This section is from the book "Modern Buildings, Their Planning, Construction And Equipment Vol3", by G. A. T. Middleton. Also available from Amazon: Modern Buildings.
In Fig. 43 is illustrated a simple apparatus which will serve to describe the chief points of the Two-pipe System from. Each radiator, it will be seen, communicates directly with a main flow and return service, and from this detail the term "two-pipe system" is obtained.
A main flow-pipe is taken from the boiler, and is carried so as to pass as near as possible to the points where the radiators are to be placed. Alongside the flow is carried the return pipe, the two pipes being alike in all general respects except for their terminating at different levels in the boiler. It is also the custom, when the pipes run in a horizontal direction, to let the flow be above the return. This is not necessary for the circulation, but has become a rule in the trade, so that when pipes are inspected the upper pipe is always assumed to be the flow.
In the example (Fig. 43) it will be seen that the main flow is branched near the boiler, for the purpose of serving some radiators on the right side of the illustration. This is quite a customary and proper plan, but when a flow branch comes as near the boiler as is shown it would be as well to let it start from the boiler direct as a distinctly separate circulation. Assuming, however, that the boiler has been delivered with but one flow-pipe tapping, then the flow branch may be arranged as illustrated. It may be asked why the main circulation has not been carried up to the radiators nearer the boiler (on the right of the illustration) and the branch to the distant ones. The reason is that distant circulations must always be favoured, and the connections, as shown, favour the circuit up the left side of the illustration. A continuing vertical pipe is always favourable to a better circulation than a tee branch with a horizontal outlet.
The returns, it will be noticed, do not branch or join into one another, but enter the boiler independently. An excellent rule to remember in this work, to ensure getting the most uniform working possible with branched circulations, is to calculate that a flow-pipe may be branched to any reasonable extent and a return pipe never. Occasionally returns have to join, but no experienced engineer does this without having some doubts as to getting the best results. It is probable that keeping the returns separate admits of branching the flow, but it is certain that keeping flows separate will not allow of joining the returns. Perhaps the rule should be, join the returns as little as possible, and avoid doing so entirely unless it be absolutely imperative. Joining the returns does not necessarily imply failure, but it hazards it and usually limits the success.
The expansion pipe is shown extending from the top of the main flow-pipe, a good place for it when it can be got there, but it will do at the head of a branch circulation, or in case of necessity it may be at any lower point on the main flow-pipe. If desired, more than one expansion pipe can be used, but it is important to note that if the expansion pipe or pipes do not occupy high points it is necessary that radiators be there to receive the air. An expansion pipe is really an air pipe, and at all high points there must be air pipes or radiators. A radiator will receive a good volume of air without ceasing to heat properly. As explained on a previous page, water discharges or releases a certain amount of air when it is heated, and in domestic hot-water works the air question requires full consideration, as the frequent changing of the water makes the supply of air almost continuous. With heating works the change of water is practically nil, therefore the collection of air, after discharging that from first heating, is very small.
Realising that the air trouble is small, after the water has been well heated and deprived of the air it carries, attempts have been made to do without an expansion pipe. Expanded water could find its way into the cold cistern, as it usually does, and safety from explosion would be had in the open cold-supply pipe and the safety valve on the boiler. The attempts have failed, however, and the writer, like others, has experienced the failure. On heating the water in an apparatus minus an expansion pipe there occurs what can only be described as a " swelling back " in the cold cistern, and if the cistern is only of moderate size there will be an overflow. It is not caused by expansion, for it occurs when the water is no more than warm, and it happens so suddenly and swiftly that the water will get over the edge of the cistern although it may be provided with an overflow pipe. No good explanation has been offered as to this phenomenon, but a reliable remedy lies in the use of an expansion pipe.
The cold supply is not shown intact in Fig. 43, as it is not always run in the same way. The important detail is that this pipe should always join the boiler (at a low point), or a return pipe at a point close to the boiler. For economy's sake the cold supply is sometimes taken into a return at a high point, but in the majority of cases this brings about trouble by locking air in the apparatus either at the time of filling or refilling. In some works the engineer has to provide the cold-supply cistern, in which case he puts in one of the least necessary size, and places it as closely over the highest radiator as he can, so as to economise in the length of the cold-supply pipe and the expansion pipe. When possible, however, he utilises the house or general cold cistern of the building, and instead of bringing down a separate cold-supply service from cistern to boiler, he takes a branch from one of the services that come down for some other purpose. This is not to be objected to, as it introduces no element of a dangerous or troublesome character.
Quite a small pipe will serve the boiler, as it uses such a small quantity of water ; but the supply cistern, if the engineer provides it, must bear in size some relation to that of the apparatus. When the apparatus is newly filled with cold water there only requires to be 2 or 3 inches depth in the cistern, and the remaining space in the cistern must be large enough to accommodate the water of expansion The expansion of water when heated from 39° to 212° Fahr. is 1 in 24- 24 gallons becoming 25 ; and as this range of temperature is by no means impossible, an allowance equal to 1 gallon for every 24 in the apparatus should be provided in the cistern. Some difficulty lies, however, in calculating the contents of the apparatus, but for arriving at a figure with reasonable quickness it may be taken that 5 square feet of radiator surface represent a gallon of water, that 4-inch pipe holds half a gallon per foot, 3-inch pipe one-third of a gallon, 2-inch pipe one-eighth of a gallon, while mixed smaller sizes hold about one gallon to 20 feet run. In arriving at the size of the cistern it is then only necessary to allow that 6 1/4 gallons occupy a cubic foot of space.