The Parsons & Law Method of Regulating Dynamo Electric Machinery (Specially Contributed by the Hon. Chas. A. Parsons.)

Progress, Volume IV, Issue 2, 1 December 1908, Page 65

The Parsons & Law Method of Regulating Dynamo Electric Machinery (Specially Contributed by the Hon. Chas. A. Parsons.)

In dynamo electric machinery (both continuous and alternating) it has been found that on account of armature reaction and ohmic loss the excitation has to be increased as the external load is increased. This is usually done by compound winding, or by alteration of the field rheostat. Further, it has been found that when iron forming part of a magnetic circuit is subjected to an alternating flux superimposed on a continuous flux, the latter is reduced, that is, the magnetic resistance is apparently incieased. According to the following method of regulation there are piovided means for the regulation of dynamo electric machinery, utilising the latter principle m a new manner The method of regulation consists in the use of a leakage path provided between the poles of a magnet, which path can be subjected to an alternating magneto motive, whereby the amount of leakage can be regulated at will by varying the current producing such alternating force. 1 have repeatedly on examining cars in for repair been able to lift off the connecting link after lenoving the leather grease bags, and on one 01 two occasions the link has actually dropped off, and the rest of the car has been m much the oame condition. I also know instances where the nut has jumped the threads on the shell, with disastrous results. An example of this came under my notice, the distorted condition of the screw threads being clearly seen, and to an observant eye the cause was at once apparent on a further glance The joint became detached when the car was coasting down a long hill. This involved serious injuries to all in the car. It was cvi' dent that whoever had put this joint up had foi gotten to insert the lynch pin or it had dropped out, which is not at all likely ; but, notwithstanding this, the nut was hard up to its work— a rather remarkable thing when one takes into consideration the vibration it would have to endure. The car had previously given trouble with the steering gear, and in each case the cause was

identical ; the joints had been adjusted up till the back and front halves of the cup were m contact with each other, thus partially encircling the ball, which was also worn, when, of course, no further adjustment was possible, and they should have been taken out and replaced by new ones at an earlier stage. It is obvious that in this case the halves of the cup, being of gunmetal, would wear rapidly, and to such an extent as to allow the ball to come out, and this is precisely what had occurred, and, to make matters worse, the sides of the shell also bulged out for the reason already noted. From the foregoing it will be evident that for a ball joint to be satisfactory it should be constructed rather more heavily than is generally the case, and I have found the following propoitions successful in practice: The thickness of the shell should not be less than one eighth the diameter of the ball, and the clearance bet.veen the ball and the inside of the shell should not exceed l-32nd in for balls of lin. m diameter, and correspondingly less for smaller sizes, whilst it is advisable, and is now a general practice, to make the halves of the cup of some hard material, case hardened steel for preference.. The depth (inside) of the cap nut varies considerably, but a suitable proportion is three-fifths the diameter of the ball. Apart from all other considerations, however, the difficulties attending the lubrication of this type of joint are serious, and I have yet to find a ball joint which is entirely satisfactory in this respect. It also consists in the method of applying this device to alternators coupled in parallel to multipolai machines, and to polyphase machines. According to one form there is provided for example in the pole piece of a dynamo or alternator a ring, preferably of laminated iron, concentric with the armature. In such an anangement the E.M.F. derived from the armature will be that from the lines of force crossing the ring. This ring is wound with a winding thiough which an alternating current is passed, wheifby the leakage due to the rine, is reduced, and thus moie lines of force will pass thiough the armature, and the voltage will be increased and compensation for armature reaction thus effected. Instead of adding a single ring, a series of such rings connected to two or more of the pole pieces may be employed, or only part of a ring may be used, the main magnets acting as the remainder. The windings on the ring or lings may be in a number of sections suited to the number of phases of the alternating current. It is not necessary that the rings entirely suriound the armature, but they may form or be attached to projections from the pole pieces, or in ans other way, provided that they act as a leakage path, carrying a certain proportion of the flux which does not pass through the armature. One form of this is well known, in which the pole pieces are provided with projections, between which the leakage path is situated, carrying a winding to which an alternating current is applied. The rings may be provided with an air gap, which may further be varied, and thus effect an adjustment to meet different working conditions. In an alternator either the exciter magnets or the main magnets or both ra&y be provided with a ring or rings, and the main current or a portion of the same, which may or may not be transformed, passed round them. As the load rises, therefore, the magnetic resistance of these rings is increased, whereby more lines of force are driven through the armatuie, and thus even with a low-power factor the voltage can be maintained or even increased automatically. Either the main magnets or the exciter magnets may be provided w ith this device, but the latter is preferred, as the inductive loss of voltage in the leakage path winding is smaller. In the case of several alternators running in parallel, equalising wires — similar to those known in conjunction with compound dynamo electric machines — may be provided to connect the various slip rings of the alternators on the various leakage paths ; or the leakage paths of the various machines may be excited either with or without subdivision or transformation by the sum of the currents of the coupled machines, that is to say, the leakage path may be inserted in the main leads away from the power station or in a transformer circuit connected to the same. In this latter case equalising wnes may be applied to the transformed circuits. It will be obvious that the use of a leakage path composed of a ring or part of a ring is not essential, as any suitable bar of iron jor any other more or less "magnetic" material connecting the pole pieces may be used.

Further, the leakage path may be divided up into portions ; also the leakage path may join the main magnetic circuit between the poles and the yoke of the field magnet, thus bridging a portion of the magnetic circuit upon which the main magnetomotive force is impressed. The Parsons and Law method of regulation further relates to the regulation of alternators, dynamos, and the like, with the object to improve the regulation of voltage of two or more phase machines, and also rotary convertors, that is machines for converting from one, two, or more phases to continuous current, and to provide means for easy adjustment of the apparatus to meet different working conditions. According to the method described above it has been found that in alternations with two or more phases the insertion of a leakage path winding in one phase only, causes an inequality of the phases, and for many reasons this inequality may prove objectionable. In the following airangements, therefore, separate leakage paths are provided governed by all or part of the phases. For instance, in a 3-phase alternator there may be three leakage paths, one connected to each phase. There may also be, in cases where a wiie is led from the star connection, another leakage path in that, if desired. Leakage paths to lotary convertors for the purpose of regulation are also described. The leakage paths may be placed, as described above, on one side of the pole pieces, or may be distributed, some between one pair of pole horns and some between the other, or leakage paths may be placed on both polar horns in series or in parallel with one another, and in the case of multipolar machines the leakage paths may bo either all in series or all in parallels, or some poles may have leakage paths connected to one phase and some to another phase. In fact, the distribution of these leakage paths may be as universal as possible but it is preferable to have approximately equal amount of leakage path between each pair of poles on the exciter or main magnets, so as to equalise the effect on the field as much as possible, and also to have equal amounts on each phase. The amount of compounding; given bj these leakage paths can be varied by varying the nu-nber of the leakage paths For instance, in some cases, if a three-phase plant be run on non-inductive load only, the windings on one or two leakage paths may be used, and the others are then short-circuited or otherwise rendered inoperative ; for a power factor of 90 per cent, two may be used, and for a power factor of 80 per cent, all three may be in use. An adjustable an gap between the leakage paths and the poles foi adjusting the amount of compounding has been refenecl to above. Simi-

larly, an adjustable air gap may be provided in one or both limbs of the leakage paths themselves, or a separate w inding may be placed on the leakage paths, having an adjustable resistance or induction or condenser in circuit with it ; by altering this resistance or induction or condenser the magnetic flux through the leakage paths can be varied, if desired. Similarly, a resistance or choking coil or condenser can be put in shunt across the leakage paths, which choking coil or resistance or condenser can be made adjustable. In cases where the current for the leakage paths is supplied from a transformer or transformers, these transformers may be made adjustable in any well-known manner. In one example, a transformer supplied with primary current may supply secondary current to the leakage path. The adjustment of the compounding may in this case be effected by shunting either the primary current or the secondary cuirent. A variable inductance resistance or condenser may be placed in the shunt if desired. Also the transformers may be used for cutting out the leakage paths as and for the purposes described above. Arrangements of current transformers and leakage paths in a three-phase system, in which each phase has one current transformer controlling a single phase leakage path, each phase winding of the alternator is connected to a common bar or lead, after it has passed through the primary winding of its transformers. This bar also forms tho neutral of the alternator. The leakage paths, which are in practice arranged behind one another on the pole pieces of the exciter in this form, aie excited by the secondary transformer windings, one end of each of which is connected to a common bar, and the other end to another common bai , the arrangement being such that the secondaries of the transformers and the leakage paths are encircled in series between the two bars In some cases it is prefeiable to combine tw o or more phases on one or more transformers, and supply one or mom leakage paths fiom these transfoimers ; for instance, in a three-phase machine the tiansfoimer 01 tiansformers will conveit from tlnee-phase into single-phase, the transfoi mci s being so arranged that each circuit of the thiee-phase cuirent gives a portion of its effect to the single-phase, and this single-phase cuirent may then be wound round one or more leakage paths. The whole of the above devices can be apphed_ to alternators of one 01 moie phases, rotaries, motor generators, or motor convertors, and, in fact, to am alternating current machinery where there are poles excited bv continuous current between which leakage paths can be put.

Messrs. L. C. Knight and Co., 50 Cuba street, have secured the contract for supplying and erecting the large storage battery for the Eketahuna Borough Council. This will consist of 120 Fritchett and Golds cells of the seven-plate type, with a guaranteed discharge of 45 amperes for three hours. This is a similai battery to that ordered for the Christchurch abattoirs by the same firm, and speaks well for this accumulator, which is recognised as one of the heaviest type of cell on the market The same firm also have tlxe order for lh& booster and switchboards. The latter will con&ist of four special polished panels, and will be made up in their workshops to meet the exact requirements of the Eketahuna Borough Council. Mr H. A. Smith will personally supervise the above work. The tallest lighthouse on British coasts is the Skenyvore. off Argyllshire. It is 240 ft. high. It contains 4308 tons of masonry, and cost £90,268.

its reaction on the platform of a scale. The faces of the upper and lower plates, A and a, were planed, and in one series of experiments both faces of the middle plate B. In other series, both faces of plate B were left just as they were received from the rolling mill, that being a very decent surface, comparatively smooth and flat. Increasing vertical pressures were applied up to a maximum of 5501b. per lineal inch of roller. The apparatus employed would not permit of a higher pressure, owing to the resistance to motion of plate B when solid rollers were applied, but under the same conditions with spiral rollers it was possible to place a pressure per lineal inch of 7001b. As these experiments were all made between rigid plates on a machine in which the pressure was applied by screws, and registered on a scale beam, it was suspected that the rigidity might be more favourable to the spiral than to the solid rollers. Therefore, an experiment was made in which the pressure was applied by weight acting on a lever and due to gravity only, with the following result —

a bearing given as low as .003, or sometimes lower, but the method of arriving at these figures is manifestly different from that employed in this ease under question. The comparison between three averages is really what is most interesting. You will note that the difference in the co-ellicient of friction as given in the results of these tests between the flexible roller and the solid roller is .0043, which shows a difference of 23 per cent, in favour of the flexible roller. The difference in the co-efficient or friction between the roller and the plain bearings seems to obviously call for further reference. It may be said that the friction of roller bearings is shown to be one-fifth to one-third of the plain bearing. It is also shown in these tests that co-efficient of friction in the roller bearings decreases as the load increases. Many other laboratory tests have been made, all showing the same general result. One or two practical tests may be of interest. I might describe briefly a test which was made in one of the plants belonging to the United Shoe Machine! y Co. The line shaft which was tested is 152 feet long, 3 inches diameter, and is supported by 20 beaiings. It is belt driven at one end from a head shaft. Belted from this shaft are counters of eighty-eight machine tools. jSTo changes were allowed in the countershafts or belts during the progress of the tests. The tests extended over a period of about one month, and were carried out to the minutest details. The summary of results, however, which is what we are most interested in, is as follows: — First the average frictional load of the shaft, with the 88 counters constantly running in white metal bearings.

These results corresponded closely to the first experiment, and prove that the method of applying pressure was not the cause of the difference in the resistance to rolling. It will be observed by the figures as given above, that the resistance to motion was much less with the spiral than with the solid rollers under both pressures employed. The reduction of rolling resistance varies in the various tests from 10 to 66 per cent., and averages 51 per cent, less with the smooth centre plate, and 23 per cent, less with the rough centre plate where the flexible rollers were used in the place of the solid one. Observations were likewise made to ascertain the thoroughness of the bearing contact between the plates and the rollers. With the so-called rough plates, the incomplete bearing of the solid roller showed indications of grooving at the points of contact, which is the usual manner of failure of anti-friction rollers. On the contrary, under the same conditions the spiral rollers showed complete bearing throughout their entire length the elasticity of the roller permitting it to follow the sinuosities of the rough plate, and maintain contact. The pressures employed in the experiments were about as high as good practice tolerates with the slowest motion, and very much higher than is permissible with the velocities in ordinary shafting. The high pressure borne by the spiral roller without permanent deformation, its low resistance to motion as compared to the solid rollers, and its elastic adaptation to inequalities of surface, were clearly exhibited by the experiments. Another experiment which was described in the "Engineering Eeview" of February, 1906, is very interesting. A special friction-testing machine of particular design was employed for determining the co-efficient of friction of the flexible roller bearings, as compared with the solid roller bearings and plain bearings. The machine was fitted with a pendulum suspended from the bearing in such a way that the same could be weighed. The force necessary to keep the pendulum in vertical position was measured by putting weights in a scale pan attached to a cord passing over a pulley. The shaft or journal was of ordinary machinery steel, and rotated by means of a belt and pulley. In beginning an experiment, a pointer on the lower end of the pendulum was brought to a zero mark exactly beneath the centre of the shaft by means of adjusting screws in the yoke holding the bearing. After the shaft began to revolve, the pointer was held to the zero mark by putting weights on the scale pan. The product of the force thus supplied, to the pendulum, by distance of the point of application from the centre of the shaft, gave the moment of friction, and dividing this by the radius of the journal, gave the fuction at the surface of the journal. Dividing this, again, by the total weight on the journal, gave the co-efficient of friction. A long series of tests of this kind were made with each one of the four different sizes of shaft, viz., 2in., 2%in., 2%in., and 3in., and the result of the entire test condensed averages, shows the average co-efficient of fiiction on the four sizes for the thiee different kinds of bearings to be as follows : — For the flexible idler bearing co-efficient of friction .. ... .019 For the solid i oiler beaimgb .0233 For the plain bearings . . 082 There have been many tests of diffeient ehaiacters made at different times to obtain the coefficient of friction in bearings of various types, and we have often seen co-efficient of friction in

Main shaft in white metal bearing 8.85 H.P. Main shaft in flexible roller bearing, 6.36 H.P. Per cent, of saving by flexible roller bearing 16.7 per cent. Second, average frietional load on main shaft only: all counter belts thrown off — In white metal bearings, 2.28 H.P. In flexible roller bearings, .80 H.P. Per cent, of saving by flexible roller bearings, 64.9 per cent. Another test which may be interesting was made on an overloaded shaft, 2in. diameter, and the object was, if possible, to avoid the fixing of a larger motor. There were thirty bearings in use on this line of shaft, and the results were as follows, as shown by the voltmeters and ammeters: —

Rated capacity of motor was 15 H P. Saving slwwn. Percentage saving b> Hyatt roller bearings, 17.4 per cent. Indicated H.P. saved by Hyatt roller bearings (30.2 in.), 2.4 H.P. Cost of 30 2-in. flexible roller bearings, £30. Net profit on investment after paying interest and depreciation, 28 per cent. Time required for saving to repay investment, 2£ years. Time required, assuming flexible rollers bearings had been originally fixed, lj years. Tests with machinery of a heavy, slow-moving order have also been made, and very fine results in favour of the flexible bearings have been shown. For instance, the Wellman-Seaver-Mor-gan Company, or, as they have recently changed their name, the Wellman-Seaver-Head, Ltd., of 47 Victoria-street, have reported that on their Open Hearth Charging Machine (which machine weighs 19 tons without its load) that this machine, when tested, fitted with roller bearings, took 18 amperes at 450 volts to travel the machine along its rails, showing a current consumption at the rate of 10.9 H.P. The same machine, fitted with the ordinary brass journals, they calculate took 19.1 H.P. to travel at the same speed. It will be seen, therefore, that for slow moving machinery, and especially a machine of this kind, which is sometimes started a great many times during the day, the saving to be effected by the use of roller bearings is enormous. Another test has been reported from Wm. Beardmore's works in Parkhead concerning a roller table serving a pair of their heavy armour plate rolls. It is stated that this roller table, after it was refitted with flexible roller bearings, is being operated with 50 per cent, less power than what it required previously, when it was fitted with bearings of the ordinary type.

With >piral H 9 lb 17 Ib. to. tors. h I Solid H 26 Ib. 34 Ib. Jo. 'ci& 1,000 lb. ,000 lb.

Total Pressure Applied. Resistance to motion of Plate B :—

Volts. Nith. white metal bearings . . . 224 With flexible roller bearings 222 Motor (running free) ... 223 Amps. 36 30 4 Watts 8064 G660 892