been consistently designed for the power-factor at which it is operating, the different parts of the system are unequally loaded. If the engines are underloaded the possible revenue is curtailed; if the engines are carrying their rated load the electrical part of the system is overloaded, with attendant troubles and poor service to the customers.

Voltage Regulation:

The second injurious effect of low power-factor is poor voltage regulation. This effect is well known and requires no extended explanation. It will suffice, in the present paper, to show how large this effect may be, considering in turn each part of the electrical system.

Considering, first, the generators-it has been the general practice up to this time to design generators for an inherent regulation of approximately 8 per cent at their rated load and 100 per cent power-factor. There is, however, a tendency toward higher regulation percentages, particularly in large units, since

considerable reduction in cost can be made without any corresponding sacrifice in practical operating performance. However, with a generator having 8 per cent regulation at 100 per cent power-factor, with the same current and 80 per cent powerfactor the regulation, in the majority of cases, will be between 20 per cent and 25 per cent, varying in different generators on account of more or less magnetic saturation. It is worth while noting that the more saturated a generator is, the better regulation it will have, particularly at low power-factors; so that, while apparently the saturated generator is the better unit on account of the better regulation, it is, in fact, the poorer unit, since it will not hold up its voltage under heavy loads so well as the more nearly unsaturated generator. This will be referred to later. At the same kilowatt load and 80 per cent power-factor (which means 25 per cent overload in current) the regulation will be between 25 per cent and 30 per cent. The importance of inherent regulation of generators from an operating standpoint is minimized by the ease with which the excitation can be increased and thus compensate for the drop in voltage. This is particularly true provided an automatic voltage regulator is installed. Similar compensation in transformers or feeders requires relatively expensive auxiliary apparatus.

In transformers of 60 cycles and in fairly large units-say between 200-kw and 1000-kw-the regulation at 100-per cent power-factor and normal rated current will be between 1.6 per cent and 0.8 per cent. With the same current and 80-per cent power-factor the regulation will be between 4 per cent and 2.5 per cent. At 25-per cent overload in current and 80-per cent power-factor, which is equivalent to normal load in kilowatts and 80-per cent power-factor, the regulation will be between 5 per cent and 3 per cent.

In small lighting transformers-up to 50-kw capacity-the regulation will be poorer than in the larger transformers at 100-per cent power-factor and very nearly the same at 80per cent power-factor. At 100-per cent power-factor the regulation of small transformers will lie btween 2.7 per cent and 1.3 per cent. At normal rated current and 80-per cent power-factor the regulation will lie between 3.9 per cent and 2.5 per cent, and at 25-per cent current overload and 80-per cent power-factor the regulation will lie between 5 and 3 per cent.

The regulation of feeders is similarly affected by low powerfactor, but the magnitude of the effect varies widely on account of the varying ratio between the resistance and reactance drops in different cases. With equal resistance and reactance drops and a total drop of 10 per cent at 100-per cent power-factor, the drop at the same current and 80-per cent power-factor will be 13.5 per cent and the drop at the same kilowatts and 80-per cent power-factor will be 16.5 per cent. With a reactance drop double the resistance drop and a total drop of 10 per cent at 100-per cent power-factor, the drop at the same current and 80-per cent power-factor will be 17 per cent and the drop at the same kilowatts and 80-per cent power-factor will be 23 per cent. With 60 cycles and a spacing of about twelve inches between wires, the reactance and resistance become equal with about No. o B. and S. gauge conductors. For larger conductors the reactance is greater and for smaller conductors the resistance is greater. These various figures, showing the magnitude of the effect of power-factor on regulation, are summarized in the following table:

While these figures are necessarily approximate, they at least indicate the importance of power-factor in maintaining good service.

It has been previously pointed out that the importance of inherent regulation of generators is minimized by the ease with which drop in generator voltage can be compensated for by increasing the field current. This is only true provided the field current can be sufficiently increased with the available exciting voltage. In any generator there is a maximum field current that can be obtained, determined by the resistance of the field winding and the available exciting voltage. If the load and power-factor require a field current in excess of this maximum, it is obvious that the voltage will fall in spite of anything the station operators can do. The field current required depends somewhat on the current load but largely on the power-factor. Unless a generator has been designed to carry loads of low power-factor the maximum field current obtainable will be insufficient to maintain the normal voltage under this load condition. If a generator has a saturated magnetic circuit the increase in field current required for loads of low power-factor is much greater than if there were less saturation. For this reason an unsaturated generator will give better operating performance, because it can carry heavier inductive loads, even though the percentage regulation would be lower than in a similar generator with more saturation. With a generator not suited for loads of low power-factor, either by reason of insufficient field current or saturation, it is possible, then, to have a serious drop in voltage in the generator, due to low power-factor.

A caution is needed, in considering the bad effects of low

power-factor, against concluding that a high power-factor is always better than a low power-factor without considering the means by which the higher power-factor is obtained. The power-factor may be raised by decreasing the inductive load or by increasing the energy load. High power-factor obtained by decreasing the inductive load will always mean a gain; high power-factor obtained by increasing the energy load will only mean a gain provided the increase in energy load is accompanied by a corresponding increase in revenue. An increase in powerfactor obtained by increasing the energy load by increasing the losses in the system simply means an increased load on the entire system with no compensating advantage.

APPLICATION OF SYNCHRONOUS MOTORS TO POWER-FACTOR

CORRECTION

Enough has been presented to show the desirability of operating a system at high power-factor. But many classes of load, particularly constant-current transformers for series arc lamps and small induction motors, particularly when underloaded, have inherently a low-power-factor, on account of their inductive effect, due mainly to magnetizing current; and with such loads, high-power-factor can only be obtained by adding to the system an additional source of magnetizing current that will supply magnetizing current to the inductive apparatus. The synchronous motor presents a practicable means for raising the power-factor in this way. That central-station managers recognize the value of the synchronous motors in this field is shown by the increasing use of synchronous motors in motor-generator sets, in industrial applications that require large units, and, in particular cases, by their use running without energy load, and entirely for their corrective effect, as synchronous condensers.

Where there is a suitable load for a synchronous motor there is now very little question regarding the advisability of its installation in cases where improvement in power-factor is a consideration. The similar use of over-excited synchronous motors, running without load as synchronous condensers, is not so generally recognized as justifiable. The expense connected with their installation is easily seen, while the benefits to be derived are not so obvious.

There are three cases, however, in which the use of synchronous condensers can be shown to be good practice:

1. In existing systems that are up against trouble due to low power-factor, which can be remedied only by the installation of synchronous condensers. This is, obviously, no case for argument. This is, undoubtedly, the first field that the synchronous condenser will occupy and, in fact, the synchronous condenser is already in the field-two large installations, at least, being under construction.

2. In proposed systems which will have a large inductive load and in which the investment for feeders and transforming apparatus will be a large proportion of the total cost.

3. In proposed systems which will have a large inductive load and in which slow-speed generating units-such as gasdriven units-will be used.

The field of application of the synchronous condenser will be largely confined to these three cases. Under all other conditions the additional capacity made necessary by low power-factor can be most economically provided by increasing the size of the main generators and other electrical apparatus, so that the entire system, engines, generators, transformers and feeders, will be proportionately loaded at the power-factor at which the system will operate.

Before it can be shown why the use of synchronous condensers is justifiable under the above conditions it will be necessary to review the principles involved in the operation and application of over-excited synchronous motors.

The action of the over-excited synchronous motor in raising the power-factor of a circuit may be looked at from two points of view. It may be considered that the over-excited synchronous motor furnishes directly the magnetizing current required by the inductive load. For example, referring to Figure I, and assuming the synchronous condenser not connected to the circuit, the magnetizing current required by the inductive load must be supplied by the main generator, and all the line between the generator and the inductive load, as well as the generator itself, will be loaded up with this magnetizing or wattless current (so called because it represents no energy and does not affect a wattmeter reading). Now, assume that the synchronous condenser is in proper operation. Then this magnetiz