The distinguishing feature of this escapement lies in the fact that it aims to drive the pendulum by applying to it a falling weight at each excursion on each side. As the weight is lifted by the train and applied to the pendulum on its return stroke and there is no other connection, it follows that the pendulum is more highly detached than in any other form of pendulum escapement. This should make it a better timekeeper, as the application of the weight should give a constant impulse and hence errors and variations in the power which drives the train may be neglected.

On tower clocks this is undoubtedly true, as these clocks are interfered with by every wind that blows against the hands, so that a detached pendulum enables a surplus of power to be applied to the train to meet all emergencies. With a watchmaker’s regulator, however, the case is different. Here every effort is made to favor the clock, vibrations, variations of temperature, variations of power, dirt, dust, wind pressure and irregularities of the mechanism are all carefully excluded and the consequence is that the special advantages of the gravity escapement are not apparent, for the reason that there are practically no variations for the escapement to take care of. Added to this we must consider that the double three-legged form, which is the usual one, is practically an escape wheel of but six teeth, so that another wheel and pinion must be added to the train and this, with the added complications of the fan and the heavier driving weight required, counterbalance its advantages and bring it back to an equality of performance with the simpler mechanism of the well made and properly adjusted dead beat escapement. Theoretically it should work far better than the dead beat, as it is more detached; but theory is always modified by working conditions and if the variations are lacking there is no special advantage in constructing a mechanism to take care of them. This is the reason why so many watchmakers have constructed for themselves a regulator with this escapement, used in the making all the care and skill of which they were capable and then been disappointed to find that it gave no better results with the same pendulum than the dead beat it was to replace. They had eliminated all the conditions under which the detached escapement would have shown superiority.

Although the gravity escapement will not give a superior performance under the most favorable conditions for timekeeping, it is distinctly superior when these conditions are unfavorable and therefore fully merits its high place in the estimation of the horological fraternity. We have instanced its value in tower clock work; it has another advantage in running cheap and poorly made (home made) regulators with rough and poor trains; therefore, it is a favorite escapement with watchmakers who build their own regulators while they are still working at the bench, before entering into business for themselves. As the price of a first-class clock for this purpose is about $300 and the cheapest that is at all reliable is about $75, it will be seen that the temptation to build a clock is very strong and many of them are built annually. Regulators with the gravity escapement are built by the Seth Thomas Clock Co., the Howard, and one or two others in this country, but they are furnished simply to supply the demand and sales are never pushed for the reasons given previously. Clocks with this escapement are quite common in England and many of them have found their way to America. It is one of the anomalies of trade that our clockmakers are supplying Europe with cheap clocks, while we are importing practically all the high priced clocks sold in the United States and among them are a few having the three-legged and four-legged gravity escapements, therefore the chances are that when a repairer finds such a clock it is likely to be either of English origin or homemade, unless it be a German regulator.

Figs. 47 and 48 show plans and side views of the three-legged escapement. Fig. 48 also shows an enlarged view of the escape wheel, showing how the three-leaved pinion between the two escape wheels, is made where it is worked out of the solid. A, B and C and a, b and c show the escape wheel which is made up of two three-armed wheels, one on each side of a three-leaved pinion marked D¹ and D² in the enlarged view of Fig. 48. The pallets in this escapement consist of the two arms of metal suspended from points opposite the point of bending of the pendulum spring and the lifting planes are found on the ends of the center arms in these pallets, which press against the three leaves of the pinion, while the impulse pins e¹ and e², Fig. 47 and 48 act directly upon the pendulum in place of the verge wire. The pallets act between the wheels in the same plane as each other. The lifting pins or pinion leaves act on the lifting planes after the line of centers when the long teeth or legs of the escape wheels have been released from the stops, F and G, Figs. 47 and 48, which are placed one on each side of the pallets and act alternately on the wheels. These pallets are pivoted one on each side of the bending point of the suspension spring. To lay out the escapement, draw a circle representing the escape wheel diameter, then draw the line of centers and set off on the diameter of the escape wheel from each side of the line of centers 60° of its circumference, thus marking the positions for the pallet stops 120° apart. Draw radii from the center of the escape wheel to these positions and draw tangents from the ends of these radii toward the center line. The point where these meet will be the bending point of the pendulum spring.

This is clearly shown at H, Fig. 47. The points of suspension for the pallets are planted on the line of these tangents and a little below the point H, where the tangents meet on the line of centers. This is done to avoid the mechanical difficulty of having the studs for the two pallets occupy the same place at the same time. The arms of the pallets below the stops may be of any length, but they are generally constructed of the same angle as the upper arms and will be all right if drawn parallel to these upper arms. They are in some instances continued further down, but this is largely a matter of taste and the lower portion of the escapement is generally drawn so as to be symmetrical.

The impulse of the pendulum is given by having pins projecting from the pallet arms and bearing upon the pendulum rod, which pins may be of brass, steel or ivory. In the heavier escapements they are made of ivory in order to avoid any chatter from contact with the pendulum rod of a heavy pendulum. These pallets should be as light as it is possible to make them without having them chatter under the impact of the escape wheel arms on the stops. They have only to counteract the force of the pendulum spring and the resistance of the air and for light pendulums this force is much less than is generally understood. Two ounces of impulse will maintain a 250-pound pendulum, but two pennyweights is more than sufficient for a fifty-pound pendulum. The reader can see that in the case of a pendulum weighing but eight to fourteen pounds, there will be a still greater proportionate drop, as the spring itself is thinner, the rod is thinner, the pendulum ball offers little resistance to the air and the consequence is that it is difficult to get the pallet arms light enough for an ordinary clock.

Watchmakers who make this escapement for themselves, to drive an eight to fourteen pound pendulum, generally make the escape wheel three inches diameter and make the escape wheel and pallet arms all from the steel obtained by buying an ordinary carpenter’s saw. The lifting planes should not be more than one-eighth its diameter from the center of the escape wheel, as where this is the case the circular motion of the center pins will be so great that the pallet in action will be thrown out too rapidly and will chatter when striking the pendulum rod. On the other hand it should not be less than one-twelfth of the diameter of the escape wheel, or the pendulum will not be given sufficiently free swing and the motion will be so slow that while such a clock will work under favorable conditions, jarring, shaking in wind storms, etc., will have a tendency to make the pendulum wabble and stop the clock. From what has been said above, it will also be seen that the necessity for slow motion of the pallet arms unfits this escapement for use with short pendulums.

The action of the escapement is as follows: The pendulum traveling to the right, when it has thrown the right pallet arm sufficiently far, will liberate the escape wheel tooth from the stop G and the pinion, acting on the lifting plane, will raise the pallet arm, allowing the pendulum to continue its course without doing any further work until it has reached nearly its extreme point of excursion, when the weight of the pallet will be dropped upon the pendulum rod and remain there, acting upon the pendulum until it has passed the center when the pallet arm will be stopped by the banking pin M¹; exactly the same procedure takes place on the left side of the escapement during the swing of the pendulum to the left. The beat pins M and M¹ should be set so that the impulse pins e¹ and e² will just touch the pendulum when the latter is hanging at rest and the escapement will then be in beat. The stops should be cut from sheet steel and the locking faces of the escape wheel arms, stops on the pallets, lifting planes of the pallets and the lifting pins should all be hardened. In some of the very fine escapements the faces of the blocks are jeweled. The arms of the inner part of the escape wheel are usually set at equal angular distances between those of the outer, although this is not absolutely necessary, and the lifting pins are set on radii to the acting faces of the arms of one of the wheels, so as to cross the line of centers at the distance from the center, not exceeding one-eighth of the radius of the wheel, for the reasons explained above.

From the comparatively great angle at which the arms are placed, the distance through which they have to be lifted to give sufficient impulse is less in this escapement than in one with a larger number of teeth acting in the same plane, as the pallets would then hang more nearly upright. This is a great advantage, as the contact is shorter. The unlocking is also easier for the same reason, and from the greater diameter of the wheel in proportion to other parts of the escapement, the pressure on the stops is considerably less. The two wheels must be squared on the arbor, so there will be no possibility of slipping. The lifting pins D are shouldered between them like a three-tooth lantern pinion. In small escapements the lifting pins are not worked out of the solid arbor, but are made as hardened screws to connect the two portions of the wheel. In tower clocks the pinion is generally made solid on the shaft J, Fig. 48. The wheel, A, B, C, is made to pass over the pinion D and is fitted to a triangular seat, the size of the triangle of the leaves, D, against the collar on the shaft. The other wheel, a, b, c, is fitted to the inside triangle of the pinion, so that the leaves, D, form a shoulder against which it fits. The pallets, E and E¹, also lie in one plane between the wheels, but one stop, F, points forward to receive the A, B, C, teeth and the other, G, points backward to receive the a, b, c teeth alternately. The distance of the pendulum top, H, or cheeks from the center of the escape wheel, J equals the diameter of the escape wheel. The lifting pins should be so placed that the one which is holding up a pallet and the one which is to lift next will be vertical over each other, on the line of centers, the third pin being on the level with the center, and to one side of it, see Fig. 48, enlarged view.

The fly is a very essential part of this escapement, as the angular motion of the escape wheel is such that unless it were checked it would be apt to rebound and unlock; consequently, a large fly is always a feature of this escapement and is mounted upon the scape wheel arbor with spring friction in such a way that the fly can continue motion after the scape wheel has been stopped. This is provided for by a spring pressure, either like the ordinary spring attachment of the fly of striking trains of small clocks, or as shown in Fig. 49 for tower clocks. This fly is effective in proportion to its length and hence a long narrow fly will be better than a shorter and wider one, as the resistance of the air striking against the ends of the fly is much greater the further you get from the center.

The pallet stud pins and the impulse pins should on no account be touched with oil or other grease of any kind, but be left dry whatever they are made of, because the slightest adhesion between the impulse pins and the pendulum rod is fatal to the whole action of the escapement. Care must also be taken that one pallet begins to lift simultaneously with the resting of the other, neither before nor after.

The gravity escapement requires a heavier weight or force to operate the train than a dead beat escapement, because it must be strong enough to be sure of lifting the pallets quickly and firmly, and also because the escape wheel having but six teeth necessitates the use of another wheel and pinion between the escape and center and consequently the train is geared back more than it would be for a dead beat escapement, with the seconds hand mounted on the escape wheel arbor. But with this form of escapement the superfluous force does not work the pendulum and it does no harm if the train is good enough not to waste power in getting over rough places left in cutting the teeth of the wheels or any jamming from those which have unequal widths or spaces. For this reason a high numbered train is better than a low numbered one, as these defects are greater on the teeth of a low numbered train and any defect in such cases will show itself.

In the gravity escapement the escape wheel must have a little run at the pallets before it begins to lift them and in order to do this the banking pins, M, M¹ for the pallet arms to rest on, should hold them just clear of the lifting pins or leaves of the escape wheel. The escape wheel should be as light as possible, for every blow heard in the machine means a loss of power and wear of parts. Of course, in an escapement a sudden stop is expected, but the light wheel will reduce it to a minimum if the fan is large enough. Particular attention should therefore be given to the length of this fan and if the stop of the escape wheel seems too abrupt, the fan should be lengthened.

Figs. 50 and 51 show the same escapement with a four-legged wheel instead of the double three-legged. In this case, where there is but one wheel, the pallets must of necessity work on opposite sides of the wheel and hence they are not planted in the same plane with each other, but are placed as close to each side of the wheel as is practicable.

To lay out this escapement, draw the circle of the escape wheel as before, make your line of centers and mark off on the circle 67½° on each side of the line of centers and draw radii to these points, which will indicate the approximate position of the stops. Tangents to these radii, meeting above the wheel on the line of centers will give the theoretical point of the suspension. One set of the lifting pins is planted on radii to the acting faces of the teeth of the escape wheel. The opposite set, on the other side of the wheel, is placed midway between the first set. This secures the lifting at the line of centers. The wheel turns 45° at each beat and its arbor likewise carries a fly.

In case the locking is not secure, the stops may be shifted a little up or down, care being taken to keep them 135° apart. In this way a draw may be given to the locking of the scape wheel arms similar to the draw of the pallets in a detached lever escapement and thus any desired resistance to unlocking may be secured. The stops in either escapement are generally made of steel and it is of the utmost importance that the arms of the escape wheel should leave them without imparting the least suspension of an impulse. Therefore, the stops and the ends of the arms should be cut away (backed off) to rather a sharp angle to insure clearance when the arms are leaving the stops. It is also of equal importance that the legs of the wheels should fall on the stops dead true. The fit of each of the legs should be examined on both stops with a powerful eye-glass, so that they should be correct and also see that when the unlocking takes place the wheel is absolutely free to turn.