Su Song's water-powered clock tower was the world's first astronomical clock. How was it possible for the whole apparatus, with its armillary sphere, celestial globe and timekeeping system, to be driven by just a trickle of water?
The passage of time is uniform, so to measure time the ancients had to devise a mechanical system with a uniform motion. Mechanically powered clocks had not yet been invented, but the ancient Chinese had the idea of using the power of water.
In Su Song's clock tower, 36 tilting scoops were mounted on a large drive wheel. A steady flow of water filled each of the scoops in turn. When a scoop filled with water, it became heavier than the checking fork and lower balancing lever which supported its weight, and turned downward like the nodding pots used for irrigation in ancient China. At this point the weight of the water remaining in the scoop turned the drive wheel, which in turn drove a train of gears to synchronously turn the observational armillary sphere on the top level of the tower, the demonstrational celestial globe on the middle level, and the timekeeping system inside the bottom level.
This mechanism sounds simple enough, but there were quite a number of problems to be overcome. Firstly, how could the water be made to run into the scoops at a steady rate? To have someone pour it directly by hand would certainly not be accurate enough, and if one were simply to pierce a hole in the bottom of a tank and let the water flow out through it, the changing level of water in the tank would alter its pressure, and this would cause the rate of flow to vary. The solution the ancients devised was to link two tanks together, the upper tank being used to maintain a constant water level in the lower one. This meant that water could be kept flowing into the scoops at a constant rate.
Another problem was how to ensure that once each scoop filled up and tilted downwards, the drive wheel would turn only as far as the next scoop. This was the clock tower's biggest secret. Apparently the mainland Chinese model of the clock tower failed to overcome this obstacle.
According to Kuo Mei-fang's research, Su Song solved this problem by using a system of balanced levers as an escapement, to alternately check and release the drive wheel:
When a full scoop tilted down and struck the trip lever, a connecting rod joined to the trip lever pulled sharply downwards on the weighted end of the upper balance lever. This raised the opposite end of the lever, jerking open the left upper lock, which was joined to it by another connecting rod. This released the drive wheel, allowing it to turn due to the imbalance between the empty scoops on one side and the full scoops on the other.
At this point the next scoop would descend onto the checking fork, and the left upper lock would re-engage, checking the rotation of the drive wheel. The right upper lock prevented the wheel from recoiling and turning backwards. When the next scoop filled with water and pushed down the trip lever, the escapement action would begin another cycle.
Because of the large diameter of the drive wheel, it moved with enough force to turn the gear wheel which drove the vertical transmission shaft. The middle and upper gears on the transmission shaft drove the timekeeping system in the bottom level of the clock tower, the celestial globe in the middle level and the armillary sphere on the top level, all in time with the regular movement of the drive wheel.
Su Song used two connected water tanks--an upper reservoir and a constant-level tank-to ensure that water flowed into the scoops of the drive wheel at a constant rate. （drawing courtesy of the National Palace Museum／photo by Hsueh Chi-kuang）
The drive wheel turned the main shaft to drive the lower gear wheel, which turned the vertical transmission shaft. （courtesy of the National Palace Museum） The middle and upper gears on the vertical shaft drove the armillary sphere, the celestial globe and the timekeeping section. （photo by Hsueh Chi-kuang）
