Pipes and the steam main

A steam main is the plumbing between the boilers and the engine, and for most of the period it was cast iron, bolted up in lengths. How to lay one in the game is in the pipe networks guide, and the blocks are in the catalogue. This page is about the real pipe, how it was made and joined, and why higher pressure moved to welded tube.

Old iron pipes left over a mine shaft at Lane End. The flange at each end is part of the pipe, not a piece added to it: the joint was made in the mould, and only the bolts were left to fit.
Old iron pipes left over a mine shaft at Lane End. The flange at each end is part of the pipe, not a piece added to it: the joint was made in the mould, and only the bolts were left to fit.
The game's cast main in its planned drawing: a one-block straight with a flange at each end, drilled for the bolts that hold it to its neighbour., iso
The game's cast main in its planned drawing: a one-block straight with a flange at each end, drilled for the bolts that hold it to its neighbour.

How a cast pipe was made

Bolland, writing for iron-founders in 1892, shows a trade that cast every pipe whole around a core, flange or socket included. The bore is the core's surface and nothing is bored afterwards (Bolland pp. 209-232).

Bolland's core for a large elbow pipe, cast standing on its end (Bolland p. 228, Fig. 193). This is a core, the part that forms the bore: bottom plate A and plate B are bound together by hook bolts through courses of brick, and the curved outside is struck in loam. The whole is built in a pit, around which the cope is built and closed. Bolland's example is an elbow 5 ft across (Bolland p. 227).
Bolland's core for a large elbow pipe, cast standing on its end (Bolland p. 228, Fig. 193). This is a core, the part that forms the bore: bottom plate A and plate B are bound together by hook bolts through courses of brick, and the curved outside is struck in loam. The whole is built in a pit, around which the cope is built and closed. Bolland's example is an elbow 5 ft across (Bolland p. 227).

Most pipe was made in one of three ways, by size and by how many were wanted. Some shops kept "six to ten men" constantly on jobbing pipes from 2 to 24 in (Bolland p. 142). Pipes up to 18 in of any form could be rammed in green sand from a kit of wedge-shaped half rings (Bolland pp. 324-332). Very large elbows were made in short lengths lying "on the flat", but a long piece with an elbow on one end was better built standing up (Bolland p. 227).

The straight pipe of the waterworks trade went on end. At the Vauxhall Foundry in Liverpool, about 1863, "large numbers of straight lengths were being swept on end in loam" for the corporation's order, and a new arrangement was being built for socket pipes 3 ft across and 10 ft long (Bolland p. 335). In loam a sweep turning on a spindle strikes the round face (Bolland p. 227), and where patterns were practicable the same foundry rammed others in green sand (Bolland p. 335). Later the "great and increasing demand" brought plants built only for pipe (Bolland p. 336). A core mixture that let a casting be run from the top was "an absolute necessity sometimes where a sound casting is required" (Bolland p. 125).

Joining it

Before the wrought-iron tube, a joint was a thing to be made tight on site, with what the fitter could caulk or bolt.

Two older cast-iron joints (Hawkins p. 248). Above, Fig. 115: the hub and spigot, with iron borings caulked into the ring, "probably the oldest kind of joint", still used for hot water and formerly for low-pressure steam. Below, Fig. 116: the flanged joint, with a wrought-iron ring A wrapped in yarn between the flanges, sometimes dipped in red and white lead, and a rust joint B caulked outside it. Facing the flanges came after this (Hawkins p. 248).
Two older cast-iron joints (Hawkins p. 248). Above, Fig. 115: the hub and spigot, with iron borings caulked into the ring, "probably the oldest kind of joint", still used for hot water and formerly for low-pressure steam. Below, Fig. 116: the flanged joint, with a wrought-iron ring A wrapped in yarn between the flanges, sometimes dipped in red and white lead, and a rust joint B caulked outside it. Facing the flanges came after this (Hawkins p. 248).

Iron borings mixed with sal-ammoniac rust together into a solid joint, and the rust joint between flanges is described in the engine books (Tredgold pp. 254-255, Farey p. 487, Bourne 1846 p. 227). Bolts draw the flanges up; the packing is what holds the steam. Hawkins's next step is the facing or turning up of the flanges, with a gasket of rubber, copper, paper or cement between them, and for some work a lip and recess to keep the gasket from being blown out (Hawkins pp. 248-249).

Wilson explains why cast iron was trusted here and not in the boiler. Cast iron "is almost invariably used, except with marine boilers, for pipes to carry the steam to the engine cylinders at the full boiler pressure" (Wilson p. 27), and the engine cylinders were of cast iron too, though he called it foolhardy to make a boiler of the same material to work at the same pressure. His reason is the fire: in the cylinder the heat is evenly spread, with no furnace and no cold draught through the doors, where a horizontal vessel with a fierce fire under it is severely strained by the unequal expansion (Wilson p. 27). The danger was always the casting itself. "A very slight hidden or surface defect, in an unsound casting, is sometimes sufficient to lead to a sudden explosive fracture" (Wilson p. 26).

A battery of boilers on one main, as Hawkins recommends it (Hawkins p. 277, Figs. 153 and 154). Each boiler has a riser from its nozzle, turned over to run backwards into the main steam pipe, with its own stop valve C. The valves have flanged ends F and F', so one can be taken out for repair without disturbing the rest of the piping. The stem should lie horizontal or nearly so, Hawkins adds, "in order that the valve may not trap water".
A battery of boilers on one main, as Hawkins recommends it (Hawkins p. 277, Figs. 153 and 154). Each boiler has a riser from its nozzle, turned over to run backwards into the main steam pipe, with its own stop valve C. The valves have flanged ends F and F', so one can be taken out for repair without disturbing the rest of the piping. The stem should lie horizontal or nearly so, Hawkins adds, "in order that the valve may not trap water".

A main also has to move. Wrought iron was reckoned to grow 1/150,000 of an inch for each inch of length and each degree, so a 100 ft pipe heated to 338 degrees wants 2.7 in of play in an expansion joint (Hawkins pp. 276-277).

Why higher pressure moved to welded tube

Wrought iron cannot be poured, so a tube of it has to be made from a strip by bringing the edges together and welding them. For a long time this was the gunsmith's job, a "skelp" bent round a mandrel and hammered up piecemeal (Marks p. 2).

Cornelius Whitehouse's butt-welding apparatus of 1825 (Marks pp. 5-6, Figs. 6 and 7, labelled for this page). A strip already bent round is brought to the point of fusion in a hollow fire, then gripped and drawn out by a chain through dies in the screw press, which press the edges together and weld them. Whitehouse's specification was for "tubes for gas and other purposes"; Marks says the process was still in use in all such tube works at the time he wrote.
Cornelius Whitehouse's butt-welding apparatus of 1825 (Marks pp. 5-6, Figs. 6 and 7, labelled for this page). A strip already bent round is brought to the point of fusion in a hollow fire, then gripped and drawn out by a chain through dies in the screw press, which press the edges together and weld them. Whitehouse's specification was for "tubes for gas and other purposes"; Marks says the process was still in use in all such tube works at the time he wrote.

Lap welding came next for the larger sizes: Russell's specification of 1845 rolls the overlapped seam against a mandrel (Marks p. 122). The processes in use were "evolved and established in and around Wednesbury" between 1827 and 1850 (Marks p. 122). By Hawkins's day the trade divided pipe at about 1 1/4 in, butt-welded below and lap-welded above (Hawkins p. 248). Marks sets the limit for steam tube at 2 in, and in some cases less (Marks p. 124).

A lap-welding line of 1918, with the white-hot pipe coming off the welding rolls at Youngstown Sheet and Tube, Ohio. The strip has been turned into a tube with an overlapped seam before this, and the seam is pressed together by a roller over a mandrel, the method of Russell's specification (Marks p. 122).
A lap-welding line of 1918, with the white-hot pipe coming off the welding rolls at Youngstown Sheet and Tube, Ohio. The strip has been turned into a tube with an overlapped seam before this, and the seam is pressed together by a roller over a mandrel, the method of Russell's specification (Marks p. 122).

Hawkins gives the welded pipe's advantages as "lightness, the ease with which various lengths can be obtained and its strength" (Hawkins p. 249). Wrought iron should have an ultimate strength of 20 tons a square inch (Marks p. 124) against about 6 for cast iron (Wilson p. 26). The weld is the weak line: scale caught in it can leave a tube "not even gas tight", and the screw cut on its end thins the wall, so every tube was given a hydraulic test before it left the works, seldom less than 300 lb a square inch for steam (Marks p. 124).

The joints of the welded pipe (Hawkins p. 249, Figs. 118 to 120). Fig. 118 is the plain coupling, a wrought-iron sleeve with tapered threads at both ends, the pipes not quite meeting in it. Fig. 119 is a pipe into a cast fitting, with a large recess inside for the tapping, which raises the friction. Fig. 120 is a union. Above 4 or 5 in, a flange union was generally used for valves and such, so the old flange survived on the new pipe (Hawkins p. 251).
The joints of the welded pipe (Hawkins p. 249, Figs. 118 to 120). Fig. 118 is the plain coupling, a wrought-iron sleeve with tapered threads at both ends, the pipes not quite meeting in it. Fig. 119 is a pipe into a cast fitting, with a large recess inside for the tapping, which raises the friction. Fig. 120 is a union. Above 4 or 5 in, a flange union was generally used for valves and such, so the old flange survived on the new pipe (Hawkins p. 251).

Period basis, and what is the game's own

  • Period: cast pipe with flanges cast on, the bore cored and never bored (Bolland pp. 209-232). The game's cast segments are made from a cast part with a core, the flanges drilled, and the bore left as cast, and this route is planned: the cast segments ship without a recipe today.
  • Period: straight pipe swept on end, large elbows lying or on end by size (Bolland pp. 227, 335). The game casts a one-block straight on end in the pit flask, and two- and three-block lengths lying in a closed flask, planned. The choice of on-end is the sand the flasks hold, a limit of the game's own.
  • Period: a rust joint caulked between the flanges, then faced flanges with a gasket (Hawkins p. 248, Tredgold pp. 254-255). The game plans the same two grades: flanges as cast, and an optional facing at the lathe that raises the pipe's limit. No book gives a pressure for either joint, so the grades are the game's own.
  • Period: wrought pipe screwed into couplings and fittings, with the flange union on the largest sizes (Hawkins pp. 249-251). The game's rolled pipe is octagonal with screwed ends and no flange, so a rolled run joins only rolled pipe and machine ports. The period flange union does not exist in the game.
  • Period: Marks's butt- and lap-welded tubes are of iron, from strip the tube maker bought rolled from the ironmaster (Marks p. 122). The game's rolled pipe is of open-hearth steel, planned to come from skelp cut off the cast bloom, one skelp to a straight, bent in a machine, with no separate weld.
  • Period: mains on rollers, short pieces of round bar or two-inch cast pipe (Overman p. 414, for blast mains). The game plans drawn supports on the same lines.
  • The game's own: pipes do not grow with heat, and a main needs no expansion joint.
  • The game's own: steam stays steam in a pipe. A run sheds temperature, but water is made only at a device that condenses. There is no water hammer and no drip pipe.
  • The game's own: an over-pressure run does not crack without warning. A pipe past its rating is given a grace of 30 seconds and then one segment goes, with no chance of a hidden flaw. See vanilla against reality.

The numbers

Pipe ratings are the game's own ladder (the scaling page); the books give materials and test pressures, not a working pressure for a pipe line. "Planned" is the rating in design. The boiler each one has to carry is on the Cornish boiler page.

In the gamePeriod figureHow it was scaledGame value
Plated pipe, bursts atNone for a pipe lineThe game's own2.5 atm
Cast pipe, rough, works / bursts atNone for a pipe line; rust and packed joints (Hawkins p. 248)The game's own, set against the Cornish boiler's working pressureToday one rating, 5.0 atm burst; planned 3.6 / 4.5 atm
Cast pipe, faced, works / bursts at80 to 90 lb, a boiler of vertical cast pipes 4 or 6 in across (Wilson p. 27)The boiler's figure, taken to the pipePlanned 6.0 / 7.5 atm
Rolled pipe, bursts atHydraulic test seldom under 300 lb, about 20 atm, for steam tube (Marks p. 124)A test pressure, not a working one; the rating is the game's ownPlanned about 20 atm; today 12 atm
Flow, plated / cast / rolledNoneThe game's own50 / 120 / 250 L/s
Strength of the metal6 tons a sq in cast (Wilson p. 26), 20 tons a sq in wrought (Marks p. 124)Not used directlyDiffers by tier only through the ratings above
Expansion2.7 in on 100 ft at 338 degrees (Hawkins pp. 276-277)Not modelledNone

Open, and what you can tell us

Bolland describes straight pipe swept on end and large work built standing, and never says why, and neither does any other book here. Plain physics offers a reason: gas and dross float, and in a mould stood on end they gather at the top end instead of along one whole side of the pipe. If you know a founder's source that states it, or that gives a working pressure for a cast steam main as a line, we would like to read it. The books at hand also do not say how the bolt holes of a pipe flange were made, so the game drills them as it drills other cast work.

Picture credits