arma

prevents the increase of the total magnetisation through the armature, and thus a constant potential is maintained. The "armature characteristic " IV., plotted in Fig. 7, IV., plotted in Fig. 7, was taken from the machine of the above sort, built with cast-iron fields. The cross-section of the field cores proved on trial to be too small, and became strongly saturated at full load, while they were quite a little under the point of saturation at no load. This curious result for a constant-potential generator was due to the increased magnetic leakage produced as the series ampere-turns on the field came up with the load. Saturation took place, as the curve indicates, when the armature furnished a current of about 140 amperes, and no possible compounding could ever make this generator produce even approximately a constant potential, with variation of load. Steel cores of the same dimensions were substituted for the cast-iron cores. Saturation did not occur in them due to magnetic leakage. The pole corners were very thin, as in Fig. 2, and the " ture characteristic " III. was obtained. The machine was then furnished with a shunt winding that produced a slightly smaller number of initial ampere-turns than Curve III. indicates as required to produce 125 volts, and with series-turns at such a number that the total number of ampere-turns on the field for any current developed by the armature is shown by the broken line drawn through Curve III. It was under these conditions that the machine performed in the manner described above, and not vary more than 5 per cent. from the normal E.M.F. on either side, or a total variation of 10 per cent. It was then almost entirely rebuilt. The armature was provided with a core that was considerably larger in crosssection, and the maximum magnetic density used in it was 11,000 lines per square centimetre, as against 20,000 used before. The lugs on the core were dispensed with and the wires wound on the surface of the core. The poles were made of cast iron, and fashioned to accord more nearly with those in Fig. 3. The air gap required 10,000 ampere-turns to set up the magnetisation through it at no load, while the armature ampere turns were 8,000 at normal output, so that series ampere-turns had only to be added to counteract the action of the ampere-turns on the armature that lie between the double angle of lead, to increase slightly the E.M.F. by the amount equal to the fall of potential through the armature caused by its resistance, and to compensate for the slight effect of the pole corners that still became saturated to a limited extent for the higher outputs. It should be remembered that the magnetic leakage that takes place between the adjacent north and south pole corners, one of which is strongly and the other weakly magnetised, plays an important part in saturating thin pole corners. It is evident that unless the "armature characteristic" is a straight line, as in Curve I., Fig. 7, that the machine cannot be made to regulate for constant potential with a high degree of refinement. The poles were again changed and shaped as in Fig. 4, when an 'armature characteristic " given in Curve I. was obtained and the proper number of shunts and series ampere-turns for a refined degree of regulation were readily decided upon. These experiments confirm what has been said above, and show how useless have been the attempts to diminish the air gap beyond

certain limits.

The "armature characteristic" curved considerably, indicating that the pole corners become saturated. It is evident, too, that the normal magnetisation in the pole corners, in addition to the magnetic leakage, which is greater there than anywhere else, produced saturation in all pole corners, even with no current in the armature. For, at full load, there were 4,000 ampere-turns on the armature, while 4,000 series ampere-turns had to be added to the field that produced 125 volts at no load to keep the E.M.F. the same. Therefore, at full load we have the same number of ampere-turns acting through the weakened pole corners as at no load, and the total amount of magnetisation has only been increased 5 per cent. to compensate for the resistance of the armature conductors. The conditions, however, are not the same, for there are just 4,000 more ampere-turns to cause magnetic leakage at the pole corners, so that on the whole the magnetisation in them is increased. This increase of magnetic density in them greatly increases their magnetic resistance, for they are saturated to begin with. It is on this account that we find the magnetisation under the weakened pole corners diminished when apparently the forces acting have not been changed. The magnetisation under the strengthened pole corners through the air gap is increased more than it is diminished by the effect of the added magnetic leakage through the 8,000 additional ampereturns, that act to produce magnetisation by this route through the armature.

In Fig. 9 the diagrams show the magnetic action of the armature of a 10-h.p., 110-volt motor, with poles fashioned as in Fig. 6. Measurements of the magnetic leakage were made on this motor, and the results indicate that the shape given to the pole corners avoided saturation in them even at full load. The double angle of lead was almost 60deg. The ampere-turns embraced by it on the armature were partially compensated for by nine series-turns on each of the consequent fields. The remainder of the armature ampere-turns that lie between the double angle of lead served to weaken the field by just the amount required to produce a constant speed. The following figures give additional data on this motor:

armature.

This is the number of field ampere-turns that exerted their magnetising force between the pole faces through the pole corners, are therefore very near zero, which is entirely The ampere-turns, acting through the weakened corroborated by the fact that the magnetisation was observed to be zero at this point. See full load curve in Fig. 9. Through the strong pole corners the ampere-turns acting were the 5,600 of the shunt ampere-turns + the 720 of the nine series-turns + the armature ampere-turns, 5,750twice the ampere-turns between the double angle of lead, ? [3 x 5,750], or 3,400 8,670, which will produce a magnetic density through an air gap of 2-64 centimetres depth of

while the actual magnetic density measured at this point was 3,950, an agreement within the possible limit of error. In Fig. 10 are given curves showing the magnetic per formance of an armature, with its conductors laid in deep narrow grooves, as shown in Fig. 12. The clearance on each side was in., making the double air gap in. Additional dimensions are as follows:

Diameter of the armature core

Length of armature core

6 in.

6. in.

Curve I. in Fig. 10 shows the distribution of magnetisation at 112 volts, no current, a speed of 1,800, and a field excitation of 2,600 ampere-turns. Curve II. shows the magnetic distribution for an output of 97 volts and 24 amperes, at a speed of 1,800, and a field excitation of 2,600 ampere-turns. Curve III. shows the magnetic distribution at an output of 40 volts and 20 amperes, at a speed of 1,800, with a field excitation of 750 ampere-turns. This same excitation, when the armature furnished no current, produced an E.M.F. of 48 volts, at a speed of 1,800 revo lutions. The poles were shaped as in Figs. 11 and 12, but modified as explained below. In making these experiments carbon brushes were used, and their position maintained at the normal diameter of commutation. An average magnetic density in the air gap of 3,400 lines per square centimetre was required to produce an E.M.F. of 112 volts. The grooves on the armature in which the conductors were placed occupied one-half of the armature surface, so that the actual magnetic density in the air gap was 1.8 times this average magnetic density. The ampere-turns required to set up this magnetic density in the air gap were •125 × 2.54 × [3,400 × 1.8]

1,520,

which is the number of field ampere-turns, whose magnetising force is impressed between the pole surfaces through

The

The total number of ampere-turns acting to produce magnetisation through these strong pole corners was the sum of the field ampere-turns that impressed magnetising force from pole face to pole face through the armature, and the ampere-turns on the armature covered by the poles. poles covered approximately 85 per cent. of the armature surface, making this value 1,500 × [2,300 × 85]=3,400. . Of this number, as was just shown, 1,000 were utilised in producing the magnetic density of 5,100 through the saturated portion of the poles. The remaining 2,400 ampereturns exerted their magnetising force in producing the average magnetic density of 5,100 through the air gap, and affords another opportunity of checking these ideas of the For the ampere-turns required to set up an average magnetic density of 5,100 through the air gap under consideration were : [1.8 x 5,100] x 125 x 2.54 = 2,300.

action of the armature on the field.

1.26

which checks with the above value as well as could be expected. When the armature furnished 24 amperes, the E.M.F. at the brushes was 97 volts, while with no current it was 112 volts. Of this drop of 15 volts, eight are accounted for by the resistance of the armature and the extra seven were caused by the saturated pole corners. By operating this same machine at an E.M.F. at which the pole corners could not saturate with normal output of

D2

(3.000 LINES, PER s༠.Z; ༥ 4iR GAP 3.000

as against 1,520 impressed by the field. Under these circumstances, the magnetism under the weakened pole corners is reversed, as is also clearly indicated by Curve II. in Fig. 10, or the curve C, E, D, F, C, in Fig. 11. This curve also shows that the magnetic density under the strong pole corners was 5,100 lines per square centimetre. Now in building this machine, six longitudinal slots, 1in. deep, were cut in each pole immediately back of the surface, which enables us to be sure of the exact densities in the pole corners for a given distance. For a depth of 14in. immediately back of the pole faces these slots took up onehalf of the cross-section of the poles. Then a density, therefore, of 5,100, really means a density of 10,200, or a strong saturation for a distance of 24in. in the cast iron of the poles. The magnetising force required to produce 10,200 lines per square centimetre through cast iron is 200 per centimetre length. Therefore, the ampere-turns required to establish this density through 24in. are

FIG. 12.

The

current, we have demonstrated for us in a very striking manner that the armature ampere-turns cannot change the total magnetisation established through the armature by the field when the pole corners do not saturate. E.M.F., with the armature current at zero, was brought to 48 volts with a separate field excitation of 750 ampereturns. Then when the armature was allowed to furnish 20 amperes the E.M.F. at the brushes dropped to 40 volts. Of this drop of eight volts seven were produced by the resistance of the armature. Yet the field is powerfully distorted by the armature current, as may be seen by reference to Curve III., Fig. 10, or the curve A, D, C, B, in Fig. 12. Even with this very great rearrangement of the magnetisation produced by the armature current, the total magnetisation set up by the field is practically unchanged. The difference of potential on the commutator between the points A B, Fig. 12, was observed to be 72 volts. This excess of E.M.F. over that which was produced at the brushes, the figure shows clearly to be due to magnetisation produced by the armature through itself and the strengthened pole corners. The points where the field is zero are at A B. They mark the diameter through which the ampere-turns encountered on the armature are just equal and opposite in action to the ampere-turns of the field that impress a magnetising force between the pole faces through the armature. A simple computation will show that this is true. The field ampere-turns that impress

12

40

The ampere-turns on the armature opposed to the magnetisation set up by the route A B are of the total number of ampere-turns on the armature (see Fig. 12) or 12 x 64 × 3 × 20 =576. 40 × 2

This is a fair agreement when we consider the accuracy with which the original data may be determined.

Mr. Esson, in his valuable paper above referred to, discussed the requisite features for a generator of constant current with closed-coil armatures, in which regulation is effected by shifting the brushes. He stated that the field should be uniform at all points under the poles, and that the arma ture cord should be saturated. These statements are a

little misleading. The magnetising force impressed by the field ampere-turns must be uniform at all points between the pole faces. This is accomplished by proportioning the poles so that the strongly-magnetised pole corners will not become saturated when the brushes have their extreme position for the development of the highest E.M.F. that the machine is to produce. The air gap is made of such a depth that the ampere-turns required to set up the magnetisation through the armature, without current, and for the production of the highest E.M.F. that the machine will be called on to give, shall be a little more than the armature ampere-turns when it furnishes its normal current. Then as long as the brushes are kept under the pole faces the non-sparking point will be wherever the brushes are placed. This will be the case whether the armature is or is not saturated. A practical demonstration is found in the following experiment: A Siemens and Halske dynamo, with magnet and armature cores, whose shape and dimensions are shown in Fig. 5, was used.

The field was separately excited with 4,000 ampere-turns on each of the sets of consequent poles. Regulation could then be effected for a constant current in the armature of 22 amperes, by shifting the brushes from no E.M.F. to 35 volts without the slightest sparking, even when metallic brushes were used. Within this limit the pole corners did not saturate. The field cores were wrought and the yokes cast iron. When the armature circuit was broken it was found that the field excitation of 4,000 ampere-turns produced an E.M.F. of 50 volts. The magnetic density in the field cores, including leakage, was only 11,000 lines per square centimetre. Therefore, of the 4,000 ampere-turns on the field not more than 200 were applied in setting up the magnetisation from pole face to pole face through the field cores. It is safe to assume, then, that of these 4,000 ampere-turns 3,800 were active in producing a magnetising force impressed uniformly over the pole faces through the armature. This same value is obtained by the method adopted in the previous cases. That is, by calculating the magnetic density in the air gap when 50 volts were developed, and then deducting the number of ampere-turns required to establish such a magnetic density through a 1in. air gap As to the armature, when it produced 22 amperes its ampere-turns numbered

the coil when its terminal bars at the commutator pass under the brush. When regulation is effected by this means, it is seen that all pole corners are alike magnetised, and at the centre of the pole faces the magnetisation is zero when the machine is short-circuited. At full output, at the highest E.M.F., the magnetisation under the one set of pole corners is almost zero, and under the other set it is at the maximum value that is ever obtained. In a generator of this type, when the poles are made stout enough at all points, the total amount of magnetisation through the armature at all loads will remain at a constant value.

What has been said of dynamos applies equally well to motors. In a motor the armature rotates in an opposite direction when field and armature currents remain the same as in a dynamo. The E.M.F. of self-induction, caused by the reversal of the current in a current has not changed, while the E.M.F. developed in the coil by the field has changed sign with the change of the direction of rotation.

The result is that the reversal of the current in an armature section must take place in a weak field of an opposite sign in a motor from what it does in a dynamo, when sparking is to be avoided entirely.

The action of the current in the armatures of multipolar dynamos and motors will be the same as that found for twopole machines.

The cost price of the light is therefore increased as the number of lamps decreases, provided that the greater part of the expenses remain the same, whatever the number of the hours of lighting.

The use of accumulators allows the energy to be stored during the hours of least load, and the station to be worked at its greatest output.

The accumulators can be used in two ways: 1. The cells can be charged by the dynamos, and the whole lighting can be supplied from them; this is the simplest process, and were it not for the high cost of cells, would be that universally employed. 2. The cells can be charged during the hours of least load, and during the hours of greatest load the cells and dynamos supply the lamps in parallel. This latter method, which is rather more complicated, requires a considerably less number of accumulators.

In the first arrangement the accumulators must supply the whole energy necessary for lighting-namely, 280,000 watt-hours. They must furnish this energy at a rapid rate of discharge-that is, in two hours. In this case the fol lowing arrangement will be adopted:

Two batteries will supply the lamps during the hours of greatest demand. Afterwards, one battery will be supplying current while the other is being charged; then this battery will supply current and the first will be charged. Under these conditions, the two batteries must together be able to supply 140,000 watts maximum at a given moment. Taking a capacity of 400,000 watt-hours for each battery, be certain of proper working.

we

shall

The price of these two batteries will be 86,000f. (£3,440). The efficiency of a good accumulator in actual work may be taken at 70 per cent. It will be necessary, * Translated from L'Electricien,

TO OUR READERS.

Coppered Staples.-A correspondent would be glad to know where he can buy large coppered staples insulated inside at the bend by means of a piece of fibre which remains in owing to the spring of the staple.

OKONITE.

The world stands almost in need of another set of interpreters-interpreters of the balance-sheets of public companies-so that some glimmer of the actual state of affairs may be understanded of the multitude. While language may be regarded as the gift of Providence to enable us to hide our thoughts, it is certain that published balance-sheets are usually intended to hide the real state of affairs. We have been especially anxious to see the balance-sheet of the Okonite Company, because our experience of its officials has been anything but satisfactory; some of them seem altogether too big for their places, and have not that suaviter in modo necessary to propitiate men with whom they are brought into contact. The balance-sheet, as we expected, is a very faulty document, and notwithstanding the payment of a dividend, does not point to successful enterprise. It may be that the profits shown have been legitimately won it may be that they are shown by a mere juggling of bookkeeping. At any rate there is no company with so large a capital that ought to go about to get a "loan," or to have such a large amount due to creditors in cash and in bills payable. The cash at bankers and in hand is cash in hand because payment has been deferred, and such deferred payments, as everybody knows, means, in the first place, paying more for goods, and, in the second place, losing all discounts. A little examination of the balance-sheet will show that the company has not a single penny of working capital, and that it is wholly in the hands of its creditors. Let us see. Capital (preference and ordinary shares)...... £336,590 Debentures... Loan

82,850 4,785 £424,225

The amount spent in buying the business, however, if we interpret aright, includes £52,291 value of plant at 1st January, 1890, and this should be deducted from the above total, as it also appears in the £82,771. We are told that the value of plant, machinery, stock-in-trade at time of purchase was £107,733, so that if the plant, etc., was worth £52,291, the value of the stock-in-trade at the time of purchase could not exceed £55,442. It is now estimated at £126,901, and if this estimate is too high it can easily be seen where the profit is made.

Going back to capital and outlay, we get a difference in favour of capital (allowing £52,291 appearing twice) of £43,620. The whole of this is used up in adding to stock, and a good deal more