Electric clocks may be divided into three kinds, or principal divisions. Of the first-class are those in which the pendulum is driven directly from the armature by electric impulse, or by means of a weight dropping on an arm projecting from the pendulum. In this case the entire train of the clock consists of a ratchet wheel and the dial work.

The second class comprises the regular train from the center to the arbor. This class has a spring on the center arbor, wound more or less frequently by electricity. In this case the aim is to keep the spring constantly wound, so that the tension is almost as evenly divided as with the ordinary weight clock, such as is used in jewelers’ regulators.

The third system uses a weight on the end of a lever connected with a ratchet wheel on the center arbor and does away with springs. One type of each of these clocks will be described so that jewelers may comprehend the principles on which the three types are built.

In the Gillette Electro-Automatic, which belongs to the class first mentioned, the ordinary clock principle is reversed. Instead of the works driving the pendulum, the pendulum drives the train, through the medium of a pawl and ratchet mechanism on the center arbor. The pendulum is kept swinging by means of an impulse given every beat by an electromagnet. This impulse is caused by the weight of the armature as it falls away from the magnet ends, the current being used solely to pull back and re-set the armature for the next impulse. Any variation in the current, therefore, does not affect the regulation of the clock, as the power is obtained from gravity only, by means of the falling weight. Referring to the drawings, Figs. 123 and 124, it is seen that each time the pendulum swings the train is pushed one tooth forward. A cam is carried by the ratchet (center) arbor in which a slot is provided at a position equivalent to every fifth tooth of the ratchet. Into this slot drops the end of a lever, releasing at its other end the armature prop. Thus at the next beat of the pendulum the armature is released and in its downward swing impulses the pendulum, giving it sufficient momentum to carry it over the succeeding five swings.

The action of the life-giving armature is entirely disconnected and independent of the clock mechanism. It acts on its own accord when released every tenth beat and automatically gives its impulse and re-sets itself. It is provided with a double-acting contact spring (see Fig. 125) which “flips” a contact leaf from one adjustable contact screw to the other as the action of the armature causes the spring to pass over its dead center. Thus, when the armature reaches the lowest point in its drop (Figs. 126 and 127) the leaf snaps against the right contact screw, the circuit is completed, the magnet energized and the armature drawn up. As the armature rises above a certain point, the dead center of the flipper spring is again crossed and the leaf snaps back against the post at the left. In the meantime, however, the armature prop has slipped under the end of the armature and retains it until the time comes for the next impulse.

In adjusting the mechanism of this type of clock the increasing pendulum swing should catch and push the ratchet before the buffer strikes and lifts the armature from the prop. The adjustment of the “flipper” contact screws (with ¹/12 inch play) should be such that as the armature falls the contact leaf will be thrown and the armature drawn up at a point just beyond the half way position in the swing of the pendulum. The power of the impulse can be regulated by turning the adjusting post with pliers, thus varying the tension of the armature spring, the pull of which reinforces the weight of the armature. Care should be taken, however, that the tension is not beyond the “quick action” power of the electromagnet. It is much better to ease up the movement in other ways before putting too great a load on the life of the battery.

The electrical contacts on the leaf and screw are platinum tipped to prevent burning by the electric sparking at the “make” and “break.” This sparking is also much reduced by means of a resistance coil placed in series connection with the magnet coil, Fig. 127, to reduce the amount of current used. If this coil is removed or disconnected the constant sparking and heat would soon burn out the contact tips.

Care should be taken to see that the batteries are dated and the battery connections are clean at the time of sliding in a new battery. The brush which makes connection with the center or carbon post of the battery is insulated with mica from the framework of the case. The other connection is made from the contact of the uncovered zinc case of the battery with the metal clock case surrounding it. The contact points should be bright and smooth to insure good contacts.

These clocks need but little cleaning of the works as no oil whatever is used, except at one place, viz., the armature pivot. Oil should never be used on the train bearings, or other parts. This clock ran successfully on the elevated railway platforms of the loop in Chicago where no other pendulum clock could be operated on account of the constant shaking.

In considering the electrical systems of these clocks, let us commence with the batteries. While undoubtedly great improvements have been made in the present form of dry battery they are still very far from giving entire satisfaction. Practically all of them are of one kind, which is that which produces electricity at 1½ volts from zinc, carbon and sal-ammoniac, with a depolarizer added to the elements to absorb the hydrogen. The chemical action of such a battery is as follows:

The water in the electrolite comes in contact with the zinc and is decomposed thereby, the oxygen being taken from the water by the zinc, forming oxide of zinc and leaving the hydrogen in the form of minute bubbles attached to the zinc. As this, if allowed to stand, would shut off the water from reaching the zinc, chemical action would therefore soon cease and when this happens the battery is said to be polarized and no current can be had from it.

In order to take care of the hydrogen and thus insure the constant action of the battery, oxide of manganese is added to the contents of the cell, generally as a mixture with the carbon element. Manganese has the property of absorbing oxygen very rapidly and of giving it off quite easily. Therefore while the hydrogen is being formed on the zinc, it becomes an easy matter for it to leave the zinc and take its proper quantity of oxygen from the manganese and again form water, which is again decomposed by the zinc. As long as this cycle of chemical action takes place the battery will continue to give good satisfaction, and usually when a battery gives out it is because the depolarizer is exhausted, for the reason that the carbon is not affected at all and the zinc element forming the container is present in sufficient quantity to outlast the chemical action of the total mass.

There are great differences in the various makes of batteries; also in the methods of their construction. It would seem to be an easy matter for a chemist to figure out exactly how much depolarizer would serve the purpose for a given quantity of zinc and carbon and therefore to make a battery which should give an exact performance that could be anticipated. In reality, however, this is not the case, owing to the various conditions. There are three qualities of manganese in the market; the Japanese, which is the best and most costly; the German, which comes second, and the American, which is the cheapest and varies in quality so much as to be more or less a matter of guesswork. We must remember that in making batteries for the price at which they are now sold on the market we are obliged to take materials in commercial quantities and commercial qualities and cannot depend upon the chemically pure materials with which the chemists’ theories are always formulated. This therefore introduces several elements of uncertainty.

In practice the Japanese manganese will stand up for a far longer time than any other that is known and it is used in all special batteries where quality and length of life are considered of more importance than the price. The German manganese comes next. Then comes a mixture of American and German manganese, and finally the American manganese, which is used in making the cheaper batteries which are sorted afterwards, as we shall explain farther on. These batteries are sealed after having been made in large quantities, say five thousand or ten thousand in the lot, and kept for thirty days, after which they are tested. The batteries which are likely to give short-life will show a local action and consequent reduction of output in thirty days. They are, therefore, sorted out, much as eggs are candled on being received in a storage warehouse, for the reason that after a cell has been made and put together it would cost more to find out what was the matter with it and remedy that than it would to make a new cell. Many of the battery manufacturers, therefore, make up their batteries with an attempt to reach the highest standard. They are sorted for grade in thirty days and those which have attained the point desired are labeled as the factories’ best battery and are sold at the highest prices. The others have been graded down exclusively and labeled differently until those which are positively known to be short-lived are run out and disposed of as the factories’ cheapest product under still another label.

When buying batteries always look to see that the tops are not cracked, as if the seal on the cell is broken, chemical action induced from contact with the air as the battery dries out, will rapidly deteriorate the depolarizer and sulphate the zinc, both of which are of course a constant draft on the life of the battery, which contains only a stated quantity of energy in the beginning. Always examine the terminal connections to see that they are tight and solid.

Batteries when made up are always dated by the factory, but this does the purchaser little good, as the dates are in codes of letters, figures, or letters and figures, and are constantly changed so that even the dealers who are handling thousands of them are unable to read the code. This is done because many people are prone to blame the battery for other defects in the electrical system and many who are using great quantities would find an incentive to switch the covers on which the dates appear if they knew what it meant. This is perhaps rather harsh language, but a good many men would be tempted to send back a barrel of old batteries every now and then with the covers showing that they had not lasted three months, if they could read these signatures.

Practically the only means the jeweler has of obtaining a good cell, with long life, is to buy them of a large electrical supply house, paying a good price for them and making sure that that house has trade enough in that battery to insure their being continuously supplied with fresh stock.

The position of the battery also has to do with the length of life or amount of its output. Thus a battery lying on its side will not give more than seventy-five per cent of the output of a battery which is standing with the zinc and carbon elements perpendicular. Square batteries will not give the satisfaction that the round cell does. It has been found in practice by trials of numerous shapes and proportions that the ordinary size of 2½ × 6 inches will give better satisfaction than one of a different shape—wider or shorter, or longer and thinner; that is for the amount of material which it contains. The battery which has proved most successful in gas engine ignition work is 3¾ × 8 inches. That maintains the same proportions as above, or very nearly so, but owing to local action it will give on clock work only about fifty per cent longer life than the smaller size.

It has been a more or less common experience with purchasers of electric clocks to find that the batteries which came with the clock from the factory ran for two or three years (three years not being at all uncommon) and that they were then unable to obtain batteries which would stand up to the work for more than three weeks, up to six months. The difference is in the quality and freshness of the battery bought, as outlined above.

In considering the rest of the electrical circuit, we find three methods of wiring commonly used and also a fourth which is just now coming into use. The majority of electric clocks are wound by a magnet which varies in size from three to six ohms; bridged around the contact points, there has generally been placed a resistance spool which varies in size from ten to twenty-five times the number of ohms in the armature magnets. See Fig. 128. This practically makes a closed circuit on which we are using a battery designed for open circuit work.

If we use an electromagnet with a very soft iron core, we will need a small amount of current, but every time we break the contact, we will have a very high counter electro-motive force, leaping the air gap made while breaking the contact and therefore burning the contact points. If our magnet is constructed so as to use the least current, by very careful winding and very soft iron cores, this counter electro-motive force will be at its greatest while the draft on the battery is at its smallest. If the magnet cores are made of harder iron, the counter electro-motive force will be much less; but on the other hand much more current will be needed to do a given quantity of work with a magnet of the second description; and the consequence is that while we save our contact points to some extent, we deplete the battery more rapidly.

If we put in the highest possible resistance—that of air—in making and breaking our contacts, we use current from the battery only to do useful work; but we also have the spark from the counter electro-motive force in a form which will destroy our contact points more quickly. If we reduce the resistance by inserting a German silver wire coil of say sixty ohms on a six-ohm magnet circuit, we have then with two dry batteries (the usual number) three volts of current in a six-ohm magnet during work and three volts of current in a sixty-six ohm circuit while the contacts are broken, Fig. 128. Dividing the volts by the ohms, we find that one twenty-second of an ampere is constantly flowing through such a circuit. We are therefore using a dry battery (an open circuit battery) on closed circuit work and we are drawing from the life of our battery constantly in order to save our contact points.

It then becomes a question which we are going to sacrifice, or what sort of a compromise may be made to obtain the necessary work from the magnet and at the same time get the longest life of the contact points and the batteries. Most of the earlier electric clocks manufactured have finally arranged such a circuit as has been described above.

The Germans put in a second contact between the battery and the resistance with a little larger angular motion than the first or principal contact, so that the contact is then first made between the battery and resistance spool, B, Fig. 129, then between the two contact points of the shunt, A; Fig. 129, to the electromagnet, and after the work is done they are broken in the reverse order, so that the resistance is made first and broken after the principal contact. This involves just twice as many contact points and it also involves more or less burning of the second contact.

The American manufacturers seem to prefer to waste more or less current rather than to introduce additional contact points, as they find that these become corroded in time with even the best arrangements and they desire as few of them as possible in their movements, preferring rather to stand the draft on the battery.

One American manufacturer inserts a resistance spool of 60 ohms in parallel with a magnet of seven ohms (3½ ohms for each magnet spool) as in Fig. 130. He states that the counter electro-motive force is thus dissipated in the resistance when the contact is broken, as the resistance thus becomes a sort of condenser, and almost entirely does away with heating and burning of the contacts, while keeping the circuit open when the battery is doing no work.

It has been suggested to the writer by several engineers of high attainments and large experience that what should be used in the above combination is a condenser in place of a resistance spool, as there would then be no expenditure of current except for work. One of the clocks changed to this system just before the failure of its manufacturers, but as less than four hundred clocks were made with the condensers (Fig. 131), the point was not conclusively demonstrated.

It should also be borne in mind that the condenser has been vastly improved within the last twelve months. With the condenser it will be observed that there is an absolutely open circuit while the armature is doing no work and that therefore the battery should last that much longer, Figs. 130 and 131. As to the cost of the condensers as compared with resistance spools, we are not informed, but imagine that with the batteries lasting so much longer and the clock consequently giving so much better satisfaction, a slight additional cost in manufacture by changing from resistance to condensers would be welcomed, if it added to the length of life and the surety of operation.

Electric clocks cost more to make than spring or weight clocks and sell for a higher price and a few cents additional per movement would be a very small premium to pay for an increase in efficiency.

The repairer who takes down and reassembles one of these clocks very often ignorantly makes a lot of trouble for himself. Many of the older clocks were built in such a way that the magnets could be shifted for adjustment, instead of being put in with steady pins to hold them accurately in place. The retail jeweler who repairs one of these clocks is apt to get them out of position in assembling. The armature should come down squarely to the magnets, but should not be allowed to touch, as if the iron of the armature touches the poles of the magnet it will freeze and retain its magnetism after the current is broken. Some manufacturers avoid this by plating their armatures with copper or brass and this has puzzled many retailers who found an electromagnet apparently attracting a piece of metal which is generally understood to be non-magnetic.

The method offers a good and permanent means of insulating the iron of the armature from the magnet poles while allowing their close contact and as the strength of a magnet increases in proportion to the square of the distance between the poles and the armature, it will be seen that allowing the armature to thus approach as closely as possible to the poles greatly increases the pull of the magnet at its final point. If when setting them up the magnet and armature do not approach each other squarely, the armature will touch the poles on one side or another and soon wear through the copper or brass plating designed to maintain their separation and then we will have freezing with its accompanying troubles.

A very good test to determine this is to place a piece of watch paper, cigarette paper or other thin tissue on the poles of the magnet before the naked iron armature is drawn down. Then make the connection, hold the armature and see if the paper can be withdrawn. If it cannot the armature and poles are touching and means should be taken to separate them. This is sometimes done by driving a piece of brass into a hole drilled in the center of the pole of the magnet; or by soldering a thin foil of brass on the armature. As long as the separation is steadily maintained the object sought is accomplished, no matter what means is used to attain it.

Another point with clocks which have their armatures moved in a circular direction is to see that the magnet is so placed as to give the least possible freedom between the armatures and the circular poles of the magnet, but that there must be an air-gap between the armature and magnet poles.

In those clocks which wind a spring by means of a lever and ratchet working into a fine-toothed ratchet wheel, or are driven by a weighted lever, there is an additional point to guard against. If the weight lever is thrown too far up, either one of two things will happen. The weight lever may be thrown up to ninety degrees and become balanced if the butting post is left off or wrongly replaced; the power will then be taken off the clock, if it is driven directly by weight, so that a butting post should meet the lever at the highest point and insure that it will not go beyond this and thus lose the efficiency of the weight.

In the cases where a spring on the center arbor is interposed between the arbor and the ratchet wheel, it should be determined just how many teeth are necessary to be operated when winding, as if a clock is wound once an hour and the aim is to wind a complete turn (which is the amount the arbor has run down) if the lever is allowed to vibrate one or two teeth beyond a complete turn, it will readily be seen that in the course of time the spring will wind itself so tightly as to break or become set. This was a frequent fault with the Dulaney clock and has not been guarded against sufficiently in some others which use the fine ratchet tooth for winding.

When such a clock is found the proper number of teeth should be ascertained and the rest of the mechanism adjusted to see that just that number of teeth will be wound. If less is wound there will come a time when the spring will run down and the clock will stop. If too much is wound the spring will eventually become set and the clock will stop. Therefore such movements should be examined to see that the proper amount of winding occurs at each operation. Of course where a spring is wound and there are but four notches in the ratchet wheel and the screw stop is accurately placed to stop the action of the armature, over action will not harm the spring, provided it will not go to another quarter, as if the armature carries the ratchet wheel further than it should, the smooth circumference between the notches will let it drop back to its proper notch.

There are a large number of clocks on the market which wind once per hour. These differ from the others in that they do not depend upon a single movement of the armature for an instantaneous winding. Thus if the batteries are weak it may take twenty seconds to wind. If the batteries are strong and new it may wind in six seconds. In this respect the clock differs radically from the others, and while we have not personally had them under test, we are informed that on account of winding once per hour the batteries will last very much longer than would be expected proportionately from those which wind at periods of greater frequency. The reason assigned is that the longer period allows the battery to dispose of its hydrogen on the zinc and thus to regain its energy much more completely between the successive discharges and hence can give a more effective quantity of current for hourly discharge than those which are discharged several times a minute, or even several times an hour. It is only proper to add that the manufacturers of clocks winding every six or seven minutes dispute this assertion.

Another point is undoubtedly in the increased length of life of the contacts; but speaking generally the electric clock may be said now to be waiting for further improvements in the batteries. Those who have had the greatest experience with batteries, as the telephone companies, telegraph companies and other public service corporations, have generally discarded their use in favor of storage batteries and dynamos wherever possible and where this is not possible they have inspected them continuously and regularly.

In this respect one point will be found of great service. When putting in a new set of batteries in any electrical piece of machinery, write the date in pencil on the battery cover, so that you, or those who come after you, some time later, will know the exact length of time the battery has been in service. This is frequently of importance, as it will determine very largely whether the battery is playing out too soon, or whether faults are being charged to the battery which are really due to other portions of the apparatus.

Never put together any piece of electrical apparatus without seeing that all parts are solidly in position and are clean; always look carefully to connections and see that the insulation is perfect so that short circuits will be impossible.

All contacts must be kept smooth and bright and contact must be made and broken without any wavering or uncertainty.

Fig. 132 shows the completely wired movement of the American Clock Company’s weight driven movement, which may be accepted as a type of this class of movements—weight driven, winding every seven minutes.

The train is a straight-line time train, from the center arbor to the dead beat escapement, with the webs of the wheels not crossed out. It is wired with the wire from the battery zinc screwed to the front plate H and that from the battery carbon to an insulated block G.

Fig. 133 shows an enlarged view of the center arbor. Upon this arbor are secured (friction-tight) two seven-notched steel ratchets, E, and carried loosely between them are two weighted levers pivoted loosely on the center arbor. Each lever is provided with a pawl engaging in the notches of the nearest ratchet, as shown. The weighted lever has a circular slot cut in it, concentric with the center hole and also has a portion of its circumference at the arbor cut away, thus forming a cam. Between these two levers is a connecting link D with a pin in its upper end, which pin projects into the circular slots of the weight levers.

The lever F is pivoted to the front plate of the clock and carries at right angles a beveled arm which projects over the ratchets E, but is ordinarily prevented from dropping into the notches by riding on the circumferences of the weighted levers. When one lever has dropped down and the other has reached a horizontal position the cut portions of the circumferences of these levers will be opposite the upper notch of the ratchets and will allow the bar projecting from F to drop into the notches. This allows F and G to connect and the magnet A is energized, pulls the armature B, the arms C D, and thus lifts the lever through the pin in D pulling at the end of the circular slot. As the lever flies upward, the cam-shaped portion of its circumference raises the arm out of the notches, thus separating F and G and breaking the circuit. A spring placed above E keeps its arms pressed constantly upon E in position to drop. The wiring of the magnets is shown in Fig. 130.

The upper contact (carried in F) is a piece of platinum with its lower edge cut at an angle of fifteen degrees and beveled to a knife edge. The lower point of this bevel comes into contact first and is the last to separate when breaking connection, so that any sparking which may take place will be confined to one edge of the contacts while the rest of the surface remains clean. (See Fig. 134.) Ordinarily there is very little corrosion from burning and this is constantly rubbed off by the sliding of the surfaces upon each other. The lower contact, G, consists of a brass block mounted upon an insulating plate of hard rubber. The block is in two pieces, screwed together, and each piece carries a platinum tipped steel spring. These springs are so set as to press their platinum tips against each other directly beneath the upper contact. The upper and lower platinum tips engage each other about one-sixteenth inch at the time of making contact. The lower block being in two pieces, the springs may be taken apart for cleaning, or to adjust their tension. The latter should be slight and should in no case exceed that which is exerted by the spring in F, or the upper knife edge will not be forced between the two lower springs. The pin on which F is pivoted and that bearing on the spring above it must be clean and bright and never be oiled, as it is through these that the current passes to the upper contact in the end of F. The contacts are, of course, never oiled.

The two weighted levers should be perfectly free on the center arbor and their supporting pawls should be perfectly free on the shoulder screws in the levers. Their springs should be strong enough to secure quick action of the pawls. This freedom and speed of action are important, as the levers are thrown upward very quickly and may rebound from the butting post without engaging the ratchets if the pawls do not work quickly.

The projecting arm, C, of the armature, B, has pivoted to it, a link, D, which projects upward and supports at its upper end a cross pin. The link should not be tight in the slot of C, but should fit closely on the sides, in order to keep the cross pin at the top of D parallel with the center staff of the clock. This cross pin projects through D an equal distance on either side, each end respectively passing through the slot of the corresponding lever, the total length of this pin being nearly equal to the distance between the ratchets. When the electric circuit is closed, and the magnets energized, B, C and D are drawn downward; the weighted end of one of the levers which runs the clock, being at this time at the limit of its downward movement, see Fig. 135, the opposite or slotted end of said lever, is then at its highest point, and the downward pull in the slot by one end of the above described crosspin which enters it will throw the weighted end of the said lever upward. The direct action of the magnets raises the lever nearly to the horizontal position, and the momentum acquired carries it the remainder of the distance. By this arrangement of stopping the downward pull of the pin when the ascending lever reaches the horizontal, all danger of disturbing the other lever A is avoided. The position is such that the top of the ascending lever weight is about even with the center of the other weight when the direct pull ceases.

Before starting the clock raise the lever weights so that one lever is acting upon a higher notch of the ratchet than the other. They are designed to remain about forty-five degrees apart, so as to raise only one lever at each action of the magnet. This maintains an equal weight on the train, which would not be the case if they were allowed to rise and fall together; keeping the levers separated also reduces the amount of lift or pull on the battery and uses less current, which is an item when the battery is nearly run down. If these levers are found together it indicates that the battery is weak, the contacts dirty, making irregular winding, or the pawls are working improperly. See that the levers rise promptly and with sufficient force. After one of them has risen stop the pendulum and see that the butting post is correctly placed, so that there is no danger of the lever wedging under the post and sticking there, or causing the lever to rebound too much. The butting post is set right when the clock leaves the factory and seldom needs adjustment unless some one has tinkered with it.

The time train should be oiled as with the ordinary movements, also the pawls on the levers. The lever bushings should be cleaned before oiling and then well oiled in order to avoid friction on the center arbor from the downward pull of the magnets when raising the levers. In order to clean the levers drive out the taper pin in the center arbor and remove the front ratchet, when the levers will slip off. In putting them back care should be used to see that the notches of the ratchets are opposite each other. Oil the edges of the ratchets and the armature pins. Do not under any circumstances oil the contact points, the pins or springs of the bar F, as this will destroy the path of the current and thus stop the clock. These pins must be kept clean and bright.

Hourly Winding Clocks.—There are probably more of these in America than of all other electric kinds put together (we believe the present figures are something like 135,000), so that it will not be unreasonable to give considerable space to this variety of clocks. Practically all of them are made by the Self Winding Clock Company and are connected with the Western Union wires, being wound by independent batteries in or near the clock cases.

Three patterns of these clocks have been made and we will describe all three. As they are all practically in the same system, it will probably be better to first make a simple statement of the wiring, which is rigidly adhered to by the clock company in putting out these goods. All wires running from the battery to the winding magnets of the movement are brown. All wires running from the synchronizing magnet to the synchronizing line are blue. Master clocks and sub-master clocks have white wires for receiving the Washington signal and the relay for closing the synchronizing line will, have wires of blue and white plaid.

By remembering this system it is comparatively easy for a man to know what he is doing with the wires, either inside or outside of the case. For calendar clocks there are, in addition, two white wires running from the calendar to the extra cell of battery. There is also one other peculiarity, in that these clocks are arranged to be wound by hand whenever run down (or when starting up) by closing a switch key, shown in Fig. 136, screwed to the inside of the case. This is practically an open switch, held open by the spring in the brass plate, except when it is pressed down to the lower button.

The earliest movement of which any considerable number were sent out was that of the rotary winding from a three-pole motor, as shown in Fig. 137. Each of these magnet spools is of two ohms, with twelve ohms resistance, placed in parallel with the winding of each set of magnet spools, thus making a total of nine spools for the three-pole motor.

On the front end of the armature drum arbor is a commutator having six points, corresponding to the six armatures in the drum. There are three magnets marked O, P and X; each magnet has its own brush marked O', P' and X'. When an armature approaches a magnet (see Fig. 137) the brush makes contact with a point of the commutator, and remains in contact until the magnet has done its work and the next magnet has come into action. When properly adjusted the brush O' will make contact when armatures 1 and 2 are in the position shown, with No. 2 a little nearer the core of the magnet than No. 1; and it will break contact when the armature has advanced into the position shown by armature No. 3, the front edge of the armature being about one-sixteenth of an inch from the corner of the core, armature No. 4 being entirely out of circuit, as brush X' is not touching the commutator.

The back stop spring, S, Fig. 137, must be adjusted so that the brush O' is in full contact with a point of the commutator when the motor is at rest, with a tooth of the ratchet touching the end of the spring, S.

Sometimes the back stop spring, S, becomes broken or bent. When this occurs it is usually from overwinding. It must be repaired by a new spring, or by straightening the old one by burnishing with a screwdriver. Set the spring so that it will catch about half way down the last tooth.

Having explained the action of the motor we come now to the means of temporarily closing the circuit and keeping it closed until such time as the spring is wound a sufficient amount to run the clock for one hour; as the spring is on the center arbor this requires one complete turn.

This is the distinguishing feature of this system of clocks and is not possessed by any of the others. It varies in construction in the various movements, but in all its forms it maintains the essential properties of holding the current on to the circuit until such time as the spring has been wound a sufficient quantity, when it is again forcibly broken by the action of the clock. This is termed the “knock away,” and exists in all of these movements.

To start the motor the circuit is closed by a platinum tipped arm, A, Fig. 138, loosely mounted on the center arbor, and carried around by a pin projecting from the center wheel until the arm is upright, when it makes contact with the insulated platinum tipped brush, B. A carries in its front an ivory piece which projects a trifle above the platinum top, so that when B drops off the ivory it will make contact with the platinum on A firmly and suddenly. This contact then remains closed until the spring barrel is turned a full revolution, when a pin in the barrel cover brings up the “knock-away,” C, which moves the arm, A, forward from under the brush, B, and breaks the circuit. The brush, B, should lie firmly on its banking piece, and should be so adjusted that when it leaves the arm, A, it will drop about one-thirty-second of an inch. Adjusted in this way it insures a good, firm contact.

The angle at the top of the brush, B, must not be too abrupt, so as to retard the action of the clock while the contact is being made. Wire No. 8 connects the spring contact, B, to one of the binding plates at the left hand side of the case; and wire No. 6 connects the motor, M, to the other. To these binding plates are attached brown wires that lead one to each end of the battery.

When the clock is quite run down, it is wound by pressing the switch key, Fig. 136, from which a wire runs to the plate. The switch key should not be permanently connected to its contact screw, J. See that all wires are in good condition and all connections tight and bright. The main spring is wound by a pinion on the armature drum arbor, through an intermediate wheel and pinion to the wheel on the spring barrel.

At stated times—say once in eighteen months or two years—all clocks should be thoroughly cleaned and oiled, and at the same time inspected to be sure they are in good order.

Never let the self-winding clocks run down backward, as the arm, A, Fig. 138, will be carried back against the brush, B, and bend it out of adjustment.

To clean the movement, take it from the case, take out the anchor and allow it to run down gently, so as not to break the pins, then remove the motor. Take off the front plate and separate all the parts. Never take off the back plate in these clocks. Wash the plates and all parts in a good quality of benzine, pegging out the holes and letting them dry thoroughly before reassembling. The motor must not be taken apart, but may be washed in benzine, by using a small brush freely about the bearings, commutator and brushes. Put oil in all the pivot holes, but not so much that it will run. The motor bearings and the pallets of the anchor should also be oiled.

Inspect carefully to see that the center winding contact is right and that the motor is without any dead points. Dust out the case and put the movement in place. Before putting on the dial try the winding by means of the switch, Fig. 136, to be sure that it is right; also see that the disc on the cannon socket is in the right position to open the latch at the hour, and after the dial and hands are on move the minute hand forward past the hour and then backward gently until it is stopped by the latch. This will prove that the hand is on the square correctly.

On account of the liability of the motor to get out of adjustment and fail to wind, from the shifting of the springs and brushes, under careless adjustment, various attempts have been made to improve this feature of these clocks and the company is now putting out nearly altogether one of the two vibrating motors, shown in Figs. 139 and 140.

In Style C, Fig. 139, the hourly contact for winding is the same as in the clock with the three-magnet motor, as shown in Fig. 138. The magnet spools are twelve ohms and the resistance coil is eighty ohms, placed in parallel, as described in Fig. 130.

The vibrating motor, Fig. 139, is made with a pair of magnets and a vibrating armature. The main spring is wound by the forward and backward motion of the armature, one end of the connecting rod, 8, being attached to a lug of the armature, 2, and the other to the winding lever, 10. This lever has spring ends, to avoid shock and noise. As the winding lever is moved up and down, the pawl, 9, turns the ratch wheel, 11, and a pinion on the ratch wheel arbor turns the spring barrel until the winding is completed.

Fig. 139.