A NOVEL clock escapement in which, paradoxically, the pendulum controls the escapement, but the pendulum is 'unaware that the escapement exists' is detailed here. A proof-of-concept model is described which demonstrates the feasibility of the escapement. In addition, two alternative modes of impulse are described.
Background
In preparing this article I felt in somewhat of a dilemma: should I write it with experienced horologists in mind, or address 'a broader audience'? I decided on the latter.
The article is concerned with precision pendulum clocks, the pendulum and escapement of which are the most important components. The pendulum per se is an excellent timekeeper, provided it is not affected in any way by variation in its ambient surroundings, such as: temperature, pressure, viscosity, density, humidity, and the stability of its support. In addition, the amplitude of oscillation of the pendulum must remain absolutely constant. Pendulum design is a vast topic in itself and beyond the remit of this article, which is concerned only with escapement aspects.
The escapement has three important functions: 1) Maintain oscillation of the pendulum. 2) Provide a means of indicating the time. 3) Control the release of energy in order to perform the aforementioned tasks. Furthermore:
In an ideal escapement, the pendulum controls the escapement, but, paradoxically, the pendulum should be 'unaware that the escapement exists'.
In a nutshell, that is the problem to be solved… a problem that has exercised the minds of horologists since Galileo proposed the very first escapement in 1641.
Many different types of escapement have been described. The reasons why escapements adversely affect the consistency of oscillation of the pendulum is mainly due to friction. The word friction, in this context, refers to any source of friction (mechanical or viscous) or, more precisely, variation in friction related effects.
More specifically, and without describing a myriad of escapements, sources of frictional and other errors include:
1. Friction as the escapement drives (impulses) the pendulum.
2. Friction as the pendulum releases (unlocks) the escapement.
-
3.Variation in driving torque – particularly in spring driven clocks.
-
4.Imprecision and friction variation in the gears driving the escapement and clock hands.
5. In the case of public clocks, external forces, such as wind rain and snow, acting on the hands.
The nearest 'relative' of the SeeSaw Gravity Escapement (SGE), is the delightfully named Double Three Legged Gravity Escapement (DTLGE), invented by Edmund, First Baron Grimthorpe, in about 1854, and used in the Great Clock of Westminster (Big Ben). 'Gravity-type' escapements obviate problems of driving torque variation. In most clocks this is due to variation of friction in the gearing and pivots. In the case of public clocks, with faces open to the elements, the wind, rain or snow acting on the clock hands contributes to torque variation.
One problem with the DTLGE is that the pendulum unlocks the escapement, and any variation in unlocking friction affects timekeeping. A second problem is that the impulse weights, which act on the pendulum, are pivoted. Again, variation in friction affects timekeeping. Nevertheless, 'Big Ben' was an accurate clock for its size. The original specification required an accuracy of one second per day. However, the time keeping of mechanical observatory clocks, of the same period, was about 20 times better, or 0.05 seconds per day.
The problem of unlocking friction was solved by Jim Arnfield, in 1987, using his Inertially Detached Gravity Escapement. As far as I know, the problem of friction, associated with the impulse weights, remained unsolved.
Basic Ideas
Before describing the SGE concept model, it will be useful to discuss the four basic ideas on which it is based:
Energy interchange:
Figures 1a - d show a simplified pendulum (blue), pivoted at two points A and B (yellow), and having LH and RH arms located at the pivot region. Towards the end of each arm is a weight (pink). The weights straddle the arms (blue), and can periodically rest on weight supports (green), at each side of the arms. When the pendulum rod is vertical, the top surfaces of the arms are just below the weight supports. Hence, the weights just rest on the weight supports.
Imagine that the pendulum is caused to swing to the left (Figure 1c). The amplitude of the swing is exaggerated for clarity, and the weights do not slide down the arms. The LH arm raises its weight clear of the weight support. The pendulum reaches its extreme left position and begins to swing to the right. As it does so it lowers the weight back on to the weight support, and the arm continues to lower. A similar action occurs to the weight on the RH side (Figure 1d).
Neglecting acceleration and retardation, the amount of energy abstracted from the pendulum, as it raises a weight, is exactly the same that the weight delivers to the pendulum, as the weight is lowered back to its original position. But there is also energy interchange as the pendulum accelerates and retards the weight. However, over the cycle of raising and lowering the weight, the net affect on the pendulum is zero. In addition, there is a small loss of energy as the weights impact the weight supports and pendulum arms (similar to that in the DTLGE). So, the pendulum continues to oscillate (at least for some time). The weights add inertia to the pendulum, so that the cyclic period will be a little longer than the pendulum without the weights. However, since the mass of the weights is constant, and if the pendulum's amplitude is constant, the periodic time will also be constant.
A similar energy interchange occurs in the DTLGE and other escapements. The important difference is that, in the SGE, there are no frictional losses. More specifically, no sliding occurs between the weights, arms and weight supports. However, friction at the pendulum pivot/suspension remains, but this is common to all pendulum clocks, and is just one of the energy losses to be 'made up' by the escapement
Impulse and reset:
In order for the pendulum to continue to oscillate, the escapement must make up the losses, that is, impulse the pendulum. Imagine that after the weights are lifted clear of the weight supports by the arms, the weight supports are lowered a little. Hence, when the arms lower the weights back to the weight supports, the weights will have moved a greater distance downwards than upwards. The difference between these two distances, together with the value of each weight, impulses the pendulum. When an arm is below a weight, that is, with the weight resting on a weight support, the weight support, and hence weight, needs to be raised (reset), ready for the next cycle. In this way, the pendulum will be 'unaware' that the weights have been reset. Lowering and resetting the weights, at just the right time, is the job of the escapement. Paradoxically, the pendulum needs to control the escapement, but must be oblivious to the escapement's existence! The key mechanism of the SGE that does this is called 'the prop'.
The prop:
Figures 2a & b show a simplified prop and weight support assembly. The upper vertical arm of the prop supports a pivoted weight support, on top of which is a weight. Incidentally, the horizontal arm of the prop is used to lock and unlock the escape wheel, and the lower arm is used to reset the prop (to be explained later). The prop is biased to turn clockwise by a counterweight.
With the weight in position on the weight support, there is sufficient frictional resistance between the end of the prop and underside of the weight support to prevent the prop from turning clockwise. As soon as the weight is lifted clear of the weight support, the resistive force is reduced, and the prop pivots clockwise against a stop (Figure 2b). The weight support lowers, and is constrained by a stop. It will be appreciated that as the weight is lifted, the weight, and weight support, move in opposite directions. In fact, the weight is quite oblivious to the fact that the prop and weight support have moved. If the weight is replaced on the weight support, resetting the prop anticlockwise raises the weight support and weight thereon. Because of the weight, there is again sufficient frictional resistance to retain the prop in the vertical position.
The weight, in this description of the prop, corresponds to the weights shown in Figures 1a - d. Hence, the pendulum is also quite oblivious to the fact that the escape wheel has been unlocked. In fact, the SGE has two props, each controlling its own impulse weight.
Escape wheel assembly:
The escape wheel assembly ensures that all the various actions describe above occur in the correct order. It is similar to the escape wheel assembly of the DTLGE, since it consists of two 'three legged' escape wheels mounted either side of three resetting rollers (to be described in detail later).
