672. Practical skill as well as some theoretical investigation has been
expended upon that part of a clock which is called the _escapement_,
the excellence of which is essential to the correct performance of a
timepiece. The pendulum must have its motion sustained by receiving
an impulse at every vibration: at the same time it is desirable that
the vibration should be hampered as little as possible by mechanical
connection. The isochronism on which the time-keeping depends is in
strictness only a characteristic of oscillations performed with a total
freedom from constraint of every description; hence we must endeavour
to approximate the clock pendulum as nearly as possible to one which is
swinging quite freely. To effect this, and at the same time to maintain
the arc of vibration tolerably constant, is the property of a good
escapement.
[Illustration: FIG. 99.]
673. A common form of escapement is shown in Fig. 99. The arrangement
is no doubt different from that actually found in a clock; but I have
constructed the machine in this way in order to show clearly the
action of the different parts. G is called the escapement-wheel: it is
surrounded by thirty teeth, and turns round once when the pendulum has
performed sixty vibrations,—that is, once a minute. I represents the
escapement; it vibrates about an axis and carries a fork at K which
projects behind, and the rod of the pendulum hangs between its prongs.
The pendulum is itself suspended from a point O. At N, H are a pair of
polished surfaces called the pallets: these fulfil a very important
function.
674. The escapement-wheel is constantly urged to turn round by the
action of the weight and train of wheels, of which we shall speak
presently; but the action of the pallets regulates the rate at which
the wheel can revolve. When a tooth of the wheel falls upon the pallet
N, the latter is gently pressed away: this pressure is transmitted by
the fork to the pendulum; as N moves away from the wheel, the other
pallet H approaches the wheel; and by the time N has receded so far
that the tooth slips from it, H has advanced sufficiently far to catch
the tooth which immediately drops upon H. In fact, the moment the
tooth is free from N, the wheel begins to revolve in consequence of
the driving weight; but it is quickly stopped by another tooth falling
on H: and the noise of this collision is the well known tick of the
clock. The pendulum is still swinging to the left when the tooth falls
on H. The pressure of the tooth then tends to push H outwards, but the
inertia of the pendulum in forcing H inwards is at first sufficient to
overcome the outward pressure arising from the wheel; the consequence
is that, after the tooth has dropped, the escapement-wheel moves back
a little, or “recoils,” as it is called. If you look at any ordinary
clock, which has a second-hand, you will notice that after each second
is completed the hand recoils before starting for the next second.
The reason of this is, that the second-hand is turned directly by the
escapement-wheel, and that the inertia of the pendulum causes the
escapement-wheel to recoil. But the constant pressure of the tooth
soon overcomes the inertia of the pendulum, and H is gradually pushed
out until the tooth is able to “escape”; the moment it does so the
wheel begins to turn round, but is quickly brought up by another tooth
falling on N, which has moved sufficiently inwards.
The process we have just described then recurs over again. Each tooth
escapes at each pallet, and the escapements take place once a second;
hence the escapement-wheel with thirty teeth will turn round once in a
minute.
675. When the tooth is pushing N, the pendulum is being urged to the
left; the instant this tooth escapes, another tooth falls on H, and
the pendulum, ere it has accomplished its swing to the left, has a
force exerted upon it to bring it to the right. When this force and
gravity combined have stopped the pendulum, and caused it to move to
the right, the tooth soon escapes at H, and another tooth falls on
N, then retarding the pendulum. Hence, except during the very minute
portion of time that the wheel turns after one escapement, and before
the next tick, the pendulum is never free; it is urged forwards when
its velocity is great, but before it comes to the end of its vibration
it is urged backwards; this escapement does not therefore possess the
characteristics which we pointed out (Art. 672) as necessary for a
really good instrument. But for ordinary purposes of time-keeping,
the recoil escapement works sufficiently well, as the force which
acts upon the pendulum is in reality extremely small. For the
refined applications of the astronomical clock, the performance of a
recoil escapement is inadequate.
The obvious defect in the recoil is that the pendulum is retarded
during a portion of its vibration; the impulse forward is of course
necessary, but the retarding force is useless and injurious.
676. The “dead-beat” escapement was devised by the celebrated
clockmaker Graham, in order to avoid this difficulty. If you observe
the second-hand of a clock, controlled by this escapement, you will
understand why it is called the dead beat: there is no recoil; the
second-hand moves quickly over each second, and remains there fixed
until it starts for the next second.
The wheel and escapement by which this effect is produced is shown in
Fig. 100. A and B are the pallets, by the action of the teeth on which
the motion is given to the crutch, which turns about the centre O; from
the axis through this centre the fork descends, so that as the crutch
is made to vibrate to and fro by the wheel, the fork is also made to
vibrate, and thus sustain the motion of the pendulum. The essential
feature in which the dead-beat escapement differs from the recoil
escapement is that when the tooth escapes from the pallet A, the wheel
turns: but the tooth which in the recoil escapement would have fallen
on the other pallet, now falls on a surface D, and not on the pallet
B. D is part of a circle with its centre at O, the centre of motion;
consequently, the tooth remains almost entirely inert so long as it
remains on the circular arc D.
677. There is thus no recoil, and the pendulum is allowed to reach the
extremity of its swing to the right unretarded; but when the pendulum
is returning, the crutch moves until the tooth passes from the circular
arc D on to the pallet B: instantly the tooth slides down the pallet,
giving the crutch an impulse, and escaping when the point has traversed
B. The next tooth that comes into action falls upon the circular arc
C, of which the centre is also at O; this tooth likewise remains at
rest until the pendulum has finished its swing, and has commenced its
return; then the tooth slides down A, and the process recommences as
before.
[Illustration: FIG. 100.]
678. The operations are so timed that the pendulum receives its impulse
(which takes place when a tooth slides down a pallet) precisely when
the oscillation is at the point of greatest velocity; the pendulum
is then unacted upon till it reaches a similar position in the next
vibration. This impulse at the middle of the swing does not affect the
time of vibration.
679. There is still a small frictional force acting to retard the
pendulum. This arises from the pressure of the teeth upon the
circular arcs, for there is a certain amount of friction, no matter
how carefully the surfaces may be polished. It is not however found
practically to be a source of appreciable irregularity.
In a clock furnished with a dead-beat escapement and a mercurial
pendulum, we have a superb time-keeper.