L be applied to the bottom of the tank and heat the water contained therein, then the water in the portion a b will be colder, and therefore heavier, than the corresponding column in the tank, and the action of gravity starts a circulation in the direction shown by the arrows. Now, such an apparatus is obviously too crude to be of practical use. A large amount of heat is lost from the exposed surface of the water, and if the water is lowered by only a slight amount, the upper connection of the pipe will be uncovered and the circulation stopped; but by far the most serious objection is, that it is impossible to carry any part of the pipe above the surface of the water. Air is always present in water, and in such an apparatus bubbles of air might easily collect at a, and impede the circulation. The latter difficulty could be overcome by arranging the pipe as in Fig. 535, with a rise to the point A, and there providing an outlet by means of the air-pipe a p.

In the application of the system to practical cases several points must be carefully observed: -

(a) The heater must be below the lowest part of the circulation-pipes.

(b) Means must be provided for the expansion of the water produced by the application of heat.

(c) Means must be provided for keeping the apparatus full of water.

(d) The circulation-pipes must rise continuously to the highest point and then return gradually to the better.

(e) Means must be provided for ridding the apparatus of air. Fig. 536 represents diagrammatically an apparatus for heating two floors one above the other. The heater is placed in the basement, and of course cannot be left open as those shown in figs. 534 and 535. At a point somewhere above the highest part of the heating-apparatus, will be placed the expansion-tank, which also acts as a feed-cistern for the apparatus; this is marked XT, and is fed with cold water from any convenient source by the pipe cw. The water in this tank never rises more than (say) 4 inches above the bottom of the tank, and actuates a float, which controls the inlet-valve just as an ordinary ball-cock does, but in this case the valve itself is preferably put outside the tank, and the tank covered over and provided with a vapour-pipe v p, which is carried into the open air over the roof. From the top of the heater is carried the rising main, which goes direct to the highest part of the building, and thence round the two floors in the manner indicated. Thi3 pipe itself, if of adequate size, would effect the heating of the rooms, but there are several reasons why it is generally inadvisable to use such pipes in houses. In order to obtain the requisite surface in the length disposable, it would be necessary to use pipe from 3 inches to 4 inches in diameter, and this would seldom be convenient in a private bouse, as there are doorways to be passed, and a large pipe is very unsightly. A small pipe can be run behind a skirting-board, or under the floorboards; branches are then taken off to each radiator as shown, and the only parts of the apparatus above the floor arc the radiators and their connections. These connections are all taken off the return-pipe, which descends to the heater and is connected to it at or near the bottom; a pipe is carried down from the expansion-tank, and has a U-shaped bend at the bottom, connected to the return-pipe close to the heater. The action of the apparatus will be as follow-. With the system full of cold water, the height of the water in the expansion-tank will be (say) 4 inches; as soon as licit is applied, the water will expand up the vertical pipe into the tank xt, and will close the ball-cock, and if it should expand sufficiently, it will pass away by the overflow; hot water will then pass round the system in the direction of the arrows. The valves are marked by an x within a circle. It will be seen that, while the whole of the system can be shut off from the boiler, it is impossible for high pressure to be got up in the boiler, as it is always in open connection with the expansion-tank; the worst that can happen is for boiling water to pass up into the expansion-tank.

Fig 536   Diagrammatic view of Low pressure Hot water Apparatus (or Heating two Floors with one Circuit

Fig 536 - Diagrammatic view of Low pressure Hot-water Apparatus (or Heating two Floors with one Circuit.

This is probably the cheapest scheme which could be devised. The main pipes might be 1½ inches in diameter, and the branches to the separate radiators | inch. The heat in each of the radiators can be readily controlled by the valve next to it, but there is no means of emptying a portion of the apparatus so as to allow of the repair of a joint or other similar work. If a second stop-valve were put on the other side of each radiator, this would allow the radiator itself to be removed, but as it would add about 7s. 6d. to the cost for each extra valve, it is not usually done. Although only two radiators are shown, it is by no means intended that the number should be so limited.

The methods of connecting radiators with the system of pipes deserve mention. It will be observed that, in the last figure, each radiator has a branch off the main, and another back into it, as shown more clearly in Fig. 537. In Fig. 538, an alternative arrangement is shown with the inlet branch off the flow-pipe and the outlet into the return. If the radiators were arranged so closely together as shown, they might work quite well under the alternative arrangement, but there is always the danger of a short circuit being set up from flow to return, so that while the first radiator would get thoroughly hot, the second might be only warm, and the third almost cold. Such a tiling could not occur with the upper arrangement, and it should always be used except where the branch off the main is so long, and feeds so much radi ating surface, that the water is at a very low temperature when it gets back.

Fig. 537   Method of connecting Radiators to one Circulation pipe only

Fig. 537 - Method of connecting Radiators to one Circulation pipe only.