The Combe Hay Caisson Lock

A talk by Adrian Tuddenham, BA(Open) and Eur Ing, David Brown, M Eng, C Eng, FRINA, RCNC on 8 March 1999

Adrian Tuddenham first explained that, in planning the Somersetshire Coal Canal at the end of the 18th century, there was too little water available from streams to operate a long conventional flight of locks to negotiate the drop of level of about l40 ft (45 metres) which occurred near Combe Hay between Paulton at one end and Dundas Aquaduct at the Kennet and Avon Canal at the other. (see map)

Robert Weldon had recently patented the `Hydrostatick Caisson Lock' which used almost no water in passing boats through it and he had demonstrated its practicality by a model in Shropshire. Three caisson locks, each able to provide a drop (or rise) of 45 feet could avoid the need to build 22 conventional locks and install a pumping engine to raise the water needed to operate them. For this reason it was decided to adopt caisson locks at Combe Hay.

David Brown showed Robert Weldon's drawing of a caisson lock (see opposite page) which consisted of a masonry chamber built in the ground with an entry at each end, one at the top, level with the upper canal and the other at the bottom of the opposite end of the chamber, level with the lower canal. Both entries were fitted with doors. In the chamber, which was full of water, there was a box, partly full of water and also with doors, at each end, into or out of which a barge could be floated when the box was either at the top or bottom of the chamber and the appropriate doors were opened. The box `was intended to be neutrally buoyant and its position was controlled by guides and chains with a manually driven lifting and lowering mechanism. The mass of the box was unaffected by the presence of a barge floating inside it and the box was clamped to either entry to the chamber when its door was open by water pressure, then enabling the adjacent door in the end of the box to be opened to let the barge in or out. The only water allowed to fall between the canal levels would be owing to leakage of seals at the lower entry.

David Brown had found that the box would probably have displaced about 270 tonnes and weighed about 170 tonnes together with the water in it, so about 100 tonnes of ballast would have been needed to give neutral buoyancy. The box could have been strong enough to withstand the pressure of 50 feet of water i.e. about 3,000 lb/sq.ft. (150 kPa) at the bottom of the chamber, if built, as
1795. The tenders for building it were invited in January 1797 construction began about 4 months later. The lock was built in about 12 months and it. was being tested by November 1797. It worked well at first, though the box did become unstable once owing to movement of the water inside it; its guidance system was therefore improved after which all seemed well .and people were taken for demonstration rides in it. However, before it could be connected to the canal, still being built, one (possibly both) side walls of the masonry chamber bulged inwards causing considerable loss of water through the masonry joints and eventually causing the box to jam between its guides. As luck would have it, that happened when members of the Canal Committee were being taken for a demonstration



Diagram of Combe Hay Caisson Lock

was likely, of oak planks 2¼ in. (57 mm) thick with bracketed wrought iron frames 12 ft 6 in (4.2 m) apart. There would have been intermediate wood frames 15 ins. (380 mm) square 30 ins. (760mm) apart.

Adrian Tuddenham recounted that the decision to adopt the caisson lock was taken in July
ride through the lock and they were incarcerated inside the box for some time, long enough to run short of air. Repair work was halted and the decision to abandon the caisson lock taken in early 1800. Only the top caisson of the three had been completed, although a start was known to have been made on the second one.

Many suggestions were then made as to how best to take the canal down the drop in level to join the Kennet & Avon Canal at the Dundas Aquaduct near Limpley Stoke. As a temporary expedient, it was decided to adopt an inclined plane like Ralph Allen's tramway from Combe Down to Widcombe Wharf on the Avon, to Combe Hay, coal was carried down the incline in boxes on trolleys and transhipped at each end. In 1802 extra finance had been raised to allow work to begin on a flight of conventional locks, with steam pumping to return water to the higher level.

The rest of the evening was absorbingly spent first in reviewing the evidence now available, in doing which Adrian freely described himself as `standing on the shoulders of giants', and then in considering the most likely site of the top caisson lock. This he argued in the light of many points of engineering design without which it would have been impossible to discriminate between otherwise plausible interpretations of surviving data and remains. Drawings and still discernible foundations of the house built later for the beam-engine of the pump, examination of the contours, soil and rock of the site, and the geometrical imperatives involved in laying out the final series of locks while keeping the inclined plane in operation, pointed strongly to the beam-bearing wall of the engine house having been built on the top of the downhill-side wall of the filled-in chamber of the caisson lock, which served as its foundation. If that indication proves correct, the position of the caisson lock will have been found. There is however much yet to be elucidated about this remarkable piece of 18th century engineering.

In discussion, members of the Somersetshire Coal Canal Society and of the Bath & Camerton Archaeological Society questioned and commented on the presentation and arguments. His thesis proved robust in the face of that examination and there was some indication that the failure of the masonry of the chamber and hence of the caisson lock could be ascribed to providing insufficient rigidity for the end supports of the arched side walls (see figure) to act as effective abutments in face of loading caused by swelling of the clay soil of the hill into which the chamber would have been cut. The failure of the lock appeared to have little to do with the principle and mechanics of its operation.

It was clear that there was a desire to hold another meeting to discuss plans for further investigation and to resolve more points about the design of the unique caisson lock at Combe Hay.

Adrian Tuddenham & David Brown