meters and figure out the bills 
 from the readings. A knowledge by them of their daily 
 consumption of current would enable them in many in- 
 stances to institute economies which would result in a 
 material reduction in their lighting or power bill. 
 
 We will suppose, foi example, that the reading of the 
 
162 AMERICAN METER PRACTICE. 
 
 meter on May ist was 986,600 watt hours, and on May zoth 
 it was 1,230,600 watt hours, the consumption of current 
 for the ten days would be 250,000 watt hours, or 250 
 kilowatt hours. To obtain the bill, the kilowatt hours are 
 multiplied by the amount charged per kilowatt hour 
 which we will suppose to be 10 cents. Twenty-five dollars 
 ($25.00) would be the required bill. If the dial face carried 
 a multiplier, the consumption would have been multiplied by 
 it and then by the price per kilowatt hour. Whether for light 
 or power the bill would be determined in the same manner. 
 
 As a 1 6 c. p. lamp, when burned for one hour, registers 
 50 watt hours on the dial face, a rough estimate can be 
 readily made by the consumer to ascertain the correctness 
 of his meter. If ten 16 c. p. lamps are turned on for one 
 hour, the meter will register it correctly, 500 watt hours. 
 A percentage fast or slow would be readily noticeable in a 
 reading of this size, and the consumer in this way can 
 satisfy himself as to the correctness of his meter. If this 
 were more universally done, much idle complaint by the 
 consumer would be avoided. 
 
 To facilitate the duties of the meter reader, a meter slip 
 carrying blank dial faces is used, and the indications of the 
 hands are marked down as they appear. This serves as a 
 check on the reading recorded, and enables a "checker," 
 when the slips are turned into the office, to verify the cor- 
 rectness of the reading. After a dial face has been in use 
 for some time, the fifth circle at the left hand may complete 
 a total revolution between readings. In that event, 
 the previous reading is subtracted from 10,000,000, and 
 the reading found on the dial when read will be added to the 
 amount thus obtained. In other words, the dial will have 
 started all over again. This frequently happens, and 
 sometimes leads to confusion in the mind of the consumer. 
 
CHAPTER XIV. 
 Value of Losses in Meters Relative to Income. 
 
 The meter has been one of the most potent agencies in 
 the development of the central station for the distribution 
 of electrical energy for light and power purposes, but the 
 history of this development is interesting from many 
 standpoints. 
 
 Steam engines and water wheels have been brought to 
 ever-increasing heights of efficiency, due in great measure 
 to the exacting conditions imposed in the various classes of 
 central station duty. In the past few years, the gas 
 engine and steam turbine have been perfected to such a 
 degree that they bid fair to dispute in some measure the 
 claim for first places as prime movers with the steam 
 engine. 
 
 On the purely electrical side, the rapid growth of the 
 dynamo and motor both in size and efficiency, and the 
 many uses to which the latter is put, have divided the field 
 into many provinces, distinct, yet united by their common 
 ancestry. 
 
 The no less rapid perfection of the arc and incandescent 
 lamp has ever widened the lighting field, until to-day 
 electricity holds first place as the illuminant of the world. 
 In the early days of electric lighting, when the commercial 
 side of the problem was first attacked, it was recognized 
 that the broad principle upon which the commerce of the 
 world is carried on constituted the only equitable and 
 profitable basis on which lighting could be done; namely, 
 the paying for exact value received. In commercial life 
 
 163 
 
164 AMERICAN METER PRACTICE. 
 
 we meter every product that is sold either by weight, 
 measure or some known standard. It is unnecessary, how- 
 ever, in ordinary transactions to do this automatically, or 
 to still further complicate matters by bringing in the 
 element of time. The commercial measurement of elec- 
 tricity, as we have seen, must combine all of these factors, 
 and many hundreds of patents testify to the work which 
 has been done along these lines. The idea of charging so 
 much per month for 16 c. p. lamp installed was a crude 
 one, but examples of it are still found in many commun- 
 ities. The lamp then becomes a crude meter, and at the 
 first glance it is seen that this is merely an approximate 
 way of securing an income and gives rates to different 
 consumers which vary over wide ranges. 
 
 The failure of many plants to earn dividends was directly 
 traceable to this method of selling current, and, except 
 where the load and hours are fixed quantities, it is universal 
 modern practice to meter all installations whether for light 
 or power. But the mere fact of placing a meter on a cir- 
 cuit does not insure immunity from loss to the central 
 station; the meter needs care and proper attention, other- 
 wise its readings are not a true record of the current con- 
 sumed in the circuit. 
 
 In the general consideration of the various engineering 
 problems of central station management, the meter is 
 usually relegated to somewhere near the tenth place in 
 matter of importance. Its use is well recognized, but the 
 necessities of its maintenance and operation are usually 
 neglected, until to-day it is the least understood of all 
 branches of electrical engineering, even by those who are 
 acknowledged authorities. 
 
 It is easy to build a meter having a commercial efficiency 
 of from 98 to 99 per cent, on half and full loads, and, at first 
 
l( UNIVERSITY 
 LOSSES RELATIVE TO 
 
 glance, this appears to be a very creditable performance, 
 but the question is what will this same meter do commer- 
 cially on from 2 to 10 per cent, of its load? 
 
 The average load on the meters of a central station lies, 
 as a rule, under 10 per cent, of their capacity; this fact 
 emphasizes the extreme importance of having the meters 
 register correctly on this load. 
 
 In taking up the discussion of the losses in the meters and 
 their relations to income, it is necessary to outline in a 
 general way the main factors of cost of generation and 
 sale of current from a central station. 
 
 The three main factors necessary to secure the suc- 
 cessful operation of a central station are: ist, Never 
 failing integrity of service ; 2d, A market for the 
 current ; 3d, Proper registration of current sold. The 
 first and second of these factors will be assumed to exist. 
 They lie outside the scope of this discussion, and may 
 be considered as the prime requisites of any commer- 
 cial undertaking. On the purely operative side of 
 the business, the management of the meter department 
 easily stands first in its effect on the economy of the sell- 
 ing of current from central stations. The details of this 
 management are set forth in Chapter XII. The cost of 
 current to the central station is determined by its operat- 
 ing expenses and fixed charges per unit of current generated. 
 The fixed charge, after the investment is made, remains a 
 fairly constant factor per unit of current generated, hence 
 economy in operating expenses becomes the chief end of 
 the central station manager. It is necessary, for the sake 
 of brevity, that we accept the conditions which we find 
 prevalent in the actual cost of generation of current with- 
 out going into details. In this connection a brief summary 
 of what is considered fairly representative cost in the art, 
 
166 AMERICAN METER PRACTICE. 
 
 as it is to-day, will be given. The cost of generating a kilo- 
 watt hour varies with the locality, cost of fuel, etc., but the 
 range may be taken from one-half to one and a half cents 
 per kilowatt hour at the switchboard. It is assumed that 
 the plant is of modern construction, with compound or triple 
 expansion engines as prime movers, and having the usual 
 auxiliaries pertaining to the best modern practice. Under 
 such conditions one-half cent per kilowatt hour would 
 represent extremely favorable conditions and one and 
 one-half cent per kilowatt hour good conditions, so that an 
 average of one cent per kilowatt hour may be taken as a 
 basis of cost for the purpose of showing the relative cost of 
 generation to income in contrast with the meter loss to 
 income. In assuming one cent per kilowatt hour as cost 
 at the switchboard it is necessary to assign values to the 
 various items which go to make up this total, so as to show 
 what percentage the different items of cost bear to income. 
 
 Fuel will be taken as three pounds per kilowatt hour at 
 $2.00 per ton, making the cost for fuel per kilowatt hour 
 .0033 cents, leaving .0067 to represent labor, maintenance, 
 oil, etc. 
 
 Let us assume an average daily output of 30,000 kilo- 
 watt hours, and the usual line and meter loss of 25 per cent, 
 between the bus bar and the total amount registered at 
 consumer's. Let us assume, also, an average net price of 
 ten cents per kilowatt hour as registered by the consumer's 
 meters. Then the net price at the switchboard is 7^ cents 
 per kilowatt hour. From the above, the ratio between the 
 cost of coal and the net amount received per kilowatt hour 
 is roughly i to 22, or the cost of coal is only 4^ per cent, of 
 the net income. 
 
 Any loss or saving in the coal pile of from 10 to 20 per cent, 
 only effects the net income by a fraction of one per cent. 
 
LOSSES RELATIVE TO INCOME. 167 
 
 Under the conditions assumed, the total cost of genera- 
 tion is 13 H per cent, of the net income, and, even if the 
 most rigid economies permissible to good practice were 
 instituted, and the cost per kilowatt hour reduced thereby 
 to .75 cents, a saving of only 3^ per cent, of the net in- 
 come would be the result. 
 
 It is the writer's intention, not to belittle any practice 
 which reduces the cost of generation, but merely to point 
 a comparison between those losses and a loss which is in 
 excess of their total; namely, the loss in the meters them- 
 selves. 
 
 It may be safely stated that in a carefully operated 
 central station the net saving in generation, which could 
 be effected by the most rigid economies, will not 
 exceed 3 to 4 per cent, of the net income. Any 
 saving in the cost of generation is limited in extent; 
 that is, can never approach closely to its limit in 
 the present illustration of 13 1-3 per cent, of the net 
 income. 
 
 In taking the total loss in lines and meters at 25 per 
 cent., the average condition of the large central station in 
 America is represented. The usual loss attributable to 
 meters is 15 per cent., and 10 per cent. is the usual line loss. 
 The loss in transmission from the switchboard to the con- 
 sumer will be made as large or small as we please, according 
 to the amount of copper used in distributing the current. 
 The 10 per cent, loss in the lines means that the generating 
 equipment must be 1 1 per cent, larger than that necessary 
 to meet the demands at the consumer's connection. 
 
 Any reduction in line loss means larger copper invest- 
 ment, any increase in line loss means larger generating 
 capacity, so that there is established somewhere an eco- 
 nomical balance which varies for different conditions, but 
 
168 , AMERICAN METER PRACTICE. 
 
 may be roughly taken as corresponding to a 10 per cent, 
 loss in the distributing system. 
 
 This line loss of 10 per cent, in the example given above 
 means that only 27,000 kilowatt hours are delivered to 
 the consumer, of which only 22,500 kilowatt hours 
 are registered after the 15 per cent, meter loss is 
 deducted. 
 
 The line loss should be included in the cost of generation, 
 since, if the line loss be reduced, a reduced amount of current 
 is generated at a reduced cost, and vice versa. 
 
 The meter loss, on the contrary, is in current actually 
 delivered and used by the consumer, but not paid for by 
 him. 
 
 The loss is of such a nature as to affect the dividend 
 directly by its amount. The general expenses of the 
 plant and system are not altered whether it is prevented 
 or not. It seems incredible that a loss of this magni- 
 tude, the prevention of which would effect a saving that 
 would offset the entire cost of generation, should be 
 tolerated. 
 
 The peculiar conditions prevalent in central station 
 practice bring about a condition of affairs which is very 
 trying to the correct registration of current. The 
 meter capacity installed usually equals the load con- 
 nected to the lines. The maximum output of a 
 central station feeding a distributing system for light 
 and power rarely exceeds 20 per cent, of the load 
 connected. The minimum output falls as low as 2 to 3 
 per cent, of the load connected, or the meter capacity. 
 Thus it is evident that the meters, as a whole, register 
 on from 2 to 20 per cent, of their rated capacity, the 
 mean load falling as low as 6 to 10 per cent, of their 
 capacity. 
 
LOSSES RELATIVE TO INCOME. 169 
 
 When the problem is looked at broadly in this light it is 
 not difficult to see why a meter having a high commercial 
 efficiency on half and full loads fails to register properly 
 on a mean load of from 6 to 10 per cent, of its 
 capacity. 
 
 As an example of the trying conditions under which 
 meters work, let us take a store having 100 i6's installed. 
 During the day, from 7 A. M. until 4 P. M., assume that 
 2-1 6 's are burned, and that from 4 P. M. to 6 P. M. an 
 average of 50 lights is used. After 6 P. M., until 7 A. M. 
 next morning, one light is left burning. If the current were 
 correctly registered, the amounts for the different hours 
 would be as follows: 
 
 7 A. M. to 4 P. M 9 hours at $0.02 $0.18 
 
 4 P. M. " 6 2 .50 i . oo 
 
 6 " " 7 A. M 13 " " .01 .03 
 
 $1.31 
 
 In commercial practice a meter of 100 light capacity 
 does not register one or two lights with any degree of 
 accuracy, in all probability not running at all on one light. 
 The loss, in this instance, per day on the small burning 
 of 31 cents would be, say, 25 cents or 18 per cent, of the 
 amount registered. This case in hand represents a preva- 
 lent condition. There are exceptions where the meter is 
 always run at a load in which it is practically correct. 
 Again, the conditions may be easily more adverse. At any 
 rate, they are sufficiently adverse to cause a general loss 
 of 15 per cent, of the current delivered, and that too by 
 meters which, on loads of 20 per cent, of their capacity, are 
 commercially correct. 
 
 The solution of this problem would be one of the greatest 
 
170 AMERICAN METER PRACTICE 
 
 benefits to central stations. That solution so far has 
 not been reached by any meter on the market, but 
 manufacturers are constantly bringing out improved 
 types of meters with higher light load efficiencies 
 and a general improvement may be looked for in a few 
 years when the older types of meters wear out and are 
 discarded. 
 
CHAPTER XV. 
 Differential Rating. 
 
 Electric current is charged for at so much per unit, with 
 discounts of varying amounts applying to the different 
 number of units used in a specified time. A careful study 
 of the load line of a central station reveals the fact that 
 between certain hours of the day a great deal more current 
 is used than at any other time. A plant must be 
 installed which will be capable of meeting the larger 
 demand, and this means that it will lie idle for a great 
 part of each day. 
 
 The investment in this required generating capacity, 
 which is only used for a short period, increases the cost of 
 furnishing current during the hours of "peak" over that of 
 the normal load. However, by charging a uniform rate 
 per unit for each 24 hours, an average rate may be struck 
 which will yield a profit on the business done. 
 
 The plan of differential rating has secured quite a strong 
 foothold, and the theory of it has been worked out along 
 two distinct lines, one to make the maximum demand 
 of the consumer at any time during the month a basis 
 of rating his charge per unit, and the other to charge at 
 a higher rate per unit for current used in the hours of 
 "peak." 
 
 The Wright demand meter is the best known device for 
 the first method of differential rating. The principle of 
 the recording thermometer is used, and the instrument con- 
 .sists of a U-shaped glass tube with a bulb at each end, 
 partly filled with sulphuric acid, and hermetically sealed. 
 
 171 
 
172 AMERICAN METER PRACTICE. 
 
 On the upper bulb a strip of platinoid is wound, this plat- 
 inoid strip is placed in series with the current flowing, and 
 is heated thereby. The heat generated in the strip ex- 
 pands the air in the bulb, and forces the liquid up the other 
 leg of the U into the remaining bulb, where it overflows 
 into a recording tube which is graduated to represent 
 amperes flowing. The instrument can be reset by allow- 
 ing the liquid to flow from the indicating tube by tilting the 
 instrument. The origin of the system was in Brighton, 
 England. 
 
 The demand meter is placed in series with a recording 
 wattmeter, and the total number of watt hours registered 
 is checked against the maximum demand to determine 
 the number of hours per day that the maximum demand 
 was used. Whatever the comparison in the case, the 
 length of time per day that the consumer could have used 
 his maximum demand is noted and the charge adjusted. 
 The number of hours of high charging are usually varied in 
 summer and winter, say a half hour per day in summer and 
 one to two hours in winter. The remainder of the amount 
 registered is charged for at a reduced rate. The indicator 
 will not record short circuits or demands of only a few 
 minutes, but takes fully ten minutes to indicate its load. 
 
 The idea is to make each consumer pay for his propor- 
 tion of plant investment necessary to meet the peak of the 
 load he uses. In other words, if the consumer use the 
 same amount of current each hour of the 24, a minimum 
 plant investment with a maximum return would be 
 the result. If the total energy generated in 24 hours 
 were used in 12 hours, double the generating and line capac- 
 ity would have to be installed to get the same revenue. 
 If the consumer use current for one hour per day to the 
 same amount that his neighbor does in 24 hours, the plant 
 
DIFFERENTIAL RATING. 173 
 
 investment for one would be 24 times as great as for the 
 other, and the same revenue would be derived from both. 
 
 But the weak spot in the Wright demand system is 
 that the consumer does not lend himself readily to the 
 solving of central station problems, and prefers to buy 
 electricity at a fixed price per unit. Again, it is no induce- 
 ment for a consumer to burn light when he does not need 
 it, no matter how cheap it is. 
 
 A modification of the Wright demand system was put 
 on the market by Edward Halsey, of Chicago. The ad- 
 vantages of his device lie in doing away with an addi- 
 tional resistance in circuit with the line and in its 
 adaptability to either two or three-wire meters. In a 
 recording meter having a magnetic drag, the armature 
 shaft is cut in two parts and connected by a sleeve and 
 ratchet coupling. The upper portion of the shaft carrying 
 the armature carries a pointer which travels over a gradu- 
 ated scale laid off on the meter disk. The upper and lower 
 portions of the shaft are coupled by a graduated spring 
 which allows the pointer to assume a position equal 
 to the torque on the armature. When this torque is re- 
 moved, the ratchet prevents the pointer from traveling 
 back to zero, thus leaving a permanent indication of the 
 maximum demand. A viscous fluid is placed in the sleeve 
 which retards the movement of the shaft in such a manner 
 as to prevent the pointer from recording momentary over- 
 loads or short circuits. 
 
 The principles of rating in this and the Wright systems 
 are the same, the difference lying only in the means used for 
 obtaining the maximum demand. 
 
 Another system, which does away with any device at all, 
 is based upon the assumption that the maximum demand 
 will be the consumer's installation or some known 
 
174 AMERICAN METER PRACTICE. 
 
 percentage of it. This system is limited in its application 
 and is not at all suitable for stores or warehouses where a 
 large number of lamps are installed, but is applicable 
 where only a small number are used. The rating is the 
 same as in the Wright system. 
 
 The two-rate meter placed on the market by the General 
 Electric Co., affords another system of differential rat- 
 ing with a different theory at its base. The meter carries 
 two dial trains and a self-winding clock for switching over 
 clutch mechanism at suitable hours for throwing the dif- 
 ferent dials on or off. For instance, if "the peak" lie be- 
 tween the hours of 5 and 7 P. M., one dial is used to meter 
 all the energy consumed during these hours and the other 
 dial that consumed in the other- 22. Such a system can 
 be carried out successfully from a mechanical standpoint, 
 but the effect of the practice has been to drive the con- 
 sumer to using some cheap illuminant during the hours 
 of high charging. 
 
 It is true that long hour burning is encouraged by the 
 cheaper rate given during the day and latter part of the 
 night. The consumer's peak may not correspond with the 
 station peak, and if such be the case, the principles of the 
 demand system are violated, as the consumer fails to pay 
 in proportion to his plant investment. 
 
 It is not believed that differential rating is a permanent 
 solution of the problem of charging for current. If every 
 consumer were an electrical engineer and were willing to abide 
 by his faith in differential rating, the scheme would be 
 extremely practical and just ; but as such is not the case, and 
 as every other commercial problem in the law of demand 
 and supply has its peaks which are worked out on a 
 fixed rate of charge instead of differential rate, he 
 demands the same with regard to his use of electric current. 
 
DIFFERENTIAL RATING. 175 
 
 If a differential rate scheme be adopted, it is necessary, 
 in order to avoid the errors of the system already named, 
 to secure a complete load curve of the consumer's burning 
 each day. 
 
 A device that would take the total number of watt 
 hours registered for a given period and break it up into its 
 component parts, telling how long the maximum load 
 was on, how long the minimum, and also when no lights 
 were burned at all each day, would give a complete load 
 curve of the consumer's burning and furnish a perfect 
 basis for differential charging. A chart ammeter in 
 series with the meter would answer the purpose, or pref- 
 erably still, an integrating movement attached to the dial 
 train. The integrating movement attached to the dial 
 train would be worked in conjunction with a clock and a 
 pointer, actuated by a cam on one of the dial spindles, 
 made to dot off any given watt hour unit consumed in a 
 certain time. 
 
 Any auxiliary device increases the meter investment 
 and also the clerical work, and all of these things must be 
 taken into account in the consideration of differential 
 rating. The problems to be solved for any central station 
 are these: 
 
 The average cost of generating a kilowatt hour. 
 
 The saving that would be effected by the installing of 
 storage batteries and the leveling of the "peaks" in the 
 load curve as against the maintenance of these peaks and 
 differential rating. 
 
 The installing of storage batteries has the two advantages 
 of leveling the load line and making the cost per kilowatt 
 for each hour of the day the same. 
 
 If differential rating be used to secure the proper in- 
 come, the cost of generating a kilowatt hour remains the 
 
176 AMERICAN METER PRACTICE. 
 
 same and we have a very unsatisfactory method of obtain- 
 ing an income. While the general scheme of differential 
 charging is to discourage the peak burning, nevertheless 
 it remains and always will remain as long as the contrasts 
 of night and day exist. 
 
 From a careful consideration of all the conditions, it 
 appears conclusively that the ultimate solution of the 
 basis of charge must follow in the lines of general com- 
 mercial practice; a fixed rate per unit. 
 
 In large communities, such as our cities of New York, 
 Boston, Chicago, Philadelphia, the whole trend of the age 
 is towards consolidation. The railroad and electric 
 interests are frequently affiliated if they do not actually 
 belong to the same people. The current for an entire city 
 may be furnished from one plant or group of plants so 
 arranged that each individual plant is worked to its highest 
 efficiency. 
 
 The ever-growing uses to which electricity is put are 
 evening up the load curves, and the individual ceases to 
 become a short hour burner. The man in his home uses 
 electricity ; when he rides down town to the office he uses 
 electricity; at the office, lights and fans contribute to his 
 comfort; when he goes to dinner he rides on the car again. 
 After dinner he dresses by electric lights to go out to the 
 theatre or club, or if he stays at home he still uses elec- 
 tricity. If he keep an automobile, it is charged from 
 12 at night to 7 A. M., so we find the individual 
 practically a 24-hour consumer of electric current. This 
 current is all furnished by a number of individuals, A. B. 
 & C., associated together in a company. It is a matter of 
 engineering ability to determine whether it pays in any 
 given set of conditions to even up the load line in a plant by 
 means of storage batteries, and this question is for A. B. & 
 
DIFFERENTIAL RATING. 177 
 
 C., to decide. The individual claims he is a long hour con- 
 sumer and entitled as such to the best rates, and there- 
 fore is not called upon to help finance the investments of 
 A. B. & C., other than to purchase current at a fixed price. 
 In the general plan of electrical development it does not 
 pay to discourage the consumer from using current at any 
 hour of the day, but rather he should be encouraged to use all 
 he can at any hour. If told that between the hours of 5 and 7 
 in the evening A. B. & C. find that, owing to poor engineer- 
 ing, they are unable to furnish him with as much current 
 as they would like and will have to charge a higher price 
 at these hours, in order to discourage its use, the con- 
 sumer naturally rebels. The consumer has worries of his 
 own, he does not care to be eternally bothered about his 
 use of electricity, but wants to purchase current at a fixed 
 price per unit as he does every other commodity he uses. 
 
 Differential rating is a temporary makeshift which is 
 bound to go before the demands of the consumer for a 
 fixed price per unit. This has been recognized by a num- 
 ber of the best managed plants in this country, and all 
 schemes for differential rating are avoided as being im- 
 practical and cumbersome. 
 
 We cite, as an instance, the policy of New York Edison 
 Company, which has for a number of years steadily de- 
 creased the price per unit with increased demand, and has 
 successfully carried on a large and lucrative business on 
 the basis of fixed charges per unit. 
 
 American practice has been slow to take hold of the 
 differential rate idea, although a number of large plants 
 are using "Wright" meters on their systems with more 
 or less success. From the trend of the times it may 
 safely be predicted that within ten years' time the differ- 
 ential rate will be a thing of the past, 
 
CHAPTER XVI. 
 Elements of Photometry. 
 
 The 1 6 candle-power lamp was, as it is yet, to a great 
 extent a meter. The flat rate system uses the 16 candle 
 lamp as a basis of charge on a supposed number of hours 
 burning. It is recognized that this is a very crude way of 
 metering current, as the factor of time, one of the most 
 important elements, is unrecorded. The metering or 
 measuring the candle power of* lamps is a determination 
 of their efficiency as regards consumption of current per 
 candle, as well as the mere determination of their candle 
 power. Hence the measuring of candle power is closely 
 allied to measurement of current and is properly included 
 in the same discussion. 
 
 The measurement of light offers many difficulties not 
 encountered in the measurement of such things as time 
 and weight. The candle is the unit to which light intens- 
 ities are referred. The origin of this unit was quite 
 natural, as the candle was one of the earliest and most uni- 
 form of light-giving sources. 
 
 But this unit, while approximately constant, is subject 
 to considerable variation, according to the size of wick, 
 quality of the fuel and height of the flame, besides the 
 influences of atmospheric conditions. The standard Eng- 
 lish candle is supposed to give unit light with a flame 1.8 
 inch high, and a consumption of 120 grains of material 
 spermaceti and wick per hour. The difficulties of obtain- 
 ing a correct reading from such a standard are great and 
 have led to the seeking of various standards, all of which 
 
 178 
 
ELEMENTS OF PHOTOMETRY. 179 
 
 are more or less open to objection. Some of the best known 
 of these may be briefly mentioned without going into a 
 detailed description. 
 
 The pentane lamp has many excellent qualities. Pentane 
 is a refined distillate of gasolene, and in its pure state is an 
 excellent fuel for a standard light-giving source. The 
 difficulty of obtaining it in a pure state is so great that it 
 has been practically abandoned as a standard. The 
 pentane is fed by a wick to the flame, which burns inside 
 a metallic chimney having a slit and a gauge for indicating 
 the height of the flame. 
 
 The Methven screen arrangement, one of the well-known 
 secondary standards, consists of an argand gas flame burn- 
 ing inside the usual glass chimney. On the outside of the 
 chimney there is placed a metallic screen with a slit which 
 allows a light intensity of two candles to pass. The flame 
 burns at a height of three inches. The varying quality of 
 the gas is the greatest objection to this light as a primary 
 standard, although it may answer very successfully as a 
 secondary standard. 
 
 The Hefner- Alteneck lamp, burning pure amyl-acetate, 
 fulfills more fully the requirements of a reliable standard 
 than any other known standard, and has been adopted 
 provisionally by the American Institute of Electrical 
 Engineers as the best candle-power standard yet 
 obtained. For a more detailed description of this and the 
 other standards, the writer takes pleasure in referring the 
 reader to a series of articles by Prof. Wilbur M. Stine, in the 
 March, April and May (1899) issues of the American 
 Electrician. 
 
 Our appreciation of the intensity of light is at present 
 confined to one organ, the human eye, and as every eye 
 has its own particular "personal equation," the comparison 
 
180 AMERICAN METER PRACTICE. 
 
 of light intensities with any standard must vary by an 
 appreciable per cent, according to the observer. Hence, 
 when we speak of the "candle-power" of any source of 
 light, it is more or less a comparative term. 
 
 There are, then, two sources of error to contend with in 
 the measurement of light, a variable standard and the 
 personal error of the observer. In spite of these difficulties 
 the commercial measurement of various light sources has 
 reached a stage where the error amounts to an insignifi- 
 cant percentage. 
 
 A few years ago we rarely heard of the photometer 
 outside of a laboratory. The advent of the incandescent 
 lamp has been the cause of the familiarization of this 
 instrument. There was little need commercially to 
 measure the candle-power of a gas or oil flame; the light- 
 giving qualities of the one are fixed by the size of the 
 orifice of the burner, quality and pressure of the gas ; those 
 of the other, principally by its size and the condition of 
 the wick. The deterioration of the candle-power of a gas 
 flame is dependent primarily on the source of supply, 
 while that of the incandescent lamp, assuming the voltage 
 constant, depends upon its age. This decrease in illuminat- 
 ing power, due to aging, it has been the chief desire of man- 
 ufacturers to overcome. Their degree of success will be 
 brought out more fully later. It is apparent that if the 
 modern incandescent lamp remained of a constant light- 
 giving quality, the factory test would be all that would be 
 needed to establish its candle-power until it was burnt out, 
 and that the deterioration of the candle-power with age is 
 the chief cause of the development of the commercial 
 photometer. 
 
 An incandescent lamp of low efficiency can be made, 
 however, which, when properly aged, will remain 
 
ELEMENTS OF PHOTOMETRY. 181 
 
 practically constant in candle-power if burned at the prop- 
 er voltage. Such lamps are tested by a primary standard 
 in a very careful manner, and can then be used as secondary 
 standards with a fair degree of accuracy ; hence the use of 
 oil and gas flames as secondary standards has given place 
 to a properly aged and tested incandescent lamp. This 
 practice has many advantages besides ease of manipulation 
 in testing, the chief of which are slow deterioration of the 
 
 FIG. 72. 
 
 standard and the same character of light as the lamp to be 
 tested. 
 
 The commercial photometer with an incandescent 
 standard is very simple and lends itself readily to the 
 unskilled handling which it frequently encounters. Two 
 excellent types of photometers designed for the com- 
 mercial testing of lamps are the Queen Standard photom- 
 eter and the Deshler-McAllister instrument, both well 
 
182 AMERICAN METER PRACTICE. 
 
 suited to the needs of the central station. Both of these 
 photometers were designed to supply the need for a rapid 
 and fairly accurate instrument. In photometers of this 
 class the time element of a lamp test is the most important 
 mechanical factor connected with the instrument, affecting 
 directly also its value as a commercial success. 
 
 The Queen Standard photometer, which is represented 
 by Fig. 72, consists of two cast-iron pedestals carrying a 
 rack on which are mounted the photometer carriage and 
 scale. The carriage is rolled to any desired position and 
 
 b 
 
 FIG. 73. 
 
 the reading taken directly from the scale. Two styles of 
 screens are used, the Lummer-Brodhun and Bunsen. The 
 former is about three or four times more sensitive than the 
 Bunsen, but the latter may be read more quickly. Either 
 may be used, but when the number of lamps to be tested is 
 very large, the Bunsen, owing to its quick manipulation, 
 has much to recommend it. 
 
 The Lummer-Brodhun screen consists usually of an 
 opaque piece of gypsum and a set of right -angle prisms 
 with their hypothenusal faces partially coinciding, and 
 
ELEMENTS OF PHOTOMETRY. 
 
 183 
 
 mirrors so arranged as to reflect the light from both sides of 
 the screen, as indicated in Fig. 73, where 5 is the screen, 
 M, M, the mirrors, and P, /?, the prisms. One prism, p, 
 transmits the light and the other, P, reflects it, giving 
 a double field of light which is viewed through the telescope 
 at E. The two beams of light are lettered a and b. If 
 they be of unequal power, the two "fields" of light present 
 a marked contrast, and only become the same when both 
 sides of the screen are equally illuminated. The field is a 
 
 round spot and that of b is a ring surrounding this spot. 
 The contrast is so marked that a balance can be obtained 
 with great nicety. % 
 
 The Bunsen screen, represented diagrammatically by 
 Fig. 74 is probably more familiar, consisting simply of a 
 piece of opaque paper, S, with a grease spot on it. Two 
 mirrors, M, M, are set at such an angle as to reflect this 
 spot and allow the observer to judge of the relative intens- 
 ities of the fields on opposite sides of it. When the spot 
 appears of equal strength in both mirrors the screen is 
 
 
184 AMERICAN METER PRACTICE. 
 
 receiving the same illumination on each side, and the 
 candle-power is read directly from the scale. 
 
 The left-hand pedestal carries the standard lamp, in 
 series with which is a regulating rheostat to enable it to 
 burn at the exact voltage at which it gives 16 c.-p. The 
 right-hand pedestal carries the lamp to be tested, mounted 
 in a revolving socket. A special arrangement allows the 
 lamp to be put in or removed without stopping the socket. 
 A controller permits the socket to be run at the desired 
 speed of 180 revolutions per minute. An adjustable 
 rheostat is also in series with the lamp to be tested, allow- 
 ing the voltage across the lamp to be made anything de- 
 sired within a small range. Between the pedestals is 
 placed a table upon which rest two voltmeters and an 
 ammeter. One voltmeter is connected across the termi- 
 nals of the standard lamp and the other across the terminals 
 of the lamp to be tested; the ammeter is placed in series 
 with the tested lamp. One voltmeter may be dispensed 
 with by putting in a throw-over switch, connecting the two 
 lamps to one voltmeter and taking readings on each 
 alternately. 
 
 The Deshler-McAllister photometer, manufactured by 
 the Electric Motor and Equipment Company, is made in 
 the portable form, but, when equipped with a rotating 
 socket and mounted on a table or bench, it becomes suitable 
 for station work. (See Fig. 75.) The instrument is about 
 five feet long and can be folded up. The working standard 
 of light is a duplex oil lamp at the right-hand end of the 
 instrument. This lamp is lighted and allowed to burn 20 
 minutes; then it is calibrated by a standard incandescent 
 lamp. Calibration is effected by placing a standard lamp 
 in the socket which the tested lamp occupies, placing the 
 screen at the 16 c.-p. mark on the scale, and adjusting 
 
ELEMENTS OF PHOTOMETRY. 
 
 185 
 
 the oil flame until a balance of "fields" is effected, 
 screen is of the Bunsen type, riding over a celluloid 
 on which the candle-power is 
 marked. The station type of 
 instrument is now fitted with a 
 rotating lamp-socket at the left- 
 hand end and an ordinary lamp- 
 socket at the right, the oil lamp 
 being dispensed with. 
 
 Each lamp has in series with 
 it a regulating rheostat for 
 adjusting the voltage. The 
 standard incandescent lamp has 
 many advantages over the oil 
 flame, and its adoption has 
 greatly improved the instru- 
 ment. Two black cloth screens 
 shield the observer's eye from 
 the rays of the standard and 
 tested lamps. A voltmeter 
 and an ammeter are either 
 placed on the bench or 
 mounted on the back of the 
 instrument in plain view. 
 For rapid work the large illum- 
 inated dial type of instrument 
 mounted directly behind the 
 carriage has many advantages, 
 being in plain view and easily 
 read by the observer simply 
 raising his eyes. 
 
 Assuming that a commercial 
 photometer, such as one of 
 
 The 
 scale 
 
186 
 
 AMERICAN METER PRACTICE. 
 
 those described, has been selected, in which the standard 
 is an incandescent lamp, the arrangement of the circuits 
 for testing may be made in several ways, three of which 
 are given, ist, A separate circuit for each lamp. 26., 
 The same circuit for both lamps, with equalizer. 3d, A 
 three-wire circuit fed by a small storage battery. These 
 three methods all assume that the secondary standard is a 
 properly aged and tested incandescent lamp obtained 
 .from some reliable laboratory. 
 
 The first method, in which a separate circuit is em- 
 
 FIG. 76. 
 
 ployed for each lamp, is roughly illustrated by Fig. 76, 
 where A is the lamp to be tested in the rotating socket, 
 and B is the standard lamp. Both lamps are fed from a 
 common two-wire service, C, D. This is preferable to feed- 
 ing the lamp from a three-wire service, as the fluctuation 
 in voltage on either side of such a circuit may not be of 
 the same period or amount and thus cause an unevenness 
 in the intensity of either light, which is avoided by using a 
 two-wire service. Variable resistances, E and F, are 
 connected in series with each lamp respectively, and so 
 
ELEMENTS OF PHOTOMETRY. 
 
 constructed that a small movement of the sliding con- 
 tact produces only a slight variation in resistance. A 
 common form of such resistance is a cylinder of insulating 
 material wrapped spirally with bare resistance wire over 
 which slides a contact ring. 
 
 In series with the lamp, A, is placed an ammeter, Am, 
 and a voltmeter, Vm, is so arranged that by means of a 
 double-throw switch it can be connected to the terminals 
 of either lamp. The voltmeter and ammeter are not 
 essential if a mere comparison of light intensity be made, as 
 any fluctuation in the service e.m.f. affects both lamps 
 alike, but if a wattage test be conducted at the same time, 
 the two instruments are indispensable. The ammeter 
 reading multiplied by the voltmeter reading gives the watt 
 reading, but if the voltage be fairly steady, a close approxi- 
 mation of the wattage may be had by simply observing 
 the ammeter and getting the watt values from a pre- 
 viously prepared table. 
 
 If the service be taken from the street mains and the 
 lamps to be tested be for service on these mains, the 
 standard at B should be of a lower voltage and the vari- 
 able resistance should be adjusted to give this lamp the 
 voltage at which it should burn. If a wattage test be made 
 on the lamps, it is extremely annoying to have the volt- 
 age remain below the normal for any length of time, and 
 for this reason it is best to have the service come directly 
 from the bur bars in the station, and adjust the voltage 
 at lamps to the proper value by means of the variable 
 resistances. The voltage at the bus bars should be kept 
 very steady, and the resistances in series with the lamps 
 after being once adjusted should need little attention. If 
 a wattage test be made on each lamp, the voltmeter and 
 ammeter are left in circuit with the lamp being tested. 
 
188 
 
 AMERICAN METER PRACTICE. 
 
 In this connection of lamps a removal of the lamp, A, 
 only causes a rise in potential at B, equal to the drop on 
 the wires, C, D, caused by the current consumed by A, 
 hence, the voltage of the standard, B, is practically un- 
 affected by the removal of A. In testing a number of 
 lamps, however, it is found that they will vary considerably 
 in current consumed, and this fact, owing to the variable 
 drop caused thereby in the resistance, E, makes the lamp 
 at A burn at a different relative voltage to B, unless the 
 resistance, E, is constantly varied to suit each lamp tested. 
 Therefore, when a comparison is made an error is intro- 
 duced, unless the resistance, E, is changed to give the same 
 relative voltage across A that B receives. To do this, two 
 
 JFio. 77. 
 
 voltmeter readings have to be taken, and, in the meantime, 
 the voltage on C, D may have varied, vitiating the results. 
 A way of overcoming this trouble is shown in Fig. 77, 
 where the same methods of connection are used, but an 
 equalizing wire connected between E and F places the 
 resistances in parallel when the two lamps are in circuit. 
 This connection is to be broken when A is removed from its 
 socket. For convenience, the breaking of the equalizer 
 connection is so arranged in practice as to be accomplished 
 
ELEMENTS OF PHOTOMETRY. 
 
 189 
 
 by the same operation that removes the lamp, A, from its 
 socket. The voltmeter is connected to the equalizer wire 
 and the wire, C, and indicates the common voltage on both 
 lamps. The wattage of the lamp, A, is the product of the 
 ammeter reading and the voltmeter reading when the 
 equalizer is closed. In both of these first two methods, 
 the candle-power over small ranges of voltage has been 
 assumed to vary on each lamp in the same proportion. 
 But this assumes the same constant for each lamp, which 
 may or may not exist. 
 
 The standard lamp and the lamp to be tested may not be 
 possessed of the same physical characteristics or the same 
 
 Equalizing Wire 
 
 FIG. 78. 
 
 ratios of graphitic carbon to base carbon in their filaments, 
 any difference in either of which would have a direct in- 
 fluence on their light-giving qualities at different voltages. 
 To eliminate all errors of this character, the third method 
 wherein a storage battery is employed particularly com- 
 mends itself to favor. 
 
 In Fig. 78,6* and D represent two terminals of a storage 
 battery of a sufficient number of cells to give double the 
 potential needed for the lamps, at 1.8 volts per cell. From 
 these cells is led off a three-wire service with two neutral 
 conductors, / and J, connected respectively to one end of 
 
190 AMERICAN METER PRACTICE. 
 
 the variable resistances, E and F, in series with the lamps ,. 
 A and B. Leads, L and AT, complete the circuit through 
 the lamps to the positive and negative ends of the battery. 
 This battery may vary in ampere capacity, but for an 
 ordinary central station a capacity of 10 ampere hours at 
 the discharge rate of i J amperes per hour is sufficiently large. 
 
 The use of the battery provides for each lamp an isolated 
 circuit of its own of absolutely constant potential. The 
 standard lamp at B is adjusted to its proper voltage and 
 remains without need of further attention. The object 
 of splitting the neutral is to remove the slight fluctuation 
 caused by the drop in this conductor when the lamp at A 
 is removed. One side of the battery alone could be used 
 for both lamps, but the variable demand upon it would 
 cause a slight fluctuation of the voltage at the standard 
 lamp, which should be entirely eliminated for quick 
 work. The voltmeter is connected through a double- 
 pole switch across the terminals of either of the lamps, A 
 and B, as in first method, and is left in circuit across A 
 when a test is being made. Occasionally a reading across 
 B is taken, to see if the voltage has varied by the dis- 
 charging of the battery. 
 
 The lamp, A, is in a circuit which receives a constant 
 potential, but the potential across the lamp varies with the 
 current taken by the lamp, hence, the variable resistance, 
 E, is so placed as to be readily under the control of the 
 operator. 
 
 The arrangement of the dark room is a matter which 
 should be carefully planned to enable quick and accurate 
 work to be done. The reading of the instruments should 
 take as little time as possible, and, for this reason, the large 
 illuminated dial type of meter placed directly back of the 
 photometer screen is the best for quick work. These meters 
 
ELEMENTS OF PHOTOMETRY. 191 
 
 may be mounted in any convenient manner, the controlling 
 switches being within easy reach of the operator. 
 
 The scale divisions of the voltmeter should be large, 
 which condition may be secured by using an instrument 
 the range of which extends only about 15 volts above and 
 below the voltage of the lamps to be tested. The ammeter 
 need not read higher than two amperes in o.oi = ampere 
 divisions. The scales of the instruments should be elevated 
 slightly above the level of the top of the photometer screen ; 
 the values can then be read at a glance. 
 
 Unless a specific test is being made, an adjustable pointer 
 on the ammeter is of great assistance in rejecting all lamps 
 above a certain wattage, as it can be set at a given value 
 and every lamp which runs the needle over this point 
 rejected at once. 
 
 The variable resistance on the lamp to be tested should 
 be within easy reach of the left hand; in fact, the opera- 
 tor's left need never leave it, as his right can be used to 
 shift the screen. 
 
 In operating, the lamps to be tested in the rotating 
 socket are removed by a boy who disposes of them accord- 
 ing to the reader's call "good" or "bad." Usually the 
 speed of the testing is entirely dependent upon the quick- 
 ness with which the boy at the sockets can handle the 
 lamps. 
 
 The reader's eyes should be shielded from the rays of 
 the standard and tested lamps, and, as it takes several 
 minutes for the eyes to adjust themselves to reading the 
 screen, the reader should not expose his eyes to extraneous 
 light between readings. In testing a large number of 
 lamps where extreme accuracy is not desired, advantage 
 must be taken of every factor which reduces the duration 
 of a test. 
 
192 AMERICAN METER PRACTICE. 
 
 THE IMPORTANCE OF PHOTOMETRY TO CENTRAL STATIONS. 
 
 The success which the electric light has attained is 
 based primarily on its intrinsic merits, its convenience, 
 cleanliness and adaptability, rendering it superior to any 
 other commercial light, but not necessarily cheaper. Its 
 march forward to the final limit of universal use depends 
 then on two main underlying qualities ; first, its superiority 
 over other lights; and second, its ability to compete 
 commercially with other light-giving sources. Almost 
 every day brings forth some new form of light, which the 
 enthusiastic inventor claims will render the electric light 
 obsolete in a very short time. The best answer to such 
 claims is found in the rapid enlargement of the electric 
 central stations in every part of the world. 
 
 The price of electricity has shown within the past few 
 years a tendency to become less, as more economical 
 methods of generation are utilized, but it is safe to say that 
 no very radical departure in present prices will come about 
 as long as the efficiency of the incandescent lamp remains 
 at its present low figure. The lamp may be considered the 
 keystone of the entire structure ; it is the final link in the 
 series of transformations of energy which take place be- 
 tween the coal-pile and the light, and on its economy (other 
 conditions being favorable) depends the earning power of 
 the central station. 
 
 The economy of the incandescent lamp must be viewed 
 from two standpoints, that of the consumer and that of the 
 central station. The light emitted from an incandescent 
 lamp filament increases rapidly as the voltage across the 
 lamp rises, and if the voltage be maintained above normal 
 for any length of time, the life of the lamp is very much 
 shortened. The higher the voltage is raised, the greater 
 
ELEMENTS OF PHOTOMETRY. 193 
 
 becomes the efficiency of the lamp; that is, the more light 
 per watt of energy it gives. 
 
 Viewed from the consumer's standpoint, the lamp 
 should not last more than a few hours for him to get the 
 best returns for his money. This would necessitate such 
 frequent renewals by the central station that the cost 
 of furnishing lamps would be prohibitive. Viewed from 
 the central station's standpoint, the lamp which lasts the 
 greatest number of hours is the best lamp; that is, it needs 
 fewest renewals, and hence is the most economical. 
 
 Between the two extremes here presented there must 
 be a mean which should combine the good qualities 
 of both conditions in the greatest possible degree. 
 The perfect lamp would maintain a constant candle- 
 power for an indefinite period with small expenditure of en- 
 ergy ; the modern commercial lamp endeavors to give a fairly 
 uniform candle-power for a limited period with an expendi- 
 ture of energy varying from three to four watts per candle. 
 
 It has been found that the long-life lamp is usually of 
 low efficiency to start with, and this efficiency becomes 
 lower as the number of hours the lamp burns increases, 
 until in many instances the efficiency becomes as low as 
 10 watts per candle. The amount of energy consumed 
 per candle renders the lamp in this condition uncom- 
 mercial and destroys the quality of its light. In other 
 words, the consumer agrees to pay for light at the rate of 
 three to four watts per candle and will not consent to pay at 
 the rate of 10 watts per candle for an inferior light. The 
 central station, then, in order to fulfil its contract with the 
 consumer, must see to it that the amount of light furnished 
 per watt is maintained at somewhere near the required figure. 
 
 Assuming the average total life of the lamp to be 2,000 
 hours, the power to be 16 candle-power at 3.1 watts per 
 
194 AMERICAN METER PRACTICE. 
 
 candle at the start, and 8 candle-power at 6 watts per 
 candle at the end of the 2,000 hours, the mean candle-power 
 and efficiency over the total period may be roughly taken 
 as 12 candle-power at 4.5 watts per candle; in other words, 
 this might be the candle-power and efficiency at the end of 
 1,000 hours. A new lamp of this quality (12 candle- 
 power at 4.5 watts per candle-power) would be considered 
 uncommercial on a circuit having close voltage regulation. 
 Why, then, should it be tolerated simply because it has 
 had the misfortune to burn i ,000 hours ? 
 
 The lamp reaches an uncommercial state at some point 
 previous to this time, and the results of a great number 
 of tests have shown that the commercial period of a high- 
 efficiency lamp is reached in about 600 hours burning. 
 The average 3.1 watt-lamp burning at an initial candle- 
 power of 16 has, at the expiration of 600 hours, a candle- 
 power of 13, and an efficiency of 3.7 watts per candle, 
 which may be considered the end of the commercial life of 
 the lamp. The average efficiency of the whole period is 
 14.5 candles at 3.38 watts per candle. 
 
 For a central station to compete successfully with 
 Welsbach and other lights, the candle-power of the 
 lamps of the entire service must be maintained at a high 
 standard. To do this successfully a system of periodic 
 renewals must be instituted, the frequency of which can be 
 roughly estimated from the consumer's bills. If a consumer 
 having 10 lamps develop a gross bill of $10 a month at one 
 cent per lamp hour, the average service per lamp is i oo hours, 
 and the renewals of lamps need only be made twice a year. 
 
 The maintenance of a clean lighting service, without any 
 waste in the lamps displaced by renewals, necessitates the 
 photometering of all the lamps returned. The lamps in 
 the consumer's premises frequently become fly-specked 
 
ELEMENTS OF PHOTOMETRY. 195 
 
 and covered with dust so that it is impossible to judge of 
 the true candle-power of a lamp until it is cleaned. As 
 the lamps must be changed by cheap labor, it is best not to 
 invest this labor with powers of discrimination, but 
 to have all the lamps in a given installation changed 
 regardless of their appearance. 
 
 Some idea of the total cost of renewals in lamps may be 
 gained for any station by proceeding as follows: In a 
 typical plant, assume 1000 16 c.-p. lamps connected, and 
 a gross income of $3,650 per annum at cent per lamp 
 hour. Then the average income per day will be $10, 
 corresponding to 1333 lamp hours a day, or 1.3 hours per 
 lamp per day. Allowing 600 hours per lamp as the com- 
 mercial life, each lamp would last 461 days, requiring the 
 renewal of about 800 lamps per year. This figure is some- 
 what higher than would be met in practice, but it amounts 
 to less than five per cent, of the gross income. The in- 
 creased income from good service would more than counter- 
 balance the cost. This policy on the part of a central 
 station would be very popular and could be used effect- 
 ively in soliciting new business and settling disputed bills, 
 not to mention the inducement it gives the consumer to 
 burn more light. 
 
 To return to the photometric features of the work, the 
 renewal of lamps should be carried out daily, and the 
 returned lamps cleaned preparatory to testing. The 
 tested lamps are divided into two classes, those testing 
 between 13 and 16 candle-power are retained as a sort of 
 second-class stock, and those testing up to 16 candle- 
 power and over are treated specially, and considered as 
 first-class stock. 
 
 If the first-class stock be in the majority, the lamps are 
 tested for this class first; if the second-class stock be the 
 
196 AMERICAN METER PRACTICE. 
 
 largest, the lamps are tested by setting the photometer 
 screen at 13 candle-power and rejecting all under this 
 candle-power. The good lamps left are then tested by 
 setting the screen at 16 candle-power and sifting out the 
 separate classes. By following this order the lamps are 
 subjected to the smallest amount of handling. 
 
 The second-class stock is preferably used for special 
 illuminations and outdoor work, but if there be not enough 
 demand in this line to balance the supply of such stock, it 
 may be judiciously issued to points near the plant where 
 a higher voltage is maintained. This second-class stock 
 has a short average life. The lamps testing 16 candle- 
 power have their stubs buffed and the glass thoroughly 
 cleaned, any defaced labels are removed and new ones put 
 on, and the lamp is made to look as good as new in every 
 respect. Such stock is issued on equality with new lamps. 
 The cost of photometering is very small ; the entire cost of 
 cleaning, photometering and buffing the stubs is about 
 J cent per lamp. 
 
 In carrying out the system of renewals the area cov- 
 ered should be divided into districts, and these districts 
 carefully gone over at regular intervals, the work being 
 so arranged that a specified number of lamps are changed 
 each day. Boys are usually employed in this class of work ; 
 no special intelligence is required, as the work consists in 
 simply removing the lamp and putting in a new one. 
 
 The system of free renewals is no new experiment. Its 
 utility has been well established, and, when practiced in 
 conjunction with a photometer, as outlined in the fore- 
 going paragraphs, it furnishes an ideal method of main- 
 taining the commercial efficiency of the entire lighting 
 system at its proper value without incurring the slightest 
 loss due to waste by such renewals. 
 
CATALOGUE 
 
 . . OF . . . 
 
 ENGINEERING BOOKS 
 
 ... ON ... 
 
 POWER TRANSMISSION 
 
 ELECTRIC LIGHTING 
 
 ELECTRIC RAILWAYS 
 
 THE TELEPHONE 
 
 THE TELEGRAPH 
 
 DYNAMOS AND MOTORS 
 
 ELECTRICAL MEASUREMENTS 
 
 TRANSFORMERS 
 
 STORAGE BATTERIES 
 
 WATER. WORKS 
 
 BRIDGES 
 
 STEAM POWER. PLANTS 
 
 VENTILATION AND HEATING 
 
 STEAM AND HOT WATER HEATING 
 
 PLUMBING. Etc. 
 
 PUBLISHED BY THE 
 
 McGratw Publishing Co., 
 
 Publishers, Importers and 'Booksellers, 
 114 Liberty Street, * ?y* New York City. 
 
McGRAW PUBLISHING COMPANY. 
 
 BOOKS AND PAMPHLETS. 
 American Street Railway Investments 
 
 Financial data of over 1,300 American city, suburban and inter- 
 urban electric railways, statistics of operation, details of plant 
 and of equipment, and names of officers. Published annually. 
 
 Vol. X, 1903 edition 5 oo 
 
 We can still supply Vols. I (1894) to IX (1902) at $5.00 each. 
 
 American Steam and Hot-Water Heating Practice 
 
 Descriptive articles upon current practice with questions and 
 answers regarding the problems involved. 268 pages, 536 
 illustrations 3 oo 
 
 American Plumbing Practice. 
 
 Descriptive articles upon current practice with questions and 
 answers regarding the problems involved. 268 pages, 536 
 illustrations 2 50 
 
 Abbott, A. V., Telephony. 6 Vols. 
 
 Vol. I Location of Central Offices 
 
 Vol. 2 Construction of Underground Conduits. 
 
 Vol. 3 The Cable Plant 
 
 Vol. 4 Open Wire Lines 
 
 Vol. 5 The Substation. 
 
 Vol. 6 Switchboards and the Central Office 
 
 Price of the set of 6 vols. $6.00; single vols. $1.50 each. 
 
 50 
 50 
 50 
 50 
 50 
 50 
 
 Baldwin, William J. Hot-Water Heating and Fitting 
 
 385 pages, 200 illustrations .... 2 50 
 
 Baum, F. G. Alternating Current Transformer 
 
 195 pages and 122 illustrations i 5 
 
 Behrend, B. A. The Induction Motor 
 
 105 pages, 56 illustrations i 5 
 
 Bell, Louis The Art of Illumination 
 
 350 pages, 125 illustrations 2 50 
 
 E'ectric Power Transmission 
 
 The best treatise yet written on the practical and commercial 
 side of electrical power transmission. Third edition. 632 
 pages, 285 illustrations, 21 plates 3 oo 
 
 Power Distribution for Electric Railways 
 
 Deals with the laying out and calculating of electric railway 
 circuits. Substation distribution is also considered. Third 
 edition. 303 pages, 148 illustrations ." 2 50 
 
McGRAW PUBLISHING COMPANY. 
 
 Billings, W. R. Some Details of Water-Works Construc- 
 tion 
 
 96 pages, 28 illustrations 2 oo 
 
 Billings, John S. Ventilation and Heating 
 
 500 pages, 210 illustrations 4 oo 
 
 Cement Industry, The 
 
 Description of Portland and natural cement plants in the 
 United States and Europe, with notes on materials and pro- 
 cesses in Portland cement manufacture. 235 pages, 132 il- 
 lustrations 3 oo 
 
 Collins, A. F. Wireless Telegraphy (in preparation.) 
 Considere, A. Reinforced Concrete 
 
 Translated from the French, with a preface and additions by 
 Leon S. Moisseiff . 2 oo 
 
 de la Tour, Boy The Induction Motor; Its Theory and 
 Design 
 
 Set forth by a practical method of calculation. Translated 
 from the French by C. O. Mailloux. 225 pages and 77 illus. . . 2 50 
 
 Diagrams of Electrical Connections, with Explanatory 
 Text 
 
 32 pages, paper 50 
 
 New and enlarged edition, cloth binding, in preparation i oo 
 
 Dobbs, A. E. Practical Features of Telephone Work 
 
 I2mo, cloth. 134 pages, 61 illustrations 75 
 
 Electrical Designs 
 
 262 pages, 289 illustrations 2 oo 
 
 Emmet, William LeRoy Alternating Current Wiring and 
 Distribution 
 
 Contains a very clear account of the principles of alternating 
 currents from the practical point of view, and of their distri- 
 bution and application to lighting and power. Second edition. 
 98 pages, 33 illustrations. i6mo, cloth I oo 
 
 Fairman, James F. Telephone Wiring i oo 
 
 Frund, H. W.-Moonlight Tables for I 903 
 
 Covers all months for 1903. The tables are printed on a large 
 sheet, 14 x 20 inches 50 
 
McGRAW PUBLISHING COMPANY. 
 
 Gerard, EricElectricity and Magnetism 
 
 With chapters by C. P. Steinmetz, A. E. Kennelly and Dr. 
 Gary T. Hutchinson. Translated under the direction of 
 Dr. Louis Duncan. 392 pages, 112 illustrations ............... 2 50 
 
 Gonzenbach, E. Engineering Preliminaries for an Inter- 
 urban Electric Railway ................................ i oo 
 
 Goodell, John Water-Works for Small Cities and Towns 
 
 281 pages, 53 illustrations ................................... 2 oo 
 
 Gotshall, W. C. Notes on Electric Railway Economics 
 and Preliminary Engineering. 
 
 A series of Notes on the Preliminary Engineering of Inter- 
 urban Electric Railways. pages, illustrated .............. 200 
 
 Hanchett, George T. Modern Electric Railway Motors 
 
 The various commercial types of electric railway motors are 
 described in full detail, including mounting and control. 200 
 pages, 157 illustrations ...................................... 2 oo 
 
 Handy Tables for Electrical and Steam Engineers 
 
 18 tables found to be useful in arriving at approximate results 
 without extensive calculation. Pamphlet, 8vo, 16 pages ...... 50 
 
 Hering, Carl Electrochemical Equivalents 
 
 The Universal Wiring Computer. 
 
 For determining the sizes of wires for incandescent electric 
 lamp leads, and for distribution in general without calcula- 
 tion, with some notes on wiring and a set of Auxiliary Tables. 
 44 pages, 4 charts ........................................... i oo 
 
 Herrick, Albert B. Practical Electric Railway Hand Book. 
 
 Handsomely bound in leather, with flap. 450 pages .......... 3 oo 
 
 Hopkinson, John Original Papers on Dynamo Machinery 
 and Allied Subjects. 
 
 Authorized American edition. 249 pages, 98 illustrations ...... i oo 
 
 Houston, E. J. Advanced Primers of Electricity. 
 
 Vol. i Electricity and Magnetism and other Advanced 
 
 Primers. Second edition, 318 pages ...................... i oo 
 
 Vol. 2 Electrical Measurements and other Advanced Primers. 
 
 429 pages .............................. . ................ i oo 
 
 Vol. 3 Electrical Transmission of Intelligence and other Ad- 
 
 vanced Primers. 330 pages ............................. i oo 
 
McGRAW PUBLISHING COMPANY. 
 
 Dictionary of Electrical Words, Terms and Phrases 
 
 Fourth edition. Greatly enlarged. 990 double column octavo 
 
 pages, 582 illustrations 7 oo 
 
 Pocket edition. 945 pages, cloth 2 50 
 
 Pocket edition. 945 pages, leather 3 oo 
 
 Electricity One Hundred Years Ago and To-day 
 
 199 pages, illustrated i oo 
 
 Houston, E. J. and A. E. Kennelly Algebra Made Easy 
 
 ioi pages, 5 illustrations 75 
 
 Electricity Made Easy by Simple Language and 
 
 Copious Illustrations 
 
 348 pages, 207 illustrations i 50 
 
 The Interpretation of Mathematical Formulae 
 
 225 pages, 9 illustrations i 25 
 
 Electrical Engineering Leaflets, Containing the 
 
 Underlying Principles of Electricity and Magnetism. 
 
 ELEMENTARY GRADE suited to the study of Electrical 
 Artisans, Wiremen and to Elementary Students. 296 pages, 
 121 illustrations, 8vo, cloth i 50 
 
 INTERMEDIATE GRADE suited to students in High 
 Schools or Colleges, and those beginning the study of Electrical 
 Engineering. 300 pages, 140 illustrations, 8vo, cloth i 50 
 
 ADVANCED GRADE suited to students taking Technical 
 Courses in Electrical Engineering. 296 pages, 121 illustrations, 
 8vo, cloth i 50 
 
 Electro-Dynamic Machinery 
 
 A text-book on continuous-current dynamo-electric machinery 
 for electrical engineering students of all grades. 331 pages, 
 232 illustrations 2 50 
 
 Recent Types of Dynamo-Electric Machinery 
 
 612 pages, 435 illustrations 4 oo 
 
 (ELECTROTECHNICAL SERIES-10 VOLUMES.) 
 
 Alternating Electric Currents 
 
 Third edition, 271 pages, 102 illustrations i oo 
 
McGRAW PUBLISHING COMPANY. 
 
 Electric Arc Lighting 
 
 Second edition. 437 pages, 172 illustrations I oo 
 
 Electric Heating 
 
 290 pages, 86 illustrations ; i oo 
 
 Electric Incandescent Lighting 
 
 Second edition. 508 pages, 161 illustrations i oo 
 
 The Electric Motor 
 
 377 pages, 122 illustrations i oo 
 
 Electric Street Railways 
 
 367 pages, 158 illustrations i oo 
 
 Electric Telegraphy 
 
 448 pages, 163 illustrations i oo 
 
 The Electric Telephone 
 
 Second edition. 454 pages, 151 illustrations i oo 
 
 Electro-Therapeutics 
 
 Second edition. 452 pages, 147 illustrations i oo 
 
 Magnetism 
 
 294 pages, 94 illustrations I oo 
 
 Ken nelly, A. E. (See Houston, E. J. and A. E. Kennelly.) 
 
 Knox, Chas. E. Interior Wiring (in preparation.) 
 Lyndon, Lamar Storage Battery Engineering 
 
 A practical treatise for Engineers. Cloth, 360 pages, 178 illus- 
 trations and diagrams, 4 large folding plates 3 oo 
 
 MailloUX, C. O. (See de la Tour.) 
 
 MaSOn, H. Static Electricity (In preparation.) 
 
 Merrill, A. E. Electric Lighting Specifications 
 
 For the use of Engineers and Architects. Second edition, en- 
 tirely rewritten. 213 pages i 50 
 
 Reference Book of Tables and Formulas for Elec- 
 tric Railway Engineers 
 
 Second edition. Flexible morocco. 128 pages. Interleaved 
 with blank pages i oo 
 
McGRAW PUBLISHING COMPANY. 
 
 Meyer, Henry C., Jr. Steam Power Plants, their Design 
 and Construction 
 
 160 pages, 16 plates and 65 illustrations 2 oo 
 
 Meyer, Henry C. Water Waste Prevention: Its Import- 
 ance, and the Evils Due to Its Neglect 
 
 Cloth. Large, 8vo i oo 
 
 Miller, Kempster B. American Telephone Practice 
 
 A comprehensive treatise, including descriptions of apparatus, 
 line construction, exchange operation, etc. Third edition. 518 
 pages, 379 illustrations 3 oo 
 
 Moisseiff, L. S. (See Considere.) 
 
 Monroe, William S. Steam Heating and Ventilation 
 
 150 pages, 90 illustrations 2 oo 
 
 More, James, and Alex. M. McCallum English Methods 
 of Street Railway Track Construction 
 
 Reprinted from the Street Railway Journal. Pamphlet. 26 
 pages, illustrated 35 
 
 Parham, E. C., and J. S. Shedd Shop and Road Testing 
 of Dynamos and Motors 
 
 626 pages, 211 illustrations 2 50 
 
 Poincare, H. Maxwell's Theory and Wireless Telegraphy 
 
 Part I. Maxwell's Theory and Hertzian Oscillations. By H. 
 Poincare, translated by Frederick K. Vreeland. Authorized 
 American edition. Part II. The Principles of Wireless 
 Telegraphy. By Frederick K. Vreeland. In one volume. 
 Cloth 2 oo 
 
 Pratt, Mason D., and C. A. Alden Street Railway Roadbed 
 
 A treatise on the construction of the roadbed, giving data as 
 to rails, method of track fastening and making joints, guard 
 rails, curves, etc. 135 pages, 157 illustrations 200 
 
 Reed, Lyman C. American Meter Practice (In press) 
 Roller, Frank W. Electrical Measurements (in preparation) 
 Shepardson, George D. Electrical Catechism 
 
 450 pages, 325 illustrations 2 oo 
 
 Skinner, Frank W. Types and Details of Bridge Con- 
 struction (I n press) 
 
McGRAW PUBLISHING COMPANY. 
 
 Steinmetz, Charles Proteus Theoretical Elements of 
 Electrical Engineering 
 
 Second edition. 320 pages, 148 illustrations 2 50 
 
 The Theory and Calculation of Alternating Cur- 
 rent Phenomena 
 
 Third edition, 525 pages, 210 illustrations 4 oo 
 
 Vreeland, Frederick K.-(See Poincare, H.) 
 
 Water Tower, Pumping and Power Station Designs 
 
 Contains 17 designs of water towers and 17 designs of pumping 
 and power stations. Cloth, quarto. 34 full-page desig/is. ... 2 oo 
 
 Wiener, A. E. Practical Calculation of Dynamo-Electric 
 Machines 
 
 A manual for Electrical and Mechanical Engineers, and a Text- 
 book for students of Electro-technics. Second edition. 727 
 pages, 381 illustrations 3 oo 
 
 PERIODICALS. 
 
 Electrical World and Engineer. 
 
 An illustrated Weekly Review of Current Progress in Elec- 
 tricity and its Practical Applications. Annual Subscription.. . 3 oo 
 To Foreign Countries 6 oo 
 
 General Index to the Electrical World. (Subject and Author.) 
 
 Vol. I. From January i, 1883, to January I, 1897. 372 pages.. 8 oo 
 Vol. II. From January I, 1897, to January I, 1903. (In prep- 
 aration.) 
 
 American Electrician. 
 
 Monthly. Annual subscription i oo 
 
 To Foreign Countries 2 oo 
 
 Street Railway Journal. 
 
 Weekly. Annual subscription 4 oo 
 
 To Foreign Countries 6 oo 
 
 The Engineering Record. 
 
 Weekly. Annual subscription 5 oo 
 
 To Foreign Countries 6 oo 
 
 Temporary Binders. 
 
 For any of the above periodicals i oo 
 
=r 
 
 OCT 17 1935 
 MAR 10 
 
v 
 
 YC 19596 
 
 1*( 
 
 r>