US2636995A - Inductive-interference preventive - Google Patents

Description

Aprll 28, 1953 J. HIBBARD 2,636,995

INDUCTIVE-IN'IERFERENCE PREVENTIVE-MEANS Original Filed Sept. 2, 1949 5 Sheets-Sheet l Communication- 5 A? "4 System L I P Pa m F lg. I.

Arcinq-Typb} Rectifier-Tube fixnode current W'TNESSES= Fig. 3. Fig.4.

I YLIoyd diHibbord. 7 I m ATTORNEY A ril 28, 1953 L. J. HIBBARD INDUCTIVE-INTERFERENCE PREVENTIVE-MEANS Original Filed Sept. 2, 1949 3 Sheets-Sheet 2 I l I I 142 l l l Test results for 0 I00 200 300 400 500 600 700 800 900 I000 00 Cycles Fig.6.

D-C Fil1er-0urrent I I i Meier-Current INVENTOR Lloyd d Hlbburd. Fig-.7. F'Q-B' BY ATTORNEY April 28, 1953 L. J. HIBBARD 2,636,995

INDUCTIVE- INTERFERENCE PREVENTIVE MEANS Original Filed Sept. 2, 1949 5 Sheets-Sheet 3 f r Communication- Sys'rem i T F i g.9.

6 6 h Commumcuhon- /l Gommumcchon- System S stem 54 4 y 8 C 86 co IAO Fig. IO.

WITNESSES:

INVENTOR W I Lloyd J. Hibbord. 72 W BY ATTORNEY Patented Apr. 28, 1953 INDUCTIVE-INTERFERENCE PREVENTIVE- MEANS Lloyd J. Hibbard, Pittsburgh, Pa., assignor to Westinghouse Electric Corporation, East Pittsburgh, Pa, a corporation of Pennsylvania Continuation of application Serial No. 119,714, September 2,1949. This application October 8, 1949, Serial No. 120,331

25 Claims. 1 s

This is a continuation of my application Serial No. 113,714, filed September 2, 1949, now abandoned, under the title, Rectifier Power-Units for Vehicles.

My invention relates broadly to preventivemeans for preventing or reducing inductive in terference in telephone-lines which are close to alternating-current power-lines in which there are harmonics in the telephone-interference frequency-range. More specifically, my invention relates to a practical and advantageous rectifierunit for energizing'the direct-current series traction-motors of a self-propelled vehicle from a single-phase high-voltage trolley.

A rectifier-powered Vehicle of the general type just mentioned. was tried out once, on a single railway-car, as described. in the tree's Railway Journal for December 19, 1914, at. page 1343, and also in the Engineer, London, for February 5-, 1915, at page 142-143. This car was taken out of service on May 14, 1915, after a 40-day period of operation in. 1914., and a 26-day period ofop eration in 1915. There is evidence that the harmonics in the transmission-circuit for this car caused telephone--disturbancesof such serious extent that the entire track-mileage over which the car was operated was equipped with booster-- trans ormers, at frequent intervals, to force the return-currents to flow through the rails, where their inductive-interference effects in: adjacent telephone-lines are'much smaller than when: the return-circuit is through the ground.

llany engineers WhOELI'G skilled in therectifier, motor, railway, and inductive-interference arts, all of which are involved, have been opposed, from time to time, to any resumption of work on rectifier-powered vehicles, since. that one experiment.

According to one aspect of my invention, when an A.C. filter is desirable, (which is almost always the case), Iuse a specialv capacitor-circuit which is connected across the secondar, rather ti :1 across the primary, of the transformer which energizes the rectifiers. from the trolley. in this manner I provide: a low-impedance" circulating path, on the secondary side, for. all. but the lowest-frequency components of the D;-C. and I probably also cause these circulated components to flow to: some extent. non-inductively through the secondary of the transformer. For both of these reasons, very little primaryvoltage is available for. circulating high-ire quency harmonics initherprimary circuit of A.-C. supply-line; This secondary connection of the filter-capacitor. is, in eilect, an. anode-to-anodc 2 connection, and it is so unconventional that I had to demonstrate that a properly designed anode-to-anocie filter-capacitor would work, would be very economical, and would have a number of other important and unanticipated advantages, as will be hereinafter pointed out.

In a rectifier power-unit l r the tractionmotors of a self-propelled vehicle, the D.-C. output-circuit of the rectifier unit is confined to the vehicle, where there is no inductive-interierence problem due to the D.C. ripple, and hence this D.-C. ripple does not need to be smoothed down to relatively low values, to avoid inductive inter- Terence, as has been necessary in D.-C. feeder or trolley-circuitsv which have heretofore been powereo; rectiners. In such a rectifier-pow red vehicle, the entire load on the 11-6. output-sin cult of the rectifier-unit consists of the D.-C. traction-motors of the vehicle, and these motors will stand a very substantial ripple while still having clack commutation, freedom from objectionable pulsating torque or vibration, and free- 010121 from overheating, as will be subsequently pointed out; and it is an important feature of my invention that I use a fairly large D.-C. ripple.

In carrying out my invention, I also usually put enough inductance in the A.-C. supply-circuit to produce an unusually high angle of overlap or commutating angle of the rectifiers, as will be subsequently explained in detail. "1 he eiz'ect or this is to decrease the A.-C. harmonics at the pense of increasing the D.-C. ripple, increasi the regulation, decreasing the power factor, and only very slightly decreasing the efficiency. The D.--C. ripple can be easily controlled by other means, as will be explained hereinafter; and the regulation and the poweractor can both be improved, and restored to something like what they would have been Without the increased angle of overlap, by the use of the seconclary-connected filter-capacitor which has already been inentioned.

It is also an important feature of invention, that even though 1' use voltage-control, I do not do it (to any substantial d layed firing,

systems n a voltage-control Delayed fir-- ing is an attractive means for obtaining voltage control for the D.-C. circuit, because its cost -s a greatly increased inductive-interference problem, and has other disadvantages, as will subsequently be explained, and hence it cannot ordinarily be used in rectifier-powered transportation-equipments in accordance with my invention. I therefore use tap-changing means, to obtain most or all of the voltage-control which may be required in my equipments. For reasons which will be subsequently explained, I preferably use secondary tap-changing, rather than primary tap-changing, so that the transformation-ratio which is applicable to my secondary-connected A.-C. filter (when used) will not be subjected to wide ratio-changes.

As a part of my invention in rare instances, I may use a tuned, second-harmonic filter which is connected across either the motor-armatures, or the entire motor-circuits including armatures, series-field windings and commutating-field windings, and sometimes also across a part of the D.-C. inductance or choke if such an element is used. This expedient, in conjunction with the other features of my invention, enables me to use a rectifier-output having enough ripple so that the A.-C. inductive-interference can be reduced to the necessary level, and if this ripp should be too much for any particular motordesign, its major component (the double-frequency ripple) can be by-passed from the motor by the shunt-connected second-harmonic filter.

An object of my invention is to provide a successiul rectifier power-unit for vehicles, embodying one or more of the various novel design-features which have just been briefly outlined, thus combining the advantages of both single-phase trolleys and D.-C. traction-motors, while removing certain previously experienced kw. capacitylimits in self-propelled vehicles, and at the same time reducing the weight, decreasing the cost, and achieving various operational advantages, as will be subsequently pointed out, besides being able to achieve a telephone-interference level which is much lower than has been obtained with single-phase series-motor electrifications.

A more generally stated object of my invention is to apply my novel secondary filter-system to any application Where it is necesssary, such as an A.-C.-motored railway-vehicle or a polyphase rectifier-system, or, in general, any load-device of a type tending to draw objectionable supplycircuit harmonics in the telephone-interference frequency-range.

With the foregoing and other objects in View, my invention consists of the combination, circuits, systems, parts, and methods of design and operation, hereinafter described and claimed, and illustrated in the accompanying drawing, wherein,

Figure 1 is a diagram of circuits and apparatusillustrative of my invention;

Fig. 2 is a simplified schematic View which will be referred to in the discussion of the A.-C. filter;

Figs. 3 and 4 are portions of oscilloscopic traces illustrating the operation, with the A.-C. filter off and on respectively;

Fig. 5 is another simplified schematic view showing a rectifier bridge which will be referred to in the discussion of my design;

Fig. 6 is a curve-diagram which will be referred to in the discussion of the tuning and the dis charge-resistance of the A.-C. filter;

Figs. 7 and 8 are portions of oscillographic 4 traces illustrating the operation, with the D.-C. filter off and on respectively; I

Fig. 9 is a diagram similar to Fig. 1, illustrating my invention as applied to a railway vehicle which is propelled by a series A.-C. motor or motors; and

Figs. 10 and 11 are diagrams illustrative of the application of my invention to polyphase rectifier systems.

In the typical form of embodiment of my invention, shown in Fig. l, a high-voltage singlephase trolley is indicated at l. Single-phasepower is supplied to the trolley l, at various spaced points, such as 2 and 3, from an A.-C. feeder l which is supplied with energy from a source or alternator 5. In order to indicate an inductive-interference or telephone-interference problem, I have included, in Fig. 1, a diagrammatic representation of a telephone-system or communication-system 6, which is in proximity to the feeder 4, or the trolley I, or both.

The trolley l is adapted to supply power to a self-propelled railway-vehicle, which is represented, in Fig. 1, by means of a pantagraph 7 and the apparatus which is connected thereto. The pantagraph l is connected to the high-voltage terminal 8 of the primary winding 9 of a step-down transformer which is provided with an iron core i8 and a secondary winding II. The low-voltage terminal l2 of the primary winding 9 is connected to ground through a primary reactor [3 which is illustrated as bein provided with a cut-out switch I4. This primary reactor [3 is provided with an iron core l5 having an air gap.

The secondary winding H of the transformer is provided with a mid-tap l6, and it is also provided, at each end, with a plurality of terminaltaps H, which are engaged by controller-points I8 and !9, respectively, which constitute tapchanging means. The mid-tap l6 constitutes the negative terminal of the D.-C. circuit. The two tap-changing terminals 18 and I9 of the secondary winding H are connected to the anodes of two rectifiers 2i and 22, the cathodes of which are connected together, to constitute the positive terminal 23 of the D.-C. circuit.

In Fig. 1, the rectifiers 2| and 22 are symbolically indicated, as being any kind of rectifiermeans, which, in truth, they may be. For most practical purposes, in the present state of the rectifier-art, these rectifiers will be arcing-type rectifier-tubes, such as ignitrons, or other vaporrectifiers, mounted either in separate tanks, or both in one tank, and having the well-known characteristic property of single-phase rectifying devices which become substantially non-conducting, once an arc has been established to initiate a conducting period, only in response to a current-decrease to substantially zero.

In accordance with my present invention, these rectifiers 2| and 22 should preferably be rectifiers which do not have any substantial amount of delayed firing, or grid-control, or phase-control, by which I mean anything beyond the amount of delay which would produce a 10% reduction in the D.-C. voltage during maximum line-current conditions, which are obtained while the vehicle is being accelerated. Theoretically, a 25 delay in firing produces a 10% reduction in the D.-C. voltage, and substantially also a 10% reduction in the power factor. In other words, each rectifier, 21 or 22, becomes conductive when its impressed voltage reaches a small value in its conductive polarity. A 10% voltage-reduction, ob-

tained by delayed firing, would possibly cut in half the required number of voltage-changing notches and taps i T on the controller and on the transfer, and this might be desirable in some cases.

Any increase in the phase-control angle, beyond the amount which would produce about a voltage-reduction, would start to build up the harmonics in both the D.C. and A.-C. Circuits, and would start to reduce the power-factor, both at an extremely fast rate. Since the A.-C. harmonics may create a very formidable inductive-interference problem, and since a poor power-factor may be so serious a handicap as to rule out the rectifier-powered motive-equipment as a practical rival of other available kinds of motive-equipment for vehicles, my present invention usually requires the use of rectifiers which do not have any substantial amount of delayed firing, as has just been defined. When an adequate A.-C. filter 4G is used, however, its beneficial eilects on the inductive-interference, the power factor, and the voltage-regulation may sometimes make it feasible to use enough firingdelay tov produce the usual or more, reduc tion in the D.C. voltage, by firing-angle control.

The direct-current circuit 23-ifi is used to energize the direct-current traction-motor equipment, which is represented for the sake of siznplicity, in Fig. 1, as comprising a single motor 24,, although it will be understood that usually a self-propelled vehicle will have a plurality of motors, which are commonly connected in some kind of series-circuit arrangement for starting, and in some kind of parallel-circuit arrangement for running conditions. The motor 2 3 is represented as having an armature 25 and a serially connected commutating-winding 26, which may be connected in the D.C. circuit 23-I6, in either polarity, by means of a reversing-switch 2?. The motor also has a series main-field winding 23.

This direct-current motor 24, or any combination of such motors, may be controlled with any of the known forms of speed-control, and I wish, therefore, that the voltage-control me hod, which is diagrammatically indicated by the tapcontrollers is and !9, shall be understood as beingv representative of any suitable or well-known form of motor-control.

The direct-current motor 24 has a certain amount of inductance, which necessarily limits theidirect-current ripples to a certain extent. In some cases, for reasons which will be hereinafter discussed, it may be necessary to still further decrease the amount or percentage of the D.-C. ripples, and in such a case, I provide a D.C. reactor or choke-coil 39, which is connected in series with the motor 24, in the D.C. circuit. This D.C. reactor is provided with an iron core 31 having an air gap. It is also shown as being equipped with a short-circuiting or by-passing switch 32, by means of which the D.C. reactor may be cut out of the motor-circuit 23l5.

In some instances, under unusually critical or diiiicult design-conditions, it may be necessary, from considerations of telephone-interference due to the harmonics in the A.-C. lines i and 4, to. cause the D.C. circuit 23l6 to have a higher percentage or ripple, than is desirable in the motor-circuit, including the armature 25, and some of all of the two serially connected windlugs, 26 and 26 thereof. Whenever this is the case, I provide a D.C. filter-circuit 33, which is by-passed around the desired parts of the motorcircuit, under the control of a switch 34. The D.C. filter, when it is used, is a tuned filter, com- 6 prising a capacitor and a serially connected inductor 36, which is shown as having an iron core 31 with an air gap. This tuned D.C. filtercircuit 33 is tuned approximately to the second harmonic, as will be discussed in detail hereinafter.

In accordance with my invention, it is usually necessary (although perhaps not necessary in every conceivable instance, when the other features of my invention are carefully observed under favorable conditions), to provide an A.-C. filter-circuit ill, which is connected on the secondary side of the transformer, as will be subsequently discussed. This A.-C. filter-circuit 4b is shown, for illustrative purposes, in Fig. 1, as being connected across the terminals SI and S2 of the secondary winding H of the transformer, preferably not across the variable-voltage taps ll which are used to energize the anode-leads iii This connection is intended to be representative of secondary-shunting connections, in general, as will be subsequently pointed out. The A.-C. filter-circuit is an untuned filter, comprising a capacitor C and a dischargeresistance R, preferably connected in series with each other, the details of which will be hereinafter discussed in full.

The successful design of'a rectifier-powered vehicle in accordance with my present invention involves a consideration of a number of factors, which will now be discussed. Attention will be given, first, to the telephone-interference factor, se this is the factor which, in the minds of most people, been the stumbling block which has prevented the development of this kind of motive-power for electrified railroad-systems.

confident that telephone-interference, in

" powered railway-vehicle, is not anyc the problem that some people may have assumed it to be. Nevertheless, to get rectifier-powered traction-motor equipments accept ed by all parties concerned, it seems to be necessary to give what seems to me to be an exaggerated attention to a far-greater-than-necessary reduction in telephone-interference. After rectifierpowered cars or locomotives have been demonstrated, to everyones satisfaction, to be capable of economical design so as to have phenoininally low inductive interference with respect to telephone circuits, I expect that some middle ground will be found, in which the many other advantages of my present invention will frequently be far more contrcliing, in the design, than considerations cf inductive interference.

For any given l'iarmonic-distribution the A.-C. line-currents (including the return-currents), the amount of telephone-interference depends upon the amount of telephone-interference coupling, which, in turn, depends a great deal upon the dis "ibution of the return-currents between the rails and ground; but to an even greater extent it depends upon the nature of the telephone wires and cables, as well as the nearness and the length of their exposure to the trolleycircuits and feeders.

The amount of telephone-interference due to any given harmonic, in a line which is exposed to a certain standardized amount of telephoneinterference coupling, is equal to the product of the line-amperes I, times a Weighting-factor T which represents the seriousness of the noiselevel produced at the frequency of that harmonic, taking into account such factors as coupling, instrument-sensitivity, and human-ear sensitivity, at the various frequencies;

The telephone-interference weighting-factor T is vastly different, for, different frequencies, as shown, by a telephone-interference weightingcurve, an approximation of which was published by Barstow et al. in 1935 AIEE, at page 1312, Fig. 6. In a more modern version of this telephoneinterference weighting-curve, which is sometimes used, the weighting-factor starts at unity at frequencies of 60 cycles and lower, the presupposed telephone-interference coupling-condition being so chosen that this is so. The weighting-factor reaches T=2.'7 at 100 cycles, T=7.6 at 150 cycles, T=1000 at 500 cycles, and rapidly increases up to a maximum of T:l2,l at 1070 cycles, after which the weighting-factor T reduces more slowlyto a value of T:500 at 4615 cycles, and to still lower values at higher frequencies, with a slight hump from T:3500 to T=3930 and back again to T=3500 at frequencies of 1950, 2950 and 335!) cycles, respectively. The weighting-curve 'I is still decreasing for frequencies above 4615 cycles, but usually such high harmonics are so small in magnitude that the inductive-interference product IT is quite negligibly small, long before such frequencies are reached.

Ihe total amount of telephone-interference, due to all of the composite frequencies of the line-current, under the previously mentioned standardized telephone-interference coupling conditions, is measured by a so-called telephone interference factor, called TIF, which is the ratio of the square-root of the sum of the squares of the weighted values, IT, of all of the sine-wave components, including the fundamental as well as the harmonics, divided by the R. M. S. value K of the current-wave.

The calculation of TIF, as just defined, has been a standard procedure in previous comparative estimates of the relative telephone-interference effects of different systems. It is based, as has been noted, upon the assumption of a standardized amount of telephone-interference coupling, which is thus a coupling which is constant for all frequencies. The actual noise-level of the telephone-interference has been assumed to be the product of TIF and the A.-C. line current, multipled by a constant factor representing the actual amount of the telephone-interference coupling.

In a railway-electrification system, however, the assumption of a telephone-interference coupling which is constant for all of the harmonics is not at all true, because the sinusoidal currentcompents of different frequencies do not divide anywhere near equally between the rails and ground, in returning from the vehicle to the source. This is due to the fact that the rails are made of iron, and thus have a rather considerable inductance, and the rail-impedance is higher for the higher frequencies. As a matter of fact, most of the high-order or high-frequency harmonics return through the ground, while the fundamental wave and the lowest-order harmonics return, to a much greater extent, through the rails. The telephone-interference coupling-factor varies greatly, depending upon the relative distribution of the return-current between the rail and ground. Thus, if all of the return-current passes through the ground, andnone of it through the rail, the coupling-factor is about 76 times higher than the coupling-factor which is obtained when all of the current return through the rail.

The thing which controls the acceptability of any distribution-circuit installation is the noiselevel of the telephone-interference during the operating-conditions when the most noise is produced, which, in a rectifier-powered railway-car. means during starting, or during the maximum permissible short-time line-current or loadingconditions. Actually, We are interested in this noise-level and not in any theoretical TIP-calculations which assume a constant, standardized, telephone-interference coupling factor. Actually, therefore, in order to properly evaluate this telephone interference noise-level, we should consider the noise produced by each of the harmonies, by multiplying the value of that harmonic-current by both the weighing-factor T and the particular coupling-factor which is applicable to that harmonic. After making such a calculation, for all of the harmonics having any material magnitude, we could then obtain the telephone-interference noise-factor by calculating the ratio of the square-root of the sum of the squares of these several noise-level products, for all of the harmonics (including the fundamental), and dividing the same by the R. M. S. value of the current.

In the following discussions, I shall refer, in the classical or accepted way, to the telephoneinterference factor TIF, which is calculated on the assumption of a coupling-factor which is constant for all of the frequencies. I do this, because this factor is relatively easily calculated, and because so much of the known data is calculated on this basis. This method of estimating the relative telephone-interference merit of my invention is pessimistic or disadvantageous to my invention, however, because, as will be subsequently pointed out, my invention results in a very large reduction in the higher-order harmonies, while accepting large values of the lowest-order harmonics. It will thus be seen that I reduce the noise-producing effects of the higher harmonics, which are the harmonics which are associated with the highest coupling-factors. When my new system is compared, therefore, with other railway-electrification systems, this point should be remembered, namely, that my noise-level is reduced, more than would be expected from a simple comparison of the TIF- values for the different systems, because I so greatly reduce the value of the high-frequency harmonics which do not readily return through the rails.

The maximum R. M. S. line-current (during accelerating-conditions) for a small rectifier motor-equipment of a small multiple-unit car, rated at 460 H. P. (horsepower), at 11 kv. (kilovolts) on the trolley, is something like 75 amperes. On a 5000 H. P. locomotive, operating on a 24 kv. trolley, the maximum R. M. S. linecurrent (during accelerating-condtions) is something like 400 amperes. This gives some idea of the range of line-curents by which the telephoneinterference factor 'IIF' must be multipled, in order to obtain the relative or comparative total noise-level of the various equipments.

There is evidence that an acceptable type of single-phase series railway-motor (using no rectifiers) has a telephone-interference factor TIF of the order of 27 under maximum-current accelerating-conditions, and something like TIF=7 under continuous-load operating conditions.

In a rectifier-powered motor-equipment, most of the telephone-interference comes from the rectification, rather than from the commuatorsegments and other factors within the motor itself. Consequently, I believe that the telephone-interference factor TIF of a rectifierpowered self-propelled vehicle is of the same order of magnitude during accelerating and normal-running conditions, other things being equal. Without any filter, a rectifier-powered tractionmotor which is properly designed and operated in accordance with my invention, with due regard to the magnitude of the D.-C. ripple, the angle of overlap, and the delayed-firing angle, (as will be subsequently explained), will have a telephone-interference factor in the neighborhood of TIF=80, or even much less. With a filter in accordance with my invention, the telephoneinterference factor TIF may be made as nearly unity as may be desired, economically obtained values of the order of TIF=7 being easily feasible.

In a theoretical case, at one extreme. in order to obtain a rippleless D.-C. output-current from a full-wave (bi-phase) single-phase rectifier, such as is shown in Fig. 1, there would have to be an infinite inductance in the D.-C. circuit, there would have to be a zero angle of overlap in the rectifiers and a zero inductance in the A.-C. input-circuit, and the A.-C. input-current would be a Square-topped or rectangular wave. It is well-known that a rectangular A.-C. wave having square-topped half-waves can be expressed mathematically by the Fourier series as e=f2 sin met) (1) where Table I.C'omponents of rectangular A.C. wave Harmonic Fre- H rmonic Fre- Order quen cy Amperes Order quency \mpercs n f n 25 1 0435 75 338 04 125 2 0370 175 1429 03 l4 225 llll 0331 275 0909 0244 325 0769 01961 375 0657 01 639 425 0588 l408 475 0526 01235 525 0476 The telephone-interference factor TIF of such a wave would be somewhat over At the other theoretical extreme, that is, for a pure-resistance load-circuit on the above rectifier, the D.-C. output-current will be half sine waves and the A.-C. input current will be a sine wave, and hence the A.-C. input-current will have no harmonics. It is well-known that a rectified sine wave has an average value Io=2I/1r=0.637l, Where I is the peak value of the sine wave, and

that this rectified sine wave has an infinite number of even harmonics, equal to wh re m is any even integer from 2 to infinity. Defining the percent ripple as one-half of the difference between the maximum and minimum instantaneous values of the D.-C. current wave, divided by one-half of the sum of said maximum and minimum values, for the purposes of the present invention, the percent ripple in a D.-C. wave of such a shape will be This gives a telephone-interference factor of TIF=1, said factor being due only to the fundamental freduency of the A.C. line-current.

All practicable rectifier-powered motor-equipments necessarily fall between these two extremes.

One of the first reouisites of a successful design, in accordance with my invention, is to use as larae a ripple, in the D.-C. circuit 23-46, as can be successfully stood by the D.-C. motor 24, or if, in some extreme ca e, the motor cannot stand enough ripple to keep the A.-C. supply-circuit harmonics down to a value which will give a satisfactorily low level of telephone-interference, then the D.-C. ripple must be made as high as may be necessary, from the standpoint of the telephone-interference due to the harmonics in the A.-C. supply-circuits I and 4, and the excessive ripple-content in the D.-C. motor 24 will have to be bv-pas ed by a 11-0. filter, such as is indicated at 33 in Fig. 1.

Tests have indicated that the curve of TIF, potted against increasing values of the ripples in the D.-C. circuit, is fairly fiat between 25 and ripple, so that any increase in the ripple, about 25 to 30%, is not usually worth while as a means for reducing the telephone-interference which results from the reduced harmonics in the A.-C. supply-line.

The arrount of ripple which can be successfully taken or tolerated by the D.-C. motor 24 depends upon whether the motor is a small motor or a large one, a slow-speed motor or a high-speed motor, and on other design-features thereof. The smaller D.-C. traction-motors usually have 4 poles, a series-wound armature, and a solid frame. The large D.-C. traction motors usually have 6 poles, a multiple-wound armature, and may have a laminated frame. A series-wound armature has as many armature-turns, between commutator-bars, as there are pairs of poles, while a multiple-wound arirature has only one armatureturn between commutator-bars. The D.-C. component of the motor-current sets up a certain flux-level in the field-circuit of the motor, and the A.-C. ripple-content of the motor-current causes the flux to rise and fall above and below this level. The change in flux, due. to this A.C. component or ripple, sets up certain transformer-volts in each of the armature-turns undergoing commutation. The multiple-wound armature, with only one turn between bars, can thus maintain black commutation with a higher ripple, in the fieldfiux, than a series-wound armature.

A 30% ripple in a direct-current motor makes very little change in flux under maximum loadconditions, because the motor is operating above the knee of its saturation-curve. At lower motorcurrent, some considerable sparking would do a negligible amount of harm, because of the smallness of the current. The trouble which results from a reasonably'large amount of ripple in'the motor does not come from commutation, but from iron-losses.

The normal 'D.-C. traction-motor is a solidfrarne machine, and if the mainfield-flux varies, in this field-frame, it will theoretically produce additional heat. I have tested such tractionmotors with rectifier-supplied currents, such as are provided in l, and I have found that when the percent ripple is 20 to 25%, the amount of motor-heating, at continuous rating, is not measurably different from the heating which is obtained on pure direct current, without any ripple. If the percentage-ripple were increased to as high as '75 it is my judgment that special cooling-precautions would havelto be taken, for .a

ripple as high as that, if the rector were a solidframe motor. If the motor-frame is laminated, the A.-C. ripple-content of the D43. current would not cause any undue heating'even atthese lar er values.

The flux-pulsations, due to the ripple in the motor-current, cause corresponding pulsations in the motor-torque, resulting in a certain amount of vibration of the motor. There is ascarcely detectable amount of increased vibration of the motor, when the ripple is changed fromin to 25%. When the ripple is still further increased, say from 25 to the increase in motor-vibration is quite noticeable, although it would probably be acceptable, so far as vibration is concerned, even at the 75% ripple-level, if other things weresatisfactory. In other words then-C. traction-motor, when powered from a rectifier would not need to be equipped with flexible gears (not shown) ,tsu'ch as are always required by A.-C. traction-motors, even when the 11-0. motor has a 75% ripple in its current.

In general, in accordance with one aspect of my preferred embodiment, it maybe stated that the total inductance of the direct-current circuit 23-5 5 should-be sufficiently small to permit at least a 25% ripple to be present during maximum short-time load-conditions such as during acceleration; although in some cases a ripple might be regarded as the minimum desirable value, or even the usual 15% ripple, which may sometimes be tolerated in the D.-'C. circuit Z3--l 5 during the maximum short-time load-conditions.

If the second harmonic of theinput-circuit 'frequency is by-passed around the motor-winding, as when the D.-C. filter-circuit '33 is used, the higher even harmonics are the only harmonics left in the motor-circuit, and these are unimportant for two reasons: first, the magnitudes of the harmonics are inversely proportional to their orders, so that the fourth harmonic, for example, is only one-half as large as the second, and so on; and secondly, the impedance of the reactance in the D.-'C. circuit, including the ll-C. choke-coil (if any) increases in direct proportion to the order of the harmonic, so that this impedance is twice as great for the fourth harmonic (for example) as it is for the second h rrnonic, and so on. The higher even harmonics rapidly disappear, and they do not ordinarily produce sufficient ripple in the D.-C. motor-current to be of any practical consequence in a rectifier powered self-propelled vehicle.

The angle of overlap, or commutatingangle, of the rectifiers 2i and 22 has an unexpectedly important eifect upon the amount of inductive-interference which is obtained from the A.-C. supply-circuits which supply a rectifier-powered traction-motor equipment of a railway-vehicle,

the standpoint of inductive-interference.

the line.

The angle 'of overlap of a two-phase rectifier, such "as that "shown in Fig. 1, is a function "of a commutation-factor which is equal to the magnitude of the "direct-currentwhich is being commutated 'or transferred from one rectifier-circuit to the-other, "multiplied by the supply-circuit inductance'up to the rectifiers 2| and 22, this inductance being the equivalent secondary-circuit inductanceL'oi the'transformer and of the entire supply-circuit back of the transformer, as seen from the secondary terminals, this product being dividedby the secondary voltage. The angle of overlapis incre-ased'when the effective transformer-inductance L is increased.

'Ifthe angle of overlap,-or commutating-angle, is increased, the A.-'C. harmonics in'the line-current begin'to be reduced, rapidly at first, but by the time the angle of overlap reaches 40, or between and 45, the slopes of the harmonic- "curves leveloffyso that further increases in the angle of overlap yield too small a benefit, in the form of reduced A.--C. harmonics, to warrant'the expenseof the necessary primary or transformerinductance, or the consequent disadvantages in the way of poorerregulation, lower power-factor, and less efficiency. I prefer to use about a 40 angle of overlap, or say between 35 and during vehicle-accelerating load conditions, or maximum short-time line-current or load-conditions, which are the-severest conditions from During continuous-rating line-current or load-conditions, when both the motor-current and the line icurrentmay-"be about of themaximum current during acceleration, the reduction .in the line-current causes the angle :of overlap to be :reduced to'somethin'g like 25 to 30. In thepower-rectifier practice prior to my invention, the maximum 'angleiof'overlap has been about 20 to 25. I prefer to make 20 to 25 the minimumangle of overlap which I ever Z1188 under maximumcurrent conditions.

In general, in accordance with one aspect of my invention, I prefer to include, in my primary circuit 15- l2l-9-l3, sufficient inductance to cause the conducting-periods of successively operating rectifiers to overlap for at least about 20 during each half-cycle, during maximum short-time line-current or load-conditions, "such as during acceleration.

My higher-than-usual angle of overlap alsoincreases the D.-C. harmonics, but, as I have already pointed out, the motor 24 can, in general, stand the increased D.-'C. harmonics, and in the relatively small number of cases when this may not be so, the second harmonics, which are the largest harmonics of the D.-C. circuit, can be bypassed out of the motor by means of the D.-C. ifilter 33.

To give some indication as to the effect of the angle of overlap on the direct-current voltage, it may be noted that a 25 to 31 commutation-angle, with no delayed firing, ,producesa 6% to 8% regulation.

It has been noted that an increase in the angle of overlap carries with it .an increase in the amount of voltage-drop or regulation in the D.-.C. voltage, and a decrease the ,power factor of the Al-C. current which is drawn from The application of the secondary-filter All, with its really considerable leading kva, makes it possible to work at higher angles of overlap than might otherwise be acceptable from the standpoint of power factor. This same leading kva, drawn through the .reactanc'e .ofthe 13 power-supply system (including the transformer), as seen from the terminals or" the filter, increases the A.-C. voltage and hence the D.-C. voltage, thus making it possible to use a higher A.-C. reactance, and thus to work at a higher angle of overlap, than might otherwise be acceptable from the standpoint of regulation. It is desirable to have some droop, or regulation, in the D.-C. voltage-curve, in order to keep the number of speed-controlling or accelerating notches to a reasonable number, but, on the other hand, too much regulation might be undesirable, in requiring the addition of too many turns on the transformer-secondary, and too high a lowload secondary-voltage in order to maintain the desired voltage during the accelerating period.

The use of an uncommonly large angle of over lap, in accordance with my invention, involves the use of an uncommonly large amount of equivalent secondary-circuit inductance L of the combined transformer and supply-system, that is, the inductance as seen from the secondary side. This means, either the use of a large primarycircuit inductance lit, in series with the supplyline, or, what is usually more economical, the building of an unusually large amount of leakage-inductance into the transformer ill, so that the primary-circuit inductance 53 can be omitted, or by-passed by the closure of the cutout switch The primary-circuit inductance could be increased, as by a manipulation or" the switch i during line-circuit load-conditions which are lighter than maximum starting-conditions, so as to prevent the commutating angle, or angle of overlap, from decreasing when the line-current decreases. When the commutating angle decreases, the telephone-interference factor Til? increases, but the total telephone-interference is the product of TIF and the line-current, so that, in general, any increase in TIE, when the linecurrent decreases, does not increase the noiselevel of the telephone-interference.

The effect of the secondary-connected A.-C. filter to, with all of its ramifications, is somewhat difficult to grasp, even on a qualitative basis; and it is practically impossible to pre-calculate, on an exact mathematical basis, without the use of empirically obtained values of the constants. An approximately equivalent circuit for the transformer-secondary, the rectifier, and the D.-C. circuit is shown in Fig. 2, corresponding approximately to Fig. l, and using the same refcreme-numerals. One of the d'ifiiculties is to determine the proper value of the effective secondary inductance to be applied to the two sec-- ondary halves or to the whole of the secondary winding ii, and the proper value of the mutual inductance between the two secondary halves.

In Fig. 2, the current-flow is shown at an instant when the upper rectifier 22 is carrying a current 1', and when the other rectifier 2! is carrying no current. Zhis current I flows from the transformer-terminal l9, through the rectifier 22 to the positive il-C. terminal 23. and thence through the D.-C. reactor 39 and the D.-C. motor 24 to the negative 31-0. terminal it, which is the mid-point of t -e transformer secondary-winding ll. At this point, the current divides. A major portion, 1, of the current flows through the upper secondary-half, from the mid-portion i6 to the upper terminal 19, because the polarity of the line-frequency secondary voltage is such that the upper terminal 19 is positive with respect to the mid-point l6, and

the mid-point I6 is positive with respect to the lower terminal 18. However, there is presumably a certain small current-component, in one direction or the other, flowing in the circuit from the mid-point l6 through the bottom half of the transformer-secondary to the lower terminal l8, and through the filter-circuit 40 back to the upper terminal 19, as indicated by the arrow 1 in Fig. 2.

The circuit-constants for this circuit vary, with each frequency of the harmonic or ripple content of the main current I, so that, in reality, a separate equivalent-circuit diagram would have to be set up for each harmonic, replacing the conducting rectifier 22 with an equivalent impedance, and omitting the non-conducting rectifier 2| altogether. There are two transformervoltages, which are produced in the lower secondary-half between the points it and it, at any harmonic-frequency. One of these transformer-voltages is an odd-harmonic voltage, which is induced by the corresponding odd harmonic of the current in the transformer, and which is in the same polarity throughout the entire secondary winding ll, so that, if the polarity of the upper terminal IQ of the top secondaryhalf is positive, for this transformer-induced oddharmonic voltage, at any instant, the polarity of the mid-point l6, which is the upper terminal of the lower secondary-half, is also positive for that secondary-half at the same instant. Be sides this voltage, there is a balance-coil voltage, which is produced by the mutual impedance between the two halves of the transformer-secondary H, and which has a tendency to maize the current I which enters at the mid-point l6 divide equally, so that it goes in opposite directions, upwardly in the upper winding-half, and downwardly in the lower winding-half.

The effective inductance of the lower windinghalf, if that can be estimated, is to be added to the filter-circuit impedance of the capacitor C and the resistance R, to determine the direction and phase and magnitude of the current I of the lower winding-half, whereas the phase and magnitude of the main divided current I, in the upper winding-half is dependent only upon the efiective inductance of that winding-half (if it can be estimated), and the resultant of the transformer-induced voltages.

The explanation just given assumes that the harmonics involved are the odd harmonics, which are the harmonics which are present in the pri mary circuit of the transformer. In the Dz-G. circuit, however, the harmonics which are present as ripples in the actual direct current I are even harmonics. The odd harmonics are obtained in the process of rectification, as will be mentioned later on, but this rectification-process occurs during each half-cycle of the fundamental wave of the primary-circuit supply. Furthermore, the effect of the filter-capacitor C cannot be properly evaluated, in any equivalent A.-C. circuit, because this filter-capacitor C is short-circuited, and completely discharges itself, during each cominutating-period, that is, twice during each line-frequency cycle, as will now be discussed.

During the commutating period, when both rectifiers 2! and 22 are carrying current, in Figs. 1 and 2, the total secondary voltage, across the terminals H3 and i9, is practically zero, being the sum of the voltage-drops in the two rectifiers, under the current-conditions prevailing in the respective rectifiers. This means that the filtercapacitor C is practically short-circuited during this commutating period, so that it becomes completely discharged, its discharge-circuit being through the anode-lead I 8, the rectifier 21, the common cathode-circuit 23, the rectifier 22, the anode-lead l9, and the discharge-resistor R, back to the capacitor C. The inductance Ln of this discharge-circuit is very small, being practically the inductance of the leads. In one particular case, this discharge-circuit inductance has been estimated at Ln=.00006 henry. If the dischargeresistance R is smaller than the critical resistance for this discharge-circuit containing the capacitance C and the inductance Ln, then the discharge of the capacitor, during each commutating-period, will be oscillatory, at a high frequency dependent upon the product OLD. If the discharge-resistance R is equal to, or larger than, the critical value, the capacitor-discharge will be non-oscillatory.

A circuit which would be somewhat easier to analyze, than Fig. 2, would be a bridge-rectifier circuit, as shown in Fig. 5, in which the transformer-secondary Winding H has substantially the same number of turns as only half of the transformer-Winding ii in Fig. 1 or in Fig. 2. In Fig. 5, the two previously described rectifiers 2! and 22, are replaced by four bridge-connected rectifiers 22', 2!, 22" and El", arran ed to receive secondary voltage from the terminals :9 and I8 and to deliver D.-C. voltage to the terminals 23' and Hi, so as to energize the motor 2 with rectified current. In Fig. at an instant when the upper terminal 99 is positive, the current I is supplied from this terminal l9, through the rectifier 22', to the D.-C. circuit 23'l6', and thence through the rectifier 22" to the other secondary terminal Hi.

It is usually considered that, from a rectifier standpoint, the bridge-rectifier circuit of Fig. 5 is equivalent to the half-wave rectifier-circuit of Fig. 2, but this equivalency of the two circuits has previously been considered without reference to the effect of the filter-circuit lil which I have added. One practical dilierence between the two circuits is that the bridge-circuit of Fig. 5 dispenses with the necessity for using one-half of the transformer-secondary which has to be supplied with Figs. 1 and 2, but this advantage in the transformer-cost is obtained at the expense of requiring twice asmany rectifier-tanks, besides doubling the rectifier-losses, and in practical circuits and rectifier-tanks are much more costly than the secondary-winding turns, so that the double-wave rectifier-circuits of Figs. 1 and 2 are much more economical.

If we neglect the eifect of the an le of overlap, which discharges the filter-capacitor C twice during each line-frequency cycle, We can calculate the effect of the A.-C. filter 38 in the bridge-circuit of Fig. 5.

I first consider the bridge-type rectifier, because it is easier to analyze for harmonics. I first assume that a certain magnitude of a particular harmonic-current IX is present in the secondary current of the transformer. Then, without my secondary filter, and assuming a negligible magnetizing current of the transformer, the ma nitude of the (secondary-voltage-base) transformer-voltage which is produced by said secondary harmonic IX is simply the impedance-drop in the equivalent supply-circuit inductance L, which is the inductance of the transformer and primary supply-circuit combined, as seen from the secondary side, or

ELZWLIQ: (2)

'16 If now, we add the secondary filter-capacitor C, with its serially connected damping-resistance R, as shown in Fig. 5, the voltage which is required to maintain the same current-harmonic L; in the input-circuit of the rectifier-bridge will be I, jwLl( R j/yC) l R-l-jwL-j/wC' jwL Rj/wC The magnitude of this transformer-voltage, with the filter in use, is

mi 2 i =*-t wL-1 wc (4) The effect of adding the filter, therefore, (still assuming that the bridge-input current-harmonic L; is fixed by the D.-C. output-circuit of the bridge), is to multiply the previous transformervoltage E1. by a factor K which is If we put for the angular velocity 21rfo of the parallel-resonant frequency In of the filtercapa-citance C with respect to the supply-line inductance L, up to the filter-terminals, we may replace the capacitance C by yielding Man e Ra g dy H w 10 w If we now assume that the discharge-circuit has an inductance Ln of .00006 henry, and that the discharge-resistance R has its critical value, then tion 8, then the value of the multiplying-factor at resonance becomes being independent of the resonant frequency in, and becoming larger with an increase in the equivalent feeder-and-transformer inductance L, as seen from the secondary side, but becoming smaller with an increase in the extremely small discharge-circuit inductance Ln.

assets In the curve-diagram of Fig. 6, I have plottedthe calculated values of the multiplying factor K against the frequency, using Equation 9, thus assuming, with Equation 9, a critical value. of the discharge -resistance R, and further assuming three difierent transformers, having equivalent secondary inductances (including thesuppl ysystem); asfollows,

We then have: the following, three equations,

2 w? 2 l 111a a In. each case, I. have; assumed three different parallel-resonant frequencies;

f :100 f =20G In Fig. 6; the three curves, for a resonant-frequency of it!) cycles, are shown at 4 is, Ms and.4 le,. respectively, the subscripts indicating the as-- SlllllECl equivalent secondary inductanoeL,inmillihenrys. As was to be expect'edifrom Equation 11, the-maximum value of the multiplyingrfactor'K, which occurs at the resonant frequency, increases R unchanged at its criti'calivaiue, the: maximum.

value of the multiplying-factor K remains unchanged; as will' be seen by comparing curves M4, 424 and 444.

In Fig. 6, I have also plotted two other curves, numbered 452 and 455, respectively, showingthe calculated eifect of using discharge-resistances R of different magnitudes, assuming, for this purpose, a discharge-capacitor C havingacapacitance of'IiOOIZ farad', with a transformer, asshown in; Fig. 1, having an. opensoircuit voltage of 23441 onthepriinary winding 9, and 1330, on the. total secondary winding H, with a 25-cycle supplyline. The 25-cycl'e reactance of the primary. circuit, including the transformer i ii, the A.-C.,sup-

ply-lines l and i, and the source 5, is 5.59ohms on. The A.C. reactance,

the primary-voltage base. on the secondary-voltage base, is

5.59:1.800 ohmsat 25.0yc1es 18 The harmonic-order" at' which the filter-capacitor is tuned is Curve 4521s plottedfromthis equation.

The -critical value ofthe' discharge resistance,

if" the assumed value of the discharge-inductance Liar-0006 is correct, would have been- R,= /%8%i-;=J=1.i1 r ohms and with this critical value of discharge-resistance, the multiplying-factor would have been flu 2 2+Q07199f- 5 Curve dicisplotted fromthis equation.

It Will. be. seen, from. curves 452 andv 450, in Fig. 6,,that theeffect of using discharge-resistances R of. different magnitudesis. as was to be anticipated from Equation 10, namely that the peak of. the multiplyingrfactor K, at. resonance, becomes smaller. whenthe discharge-resistance R is increased.

It will be. noted that I started out by plotting curves 4'], 42 and. 44, with their various subscripts, using equations. which were calculated for. abridge-rectifier. circuit,,. as shown in Fig. 5, and then, I. used the. same. formulas. as ifthey applied to a double-wave rectifier-circuit such as. is. shown inll'ig.v 1. or Fig. 2, because curves 452. and. 450..5/81'6 plotted. for the type.- of rectifiler-connection whiohisshowninFig. 1. Inso applying the formulas,.I have; neglected the effect of. theshort-circuiting, of thefilter-capacitor C during thecommutation period, ,as already pointedout; andl. have alsoignoredthealmost insuperable problemsrelative to the partially noninduotive division. Ofthe-DJ-C. harmonics in the two halves.oflthesecondary winding l l in Fig. 2, and, also. the. well-nigh. insuperable problemv of converting, the rectified. even-numbered. harmonies of the. D.-C.. circuit. into. theequivalent.

the actual values of the multiplying-factor K, which were obtain d at different odd-numbered harmonics of the line-current, these actual values being indicated by small circles drawn on the diagram. There appear to be two respects in which the double-wave rectifier, with a midtapped secondary, as shown in Fig. 1, departs from the theoretical formulas for the bridge-rectifier circuit of Fig. 5. Apparently, the periodic short-circuiting of the filter-capacitor, during the commutating periods in Fig. 1, makes the circuit operate, particularly in the vicinity of the resonant frequency and at lower frequencies, as if the capacitor were slightly smaller than it actually is, this effect being indicated by the apparent shifting of the resonant-frequency to a slightly higher value than the calculated value.

A second respect in which the actual values obtained for Fig. l depart from the values which were calculated from the formula, as shown in curve 452 of Fig. 6, is to apparently somewhat reduce the peak of the resonance-curve from the value which is shown for the calculated curve 4-52 in Fig. 6. I believe that this effect is due to the changes in the current-division which was discussed in connection with Fig. 2 and the divided currents I and I" which flow in apparently opposite directions in the two halves of the transformer-secondary. At the higher frequencies, the voltage-drop across the filter-capacitor C approaches very small values, so that the two secondary terminals l3 and is in Fig. 2 are practically short-circuited by the filter-capacitor C,

for harmonics of high orders. This short-circuiting effect at high harmonics or frequencies would not have been obtained if the filter-circuit 43 had been tuned, that is, if it had contained any substantial amount of inductance in series with the capacitor C and the resistor B.

At lower frequencies, the impedance-drop across the capacitor C, in series with the equivalent inductance of one-half of the transformersecondary, has a finite value, and varies in both magnitude and sign (or phase), with diilerent low frequencies. These effects are practically impossible to precalculate, with any degree of ac curacy. Fortunately, the effects are favorable to the circuits shown in Figs. 1 and Z, in apparently reducing the sharpness of the peak which represents the maximum value of the multiplying-factor K by which the low-frequency harmonics are multiplied, as a result of adding the A.-C. filter-circuit 4!] in Fig. 1.

A better insight into the operation of my invention will be had by a study of the reproductions of oscillographic records which have been made, as shown in Figs. 3, 4, '7 and 8.-

Figs. 3 and 4 show the results, first without the A.-C. filter 4b of Fig. 1 (in Fig. 3), and then the results obtained with the filter connected (in Fig. 4). The apparatus used for these oscillograms was the same as that for which curve sea was drawn in Fig. 6, except that the discharge-resistance R had a value of 1.063 ohms, instead of the value 2.084 ohms, which was the case for said curve 452. The discharge-resistor value R=1.063 is thus less than the calculated critical value of discharge-resistance, which has been shown to be Rc=l.4l4 ohms, as shown in Equation 17, and hence the capacitor-discharge, during the commutating period, should be oscillatory, as indeed it was, as shown in Fig. 4, as will be pointed out directly.

' Figs. 3 and 4 each show four simultaneously obtained traces, showing relative phase-positions of line-current (or the current in the primary lead 52) the A.-C. filter-current (or the current in the filter-circuit it); the secondary voltage (appearing across the terminals 118 and i9); and finally the combined anode-currents (or the currents in the two anode-leads l3 and iii). The overlap-period or commutating angle is shown by the zero values of the secondary voltage, one such period extending from a to b in each of said Figs. 3 and 4. It will be noted that the filter-current is zero, in. Fig. 3, because the filter is off or disconnected. In Fig. 4, it will be noted that the discharge-current of the filter-capacitor, at the beginning of each commutating period ab, is oscillatory, as indicated by the change in the polarity or direction of the filter-current before it subsides to zero, during this commutationperiod. This discharging-period of the filtercapacitor extends from a to c in Fig. 4. It will be noted that the filter is inactive during the rest of the commutating period, that is, from c to b, and then the filter becomes active again during the energy-transfer periods of the respective rectifiers, as from b to d in Fig. 4.

The efiect of the A.C. filter, in magnifying or enlarging the low-frequency harmonics, in both the A.-C. line-current and the 11-6. current (which is a rectification of the anode-current), is plainly shown in Figs. 3 and 4. In fact, the shape of the line-current, in Fig. 4, with the filter on, looks, to the eye, much worse than the shape of the line-current in Fig. 3, with the filter off. This is because the low odd harmonics, which are readily detected by the eye, are actually worse in Fig. 4 than in Fig. 3, but thehighfrequency harmonics, which can be properly evaluated only by meticulously careful analysis, fall ofi very rapidly, in Fig. 4, as the harmonicorder increases, so that these harmonics practically vanish, at orders of about the 39th, corresponding to 795 cycles on a 25-cycle supplyline.

Herein lies the explanation of the efiicacy of my A.-C. filter-circuit cc, namely that it strongly and powerfully suppresses the harmonics of the.

A.-C. line-current, starting at a frequency not greatly above the resonance-frequency of the ii"- ter. This is important, from a telephone-interference standpoint, because of the extremely rapid rate at which the value of the interference weighting-factor T increases, particularly in the frequency-range from say about 200 cycles, where T=16, to 1070 cycles, where T=l2,l0(l.

The efiect of the omission or inclusion of the D.-C. filter-circuit 33is illustrated in Figs. 7 and 3, which give portions of simultaneous oscillographic traces showing the relative phases of the D.-C. filter-current (in the circuit 33), and the motor-current Idc (in the circuit 2528-2l). In Fig. 7, the D.-C. filter was off, so that the filter-current was zero, and it will be seen that the percent ripple in the motor-current was 29%. In Fig. 3, the D.-C. filter was connected in circuit, in parallel to the motor, and it will be seen that the filter draws a practically sinusoidal doublefrequency current, and that the percent ripple in the motor-current is thereby reduced to 9%. This reduction is brought about, because the second harmonic is much the largest of all of the harmonics in the D.-C. circuit 23 |e, and when this harmonic is removed, the D.-C. ripple is always reduced to relatively low values, as has previously been discussed. The commutating period in Figs. 7 and 8 has been indicated by the vertical lines a and b, as in Figs. 3 and 4.

It will. be noted that the. percentage; ripple in.

'7, without. the use of the D.-C. filter, Was quite within the usually acceptable limits, which have been indicated as approximately between 25 9, some motors may take a 40% ale, or even up to the vicinity of 75% ripple, .icut undue distress, whereas, in other cases, where r ctor-des n is more critical, in one way or another, a p of less than 25% may be required, has previously been indicated. The tuned Dx-C. filter asshown in Fig. l, and as shown. the comparison of Figs. 7 and 8, provides a me whereby this D.-C. ripple can easily be reduced. to any value which may be re l by the particular type of, direct-current rho r which is used, while at, the same time permittigthe overall ripple,in the.D.-C. circuit 23-l6 to have as. large a value as may be 116ESSlb ted from the standpoint of an acceptable inductive telephone-interference factor TIF.

It seems to be necessary, Or at least extremely important, that the.D.-C.. filter 33 should be a tuned. filter, including an inductance 35 as well as a. K A.-C. filter which is untuned, in the. sense that it s a capacitance C but substantially no included in the filter-circuit W. I i usin an untuned D.-C. illter, conly a capacitor without an int l; in accordance. with a previously shunting the outputecircuit, of.

filter-capacitor for smoothing the 11-0. voltage, and I have i. invar ably ran into difficulty, when. cult on the rectifier consisted of a without any substantial inductance in series with it, would invariably or frequently resonate in parallel-resonance with the parallel-connected motor-inductance at some harmonic-frequency, thereby producing undesirable effects.

In the usual practice of my invention, the A.-C. supply-circuit will be given a certain amount of inductance, which is usually most economically built into the transformer it, which Will produce larger-than-usual angleof overlap, or commutating angle, of the rectifiers, as has previously been discussed, while the: D.-C. circuit. 23-16. 1

will usually be designed with. av smaller-thanusual inductance (including the inductance of the motor itself, plus the inductance of any serially added inductor which may be used). In general, it will be found economical to operate the motor with as high a percentage of ripple as the motor will take, without undue distress, and in some cases, if the motor is rather unusually sensitive to ripples, it may be necessary, from the standpoint of telephone-interference, to provide the overall DJC. circuit 3lli$ with more ripple than the motor can take, and to remove the excess ripple from the motor-circuit by means of the second-harmonic D.-'C. filter 33.

With the two above-stated design-features, namely a high primary inductance (or a high supply-circuit inductance up to the rectifiers, which causes a large angle of overlap) and a small D.-C. inductance (or high ripple-percentage), I believe that the telephone-interference factor TIE will frequently be found to be Within acceptable limits, but whenever this is not so, my .A.-C. filter til provides a means whereby the telephone-interference factor TIF may be ecoquired.

The importance and the advantages of my anode-to-anode; or secondary-connected A.-C. filter at areso, great, however, that I believe that this filter will usually be advantageously used, even when it is not requiredv from. the standpoint of reducing telephone-interference. I believe it to be an. important part of my invention tov use this A.-C. filter, even aside from strict adherence to. the preferable limits regarding the commutating angle and the D.-C. ripple, and even regard less of telephone-interference requirements. Some of the advantages of the A.-C. filter 40 may be briefly indicated as follows.

The anode-to-anodc or secondary-connected filter ii! delays the building up of the anodevoltage at the beginning of each. conductingperiod, while the filter-capacitor is charging, as shown at 16 in Fig. 4, thereby allowing more deionizing time, which reduces. backfires to an extentwhich is almost: unprecedented in rectifierpractice.

I have found that the required A.-C. filter 42!, for red more efiective, on thesecondary side rather than the primary side of the transformer (particularly when a high supply-circuit reactance is used, and built into the transformer), that a secondary filter-capacitor of only one-sixth the size would do the work of a given primary filter-capacitor. There are several reasons for this.

In the first place, the cause of the A.-C; linecurrent harmonics is the rectified differenceripples which result from the reduction of the 11-0. ripple from the theoretical ripple-content of a rectified sine-wave; that is, the rectified difference between thetheoretical ripple-content of the D.-C. load-current, plus any A.-C. inputharmonicsthat are caused by or during the commutating-angle. These difference-ripples in the D.-C. circuit are rectified by the two rectifiers in such a manner that a polarity-reversal of the difference-ripple occurs, not at each half-cycle of each of the component ripple-frequencies, but at each half-cycle of the primary wave, or during the respective conducting-periods of the two rectifiers. The harmonics in the A.-C. secondary current in the output-leads I8 and ill of the transformer result from a Fourier-series analysis of these rectified difference-ripples.

I a low-impedance filter-path 4i? for these recd creme-ripples is provided in shunt relationv to the secondary side of the transformer, then the only harmonic-voltage which is built up in the transformer is the voltage-drop due to the flow of these rectified difference-ripples through the small impedance of this filter-path id. This path has a-loWer impedance than would be provided by any prinmry-side filter, because then the harmonics due to the rectified differchoc-ripples would have to pass through the inductance of the transformer before reaching a primary filter-capacitor, and the total impedance of the filter-path would be much greater than that of the secondary filter. Furthermore, when the filter-capacitor is placed across the secondary side of the transformer, the inductance of the transformer acts as a bufier between the filtervoltage and the line-currents, thus reducing the magnitude of theharmonics which are circulated in the A.-C. line by any given filter-voltage.

If the harmonic-reduction requirements are such as to require a large filter-capacitor, having a large lava. rating, a serially connected resistor ing the A.-C. harmonics, is so muchis necessary in order to prevent, or at least reduce, the oscillations in the discharge-current of the capacitor during the commutating periods of the rectifier. These oscillations, if large enough, may produce destructive excessive voltages and may render the rectifiers unstable and short-circuit them. Said oscillations also add to the harmonic content of the anode-currents, and are reflected into the line-current. If the required size of filter-capacitor is intermediate in value, the amplitude of oscillation of an undamped discharge may be so small that it would not pay to accept the power-loss incident to adding a series resistance to the filter-capacitor, so far as the capacitor-discharge conditions are concerned.

The damping-resistance R. for the parallel or A.-C. filter if! should always have 1 R losses not exceeding 1% to 2% of the kilowatt output-rating of the rectifiers, and usually much lower resistance-losses will be involved.

However, the beneficial effect of the filter, in eliminating the higher harmonics in the transformer, will also eliminate the very substantial iron-losses which would otherwise be produced in the transformer-iron by the fluxes due to the eliminated higher harmonics, resulting in an energy-saving which may, under some conditions, equal or exceed the resistance-loss in the damping-resistance of the filter.

Instead of connecting the filter-capacitor across the two terminals Si and S2 of the secondary winding, it could be connected, preferably symmetrically, across any other number of secondary-turns, thus changing the voltage of the capacitor. By symmetrical, I mean, with the mid-point of the connection coincident with the mid-tap of the secondary. There is evidence to indicate, particularly when the A.-C. filter is connected across secondary tap-connections that are symmetrical, that, due to the mutual coupling between the two secondary-halves, the inductance of the secondary winding at one end seems to be equal and opposite to the inductance of the secondary winding at the other end, so that these end-inductances apparently do not add any net inductance in the filter-circuit, over the inductance which is present when the filter-capacitor is connected directly from anode to anode.

There are certain practical limits as to the size of the A.-C. filter-capacitor C. If it is to small, it will be in parallel-resonance with the equivalent secondary inductance L of the transformer (including the primary-circuit inductance) at too high a frequency, so that the effect of adding the filter-capacitor will be to magnify or enlarge harmonics which are of a high enough order to involve rather large telephone-interference weighting factor T, thus producing a rather large increase in the telephone-interference due to the products of the harmonics in question, times the proper weighting-factor T, and this increase in interference may even be more than the decrease which results from the cutting 01f of still higher harmonics. In many cases, a resonant frequency of the order of the 6th or possibly the 8th harmonic may be a satisfactory optimum figure to name, particularly on a -cycle system. Expressed in frequencies, the C and L, as hereinabove defined, would then be optimally tuned to resonate at something of the order of 150 cycles, although possibly they might resonate at a frequency as high as 200 cycles, particularly under maximum line-current load-conditions, during acceleration.

The eifective value of the equivalent secondary inductance L, or the effective supply-circuit indu'cta'nce up to the rectifiers, is subject to a certain amount of variation, depending upon themsition f the vehicle between the successive feed-in points 2 and 3 of the trolley l, and also dependent upon the length of feeder l back to the source 5. This variation in reactance may amount to anywhere from 0.2 ohm to something like 6 ohms, on a primary-voltage basis, although ordinarily the reactance-variation is not as great as this. If the reactance-variation is small enough to keep the filter-tuning well away from both the 5th and 7th harmonics (on a 25-cycle system), with an average tuning of about the 6th harmonic, this arrangement has been found to be advantageous. However, a variation in the tuning, from the 6th, through the 7th, and into the 8th harmonic, would be quite acceptable. If a sufficiently large damping resistance R. is used, as will be subsequently pointedout, a much smaller filter-capacitor C may be used.

If a larger filter-capacitor C were used, tuned, say, to the 4th harmonic (in a 25-cycle system), or to a frequency of cycles, the fundamentalfrequency 'resistance losses (in any given size of the series discharging resistance R) would be larger, being proportional to the reciprocal of the third power of the frequency-ratio of the tuned frequencies which are being compared. I consider 100 cycles, or the 4th harmonic on a 25-cycle system, the minimum resonance-frequency to which my filter-capacitor C should'be tuned.

As has been indicated before, the filter-capacitor C may be tuned to a much higher frequency than 200 cycles (or the 8th harmonic on a 25- cycle system), particularly if an adequate damping resistance R is used. In general, whatever the size of the filter-capacitor C, there is an op timum value of damping resistance R. Thus if, for any given value of the filter-capacitor C, the value of the damping resistance R is varied, the inductive-interference noise-level may be plotted against the damping resistance R, and a value of B may readily be found at which the noise-level is the lowest, or at which the noise-level is decreasing so slowly, with increasing values of R/ that no further increase in B would be economically justifiable. If a smaller filter-capacitor C is used, it will draw a smaller filter-current at the line-frequency (25 cycles, for example), and hence a larger value of damping resistance B may be used, without prohibitively increasing the PR losses in said resistance.

When the value of the damping resistance R is increased, the sharpness of the resonance at the tuning-point is decreased, so that the multiplying-factor K, at resonance, becomes much smaller, approaching unity, as will be seen from Equation 10. Consequently, with a suficiently large damping-resistance R, the A.-C. filter 49 will not greatly increase the telephone-interference noise due to the lower-order harmonics, up to a value somewhat higher than the tuningpoint, while the filter will still reduce the noise due to still higher harmonics.

It has been discovered that some of the lineharmonics which are caused by the railway-car or locomotive seem to resonate with the trolleycapacitance, In one instance, a pronounced effect of this kind has been observed at approximately 280G cycles. My A.-C. filter 48 is quite effective in reducing this trolley-resonant harmonic, even when the filter is tuned to such a high frequency that it would otherwise seem to be of no noise-reducing value. U p

I have found that, when an optimum choice of the resistance-value R is made, my A.-C. filter is effective in making some reduction in the telephone-interference noise-level when even a small filter-capacitor C is used, tuned to about'the 12th harmonic on a 2 -cycle system, orSOO cycles; and the evidence seems to indicate that a very much smaller capacitor would be effective, tuned to the 24th harmonic (609 cyc es), or even to the 36th harmonic (990 cycles), 01' higher, provided that an adequatelylarge damping-resistance R is used, in each case.

In addition to reducing the telephone-noise interference, my A.-C filter (it, if adequately damped, also reduces or eliminates the radio-interference harmonics.

I have heretofore discussed the value of the damping resistance R from the standpoint of its critical value during the discharge of the filtercapacitor C. There is evidence that clamped or partially damped charging-conditions for the filter-capacitor C may also be desirable, particularly from the standpoint of radio-interference, referring to the conditions which exist when the line-frequency voltage is again applied to the capacitor C at the end of each commutatingperiod of the rectifiers. During this chargingperiod, the effective inductance in series with the capacitor C is the equivalent combined transformer-and-line inductance L, as viewed from the secondary side of the transformer, and the critical value of the damping resistance R, during these charging conditions, is

Rea Z76 (19) which is much larger than the critical damping resistance on discharge, as given in Equation 8.

I believe that the damping resistance R should have a value between the limits of the critical value on charge, and say about half of the critical value on discharge, with the higher values of damping resistance prevailing at the higher tuning-frequencies or smaller values of the capacitor C, and with the lower values of damping resistance prevailing at the lower resonance-frequencies or when a large filter-capacitor C is used.

The foregoing discussions of the secondary filter Ml have been based upon the assumption that a fixed transformation-ratio, or primary-to secondary voltage-ratio, is applicable to this filter, notwithstanding the use of the variable taps I! to control the secondary Voltage which is applicable to the anode-leads iBi9 of the rectifiers. fixed filter-voltage ratio is usually desirable, because any change the voltage-ratio which is applicable the filter-capacitor C will change the resonant frequency in inverse proportion to the voltage-change. Thus, if the filter-capacitor C, when connected across the full secondary voltage of the transformer (or any other voltage), is in resonance with the transformerdnductance L at the sixth harmonic, or 150 cycles, then a change in the filter-capacitor connections, to secondary taps i'l having a voltage equal to 14% of said voltage, will change the 'esonant-harmonic order from 6 to 6/.14=42.9, and will change the resonant frequency from 150 cycles to l/.l4: 071 cycles, which is substantially the frequency at which the telephone-interference weighting-factor is a maximum, or T=12,000 approximately. Since the filter-voltage should preferably, therefore, be fixed, it is obvious that any voltage-changing taps that are used should be on the secondary rather than the primary side of the transformer.

In practical railway-electrifications, in rectifierpowered equipments ranging in size from the smallestmultiple-unit car to the largest locomotive, the anode-to-anode running-voltage may vary from something like 300 volts to something like 1300 volts, more or less, while the equivalent secondary reactance L may vary over a range extending from something like .62 to .002 henry, more or less, depending upon th primary or linevoltage, the rating, the frequency (such as 25 or 60 cycles), and other variables. Ehe A.-C. filtercircuit generally takes a current 16 of the order of perhaps anywhere from 1 or 2 amperes to possibly as high as 500 ainperes, with perhaps a representative value around amperes. The resistanc'e-losses through the damping-resistor R are equal to 10 which is subject to a rather wide range of variation, with the resistancelosses increasing rapidly as the size of the filtercapacitor C is increased, that is, as the resonantfrequency of the filter is decreased.

With preferred amount An-C. filter, I save enough weight and cost in traction-motors (including the gearing), and in the transformer and the control, to more than pay for the rectifier, the filter and the associated equipment. This makes the rectifier motive-power equipment favorably competitive with all other types of electric motive-power equipments for self-propelled vehicles.

The secondary-connected filter-capacitor C of a rectifier-powered traction-motor equipment is completely discharged each half-cycle, during the commutating period when the secondary voltage is very nearly zero, being equal to the sum of the voltage-drops of two rectifiers. Hence the secondary-connecter filter-capacitor C has no opportunity to build up cumulative resonance-voltages at any except the highest frequencies, which are not a problem. In this respect, the rectifier motive-power unit has a distinct advantage over alternating-current motive-power equipments such as the single phase 'tractiommotors, which have quite a few line-harmonics, but which cannot be equipped with filters which are discharged every half-cycle, thus requiring resonating filters to reduce the line-harmonics, these resonating filtersbeing much heavier, and much more costly to build and maintain, than my non-resonating filter C. in this respect, the secondary connection of the filter, in a rectifier power-unit, has an advantage over the primary connection of the filter, in the same kind of rectifier power-unit, because the inductance of the transformer would be interposed between a primary-connected filtercapacitor and the short-circuited secondary terminals, during the ccnimutating periods, so that the prim'ary-cormected capacitor would not be as effectively and completely discharged each halfcycle.

My anode-to-anode or secondar filter-capacitor C also has a very distinct advantage in at least partially counter-balancing the principal objections to the large angle of overlap which I prefer to use, these objections being the lowered power factor and the increased regulation which is produced by the large equivalent inductance L which is required in order to maintain a large angle of overlap. The filter-capacitor C draws a substantial leading current, which directly improves the power factor, and which, flowing through the large inductance L, produces an additive voltage-component to the line-voltage, thereby improving or reducing the regulation.

An important advantage of my rectifier-powered motive-system is that it works just as well,

or better, on a GO-cycle supply, than on the usual 25-cycle (or less) frequency which is common in railway-electrifications. The 60-cycle supply has many advantages, which are well known.

A very important advantage of my rectifierpowered traction-equipment is that it combines the principal advantages of both the single-phase trolley-system and the direct-current railwaysystem. The chief asset of the single-phase trolley-system lies in the economics resulting from its high voltage, and in its flexibility. If we increase the trolley-voltage, we reduce the inductive-interference IT values in inverse proportion to the first power of the voltage. The chief asset of the direct-current railway-system lies in the simplicity and the flexibility of voltage-control of its direct-current traction-motors. My present system combines all of these advantages.

As compared to the single-phase series-motor system, a rectifier motive-power system is lower in cost and weight, occupies less total space, has higher running-efiiciencies, a higher power-factor, less standby-losses, and as small an amount .of inductive interference as may be desired. I

believe that the reduced maintenance of the D.-C'. traction-motor and its gearing should easily more than overbalance the added maintenance of the rectifier-equipment. The direct-current motors of the rectifier-powered vehicle have much higher values of tractive-effort than any single-phase motor, and the rectifier motive-power equipment has better anti-slip features, and can be safely applied at higher adhesions, than is permissible for other types of motive-power. The net results of these tractive effort advantages means that fewer axles will be needed for a rectifier powerunit than for a single-phase motor-equipment, for a given duty-cycle, and that more horsepower per axle can be used on a rectifier-locomotive than has ever before been contemplated on any type of locomotive. Moreover, the rectifier motive-power unit can be used under a 60- cycle trolley, or it can be used interchangeably between a 25-cycle trolley-zone and a 60-cycle trolley-zone, or between an A.-C. and a D.-C. zone.

Similar advantages are to be found for the rectifier power-unit when it is compared with the motor-generator power-unit for self-propelled vehicles.

I have so far described my invention in its application to a rectifier-powered railway-vehicle, which is the first application for which my invention was designed. It will be noted that I have provided a novel rectifier-design which, in some cases, avoids inductive-interference difiiculties without requiring a filter; and I have also designed a novel filter-design for connection to the secondary circuit of the transformer which energizes the harmonic-generating device from the supply-circuit which is subject to inductive interference. The general principles of my invention are susceptible of more general application, in circuits and systems other than the rectifier-powered railway-vehicle which I have thus far illustrated. Some of these more general applications of my invention are exemplified in Figs. 9, and 11, which will now be described.

Fig. 9 is a view somewhat similar to Fig. 1, except that it shows my secondary filtering-device applied to a railway vehicle in which the load may be any one of a number of known kinds or embodiments of an alternating-current motor or motors 24', which do not use any rectifier.

28 The A.-C. railway-motor 26 is an alternatingcurrent series commutator-motor, a known form of which is illustrated, comprising an armature 25', a serially connected interpole field-winding 26', and a serially connected reversing-switch 21 which connects the main field-winding 28' in series with the motor. These motor-circuits are energized in accordance with any suitable variable-speed control-system, which is symbolized, in Fig. 9, by having the motor-circuit energized between the secondary-terminal S2, and a variable tap 58 on the secondary winding H of the transformer it.

The series alternating-current motor 25 of Fig. 9 may be regarded as being characteristic or symbolic of any load-device of a type tending to draw objectionable supply-circuit harmonics in the telephone-interference frequency-range. As previously indicated, such a. motor produces a considerable amount of telephone-interference in the supply-circuits tl, although the intensity of the telephone-interference noise, due to A.-C. traction motors, has heretofore been less than was obtainable in any previously known rectifier-powered railway-vehicle. My untuned secondary filter-circuit 40, comprising a capacitor C and a damping resistance R, precisely as shown in Fig. 1, and with substantially the same limitations, is applicable also to A.-C.-motored railway-vehicles, as shown at $2 in Fig. 9, and when so used, it very materially reduces the noise-level of the telephone-interference caused by the harmonics which are produced in the A.-C. motor 2 3'.

The principles of my improved rectifier-designs, which have heretofore been described as single-phase systems, may also be applied to polyphase rectifiers, either with or without secondary filters. The application of my secondary filters to two exemplary polyphase-rectifier installations are shown in Figs. 1c and 11.

In Fig. 1G, I have shown the application of my secondary filters to a three-phase rectifier-assembly 21A, MB and 250, which is energized from the zigzag secondary windings ll of a polyphase transformer 9". The primary-winding phases A B 0, of this transformer, are energized from a polyphase supply-line 54 which is in inductive-interference coupling with respect to a communication-system 6. The secondary windings H comprise two separate windings for each phase, as indicated by the letters A, A, B, B", C and C", as will be obvious. The return-circuit 16 for the D.-C. load-device 55 is brought back to the star-point of the zigzag secondary winding H as is usual, as shown, for example, in a Bosch Patent 2,106,826 of February 1, 1938.

In Fig. 10, my secondary filter-circuits W3 and We are connected, in three delta phaseconnections, across the respective secondary terminals HA, lie, and Ho, which energize the three rectifiers 21A, 2L2, and 2lc. During the commutating-periods of the rectifier-tubes 25A and 2 is, the filter 18A is short-circuited and discharges. The other filters tiie and Mic perform in the same way during the commutating-periods for the other pairs of rectifiers.

Fig. 11 illustrates the application of my invention to a double three-phase rectifier-system of a well-known type, such as is shown, for example, in the Atherton Patent 1,979,660 of November 6, 1934. In Fig. 11, the secondary windings of the polyphase transformer 9 are connected in two three-phase stars A, B, C and A, B", C, with the two star-points connected by :an interphase transformer :IT. Three of the :six :rectifiers, marked 2.1 1, 21h and .2 l3, are connected in series with the respective terminals H1, H2, and its of the first ha'lf ofthe second ary windings, while the other three rectii'iers, marked 221, 222, Land 2'23, are connected across the respective terminals H1", .542", and H3" of the second half of secondary windings. The direct-current load till is connected between the common cathode bus 23 and the return-conductor 5-5", which goes .to the midpoint of the interphase transformer IT.

In applying my secondary :iiltcrs to the double three-phase .ectifier-system shown in Fig. 11, three of the filters, marked llh, 92, and tea, are connected .in delta across the anode-leads -l l;1'., Hz, and .113 of the first group of rectifiers 2111 212 and 2.13 which are commutating, one with respect to another. Thus, during the commutation-period for the 'rectiiiers 211 and Zia, the filter l'fii' is shcrt circuited and disc and so on. In mam-lei", three other filters, lii'l, '48s", and 503', are connected in delta across the the three anode-leads H1", H2", and lie of the other three rcctifiers 221., 22:, 223.

In all the embodiments of my invention, the preferred design-constants and limits, governthe choice of the filter-capacitors C and the damping-resistors R, are the same as has been particularly described, in considerable detail, in connection with the application of my invention to single-phase rectifier-powered railway-vehicles. In the case of the .polyphase rectifier- .systems of Figs. and '11, the rectifier-design, as to commutating-angles, substantially undelayed firing", and the ripple-content in the circuit 2616 or 2316, .follow the principles which have been outlined in connection with the single-phase rectifier-applications of invention.

While I have described my invention in detail, in connection with :a single preferred form of embodiment, for illustrative purposes, and While .Ihave cited specific values, and ranges :of values, of the various physical constants which are preferred, :or which might be used, I wish it to be understood that I am not limited, in every respect, and in 'every'possible form of embodiment of my invention, to all of these precise details, as it is obvious that many changes may be made, by way ofthe substitution or equivalents, or the omission or addition of one feature or another, or the Iii-evaluation of certain constant, without departing from certain aspects of the essentialfeatures of my invention. I desire, therefore,

that the appended claims shall be accorded the broadest construction consistent with their language.

I claim as my invention:

1. A rectifier-power transportation equipment comprising the combination, with a single-phase trolley supply-circuit which is subject to telephone-interference considerations, of a self--propelled vehicle comprising: a transformer neans having primary and secondary circuit r-ieam; a direct-current load-means including a series direct-current tractiomrnotor for said vehicle; a rectifier-assembly for energizing said direct-cur rent load-means from said secondary circuitmeans, said rectifier-assembly comprising a plurality of rectifier-means, the output circuit of said rectifier-assembly being substantially free of telephone-interference considerations; connection-means for energizing, said primary circult-means from said supply-circuit, the com- .bined transformer and supply-circuit inductance being suflici'ent to cause the conducting periods of successively operating rectifier-means to overlap for at least about 20 during each halfcycle during maximum load-conditions; and an untuned filter-means, connected in parallel-circuit relation to these'condary circuit-means of the transformer-means and in parallel-circuit relation "to successively operating rectifier-means, and comprising a capacitor and a damping resistance, said capacitor having such capacitance as to be in parallel resonance with the effective secondary-voltage-base inductance of the transformer and supply-circuit combined, at a freenemy in the range between 100 and 900 cycles, during at least some operating-conditions.

2. The invention as define-d in claim 1, charaoterized by said overlap being at least about 40.

3. A rectifier-powered equipment comprising, in combination: an alternating-current supplycircuit which is subject to telephone-interference considerations; a transformer-means having pri. mary and secondary circuit-means; a directcurrent load-means; a rectifier-assembly for energizing said direct-current load-means from said secondary circuit-means, said rectifier-assembly comprising a plurality of rectifier-means, the output circuit of said rectifier-assembly being substantially free of telephone-interference considerations; connection-means for energizins; said primary circuit-means from said supplycircuit, the combined transformer and supplycircuit inductance being sufficient to cause the conducting periods of successively operating rectifier-means to overlap for at least about 20 durii each half-cycle during maximum load-conditions; and an untuned filter-means, connected in parallel-circuit relation to the secondary circuitqneans of the transformer-means and in parallel-circuit relation to successively operating rectifier-means, and comprising a capacitor and a damping resistance, said capacitor having such capacitance as to be in parallel resonance with the effective secondary-voltage-base inductance of the transformer and supply-circuit com bined, at a frequency in the range between and 900 cycles, during at least some operatingconditions, and said damping-resistance having a value between the critical value of dampingresistance on charge and about half of the critical value of Clamping-resistance on discharge.

4. A rectifier-powered equipment comprising, in combination: an alternating-current supplycircuit which is subject to telephone-interference considerations; a transformer-means having primary and secondary circuit-means; a directcurrent load-means; a rectifier-assembly for energizing said direct-current load-means from said secondary circuit-means, said rectifier-assembly comprising" a plurality of rectifier-means, the output circuit of said rectifier-assembly being substantially free of telephone-interference considerations; connection-means for energizing primary circuit-means from said supply-circuit, the c lllblllfid. transformer and supply-cirncluctance being surlicient to cause the con- '5 periods of successively operating recticans to overlap for at least about 20 dureach half-cycle during maximum load-com ditions; and an untuned filter means, connected in p rallel-circuit relation to the secondary circuit-means of the transformer-means and in n l circuit relation to successively operating learner-means, and comprising a capacitor and a damping resistance, said capacitor being large enough to draw a current of at least an ampere,

rent load-means; a rectifier-assembly for energizing said direct-current load-means from said. secondary circuit-means, said rectifier-assembly comprising a plurality of rectifier means, each becoming conductive when its impressed voltage reaches a relatively small value in its conductive polarity; connection-means for energizing said primary circuit-means from said supply-circuit;

an untune-d filter, connected in parallel circuit relation to the secondary circuit-means of the transformer, and comprising a capacitor and a damping resistance; and a tuned filter connected in shunt-circuit relation to at least some portion of the load-means.

7. A rectifier-powered transportation equipment comprising the combination, with a singlephase trolley supply-circuit which is subject to telephone-interference considerations, of a selfpropelled vehicle comprising: a transformer having a primary circuit and a secondary circuit; a series direct-current traction-motor for said vehicle; a rectifier-assembly for energizing said direct-current motor from said secondary circuit, said rectifier-assembly comprising a plurality of rectifier-means, each becoming conductive when its impressed voltage reaches a relatively small value in its conductive polarity; and connectionmeans for energizing said primary circuit from said supply-circuit, the combined transformer and supply-circuit inductance being sufiicient to cause the conducting periods of successively operating rectifier-means to overlap for at least about during each half-cycle during maximum short-time load-conditions.

8. The invention as defined in claim 7, characterized by said rectifier-assembly comprising at least two single-phase rectifying devices of a type which becomes substantially non-conducting, after a conducting period, only in response to a current-decrease to substantially zero.

9. The invention as defined in claim '7, characterized by said overlap being at least about 40.

10. A rectifier-powered transportation-equipment comprising the combination, with a single phase trolley supply-circuit which is subject to telephone-interference considerations, of a selfpropelled vehicle comprising: a transformer having a primary circuit and a secondary circuit; a series direct-current traction-motor for said vehicle; a rectifier-assembly for energizing said direct-current motor from said secondary circuit, the output-circuit of said rectifier-assembly being substantially free of telephone-interference considerations and having a total inductance which is sufficiently small to permit at least a 20% ripple to be present in some part of said circuit during maximum short-time load-conditions, said rectifier-assembly comprising a plurality of rectifier-means, each becoming conductive when its impressed voltage reaches a relatively small value in its conductive polarity; and connection-means for energizing said primary circuit from said supply-circuit, the combined transformer and supply-circuit inductance being sufficient to cause the conducting periods of successively operating rectifier-means to overlap for at least about 20 during each half-cycle during maximum short-time load-conditions.

11. The invention as defined in claim 10, characterized by said rectifier-assembly comprising at least two single-phase rectifying devices of a type which becomes substantially non-conducting, after a conducting period, only in response to a current-decrease to substantially zero.

12. The invention as define-d in claim 10, characterized by said overlap being at least about 40.

13. A rectifier-powered transportation-equipment comprising the combination, with a singlephase trolley supply-circuit which is subject to telephone-interference considerations, of a selfpropelled vehicle including: a transformer having a primary circuit and a secondary circuit; a series direct-current traction-motor for said vehicle; a rectifier-assembly for energizing said direct-current motor from said secondary circuit, the output-circuit of said rectifier-assembly being substantially free of telephone-interference considerations and having a total inductance which is suificiently small to permit at least a 25% ripple to be present in some part of said circuit during maximum short-time load-conditions, said rectifier-assembly comprising a plurality of ectifier-means, each becoming conductive when its impressed voltage reaches a relatively small value in its conductive polarity; and connection-means for energizing said primary circuit from said supply-circuit, the combined transformer and supply-circuit inductance being suificient to cause the conducting periods of successively operating rectifier-means to overlap for at least about 20 during each halfcycle during maximum short-time load-conditions.

14. The invention is defined in claim 13, characterized by said rectifier-assembly comprising at least two single-phase rectifying devices of a type which becomes substantially non-conducting, after a conducting period, only in response to a current-decrease to substantially zero.

15. The invention as defined in claim 13, in combination with a harmonic-reducing filter connected in shunt-circuit relation to at least some portion of the motor-means, said filter having a greater shunting-eifect on harmonic currents than on currents of the fundamental frequency.

16. The invention as defined in claim 13, in combination with a second-harmonic filter connected in shunt-circuit relation to at least some portion of the motor-means.

1'7. The invention as defined in claim 13, characterized by said overlap being at least about 40.

18. In combination: a single-phase supply-circuit which is subject to telephone-interference considerations; a transformer having a primary circuit and a secondary circuit; connectionmeans for energizing said primary circuit from said supply-circuit; a load-means energized from said secondary circuit and being of a type tending to draw objectionable supply-circuit harmonies in the telephone-interference frequencyrange, said secondary circuit and load-means be ing otherwise substantially free or" telephoneinterference considerations; and an untuned filter; connected in parallel-circuit relation to the secondary circuit of the transformer, and comprising a capacitor and a damping resistance, said capacitor having such capacitance as to be in parallel resonance with the effective secondaryvoltage base inductance of the transformer and supply-circuit combined, at a frequency in the range between 100 and 900 cycles, during at least some operating-conditions, and said damping-resistance having a value between the critical value of damping-resistance on charge and about half of the critical value of damping-resistance on discharge.

19. A rectifier-powered transportation-equipment comprising the combination, with a singlephase trolley supply-circuit which is subject to telephone-interference considerations, of a selfpropelled vehicle comprising: a transformer having a primary circuit and a secondary circuit; a series direct-current traction-motor for said vehicle; a rectifier-assembly for energizing said direct-current motor from said secondary circuit, said rectifier-assembly comprising a plurality of rectifier-means, each becoming conductive when its impressed voltage reaches a relatively small value in its conductive polarity; connectionmeans for energizing said primary circuit from said supply-circuit, the combined transformer and supply-circuit inductance being sufiicient to cause the conducting periods of successively operating rectifier-means to overlap for at least about 20 during each half-cycle during maximum short-time load-conditions; and an untuned filter, connected in parallel-circuit relation to the secondary circuit of the transformer, and comprising a capacitor and a damping resistance.

20. The invention as defined in claim 19, characterized by said overlay being at least about 40.

21. A rectifier-powered transportation-equipment comprising the combination, with a singlephase trolley supply-circuit which is subject to telephone-interference considerations, of a selfpropelled vehicle comprising: a transformer having a primary circuit and a secondary circuit; a series direct-current traction-motor for said vehicle; a rectifier-assembly for energizing said direct-current motor from said secondary circuit, said rectifier-assembly comprising a plurality of rectifier-means, each becoming conductive when its impressed voltage reaches a relatively small value in its conductive polarity; connection-means for energizing said primary circuit from said supply-circuit; an untuned filter, connected in parallel-circuit relation to the secondary circuit of the transformer, and comprising a capacitor and a damping resistance; and a harmonic-reducing filter connected in shunt-circuit relation to at least some portion of the motor-means, said filter having a greater shunting-effect on harmonic currents than on currents of the fundamental frequency.

22. A rectifier-powered equipment, comprising, in combination, a direct-current load-means, a rectifier-assembly comprising at least two singlephase rectifying devices of a type which becomes substantially non-conducting, after a conducting period, only in response to a current-decrease to substantially zero; a pair of single-phase power-supply leads, connected to said two rectifying devices, respectively, at rectifying-device terminals of one polarity; and circuit-means for connecting the other terminals of said two rectifying devices to one terminal of said direct-current load-means; the supply-circuit inductance, up to said first-mentioned terminals of said two rectifying devices, being suflicient to cause the conducting periods of said two rectifying devices to overlap for at least about 20 during each half-cycle during maximum short-time load-conditions, in combination with an untuned parallel-resonant filter, connected in parallel-circuit relation across said power-supply leads, with at least a substantial part of said supply-circuit inductance on the power-supply side of said parallel-filter connection, said parallel-resonant filter comprising a capacitor and a damping resistance, said capacitor having such capacitance as to be in parallel resonance with the supplycircuit inductance as seen from the terminals of said parallel-resonant filter, at a frequency in the range between and 900 cycles, during at least some operating-conditions.

23. The invention as defined in claim 22, characterized by said overlap being at least about 40.

24. The invention as defined in claim 22, characterized by the direct-current load-circuit having a total inductance which is sufficiently small to permit at least a 20% ripple to be present in some part of said direct-current load-circuit during maximum short-time load-conditions.

25. The invention as defined in claim 22, characterized by the direct-current load-circuit having a total inductance which is suificiently small to permit at least a 25% ripple to be present in some part of said direct-current load-circuit during maximum short-time load-conditions.

LLOYD J. HIBBARD.

References Cited in the file of this patent UNITED STATES PATENTS Number Name Date 1,699,153 Mittag Jan. 15, 1929 1,718,515 Alexanderson June 25, 1929 1,755,859 Baker Apr. 22, 1930 1,758,680 Andre May 13, 1930 2,008,519 Smith July 16, 1935 FOREIGN PATENTS Number Country Date 218,285 Switzerland Mar. 16, 1942 441,491 Germany Mar. 10, 1927

Images