Gate valves should always be used in connection with hot water piping, although angle valves may be used at the radiators. There are several patterns of radiator valves made especially for hot water work; their chief advantage lies in a device for quick closing, usually a quarter or half turn being sufficient to open or close the valve. Two different designs are shown in Fig.. 44 and 45.

Valves And Fittings 1000126

Fig. 46.

It is customary to place a valve in only one connection as that is sufficient to stop the flow of water through the radiator; a fitting known as a "union elbow" is often employed in place of the second valve. (See Fig. 46.)

Air Valves

The ordinary pet-cock air valve is the most reliable for hot-water radiators, although there are several forms of automatic valves which are claimed to give satisfaction. One of these is shown in Fig. 47. This is similar in construction to a steam trap. As air collects in the chamber, and the water line is lowered, the float drops, and in so doing opens a small valve at the top of the chamber which allows the air to escape. As the water flows in to take its place the float is forced upward and the valve is closed.

All radiators which are supplied by risers from below should be provided with air valves placed in the top of the last section at the return end. If they are supplied by drops from an overhead system the air will be discharged at the expansion tank and air valves will not be necessary at the radiators.

Fittings

All fittings, such as elbows, tees, etc., should be of the "long turn" pattern. If the common form is used, they should be a size larger than the pipe, bushed down to the proper size. The long turn fittings, however, are preferable.

Pipe Sizes

The size of pipe required to supply any given radiator depends upon four conditions; first the size of the radiator, second its elevation above the boiler, third the length of pipe required to connect it with the boiler, and fourth the difference in temperature between the supply and return. The following illustration will serve to make these points clear.

Pipe Sizes 1000127

Fig. 47.

If we should take a glass tube of the form shown in Fig. 48, fill it with water and hold it in a vertical position, we would notice that the water remained perfectly quiet; now if the flame of a lamp were held near the tube A and a few drops of coloring matter were poured into the tube, we would find that the water was in motion, and the current would be in the direction shown by the arrows. While the water in both tubes was at the same temperature, the two columns were of the same weight and remained in equilibrium. If, however, the water in column A is heater, it expands and becomes lighter than column B, and is forced upward by the heavier water falling toward the bottom of the tube. The heated water flows across the top and into B where it takes the place of the cooler water which is settling to the bottom. As long as there is a difference in the temperature of the two columns this action will continue. If now we replace the lamp by a furnace, and connect the two columns A and B at the top by inserting a radiator, we shall have the same illustration in practical form as utilized in hot water heating. (See Fig. 49).

Pipe Sizes 1000128

Fig. 48.

Pipe Sizes 1000129

Fig. 49.

The heat given off by the radiator always insures a difference in temperature between the columns of water in the supply and return pipes, so that as long as heat is supplied by the furnace the flow of water will continue. The greater the difference in temperature of the water in the two pipes, the greater the difference in weight, and consequently the faster the flow. The greater the height of the radiator above the heater the more rapid the flow, for the difference in weight between two columns 1 foot high and two columns 10 feet high is ten times as great and if there were no friction in the pipes the flow would be directly proportional to the elevation of the radiator above the heater. The quantity of water discharged by a given pipe under constant pressure varies inversely as the length of pipe; that is, if a pipe 100 feet long will discharge 10 gallons per minute under a given pressure, it will discharge only half as many gallons if the length is increased to 200 feet, the pressure remaining the same.

As it would be a long process to work out the required size of each pipe for a heating system, the following tables have been prepared, covering the usual conditions to be met with in practice.

Table III gives the number of square feet of direct radiation which different sizes of mains will supply for varying lengths of run.

TABLE III.

Size of Pipe.

Square Feet of Radiating Surface.

100 ft. Run.

200 ft. Run

300 ft. Run.

*

400 ft. Run.

500 ft. Run.

600 ft. Run.

700 ft. Run.

890 ft. Run.

1000 ft. Run.

1

30

1 1/4

60

50

1 1/2

100

75

50

2

200

150

125

100

75

2 1.2

350

250

200

175

150

125

3

550

400

300

275

250

225

200

175

150

3 1/2

850

600

450

400

350

325

300

250

225

4

1200

850

700

600

525

475

450

400

350

5

1400

1150

1000

700

850

775

725

650

6

1000

1400

1300

1200

1150

1000

7

1706

1600

1500

These quantities have been calculated on a basis of 10 feet difference in elevation between the center of the heater and the radiators, and a difference in temperature of 17 degrees between the supply and return.

This table may be used for all horizontal mains. For the vertical risers or drops, table IV may be used. This has been computed for the same difference in temperature, and gives the square feet of surface which different sizes of pipe will supply on the different floors of a building, assuming the height of the stories to be 10 feet. Where a single riser is carried to the top of a building to supply the radiators on the floors below, by drop pipes, we must first get what is called the "average elevation of the system" before taking its size from the table. This may be illustrated by the following diagram, (see Fig. 50).

In A we have a riser carried to the third story and from there a drop brought down to supply a radiator on the first floor. The elevation available for producing a flow in the riser is only 10 feet, the same as though it extended only to the radiator. The water in the two pipes above the radiator is practically at the same temperature and therefore in equilibrium, and has no effect on the flow of the water in the riser. (Actually there would be some radiation from the pipes, and the return, above the radiator, would be slightly cooler, but for purposes of illustration this may be neglected). If the radiator was on the second floor the elevation of the system would be 20 feet (see B), and on the third floor 30 feet, and so on. The distance which the pipe is carried above the first radiator which it supplies has but little effect in producing a flow, especially if covered, as it should be in practice. Having seen that the flow in the main riser depends upon the elevation of the radiators, it is easy to see that the way in which it is distributed on the different floors must be considered. For example, in B, Fig. 50, there will be a more rapid flow through the riser with the radiators as shown than there would be if they were reversed and the larger one were placed upon the first floor.