There should always be a degree plate, with a pointer to it on the pendulum wherever it can be most conveniently seen. The length of 4 (degrees) on the plate is always .07× its distance from the top of the pendulum; and as the only use of the degree plate is to see that the pendulum keeps the same arc, or to see how it varies, no very great accuracy is required in the degrees. Pendulum springs.
—Various opinions have been propounded on the proper strength for them, but none on any solid grounds, and some on altogether mistaken ideas of the duty of the spring. Probably the thinner it is the better, provided it is thick enough not to be bent too sharply or strained by the weight of the pendulum.
Every now and then a clock is troubled with the disease of spring-breaking over and over again. Sometimes this proceeds from the lower edges of the upper chops being left sharp, which in time cut the spring; and perhaps more frequently from some unseen inequality in the fixing; and it is better to let the pendulum hang at first with the lower chops a little loose, and only to screw them up after the pendulum has got a square bearing, and to take care not to put it out in doing so.
It is wonderful how much more trouble people will take to do things wrong than would serve to do them right, and it is almost incredible that some clock-makers send out their best clocks with the lower edges of the upper chops not horizontal but rounded into a circular arc.
The slightest reflection would show anybody that springs so fixed must tend to ‘buckle,’ or bend not in a straight line, at every vibration. And others make the upper chops, which have the strain of vibration, thinner and weaker than the lower ones which have no strain. There should always be a block of wood close under the bob of a heavy pendulum, to prevent the top from breaking the crutch and pallets if it falls from the spring breaking.
A spring does not bend only at one point as a string does; and therefore a pendulum bob hung by a spring does not move exactly in a circle, but in something more like a cycloid described with that radius of curvature, as in fig. 8, p. 23; and it has often been attempted to make springs which would render the pendulum absolutely isochronous for all such arcs as it is likely to swing.
Possibly the thing could be done, if it was worth doing; but both I and some other persons who have spent a great deal of time on experiments have come to the conclusion that it is not, for reasons will appear when we come to consider the effect of the escapement on the time of vibration.
Indeed I may say at once, with respect to the only escapement I have yet described and all others on the recoiling principle, that the circular error, which it is the object of these spring contrivances to correct, is already more than corrected by those escapements: for the clocks never lose, but gain, when the arc of the pendulum increases; so that the circular error is actually useful for counteracting the escapement error, and the clock goes better than it would with a perfect cycloidal pendulum, or any equivalent contrivance.