Mudge, Cumming, Kater, Gowland- names by now almost forgotten- had their part in the history of horology.

Each of them invented a gravity escapement which gave no evidence of great results. The turn up was given by Bloxam who improved this type of escapement, ...not definitely, anyway.

The escapement he conceived was structurally delicate and , if moved by a large weight, with a tendency to let the wheels run freely with no control causing imaginable damages .

By the way the main feature of the gravity escapement is to be able to store a great quantity of energy in the wheels train , the quantity given through the impulse to the pendulum remaining constant.

It is obvious that the force needed to move the large hands of a tower clock when the pivots have been just oiled in a summer day, is different if compared to the one needed in a winter day when the oil might have frozen and the hands loaded by ice might be under strong wind gusts .

Therefore, if the change of energy to keep the clock in motion in the two extreme conditions described above coincides with a proportional change of the energy given by the impulse to the pendulum, the clock will perform disastrously.

With just small changes to the escapement created by Bloxam, Lord Grimthorpe improved it and made it so reliable to become universally used in tower clocks .

It is so inexplicable to understand the reason this escapement had been rarely used in regulators.

Someone said that a regulator being never under weather intemperance doesn't require the reserve of energy needed by tower clocks. To me this is insufficient reasoning , as Grimthorpe's escapementadvantages & results are the same of those of Graham escapement (which is usually used in regulators).

Gravity escapement doesn't suffer of lubricating oil thickness, the wheels can eventually be cut with little care and, not least, the energy surplus, when the anchor runs into its stop, makes a ticks noisily.

This feature was not underestimated in astronomic observatories, where researchers watched the motion of the stars and,with their eyes pasted to the telescope, might report the precise time of stars events just "listening to" the clock.

Graham escapement needs a train of wheels nearly perfect and accurate. Its escape wheel must be as light as possible and therefore very delicate.

The lubricating oil decaying , gives wearing matters which at last turn in a change of the pendulum amplitude with (and how many ) drawbacks. True is also that in gravity escapement the pendulum is in contact with the impulse arms for almost its oscillation.

This is exactly the contrary to what should happen. Everybody knows that a pendulum performs better if completely free from any disturbance. The explanation of this apparent contradiction lies in the fact that gravity escapement gives to the pendulum an impulse virtually constant.

Therefore we can say, if we are not too fussy, that the deviation for the escapement ,the deviation for circular error and all other matters owning to an harmonic "disturbed" vibration do not suffer errors and the pendulum maintains a constant oscillation .

To make my clock visually more attractive, the escape wheel has been moved in front of the clock. The rod pendulum is of Invar. (Invar from invariable - a league with low expansion coefficient; mine is 1.43 x 10^-6) and compensated for temperature change by a brass cylinder; There is no compensation for atmospheric pressure changes.

The oscillation amplitude is 100' each part. You can see in the picture, along the rod , the small circular plate which carries the weights used to change the oscillating time.

Usually, we are accustomed to make this correction regulating the screw which is under the bob. To do this we must stop the pendulum but this is unacceptable for a regulator, as this might cause considerable errors.

The change of pendulum length and therefore of its oscillating time, is obtained moving its center of gravity adding or removing small weights while the pendulum is in motion.

Lord Grimthorpe suggested as adequate to keep in motion a regulator with his escapement, a weight of 25 pounds (around 12 kg); 6 kg are enough to my clock.

The movement is housed in a case built by a friend amateur wood maker and fixed to the wall. It was useless to explain to my home authority how solid should be this fixing (I had in my mind the adequate one and you can figure out which one) but I must satisfy myself by four 8 mm bolts fixed into the thin wall of my house.

To check the error of the clock I used a digital quartz chronometer. Each 24 hours I controlled the deviation of its time by the quartz chronometer. I therefore recorded the average of three or four readings and I could get the error in hundreds of second (table of the daily errors is in seconds).

I'm not able to explain the best result (0.00), the only comment is that the error was recorded on the 13th day and in Italy "13" is a very lucky number . The worse one - 0.28 was recorded the day after a storm. Might the reason to be found in atmospheric pressure changes during the day ? These sixteen days average result is 0.02 seconds.