236 ABSTRACTS OF PAPERS

It is concluded that the LC oscillator should be regarded ascomplementary rather than a rival to the crystal oscillator.

In discussing the form of regenerative amplifier to be employed,the importance of phase and phase-stability is mentioned, andreference is made to a spurious temperature coefficient of fre-quency which may be caused by incorrect phasing. BriefSections deal with the precautions needed to minimize the harm-ful effects of harmonics, space charges, heater-cathode capacitance,microphony, switch-on changes in valves, and faulty design ofbuffer stages.

A Section is devoted to the LC resonator, and the choice ofsuitable lumped components is surveyed from the points of viewof temperature, humidity, pressure and mechanical vibration.This Section points out that whilst many earlier workers haveassumed that temperature effects only need be considered athumidities well below the dew-point, this is by no means thecase; indeed, it now appears that progress in the direction ofhigher frequency stabilities is mainly limited by the changes of

capacitance arising from variations of humidity. The mechanismof these capacitance changes includes adsorption effects whichare serious and not easy to forecast quantatively.

The permittivity of moist air (which may form the effectivedielectric of a capacitor of large air-gap) is however amenable tocalculation, and the Figure indicates quantities applicable to apressure of 760 mm Hg.

The analysis from which this diagram is obtained includes a for-mula from which diagrams for other pressures can easily be drawn.

In discussing the subject of mechanical vibration, a distinctionis made between the requirements of transport and those of use,and it is suggested that mere massiveness of design is a mistake.

The design of fixed and tapped inductors is considered, andone solution of the problem of the temperature compensationof tapped inductors is indicated. Some experimental resultswith an inductor for the frequency range 190-570 k/cs show thateffective temperature coefficients less than 5 in 106 per deg C areobtainable on all tappings.

THE GENERATION AND REGULATION OF ELECTRICAL POWER IN AIRCRAFT:DESIGN FEATURES OF GENERATORS AND THEIR CONTROL

By I. O. HOCKMEYER, B.Sc.Tech., Member.*

A SURVEY OF

(ABSTRACT of an Installations Section paper which was published in February, 1946, in Part II of the Journal.)

HISTORICAL REVIEWThe first part of the paper records the development of the

generation of d.c. power in aircraft by windmill and engine-driven generators from 1914 to the present day. Some mentionis made of generators designed for power supply to radio equip-ment, as distinct from general power services. These generators,which include high-voltage d.c. machines and high-frequencya.c. machines, have been combined with low-voltage d.c. machinesboth in tandem and with a common magnet system. The firstgenerator fitted to British aircraft was a windmill-driven 600-volt60-milliamp machine, but windmill drives were discarded infavour of drives from the main engines, in 1934. Engine-drivengenerators have increased in capacity from 12 volts 500 watts to24 volts 6 kW, the corresponding weights being 22 and 56 lb.The voltage originally adopted for the d.c. general-service supplywas 12 volts; this was subsequently changed, in 1936, to 24 volts(nominal), in order to save weight in commutators and cables.

On the introduction of engine-driven generators, the radiopower supplies were provided by rotary transformers fed fromthe general-service system. However, when, in 1939, radioequipment required an a.c. supply, engine-driven generators weredesigned to give an output of 500 watts at 80 volts, 1 200-2 400 c/s (proportional to engine speed), single phase. Con-siderable improvements in the design of these inductor alternatorsare recorded in the paper; the original 500-watt machine weighed29 lb, whereas the latest 1 200-watt generator weighs 20 lb.

GENERATORSThe second part of the paper discusses features of generator

design which have called for special consideration, or have beenthe subject of failure. The speed of an engine-driven generatormust necessarily be proportional to the speed of the main engineand will vary over a corresponding range. The speed rangeadopted for the first generator was 3 800-6 000 r.p.m., thoughthe speed could rise to 7 500 r.p.m. under aircraft dive condi-tions, and an overspeed test of 9 000 r.p.m. was applied. It wassubsequently found necessary to reduce the minimum figure to3 000 r.p.m. in order to maintain output at most economical

* Ministry of Aircraft Production.

cruising speeds. For large aircraft, however, a range of 3 000-4 800 r.p.m. (6 000 r.p.m. dive) has been adopted.

The most important item of electrical design discussed isassociated with the phenomenon of rapid brush wear, whichoccurs when aircraft operate at over 25 000 ft. No completetheoretical explanation has yet been accepted, despite a vastamount of research work conducted mainly in America. Apartial cure has been found by limiting the brush current-densityto 140 amp/in2 and by the use of special treatments, includingthe impregnation of the brushes with inorganic materials—forexample sulphur—and organic materials such as oils.

The general-service system includes an accumulator connectedacross the lines. The generator is switched on to the linethrough a cut-out which opens when reverse current flows.Cases are described where the generator polarity reversed, owingto the effect of this reverse current. It is explained how thisreversal will occur with compound-wound generators, or whenthe brush position is set back from neutral, owing to the currentthrough the machine building up a reverse field greater than thatdue to the diminishing shunt field.

Reference is made to the importance of the saving of weightin aircraft electrical equipment. In this connection it is evidentthat efficient methods of cooling generators must be adopted;consequently use is made of the aircraft slip stream. Earlyengine-driven generators were jacket (frame) cooled, but morerecent types have been "through" cooled (pipe-ventilated). Thechange was made following a relaxation in the flameproofing re-quirements, when it was appreciated that brush sparking couldnot ignite an explosive mixture in the engine nacelle: the generatorwould be scavenged by clean outside air before the generatorvoltage had built up. Tests and experience indicated that an airfilter was an unnecessary refinement. The quantity of coolingair required varies between 50 and 100 ft3/min, according to thetype of generator. The effectiveness of cooling air varies withaltitude; as altitude increases, the weight of air which passesthrough the generator is reduced, but, conversely, the ambienttemperature falls. Up to 35 000 ft these opposing tendenciesapproximately balance; above 35 000 ft cooling is less effective.

Engine-driven generators are flange-mounted. The same

ABSTRACTS OF PAPERS 237

flange, with a 4f in diameter spigot, has been adopted on thewhole range of generators for the R.A.F. For the larger sizes,however, the number of fixing holes has been increased from 4 to8 and the shaft end has been changed from $ in, with 6 shallowsplines, to i in with V-serrations. The retention of the originalfixing-bolt-hole pitch circle (6 in) necessitates the adoption of anecked frame to provide access to the nuts, thus introducing amechanical weakness and increasing generator weight.

The ball bearings of engine-driven generators have been thecommonest cause of failure, which has been attributed to excessiveradial loads due to the type of drive employed and to over-greasing. The remedies described in the paper include a changeto roller bearings at the commutator end, the selection and useof special lubricants, the control of diametrical bearing clearanceand, finally, the prohibition of certain types of bearing cage.

Aluminium alloys have been used for the end-frames at thedriving end of generators, though the design has been influencedby the provision of the neck, to which reference has already beenmade. In most designs the neck formed part of the casting,but in one case the yoke of the machine was "bottled" (Fig. 1)

line voltage required for consuming devices, and at the same timepermit of adequate control of accumulator charging current. Italso describes various systems of paralleling and their effect online voltage.

The first wind-driven generator and the first engine-drivengenerator were both regulated on the same principle. Anauxiliary 6-volt output was provided in the former for this pur-pose; the latter was a compound-wound machine with an addi-tional "control" field winding, connected in series with theaccumulator and acting in opposition to the shunt and seriesfields. The series field catered for change in- load, and thecontrol field for change in speed, the charging current being afunction of speed and increasing from zero at "cut-in" speed to6 amp. at 6 000 r.p.m. (The polarity reversal problem did notarise with this system.) This system had to be discontinuedowing to its failure to maintain accumulators in a chargedcondition (Fig. 2).

Some early wind-driven radio d.c. generators relied on manualvoltage control. This was not unreasonable as the armaturespeed was kept constant by the use of a governed variable-pitch

ACKING WASHER

AIR PIPE INLET

CIRCLIP

TERMINALCOVER

NNER BEARINGCOVER

:R BEARINGEND PLATE

FELT

• RETAININGWASHER

7INNER BEARING COVER

-AIR PIPE OUTLET

Fig. 1** Cfrown copyright: reproduced by permission of the Controller of H.M. Stationery Office.

and the end-frame was in the nature of a flat plate. For thecommutator end-frame, both steel pressings and aluminium-alloy castings have been used. In some designs the commutatorand brush gear were surrounded by an extension of the yoke,and in others by the end-frame.

VOLTAGE REGULATION

The third part of the paper deals with voltage regulation insome detail, and, more briefly, with various types of regulator.It shows how the control of the system voltage has been condi-tioned by the inclusion of an accumulator, and why it has notbeen possible to devise a system which would give the constant

windmill, or by interposing a governed slipping clutch betweengenerator and windmill. Later, the conventional type of third-brush dynamo was introduced, but as this was suitable only forsmall machines it soon gave place to a regulator of the Tirrilltype. The Tirrill regulator was discarded, mainly on accountof the radio interference which it set up, and because its contactswere liable to pitting, burning and welding.

The self-regulating generator system mentioned above wassuperseded, for engine-driven generators, when the carbon-pilevoltage regulator became available. This type of regulator hasalso been used for the control of engine-driven inductor alter-nators. Two basic types have been adopted for Service use,

238 ABSTRACTS OF PAPERS

Fig. 2

namely the Stone and, later, the Newton type. In both types, apile of carbon discs is connected in the generator shunt-fieldcircuit; the resistance of this pile is automatically controlled bythe generator output, by the excitation of a magnet to balancethe spring which compresses the pile. The magnetic force in theStone type is utilized through a torque-motor device, whereasthe Newton regulator is of the nature of a direct-acting solenoid.In both, the art of design lies in the means of balancing magnetpull against mechanical force. Stabilizing against hunting isaccomplished in the Stone regulator by the use of a dashpot,and in the Newton type, where necessary, by special stabilizingwindings in the regulator.

an accumulator can reach a condition of age and charge atwhich it may accept, indefinitely, a heavy charging current,causing violent gassing and possibly explosion of the accumulator!These schemes are designed to prevent such an occurrence. Oneis for a single-generator installation; no separate current-limitingregulator is employed, but a coil is wound on the regulator inseries with the accumulator. Line voltage is thus unaffected byload changes, but varies slowly with accumulator condition.The other scheme, for the multi-engined case, comprises onemain voltage regulator for each generator, each with a loadseries coil to provide stable parallel operation. The regulatorvoltage coils are paralleled and connected to the pile of a master

Fig. 3

Reverting to the effect of the accumulator on line voltage, itis pointed out by the author that the line voltage is directly relatedto the accumulator in the third-brush and self-regulating systemsof regulation. The Tirrill and carbon-pile regulator schemes are,of course, generally similar in this respect, though it is to thelatter that most thought has been given, since other complica-tions, such as paralleling of generators, have arisen since itsadoption. A carbon-pile regulator can control the voltagebetween ± 3 % of the specified value. In an aircraft, however,the line voltage will drop to the value of that of a dischargingaccumulator when the aircraft engines are not running. Thisdifficulty has never been overcome in practice. When the Stoneregulator was introduced the relative generator, accumulator andload capacities were such that conditions could arise in whichthe generator could be seriously overloaded; a current-limitingdevice was therefore introduced. This took the form of a seriescoil on the regulator which lowered the line voltage with increaseof load current.

This, however, had certain limitations; on the Newton regu-lator, current limitation was provided by the inclusion of a com-plete additional unit which lowered the voltage on overload only.By this time paralleling of generators was a requirement, andstable parallel operation was obtained by providing a series coilon the voltage regulator to give a sloping characteristic at theexpense of constant voltage (Fig. 3).

The latest schemes show some improvement, but still permitthe accumulator to vary the supply voltage. It is assumed that

regulator which is controlled by the line voltage. This masterregulator has a coil in the accumulator circuit so that the resultis similar to the single generator case (Fig. 4).

Fig. 4

CONCLUSION

The paper concludes by suggesting that improved power/weightratios and reliability will be achieved in future, and by welcomingthe introduction of an aircraft auxiliary generating plant as asuccessor to the accumulator, with the possibility of obtaining atruly constant-voltage system.