US730215A - Method of shaping rotary fluxes and operating transformers. - Google Patents

Description

No. 730,215. I l I PATENTED JUNE 2, 1903. v

M. LEBLANG.

METHOD OF SHAPING ROTARY FLUXES AND OPERATING TRANSFORMERS.

' APPLICATION nun APR. 5, 1902. 1V0 MODEL. a sums-sum 1.

% W A Maker/(Bel eZZOfl I PATENTED JUNE 2, 1903. M. LEBLANG. METHOD OF SHAPING ROTARY FLUXES AND OPERATING TRANSFORMERS.

APPLICATION FILED APR. 5, 1902.

N0 MODEL.

6 SHEETS-SHEET 2.

BJZM%47 ml: Noam: PETERS cu. wuovouma. wnsnmmomp c. v

R No. 7305215.

PATENTED JUNE 2, 1903. M. LBBLANG. METHOD OF SHAPING ROTARY FLUXES AND OPERATING TRANSFORMERS.

APPLIOATION I ILBD APR. 5, 1902. V

' e SHEETS-SHEET a.

N0 MODEL.

Tm: roams Prrzns 00., wovouma. wAsmncmu. 0. c.

PATENTED JUNE 2, 1903..

M. LEBLANG. METHOD OF SHAPING ROTARY FLUXES AND OPERATING TRANSFORMERS.

APPLICATION FILED APR. 5, 1902.

6 SHEETS-SHEET 4 H0 MODEL.

.llllllllll-l 7 r r 1 I a I I I I i a @0677757; JWO Lor /cal 6 Z Z0770 w: mums wnsns co. PHomuTHu wAsnmcYoN, 11cv No. 730.215. PATENTED JUNE 2,1903.

M. LEBLANO. METHOD OF SHAPING ROTARY FLUXES AND OPERATING TRANSFORMERS.

APPLICATION FILED APR. 5, 1902.

N0 MODEL. 6 SHEETS-SHEET 6.

Inmi

IIIIIIH lllllll wurlalealafic 3 0 9, 1 .2 E N U ..J D E T N E T A P M. LEBLANG METHOD 0? SHAPING ROTARY FLUXES AND OPERATING TRANSFORMERS.

APPLICATION FILED APR. 5, 1902.

cm H H H S S T E E H S B B0 MODEL.

m 52 ay,

a mm Paws co. vnmoummw swmmm,

NO. 730,215. Patented June 2, 1903.

NITED. STATES PATENT- OFFICE.

MAURICE'LEBLANC, OF PARIS, FRANCE, ASSIGNOR TO GEORGE \VESTING- HOUSE, OF PITTSBURG, PENNSYLVANIA.

METHOD OF SHAPING ROTARY FLUXES AND OPERATING TRANSFORMERS.

SPECIFICATION forming part of Letters Patent N 0. 730,215, dated June 2, 1903.

Original application filed January 4, 1901, Serial No. 42,038. Divided and this appication filed April 5,1902. Serial No. 101,499. (N model.)

To aZZ whom it may concern.- seek to determine the character of the flux, Be it known that I, MAURICE LEBLANC, a which is represented by the first term of this citizen of the Republic of France, and a resiseries. In all that precedes it is to be underdent of Paris, France, have invented certain stood that I have been considering the flux in new and useful Improvements in Methods of the ring at a given instant of time. I have as- Shaping Rotary Fluxes and of Operating sumed that the magnetic flux which exists in 'lransformers,More Particularly of the Recti the ring at a given instantof time has befying Type, of which the following is a specicome, as it were, frozen in the ring, so that fication. we may look at it and examine it. I have IO Let us consider an iron core, preferably of not in that which precedes been considering the ring type, aboutwhich circuits arewound. such variations in the flux as are brought Let us,furthermore,eonsiderthatwe pass into about in successive intervals of time, but

, these circuits currents of any kindsuch, for have merely been considering the variations instance, as monophase or polyphase curof magnetic flux in the core or ring as you [5 rents-or, if you please, let us pass therein pass from point to point of the ring in a given continuous currents in such a manner that instant of time. To understand, then, what the leads through which the continuous curkind of aflux is represented bythe first term of rents are fed to the circuits about the ring the above developmentnamely, by the term move around the circumference of the ring. cp sin. ('w/ '1) ve must start with the angle 20 In each of these cases there will be developed to from an initial value /3,, and let it increase, in the iron of the ring a rotating magnetic going counter-clockwise through three hunflux which, generally speaking, will be comdred and sixty degrees of arc. \Vhen the pound in character. angle 10 is equal to [3 the sine of the angle To understand what I mean by the com-1 w- /i, is zero, which means that at the point 25 pound character of the liux,let us consider of the ring represented by [3 there are no the flux in the ring at a given instant of time, lines of force traversing a radial cross-secjust as if this condition were permanent. Let tion of the ring. As the angle w increases us, furthermore, consider a section .made beyond [5,, sin. (iv-J5 increases according to though the ring by a radial plane perpendicthe sine law until w reaches 90+/J,. At this 0 ular to the plane of the ring, which radial pointthe sine of the angle isamaximum. WVe 8o plane makes an angle 10 with the vertical di therefore find that the number of lines ameter of the ring. Let us designate by @the of force constituting the flux represented byintensity of the magnetic flux which trav- 4), sin. (w [5 is zero at one cross-section of erses the section of the ring above referred the ring and that this number increases grad- 5 to. \Ve may consider the flux as a funcually in amount and reaches a maximum at a tion of the angle 10 and may write 'flw); cross-section displaced from the'zero cross but the function 4 attains the same value sectionbyninetydegrees. Stillmoving couneach time the angle to is increased by 271. ter-clockwise around the ring, the number of lVe can therefore represent the function lines of force in a given cross-section gradu- 0 flw) by a development according to the Foually diminishes until we reach the cross-sec- 9o rier theorem by designating by cp (p e1 contion one hundred and eighty degrees removed stants representing fluxes and by e, /J [i from the zero cross-section, in which the numconstants representing differences of phase, ber of lines of force is again zero. Moving as follows: on around the ring in a counter-clockwise 45 j p 3 direction the number of lines of force 1n any 5 l) f 6 radial cross-section of the ring again increase; q but this time the lines of force run in the opl lVe assume the constant term of the develposite direction. Their arithmetical number opment to be zero, which is always the case reaches a maximum in that cross-section of 50 in the application of my invention. Let us the ring which is two hundred and seventy 10o degrees removed from the original zero crosssection, after which, their direction still remaining the same, they diminish in number until we have zerolinesof force, when the starting-point is reachedthat is, considering the ring divided into four quadrants, the lines of force in the first and second quadrant run in one direction in the ring and in the third and fourth quadrant they run in the opposite direction in the ring. Considering the number of lines of force in any radial cross-see tion of the ring, they are zero at two diametrically opposite points and are at maximum at two other diametrically opposite points displaced from the first points by ninety degrees. Between the zero and maximum points the number of lines of force in any radial crosssection increases gradually in accordance "with the sine law.

The flux which I have just described and which is represented by (p, sin. (w [5 I de fine as a simple or pure sinusoidal flux of two branches; the flux which is represented by 9 1,, sin. (a 10-51.) as a flux of 2 branches.

Now the object of my invention is to take any compound magnetic fiux and to obliterate therefrom any given element or elements of the flux which I select. Thus it is particularly useful in a variety of electrical apparatuses to have a pure sine flux of two branches, and such flux only. In fact, a uniformly-rotating sine flux of two branches, as above defined, represents the pure or ideal form of a rotating field of constant intensity. In accordance with my invention I am enabled to screen. out or obliterate all of the sine fluxes of a greater number of branches than two, given in the above expression, and to leave merely that element of the flux which I have called a two-branch fiux and which produces the uniformly-rotating constant magnetic field. I leave, in other words, the flux which is represented by the first element of the development above given, which produces the constant rotating field, and screen or obliterate the fluxes corresponding to each and all of the other elements of the development above given, being the terms of the higher orders in the series, and which would tend to disturb the constancy of the field. Put in different words, I take the compound fi ux which exists in the apparatus and I screen out all of the disturbing elements of that flux and leave only the pure sine flux of two branches, which produces a strictly-constant and uniformly-rotating magnetic field. These objects I effect by a system of compensating windings, which will hereinafter be fully described.

Manifcstly, the general principle of the invention as I have thus explained it has a variety of applications. Thus, for instance, I can apply it to a polyphase transformer in which polyphase currents of a given order are transformed into polyphase currents of the same or of any other given order. In accordance with my invention. I would in such transformer obliterate any magnetic flux 01" electromotive force generated therein which is not of the desired phase. Thus, for instance, if I should transform a biphase current into a polyphase current of a larger number of phases the successive electromotive forces of the secondary polyphase system may not, in the absence of my invention, be symmetrical or equidistant in phase, and. thereby the secondary system will not be balanced; but by the use of my compensating windings the system is forced to be balanced.

But what is probably the most important ap-- plication of my invention is found in its application to rectifying-transformersthat is to say, transformers in which a monophase or polyphase current is transformed into a continuous current by connecting the secondary windings to the segments of a commutator or in which, Vice versa, a continuous current can be transformed into a monophase or polyphase current by feeding the continuous current through the commutator to what were before the secondary circuits and by tapping monophase or polyphase currents from What were before the primary circuits.

In rectifying -tranformers as heretofore constructed the secondary coils are usually entirely embedded in iron, and the electromotive forces of self-induction of commutation are enormous. In fact, the electromotive forces of self-induction, practically speaking, make commutation impossible. By the use of my compensating windings, however, regardless of the sudden change of current during commutation by the short-circuiting of the coils, the electromotive forces induced in the successive coils correspond to a symmetrical polyphase system. It is thus possible to commutate a closed magnetic circuit-transformer without meeting any excessive eleetromotive force of self-induction. I have therefore chosen such rectifying-transformers in order to illustrate the principle of my invention.

In the drawings, Figure 1 shows a vertical cross-section of a rectifying-transformer as it is made in accordance with my invention. Fig. 2 shows a front elevation thereof. Fig. 53 shows a rear elevation of the same. Fig. 4 shows a section of the ring-core. Fig. 5 shows a diagram of the windings. Fig. (5 shows a detail of a mode of connectingthe secondary circuits. Figs. 7, 8, and 9 show details of the centrifugal sliding coinimitator-brush and its holder. Figs. 10,11, and 13 show details of a modified form of the centrifugal sliding commutator-brush and its holder. Fig. 12 shows a diagram of the system.

The rectifying-transforiner is composed of a transformer-ring 1, which is generally fixed and which is shown separately in Fig. 4. Added to this is an annular commutator 2, shown-to the left of Fig. l, and which in this case is stationary. \Vithin this commutator is rotated a spider a, carrying brushes S,whose outer faces make and maintain contact with the inside surface is secured to a shaft 25, which has its bearings 1 in the frame of the transformer and has keyed to it the spider 4, to which the brush-carriers are secured. The field-magnets 6 of the small motor are mounted on a common yoke (31,

which is rotatably mounted within the annularbearing 62, which is shown as in one piece with the supportin g-frame of the transformer.

From the yoke 61 extends a toothed sector 7, which meshes with a pinion orsector 8, secured to a shaft 9, extending to the commutator side of the machine, where it may be formed with a squared end to receive a crank by which the shaft may be turned. In this manner the field-magnets of the small biphase motor are made adjustable circularly. To any shift given to the field corresponds a lead or lag in the set of the brushes on-the commutator 2 of the transformer. The same result is thus obtained as is ordinarily obtained in the case of a rotary commutator by shifting the stationary brushes with respect thereto.

Considering more in detail the spider l, which carries the brushes 3, it is seen that the number of arms in this spider are made even. In the case shown theirnumber is eight.

with each other and with a ring 10 in Fig. 2,

and the even-numbered arms are electrically connected with each other and with the ring 11 in Fig. 1. Upon the rings 10 and 11, which have just been described as electrically connected with the odd and even numbered arms of the spider, bear fixed brushes 12 12-5, and these brushes in turn are connected to binding-posts marked plus and minus respectively, in Fig. 2. These two bindingposts are the terminals of the continuous-current circuit of the rectifying-transformer. On the other hand, the number of terminals of the alternating-current circuit which enters orproceeds from the rectifying-transformer varies according to whether the apparatus is to transform or yield monophase, biphase, or multiphase currents; but in all cases the alternating-circuit terminals will be displaced around the circumference in a symmetrical manner. In the case represented in Fig. 2 there are four alternating-circuit terminals 1- 15, 16, and 17, which means that the apparatus is to utilize or yield biphasc currents.

The apparatus thus far described is seen to consist, in the main, of a transformer-frame having terminals for the alternatin current,

a fixed hollow annular commutator, on the inside of which sets of brushes revolve which take oif the continuous current, and of a small biphase motor, which serves to 'drive these brushes' The field of the biphase motoris shiftable to produce the effect ofa shift of The odd-numbered arms are electrically conn ected the brushes bearing on the commutator for lead-or lag. The field of the biphase synchronous motor is supplied with direct currents. The armature of the biphase motor receives its alternating driving-current through the brushes 22 23 24, which are shown in Fig. 3 and which bear on the rings marked with the same numbers shown to the right in Fig. 1. The brushes 22 23 24 are in turn connected to a source of biphase current in a manner which will be described later on. The number of pairs of poles of the biphase motor is equalto the number tator-segments to behereinafter described. This number may be greater than the number of poles of the transformer, whichinthe present case is taken as two. To the field-pole 6 of the biphase motor I secure a magnetic shield 63, which maybe of the squirrel-cage type shown, the construction of which is more fully illustrated in the prior patent to Hutin and Leblanc, o. 529,272, of November 13, 1894. The functions of the magnetic shield 63 will appear farther on.

A crank 25 (indicated in Fig. lines) may be used for starting nous motor by hand. v

I come now to describe the ring-core of the transformer and the windings which it carries. To this end particular attention is called to Figs. 4 and 5. The body of the transformer; ring 1 is composed of a series of superposed segmental iron laminze 27. In the case shown these segments are quadrants, and together they constitute a closed ring. Inside this ring there is a toothed ring 26,which is also built up of superposed iron laminae. In the case shown the laminae 26 are full circles and are pro- 12 in dotted the synchrovided with teeth extending radially outward.

Upon the segments of the outer iron ring 27 the bobbins ofthe transformer are strung, whereupon the segments may be secured together to form the magnetically-closed core. Th ereupon the toothed ring 26 is slipped inside the ring 27 in such a manner that the bobbins upon the ring 27 lie in the slots between the teeth of the ring 26. .The lines of force which constitute the magnetic fluxes which may be generated in the iron structure 26 27 may be represented by the broken lines in Fig. 4. It will be seen that these lines of force lie .lmost wholly within the iron-and that the air-gaps which they have to traverse are of small width.

Each bobbin which is strung on the transformer-core is composed of four separate coils. The inner coils -P in Fig. 4 I will treat as the primary coils of the transformer.- The coils S of Fig. 4 will be treatedas the secondary coilsof the transformer. The two sets of coils it and n are what will be hereinafter called the sine and cosine compensating coils, respectively, of the transformer.

Examining Fig. 5, the manner of intercon of groups of commu- IIO the secondary coils, the ratio of these num ber of turns on the primary and secondary coils giving the ratio of transformation, as is well understood.

There are shown in Fig. twenty commutator segments each. twenty segments in each group are numbered consecutively 1,2, 3, to 20. The four segments 1 are electrically interconnected, as shown, by the Wires on the outside of the commuta tor in Fig 5. So the four segments 2 of the 5 four groups of four commutator groups are electrically in te'rconnected. In general like numbered se ments of the four commutator groups are electrically interconnected.

The secondary coils S S to S are connected in series, as shown in Fig. 5. One end of the secondary coil S is tapped to one of the four commutator-segments l and is thereby electrically connected to each of the four commutator-segments l. The corresponding end of the secondary coil 8,, which is'the other end "of the secondary coil 8,, is electrically tapped to one of the four commutator-segments 2 and is thereby electrically connected to each of -the four commutator-segments 2, and so on.

The primary coils P P to P are also connected in series. The terminals of the alternating-current leads are connected to the primary winding thus formed at points equally displaced around the circumference. In the case shown'the terminals 14, 15, 16, and 17 of a biphase circuit are connected at the points shown in Fig. 5, which are ninety degrees removed from each other. If a monophase current is to befed to the primary coils, the two terminals of the monophase circuit may be connected to the points 15 and 17, which are removed from each other by one hundred and eighty degrees.

In all that precedes I have assumed a Gramme winding; but manifestly any other suitable type of winding may be employed.

So far as above described it will be seen that I feed a monophase or polyphase current to the primary winding P, and thereby generate alternating currents in the secondary windings S, which are tapped to the commutatorsegments, from which I may take a continuous current by means of the rotary brushes which bear against the commutator-segments as already described. Should I desire, 110W- ever, to transform alternating current of any number of phases into alternating currents of the same or a different number of phases, it is evident that I can dispense with the com mutator and brushes and that I can feed alternating monophase or polyphase currents to the primary windings and take alternating The r currents from the secondary winding by sim-' ply tapping these secondary windings at points symmetrically displaced around the circumference of the ring. Thus, for instance, my invention is peculiarly adapted for trans' forming polyphase currents of any number of phases into polyphase currents of a very large number of phases, and this can be done by tapping the secondary circuit directly, thus dispensing with a commutator. It will also be evident without further description that the primary and secondary circuits are interchangeable and that the transformer is reversible. Nor does it seem'necessary to mention that I can use a revolving commutator and stationary brushes instead of the stationary commutator and the revolving brushes, which I have shown. So too I may substitute for the type of motor which I have shown any other suitable type of motor.

I now come to a detailed description of the twosets of compensating coils n n and their interconnection, which are diagrammatically indicated inside the circle formed by the pri mary coils P of Fig. 5. The arrangement of these coils and their operation constitute the principal feature of the present invention.

The two compensating circuits, each ofwhich is compound-Te have now to consider the two compensating circuits, each of which is compound. The coils composing these windings as I have shown them bear a certain resemblance to the coils of a Gramme ringwinding, but this is not essential.

We assume that the number of slots in the iron ring is represented by 2K, where K is an even integer which is made as large as the requirement of construction permits. We number these slots in order from 1 to 2K in the same direction around the ring, assuming for convenience that this direction is counter-clockwise. The slots numbered from 1 to lie in the first quadrant, those from %+1 to K lie in the second quadrant. Those from K+1 to lie in the third quadrant,

and those from i +1 to 2 K lie in the fourth quadrant.

I have already stated that in each slot 01: there is a primary coil Pr and a secondary coil Sac. These are the ordinary primaryand secondary coils of the transformer. In addition to these two coils, however, I place in each slot as two other coils, one of which is to be included. in the sine compound compensating circuit, and the other of which is to be included in the cosine compound compensating circuit. Designating by 11. the number of convolutions or turns of the coil in the slot which is to be included in the sine compensating circuit and by '21 the number of convolutions or turns of the coil in slot 00 which is to be included in the cosine compenwhere "u is some arbitrary constant. I may for convenience speak of the coils n,- as the sine coils and of the coils a, as the cosine coils.

The manner of interconnecting the sine coils and the cosine coils is diagrammatically The sine coils are indicated by full lines. The cosine'coils are indicated by broken lines. The sine coil and cosine coil in the same slot are indicated by a full line and a broken'line, respectively, lying adjacent and parallel to each other. The first quadrant is taken as running from the upper end of the radius, extending vertically upward and proceeding counterclockwise to the horizontal radius extending to the left. It is seen that each sine coil in the first quadrant is connected to the sinecoil in the second quadrant, which is removed therefrom by ninety degrees. So each sine coil in the fourth quadrant is connected to the sine coil in the third quadrant which is removed therefrom by ninety degrees. Similar remarks apply to the cosine coils. Thus, for instance, one end of each sine coil indicated in Fig. 5.

of order so in the first quadrant is connected to the corresponding end of the sine coil in the second quadrant which is removed therefrom by ninety degrees-that is to say, the outer end, for instance, of a sine coil 00 in the first quadrant is connected to the outer K end of the sine coil :0 111 the second quadrant, being in this case the sine coil of order 50+ 5, since 2K is equal to twenty. So too one end of each sine coil in the fourth quadrant is connected to the corresponding end of the sine coil in the third quadrant, which is removed therefrom by ninety degrees that is, the outer end, say, of a sine coil. of ordera; in the fourth quadrant isconnected to the Ix outer wire of the sine coil'of order a; in the third quadrant. The free ends of the sine coils in the first and fourth quadrant are now connected to the ring I, as shown in Fig. 5, and the free ends of the sine coils in the second and third quadrants are connected to the ring II, as shown in Fig. 5. It is therefore seen that the sine compensating circuit, in which are included all the sine coils, is composed o f a n etwork of K-parallel bran ches,each branch containing two sine coils in series, the two coils in any given branch having their It order displaced by or by ninety degrees,

and the two sets of ends of these K-parallel branches being connected to the rings I and II, respectively. Similar remarks apply to the cosine coils and to the secondcompensating circuit, which includes them. One end of each cosine coil of order so in the first quadrant is connected to the corresponding end of the cosine coil in the second quadrant which is removed therefrom by ninety degrees. So to one end of each cosine coil in the fourth quadrant isconnected to the corresponding end of the cosine coils in the third quadrant therefrom by ninety dewhich is removed grees. The free ends of the cosine coils in the first and fourth quadrants are connected to the ring IIIof Fig. 5, and the free ends of the cosine coils inthe second and third quadrants are connected to the ring IV of Fig. 5. It is therefore seen that the cosine compensating circuit, inwhich are included all the cosine coils, is composed of a network of K-parallel branches, each branch containing two cosine coils in series, the two coils in any given I branch having their order displaced by or by ninety degrees, and the two sets of ends of these K-parallcl branches being connected to the rings III and IV, respectively. .The rings I II III IV are connected to the ter1ni nals 18 10 20 21. (Shown in Figs. 2 and 3.) Let usonce for all mark the circumferential points on the. soft-iron ring by. the angles which radii drawn through these points make with the verticalradius of the ring. Thus the vertical or top point on the ring, as seen in Fig. 5, for instance, will be marked 0, the point on the horizontal diameter to. the left will be marked the point on the vertical radius extending downward will be marked 180, and the point on the horizontal diameter to the right will be marked 270. Let us now consider that the magnetic flux in the soft-iron ring at any given instant of time is of such a character that the intensity of the flux at any radial cross-section of the ring, or, what is the same thing,

the number of lines of force which traverse any cross-section of the ring made by aradial plane perpendicular to the plane of the ring, may be represented as the sine of the angle, which the radial plane under consideration makes with any given or fixed radius of the ring, which is selected as the base or. startingpoint at the. given instant of time. To fix ideas, we assume that this base at the given instant of time under considerationis marked by three hundred and forty degrees, or, what is the same thing, by minus twenty degrees, The flux which we will at the given instant of time suppose to exist .in the iron ring will thereafter be zero at the point marked by minus twenty degrees on the ring. It will grow more and more dense as we pass through hundred and sixty degrees on the ring. Passing along the ring in the same .directionthe.

flux will now flow in the opposite direction, but increase in arithmetical magnitude until it reaches its negative maximum at two hundred and fifty degrees on the ring, whence its arithmetical magnitude will again diminish until at the angle of three hundred and forty degrees, or minus twenty degrees, the flux is again zero, and the rate of increase and decrease of the flux as you pass around the ring will follow the simple sine law. We now assume that this simple sine flux, which at the given instant of time under consideration is of zero value at minus twenty degrees on the ring, is rotated at a uniform velocity through an infinitely small angle. The value of the flux at each given point on the ring is, as I have before stated, proportional to the sine of the angle or are which runs from the point underconsideration to the point chosen as the base, which in this case is minus twenty degrees and which in the general case we can take as minus a degrees of arc. The electromotive force which will be generated by the rotation of this flux at this given instant of time in the sine coil found at the point of the ring under consideration will be proportional to the product of the number of turns of wire in the sine coil under consideration and to the variation of the number of lines of force which pass through the ring. As the number of lines of force at any point of the ring is measured by the sine of the are which separates that point from minus twenty degrees or minus (4 degrees, the variation of this num ber of lines of force will be proportional to the cosine of the angle which separates the point under consideration from the base-line, for it is known to any student of the calculus that the differential coefficient of the sine of an angle with reference to the angle is equal to the cosine of the angle.

lVithout here giving the calculation it can be shown that the electromotive force which is generated in any one of the K -parallel branches of the compensating sine circuit is,

independent of the quantity ac. This means that at any given instant of time the electromotive forces generated in each one of the K- parallel branches of the compensating sine circuit by the rotation of the simple sine flux are equal in magnitude and phase. The magnitude of the electro'motive force generated at any given instant of time in each and all of the K-parallel branches of the compensating sine circuit is, in fact, proportional to the sine of the angle which the Zero-point of the sine flux at the given instant of time makes with the point on the ring which we have designated as zero and from which the number of turns of wire in our sine coils are determined. It is thus seen that the rotation of this sine flux produces at one given instant of time zero electromotive force in all of the K- parallel branches of the compensating sine circuit. At any other instant of time there will exist an equal electromot'ive force in each of these parallel branches and in the same direction in all, which electromotive force is measured by the sine of a uniformly-increasing an gle-na1nely, the an glewhieh represents the velocity of rotation of the sine flux. \Ve can prove the same thing for the cosine coils comprised in the compensating cosine circuit. lVe [ind that the eleetromotive forces in each of the K-parallel branches of the compensating cosine circuits generated by the rotation of the pure sine flux are equal and in the same direction at any given instant of time and that the magnitude of the electromotive forces in any or all of the K branches of the compensating cosine circuit will be represented by the cosine of the angle which measures the rotation of the sine flux from any given startingpoint, the starting-point being the same as that chosen when considering the compensating sine circuit.

Let us now assume that the rings I and II of Fig. 5 are not connected to each other and that the rings III and IV of Fig. 5 are not connected to each other. It is plain from what has been said that the rotation of a simple sine flux in the iron core will generate no currents of any kind either in the compensating sine circuit or in the compensating cosine circuit. There will simply be generated a series of equal and equally-directed electromotive forces in the parallel branches of the compensating sine circuit, which means that no current will flow from anyone of these branches to any other of these branches. A similar remark applies to the compensating cosine circuit. Should we, however, connect the rings I and II to two terminals of my biphase motor before described and the rings III and IV to the other two terminals of this biphase motor, it will be seen that sine and cosine currents, or, what is the same thing, two quarterphase currents, will besupplied to the motor; but as each one of the K branches of the compensating sine circuits has an equal electromotive force at each instant there will still be no flow of current from any one of the K compensating sine branches to any of the other K compensating sine branches and the balance or equilibrium of the system will not be disturbed. A similar remark applies to the cosine coils. Such connection of the rings I and II and of the rings III and IV to the terminals of the bi phase motor will merely mean that the energy of driving such motor would be taken from the compensating-coils without in any way interfering with the action of these coils so far. as the rest of the system is concerned. I may therefore drive a biphase motor, when I need such a motor in my system, by c011- necting the terminals I, II, III, and IV to the terminals of the motor, or I may drive a biphase motor from any other independent source of current.

It is well understood in matters of this kind that when we speak of giving a coil a certain number of turns corresponding to the sine of an angle we give zero turns to the coil when the angle is zero and that we wind the coils in opposite directions when the sine becomes negative. It is unnecessary for me to enter into any minute details as to the direction of windings of the various sine coils and of the connection of these sine coils in pairs, as this will all be evident to any electrician who understands the above and who sees that the result aimed at is an electromotive force in each of the K branches of the compensating sine circuit which shall be equal in magnitude and phase at any given instant of time. The corresponding remark applies to the cosine coils.

Assuming now that there is I no external connection between the terminals 1 and II and between the terminals III and IV, it will be manifest that the uniform rotation of what we have termed a simple sine flux in the iron ring, whether this rotation be clockwise or counter-cloclnvise, will cause no currents of any kind either in the compensating sine branches or in the compensating cosine branches. Electromotive forces will be generated in the various parallel branches of these compensating circuits, but these will all be equal in magnitude and phase, so that no current will flow from one branch to another. Therefore it follows that the uniform rotation of a simple sine flux or a pure sine flux in the iron ring will have no eilfect on the compensating circuits and the compensating circuits will have no effect on the uniformlyrotating sine flux. Each will exist precisely as if the other were not present. The compensating circuits will in no way tend to deform or distort the uniformly-rotating sine flux. On the other hand, it can be shown that any other type of flux than the uniformly-rotating simple sine fluxsuch, for instance, as a flux of Zn branches, as before definedwill cause the generation of electro motive forces in the compensatin g branches which will not be equally directed, but will be such that to any electromotive force in one branch of a compensating circuit will correspond an electromotive force in the opposite direction in another branch of the same compensating circuit, so that each branch of a compensating circuit willbe shortcircuited by another branch in the same compensating circuit and current will flow from one to the other. This may be stated concisely in this fashion: For a simple sinusoidal flux of two branches the compensating branches act as open circuits, while for a flux of more than two branches the compensating branches short-circuit each other. Putting this in another way, it means that the compensating circuits will act as magnetic screens for all types of fluxes in the iron ring except the simple sine flux and will squeeze out and destroy these fluxes; but the simple sine flux will be allowed to move in the ring just as if the compensating circuits were not present.

From what has been said above it will be clear that I take a rotating flux of any character-that is, a compound flux or one composed of a number of separate dephased components or of a number of components of different periodsand that I shape this flux by cutting therefrom one or more preselected components. More specifically stated, I take a compound rotating flux composed of a number of components of different periods and I shape the flux by obliterating all but the pure sine flux of two branches, which cor-- responds to a uniformly rotating constant magnetic field. It follows therefrom that the intensity of this two-branched sinusoidal rotating flux will depend only on the effective Voltage developed between the terminals of the primary circuit or, in case of a rectifying transformer, on the voltage of the continuous current measured between the brushes. The intensity of this two-branch flux will be independent of the intensity of the currents which traverse the primary circuits. The result may be compared to that which would exist in the ordinary type of transformer for transforming monophasc alternating currents into monophase alterand secondary nating currents when such transformer is operating under ideal conditions. Generally speaking, the ratio of the electromotive force in the secondary to the electromotive force in the primary circuit in such transformer is determined by the ratio of the number of turns of these two circuits; but this is only true under ideal conditions. lVhen the load comes on and the currents in the transformer increase, the amount of magnetic leakage in the transformer varies, the normal flux is distorted, and the ratio of transformation no longer remains the same and is no longer solely determined by the number of turns in the primary and secondary circuits, but varies with the load.

My transformer when the two-branch sinusoidal rotating flux is produced therein may be said to act as a perfect transformer in l which the Variation of current strength, so long as there is no variation of impressed voltage, produces no change in the flux, and consequently no change in the ratio of transformation. A constant potential at the termi nals of the primary will therefore give a constant potential at the terminals of the secondary independent of the load. As before pointed out, such a rotating two-branched flux is peculiarly useful in a'rectifyin g transformer, since the coils which are short-circuited by commutation are not allowed to produce fluxes which interfere with the twobran ched sinusoidal flux, andare therefore not allowed to interfere with the transformation.

The transformation of monophase cm"- rents.The system of compensating circuits which I have above described will permit the formation in the transformer-ring of a sinusoidal two-branched flux rotating with uniform velocity; but these compensating windings have not the power of selecting between two sinusoidal two-branched fluxes, one of which is rotating in one direction and the other of which is rotating in the opposite dipoints of the primary winding of my trans-' former a monophase alternating current this monoph'ase current will produce in the trans former-rin g an oscillating flux which may be considered as the resultant of two sinusoidal fluxes, one moving to the right and the other moving to the left in the ring, or, what is the same thing, other moving counter-clockwise in. the ring.

The flux whichmoves clockwise in the ring may be considered as the sum of a number of.

component fluxes of different branches, as above defined. The compensating circuit, as above described, will obliterate all the componentsbut the simple sine or two-branched component. So the flux which moves counter-clockwise inthe ring will be composed of a number of component fluxes. The compensating windings, as above described, will obliterate all but the pure sine or twobranchedL component. It follows that if I feed a monophase current to the primary winding of my transformer supplied with compensating circuits its ring will be the of two sine two-branched fluxes, one rotating clockwise and the other rotating counterclockwise. In this connection I call attention to United States Patent No. 630,233,

August 1, 1899, to Hutin and Leblanc, page 3', line 23, (to. Now it isnecessary to'oblitcrate one of these two rotating fluxes and to leave in the ring a single rotating fiux,offa simple sine two-branched character, rotating in a given direction. In other words, it is necessary to obliterate one of the two rotating two-branched fluxes. This I do by connecting the compensating circuits of the'trans former to the armature-circuits of syn- V chronous biphase motor the field-circuits of which are supplied with direct currents and to the field-poles of which a magnetic shield is applied, as hereinbefore described. That is'to say, I connect the two terminals of the united sine compensating circuits to the two terminals of one of the biphase armaturewindings and I connect the two terminals of the united cosine compensating circuits to the two terminals of the other biphase armature-winding of the synchronous motor. The

sine and cosine compensating circuits terminate in the rings I II III IV and are connected to the terminals 18, 19, 20, and 21, respcctively, which thus become the real terminals of these circuits. ture-circuits of the synchronous motor terminate in,the rings 12, 13, and 1%, as shown in Fig. 12. As in the case of other biphase ciii'cuits, so here I may use a common return. I therefore connect the terminals 18, 1. 20, and 21 of the sine and cosine compensating circuits to the rings 12, 13, and 14, which'are the terminals of the biphase armature-cin one moving clockwise and the 'what has above been saidj speaking, the various points of these com- The biphase armacuits of the synchronous motor, as shown in Fig. 12.

circuit I have in one of the armature-circuits of the motor, and whatever currents I have in the cosinecompensating circuit I have in the other armature-circuit of the motor; but if I am feeding monophase currents to my transformer the compensating circuits will be the seat of eleetromotive forces corresponding to two oppositely-rotating magnetic fields. Therefore the armature-circuits of the motor will be the seat of electromotive forces corresponding to two oppositely-rotating magnetic fields. One of these fields will move in a direction opposite to that of the rotation of the armature, with a velocity of the armature, and will therefore be stationary in space. It will not be affected by the magnetic screen in the motor. The other rotating field will move witlrreference to the armature and in the same direction as the armature with a velocity equal to that of the armature. It will therefore move in space, or with reference to the magnetic shield, at a velocity equal to twicc that of the armature. The magnetic shield will therefore consume and thus obliterate this rotating field. There will thus be left in the armature-circuits of the motor simply the currents corresponding to a single rotating field. It follows that the compensating circuits will also be the seat of currents corresponding to a single rotating field moving in a given direction and that the field rotating in the opposite direction will have been obliterated.

The fmecham'cal connection of the tramsforme-r wm(lmgs.The compensating cir- V cuits are interconnected in a manner which from the electrical standpoint is clear from Mechanically pensating circuits are connected to the points ninety degrees removed therefrom by the radial bars 28 of Fig. 1; nor is there any peculiarity about interconnecting the groups of connnutator-sections-that is, of connecting to each other the commutator-segments marked 1 in Fig. 5, of connecting to each other .the commutatorsegments marked 2 in Fig. 5, and so on. These interconnections are effected by means of the ordinary c011- necting-bars 29.

There remains to describe the connections of the secondary circuits to the commutatorsegments. It is generally necessary in practice that these connections should traverse the iron parts of the machine-frames. If the currents transmitted along these connections become intense, they may develop fluxes of importance in the metal castings which surround them, and these fluxes may exert an objectionable influence 011 the commutation or may bring about a development of unnecessary heat/ This inconvenienceI remedy in a manner represented in Fig. 6.

Each secondary spool of the transformer is It follows from this that whatever currents I have in the sine compensating built up of two spools having the same number of turns, but in which the wire is of half the cross-sectionthat is to say, instead of having a single secondary coil in a given slot 1 have two coils in the slot, one wound clockwise and the other wound counterclockwise. In Fig. 6 the spools which are wound clockwise are represented by an undulating continuous line and the spools which are wound counter-clockwise are represented by an undulating dotted line. The portion of the machine-frame through which the connections from the secondary coils to the commutator-segments run is indicated by the annulus 7 0. The points marked 1 2 3 to 12 on the outside of this annulus are supposed to represent the commutator-segments in the case that there is but one group of these commutator-segments. Should there be a number of groups of these commutator-segments, then the points 1 2 3 to 12 on the outside of the machine-frame '70 of Fig. 6 would run to the commutator-segments of a given group, it being understood that the commutator-bars of one group are interconnected to the commutator-bars of the other groups by the connections 29.

To fix ideas, let us consider the secondary coil represented by the full undulating line 1 2 in Fig. 6 and the secondary coil, which is associated with it, in the same slot and which is indicated by the broken undulating line 7 8 of Fig. 6. The currents of these two coils will be in opposite direction. This means that the currents which traverse the two connections going through the aperture in the machine-frame 70, which come from the points 7 and 1, respectively, will be traversed by equal current in opposite directions. The same remarks apply to any other pair of associated connections which pass through an aperture in the machine-frame 70. In each case such pair of associated connections will respectively be traversed by currents in the opposite direction, so that these currents cannot give rise to any fluxes in the iron frame of the machine which they traverse.

The rotating centrifugal brushes. -Although it is not necessary in the use of my compensating circuits, I have described the transformer as provided with a hollow commutator, inside of which the brushes rotate. In order to have the brushes bear against the commutator-segments with the requisite pressure, I prefer to use the construction shown.

in Figs. 7 to 9, inclusive, or the modified construction shown in Figs. 10, 11, and 13.

Examining Figs. 7 and 8, there will be found a brnsh-holdera, which is fixed to or mounted on the end of the spider-arms of Fig. 1. The carbon commutator-block 3 slides radially within the brush-holder a and is electrically connected to the arm 4 of the spider by a flexible wire or cable f, one end of which is sccured to the arm 1 by the binding-screw '0 and the other end of which is secured to the centrifugal action by their weight;

much as the pneumatic carbon-brush block 3 by means of a copper connection preferably electroplated to the bottom of the carbon block. In order that the flexible cable f may reach the carbon brush 3, the brush-holder a is formed with the slot r. Ihave'for convenience constructed mybrushholder a to receive two sliding carbon blocks 3. The centrifugal force caused by the rotation of the spider-arms 4: will press the carbon block 3 against the inside surface of the hollow commutator with the required force.

The centrifugal force will act as a perfect I spring and without inertia. The dimensions of the commutator will be such that the angular speed of the brushes will produce a pressure against the walls of the commutator of the'desired value when the brushes are new. WVhen the brushes have been used up to a certain degree and have grown lighter in weight, the decrease in pressure will be compensated for by the introduction of metallic masses of suitable weight at the bottom of the holder a. These weights will press against the bottom of the brushes and will add to the It will be desirable to build up the carbon brushes of layers Z Z, as shown in Fig. 9, which are glued together by a paste of small conductivity.

The brushes which I have so far described are well adapted for use when the transformer is to transform alternating currents into continuous currents; but when the transformer is to transform continuous currents into, al-' ternating currents I use the construction shown in Figs. 10, 11, and 13. In this construction the spider-arms 4 are connected by an annular plate J just as the spokes of a wheel are connected by the rim. A pneumatic tube P rests within the groove in this rim J very tube of a bicyclewheel rests within the rim of the wheel. Be-

tween any two adjacent spider-arms 4: the.

tube is held down on the rim by segmental plates, which it is unnecessary to show. At the extremities of the spider-arms 4 I fasten plates J. (Shown in section in Fig. 10 and in plan in Fig. 13.) These plates J are formed with preferably two brush-holders a. Inside of these brush-holders slide sockets g,in which are frictionally held the carbon blocks 3. The socket g is formed with a slot r for the passage of the cable), and beneath the block 3 there is a wedge h, which serves to raise the block 3 with reference to the socket g when the block 3 begins to wear. By inflating the pneumatic tube to a desired degree the brushes or blocks 3 are caused to press against the commutatorsegments with the desired pressure, even when the brushes are not rotating and there is no centrifugal force. This enables the construction of Figs. 10, 11, and 13 to be used when the brushes have to carry a large current when starting and before their centrifugal force comes into play, which is the case when direct current is transformed rents.

The starting ofthc rectifytag-transformer. M

into alternating our- The principal circuit connections (shown in Fig. 12) have been described above. It is merely necessary to add that the wires 00 u, coming from the brushes 10 ll, carry continuous current and that there is a voltmeter V branched across them. The wires 50 u are connected to the wires 00 a by means of a pole-changing switch X. There is an ammeter A in the wire it. There is a switch I, which may either connect the field-windings of the 1n otor M to the continuous-current leads an a or which may close the field-circuits of this motor in short circuit upon themselves. Finally, there is a switch I connecting the alternating-current leads to the alternatingcurrent side of the transformer.

If it is desired to transform alternating current into continuous current, the pole-changing switch X is placed in its open-circuit position and the switch I is placed in the position in whichit short-circuits the field-winding of the motor M. The alternating currents being supplied to the transformer by closing switch I, the motor M is started by hand in making use of the special crank or key. It begins to run as an induction-motor and attains a speed approximating synchronism. \Vhen this approximately-synchronous speed is attained, the switch I is manipulated so as to send into the field of the motor M the current which is being rectified by the brushes of the transformer. The motor then instantly synchronizes. In order to determine the direction of this continuous current, the Voltmeter V is examined and the pole-changing switch X is manipulated to throw the continuous current onto the line ac a in the proper direction.

In order to transform continuous current into alternating currents, the switches I and X are opened, while the switch I is turned to connect the field of the motor M with the leads a: u. The armature of the motor is then brought up to speed by hand and the switch X is closed. As soon as this is done the motor synchronizes and the switch I may now be closed to send the alternating currents to line.

I repeat that while I have shown my invention as applied to a rectifying-transformer the broad features of my invention and certain of the constructions which I have shown are useful in other connections. Thus, for instance, I have already pointed out that I may use the broad feature of shaping a mag netic flux on transformers for converting polyphase-currents into other polyphase-currents.

Other uses of my invention will readily suggest themselves. I may, for example, use my compensating circuits as the armaturewindings of polyphase-generators. If it be desired to generate threephase currents each of the components of which is sinusoidal and which are accurately displaced from each other in phase by one hundred and twenty degrees, I use three of my compensating circuits as the armature-circuits of the generator. I would take three such compensating circuits, each one of which is identical with either one of the compensating circuits de scribed above, and I would superpose these three circuits on the armature-body in such a manner that the zero-coils of each compensating circuit are displaced from the corresponding zero-coils of the other two compensating circuits by one hundred and twenty degrees. These coils would be tapped to the mains in a manner well understood.

This application is a division of my application, Serial No. 2,038, filed January 4, 1901.

IVhat I claim is- 1. The method of shaping a rotary complex magnetic flux which consists in suppressing a predetermined component or components of the same, substantially as described.

2. The method of shaping a complex rotary magnetic flux which consists in suppressing a predetermined periodic component or components of the same.

3. The method of producing a simple rotating flux from a complex rotary flux which consists in suppressing all but a single component of the same, substantially as described.

4. The method of producing a sinusoidal rotating flux which consists in energizing the core by periodic electrical impulses which tend to produce in the same a complex magnetic flux, and suppressing the formation of all but a single component of the same, substantially as described.

5. The method of producing a sinusoidal rotating flux of two branches, which consists in energizing the core by periodic electrical impulses and suppressing the formation of all fluxes of higher orders, substantially as described.

6. The method of producing a sinusoidal rotating flux of two branches, which consists in energizing the core by periodic electrical impulses which tend to produce in the same a complex magnetic flux, and consuming all but one component of the same in sinusoidally distributed circuits on the core, substantially as described.

7. The method of transforming alternating currents of a given number of phases into direct currents or alternating currents of another number of phases or Vice versa, which consists in distributing one kind of these currents through the primary winding of atransformer, suppressing the formation in the core of all but one rotating sinusoidal flux of two branches, and drawing from the secondary winding the desired current, substantially as described.

8. The method of avoiding the disturbing effects of the currents due to the electromotive forces of self-induction in the windings of iron-cored transformers, which consists in suppressing the formation of fluxes by these currents, substantially as described.

the secondary winding, substantially as described.

11. The method of adjusting the set of the motor-driven brushes of a commutator, which consists in correspondingly adjusting the position of the magnetic field of the motor, substantially as described.

In testimony whereof I have signed my name to this specification in the presence of two subscribing Witnesses.

MAURICE LEBLANO.

\Vitnesses:

F. T. CHAPMAN, EDWIN S. OLARKsoN.

Images