The object of this chapter is to acquaint the reader with a general working knowledge of the subject. Few laboratories are equipped with facilities for handling molten metals, so that the experimenter will often be obliged to have casts made by commercial foundries; but they are usually equipped for him to make his own patterns from which the casts are made. He will find it economical to do so, for the cost of having the patterns made may be many times the cost of the casts. In order to construct his patterns with intelligence, he should have an understanding of current foundry practice.

As soon as the investment has set, pull out the supporting rod, and around the hole, or sprue, carve a small funnel to receive the metal (Fig. 5). Be careful to remove any bits of material that fall down the sprue. Place the whole thing in a ring stand and heat with a Bunsen burner. Heat until the whole mold is red hot, to make sure that the last traces of wax and moisture are driven off (Fig. 6).

The mold is now completed and ready to receive the metal. There remains the problem of getting the metal to flow into the mold. When the sprues are tiny and the masses of metal small, the surface tension may prevent the metal from flowing. Also, the trapped air in the mold will be a real obstacle. Dentists use a vacuum method for getting the air out of the mold. A metal disk with a hole in the center communicates with a small tank by means of a pipe. The pipe has a stopcock, which is kept closed until the metal is melted. A small hand pump is attached to the tank for reducing the air pressure to about one-half to one-fourth atmospheric pressure. The mold is placed on the disk, and the metal is melted in the funnel on the top of the mold with a blast lamp. When the metal is completely molten, the cock is opened, and the air in the mold is drawn out through its porosity, allowing the metal to flow in. Of course, air leaks around the cast, but that does not matter, since the object is to reduce the air pressure in the mold for only a few seconds. This method will make the finest of casts.

Fig. 7 shows a setup of the type described above, which can easily be assembled with the aid of a bicycle pump and a few accessories.

Another method is to force the metal into the mold with steam. The metal is melted in the top of the mold as before. When it has melted, a large piece of moist clay is quickly pressed down onto the mold. The steam generated will force the metal into the mold. When this method is used, the mold should be placed on a perforated plate to allow the air to escape at the bottom (Fig. 8).

Still another method introduces the metal into the mold by means of a centrifuge, which need be only a simple device consisting of a bar pivoted to rotate in a horizontal plane. A spring is arranged to rotate the arm, and a movable trigger acts as a stop to prevent the rotation until the metal is molten. The mold is placed in a holder at the end of the arm with its sprue facing the pivot. The crucible, a small trough of firebrick or any other suitable material, is mounted on the arm with the end of the trough adjacent to the sprue. The procedure is to cock the spring, place the mold in the holder, place the metal in the trough, and heat it with a hand torch. When the metal is ready, the trigger is pulled, and the whole arm spins on the pivot, the centrifugal force carrying the metal into the mold. For this method the bottom of the mold should be made especially strong; otherwise it may give way and the molten metal be thrown about the room. This method is used by commercial jewelers as well as by some dentists. Fig. 9 shows a centrifuge of the type described. For casting the object illustrated in Fig. 1, this method should probably be used, since the top of the mold would hardly be large enough to contain the required amount of metal.

In any of the foregoing methods metals such as gold, silver, copper, brass, and so forth, should be liberally sprinkled with borax as they are fused. This treatment prevents oxides from forming. Metals such as lead, babbitt, solder, and so forth, may be kept covered with powdered charcoal for the same purpose.

Patterns for sand casting. Sand casting from permanent patterns of metal or wood is the method commonly employed for all manner of mechanical parts, irrespective of size or of metal. Die casting is the only other method of importance, but it is restricted to commercial work in which the otherwise prohibitive cost of the metal dies is absorbed in the tremendous quantity of casts to be produced. In making the patterns for sand casting, the first step is to make a careful drawing of the object (Fig. 10). Over this drawing a second one should be made (Fig. 11), which is the pattern drawing. It can be made on thin paper. For future convenience it is well to dimension the pattern drawing copiously.

As an object for the illustrations that demonstrate sand casting, a polar axis for a telescope has been chosen. The axis is designed to have babbitt bearings. Such a design eliminates the necessity for chucking the whole axis in a lathe and boring it–a job to be accomplished only in a very large lathe. Figs. 10 and 12 show drawings of the parts required, and Fig. 24 shows the completed mounting.

In planning the pattern, the first thing that must be determined is the plane of division of the mold. This plane should pass through the object in such a way that both parts of the pattern can be drawn from the sand mold. The plane must intersect the object in such a way that all points on the object, when they are projected normally to the plane, will fall on or within the intersection. It is to this plane that all questions of "draft" are referred. If there is no single plane that will fulfill the requirements, one must be chosen which will come the nearest possible, and the parts that do not conform will have to be accomplished with cores.

The pattern will be divided along the plane of division unless the plane happens to coincide with one surface of the object, in which case the pattern will be in one piece (Fig. 12) Since the usual cases are those in which the pattern must be divided into two halves, we will describe the procedure for a two-piece pattern. The designer need not be limited to the two-piece pattern, but the multi-piece pattern is so seldom employed that it is not within the scope of this chapter to treat of it. The reader is referred to books on commercial founding.

Since parallel-sided objects can not be withdrawn from sand without the friction of the sides destroying the walls, it is necessary to taper the sides slightly. This taper is called "draft," and it may be as little as 1/2 for fine commercial work. In most work, however, 3 is regarded as the proper draft angle. Round objects need no draft when the plane of separation passes through the axis. When the requirements of the object are such that there can be no draft on one of the faces, that is, when that face must be at a right angle with the plane of separation, then the draft on all opposite faces should be doubled (Fig. 11).

If there is a projection on one of the pieces of the pattern that does not touch the dividing plane, then between it and the dividing plane must be provided a separate block which can be removed after that half of the mold is made. This block is called the "false core," and the volume which it occupies in that half of the mold will be replaced by an equal volume of sand in the other half of the mold. The false core must also have draft. It will be seen later in the description of the process for making a mold that false cores must be used on only one piece of the mold, and that this must be the piece which has the holes to receive the pegs that hold the parts of the pattern in alignment. Since this is the part that is molded first, it must be capable of lying flat, with the plane of separation in contact with a table. If false cores are required on the other piece of the mold, they must be regarded as true cores. The blocks must be fixed in place and a core box made for them as will be described later in this chapter.

If there is to be a hole through the object, bosses must be put on the pattern wherever the hole comes through the surface of the object. These bosses are to form s6ckets in the mold for carrying the ends of the core, thereby preventing its floating in the molten metal. The bosses must have draft angles to conform to the rest of the pattern. Separate drawings should be made of the cores, the drawings to include the bosses just mentioned (Fig. 18).

A core is any piece of sand or mold material that is molded separately and inserted in a mold, thereby forming a hole or cavity in the finished cast. If the core comes to the surface of the pattern at only one place, the boss to receive it must be long enough to support the core as a cantilever. Should this prove impractical, projections may be made on the core which will bear against the inside of the mold and carry the core. These projections will leave holes in the finished cast which will have to be plugged. Still another way is to use metal supports to carry the core. The supports are in the form of pins with broad heads and crooked shanks, and they are pressed into the sand by the molder. They weld with the metal of the cast. It is well to remember that not only must the core be supported against gravity in the empty mold, but it must also be restrained from floating in the molten metal when the mold is filled.

Since nearly all metals shrink in solidifying, patterns have to be made enough larger to compensate. The amount of shrinkage varies with the metal. A list of shrinkage scales is given in Table I.

For patterns that are to be cast many times, it is usual to reproduce the original wooden pattern in aluminum or some other metal, and to use this metal pattern for making the molds. In such case the original wooden pattern must be made large enough to allow for shrinkage of both metals. The allowance should be the sum of the shrinkages of both metals. Metal patterns are better for repeated use, since commercial foundries are quite rough on the wooden ones. Furthermore, several metal patterns may be reproduced from the first wooden one, and from these many molds may be made at a time. Thereby the cost of production is reduced.

The rate of solidification of metal in the mold is a function of the thickness of the metal. Since the greater part of the shrinkage occurs at the instant of solidification, it is apparent that unequal thicknesses of metal in the same pattern win cause a certain amount of distortion and warping. For this reason patterns are usually designed to have as uniform a thickness throughout as possible. Of course, the warping may be unimportant, since small projecting lugs on large casts will be carried by large masses of metal. Nevertheless, care should be exercised in designing the patterns to preserve the equality of thickness. Metals cooling from the molten state do not necessarily shrink evenly as the temperature falls. White cast iron, for instance, shrinks a while, then expands a little, and then continues shrinking. Gray iron expands twice and phosphorous iron three times. Thus, in casts of uneven thickness one part may be shrinking while another is expanding, with the result that the casts will be under stresses. In fact, the casts may be broken by the stresses. The hand wheels that are used to tighten the brakes on freight cars are made with spiral spokes, because the uneven cooling of the thick hub and the thin rim sets up stresses in the spokes which might break them if they were straight.

Pure metal, such as aluminum, copper, tin, or zinc, may be cast in almost any thickness. The case is different with the alloys, especially with those alloys in which there is a great difference between the melting points of the constituent metals. The alloys cooled slowly have time to grow large crystals. In fact, some of the constituent metals may crystallize out, so that thick alloy castings may be weak and full of crystal pockets. The best thickness for alloy castings seems to be between 3/16 and 5/16 inch. Even 1/8 inch is not too thin. All metals, pure or alloyed, must be cast in a thickness sufficient for the metal to reach all parts of the mold before it sets. However, this matter is not too important, for a good molder can arrange his gates to insure that the molten metal reaches all parts of the mold before it solidifies.

Patterns are usually made of wood–white pine, sugar pine, or mahogany. The wood should be clear and well seasoned. If the pattern is in two pieces (not including false cores), they must be pinned together in such a way that they can be separated and reassembled by the molder. In making the pattern, it is well first to make the division part with its pins. Wooden dowels may be used, or special pins and sockets, which can be purchased from a hardware dealer (Figs. 13 and 14).

In general, the construction of patterns follows the usual practice in carpentry and cabinetmaking. The parts may be glued together at will, and all the tricks of joinery may be employed, provided the finished result is the desired shape. The exterior surfaces of the pattern should be carefully smoothed. Any roughnesses or irregularities in the surfaces, particularly in the draft surfaces, will mean that the molder will have to rap the pattern vigorously to free it from the mold, with the result that the mold, and consequently the cast, will be larger than designed.

Each piece of pattern, including the false cores, should be provided with a rapping plate, which is set in flush with the parting face (Fig. 15). These plates can be purchased from a hardware dealer. They provide a hole for rapping and a hole for a lifting screw, or handle, which the molder will use in withdrawing the pattern from the sand.

In making casts it has been found that sharp, internal corners and dihedral angles are a source of trouble. The sharp edges of the mold may crumble, or strains in the metal may cause the cast to break at these corners. To avoid such trouble, it is the practice to round the corners, and the corners so rounded are called "fillets." They may be carved in the pattern, but it is easier to make them of wax. The wax can be purchased at hardware stores. It comes on spools or in strips in the form of an extruded ribbon of the proper profile. Before the ribbon is used, the pattern must be shellacked. The ribbon, cut into convenient lengths, is rubbed into place with a hot tool that has a ball-shaped end (Fig. 16). The fillets are sold in sizes according to the radius desired, and the tools can be made or purchased to correspond. After the fillets are finished, the whole pattern is again shellacked. In patterns that are to be used only once the fillets may be made of plastiline or plasticine, which can be purchased at an artists' supply house. Only the hardest grade should be used. It will stick quite well to the shellacked pattern, and may be modeled with the fingers or with wooden modeling tools. It should be shellacked when it is finished. For larger work, leather fillets can be in many sizes. They are glued directly to the wooden pattern, finished with sandpaper, and shellacked.

If there are no cores, clear shellac is used over the whole pattern; but if there are cores, the projections which correspond to the ends of the cores are left in the clear shellac finish or are painted red, while the rest of the pattern is finished with shellac and lampblack. The shellac and lampblack are usually mixed to give a dead black finish. The mixture is liberally thinned with alcohol. The reason for this color distinction is that it tells the molder where the cores belong (Fig. 17).

If the pattern has cores, the next problem is to make the core boxes. These are essentially wooden molds, in which the cores are cast. The requirements for draft and shrinkage are the same as for the patterns. If the core requires a two-piece mold, as for cylindrical shapes and the like, the plane of division need have no relation to the plane of division of the original pattern. There is no need to pin the two sides of the core box together, for the two halves of the core are made separately- and are stuck together after they are baked. When the two halves are alike, it is necessary to make a box for only one side. If simple cylindrical cores are required, it is unnecessary to make the boxes at all, for most foundries have stock boxes, or even stock cores. In general, core boxes are carved from solid wood, and they are frequently more complex than the pattern itself (Fig. 19).

The core in the illustration is shown to be made in two pieces, because the surface in the mold which it has to fit has draft. If the pattern had no draft on this top face, the core could easily be in one piece.

Core boxes are made of sugar pine or any other clear and workable wood, and they are finished with shellac like the patterns. One point is important: Since the pattern shows only the ends of the core, and since the core may be unsymmetrical longitudinally, it is well to make the ends different, so that the molder will get the core in the proper orientation.

The process of making cylindrical core boxes, as it has been practiced for years, is very interesting. A core-box plane is used, the face of which is two surfaces at 90 to each other. The plane iron comes through the edge of these surfaces and is sharpened to conform to them. Two parallel lines are drawn on the surface of the stock that is to be used. The plane is made to cut away all the wood possible without removing the two lines. The 90 angle automatically generates a semicircular groove (Fig. 20). The plane can be used to generate conical shapes as well.

Sand casting. The procedure for making the mold is as follows: The molder first separates the pattern and sets the part that has no pegs with its separation plane face down on the table, putting the false cores in place. Around this piece he puts a wooden or metal frame called a "drag." The "drag" and the "cope" together constitute the "flask," which is a framework to contain the sand forming the mold. The cope and the drag are open rectangular boxes without lid or bottom. They are fitted with crossbars to help hold the sand. There are three sockets in the rim of the drag and three corresponding pins in the rim of the cope which permit them to be separated and reassembled in the same relation as shown in Fig. 21(a).

The molder next sprinkles molding sand through a riddle, or sieve, until the drag is half full. Molding sand is a mixture of fine clear sand and a small amount of clay, and sometimes a little powdered charcoal or graphite. It has been moistened until a handful compressed will retain the print of the hand and will form a fairly firm piece, but moistened not so much but that it can be shaken through a riddle of about 1/4-inch mesh, the riddle being shaken vigorously. Molding sand is used over and over again with only the addition necessary to replace the inevitable losses.

When the drag has been half filled, the sand is carefully tamped around the pattern. A wooden implement is used which is about the shape and size of a dumbbell except that one end is fiat and the other is a blunt, truncated wedge as shown in Fig. 21(b). More sand IS added and tamped until the drag is level, full, and firmly tamped. The sand is then perforated with a thin metal wire to assist the escape of steam and gases that are given off when the metal is poured in. This is illustrated in Fig. 21 (c). The drag is now picked up bodily and inverted on the table, exposing the separation face of the pattern. The false cores are removed as shown in Figs. 21(d) and 21(e).

The other half of the pattern is next placed on the part already molded, the pegs insuring alignment of the parts. Dry sand is sprinkled over all the exposed sand of the mold as illustrated in Fig. 21(f) to prevent the two halves of the mold from sticking together. The cope is placed on the drag. It is filled with sand and is tamped, as was the drag, as shown in Fig. 21(g). The sand is pierced many times with a thin wire, as in the case of the drag, in the manner illustrated in Fig. 21(h).

At a point in the sand well clear of the pattern, a sprue is cut deep enough to reach a little below the separation plane. It is cut with a piece of thin-walled brass tubing which is pressed gently into the sand and removed, bringing with it a plug of sand as shown in Fig. 21(i). The hole left is about an inch in diameter, large enough for casts of 10 to 100 pounds. It is well to cut the hole an inch at a time. It is not cut to communicate directly with the pattern, for the force of the descending metal might injure the mold. A trough is cut around the top of the hole as illustrated in Figs. 21(j) and 21(k), and into this the metal will be poured.

The cope is now carefully lifted off the drag and laid beside it, face up. A rod is inserted in the rapping plate and rapped smartly in all directions to break the adhesion of the sand as shown in Fig. 21(1). A handle is screwed into the rapping plate, and the pattern is carefully withdrawn from the sand as illustrated in Fig. 21(m).

A channel, or "gate," is next cut along the parting plane from the mold to the sprue. The reason for not cutting the sprue to communicate directly with the mold is that the sudden rush of heavy metal directly into the mold might injure it. The horizontal gate breaks the fall of the metal as shown in Figs. 21(m) and 21(n).

The mold should be closely inspected for broken corners and edges that must be carefully mended and modeled with steel molder's tools. Bits of sand that have fallen in are brushed away with a soft brush or are blown out with a bellows. A cloth bag half full of powdered graphite is shaken over the mold, and frequently powdered graphite is painted on the surface of the mold with a soft camel's-hair brush as shown in Fig. 21(o). In iron and steel castings this coating of graphite is responsible for the very hard surface layer. Its function is to harden the surface of the mold, and also partly to fill the grainy surface of the sand.

The pattern is removed from the cope as illustrated in Fig. 21 (p) in the same manner as from the drag. If the cast is complicated or very large, a riser or several risers are cut in the cope. They are like the sprue, except that they are cut from the highest parts of the mold. In complicated molds they assist in carrying off the entrapped air; in large molds they serve to collect the slag that rises to the surface, and to provide reservoirs for extra metal that will flow back into the mold as the metal within cools and shrinks. This is shown in Fig. 21(q).

Cores are made by filling the core boxes with a mixture of coarse sand and a binder. The core boxes are tamped full and leveled off with a straightedge. They are then inverted on sheets of metal, rapped, and removed, leaving the halves of the core, which are baked for a few hours in an oven. The baked halves are stuck together with the mixture of which they are made, or with mucilage or paste, and are baked again (Fig. 22). Pure silica sand which has passed through a 50-mesh screen and has been retained on a 70-mesh screen should be used. Many substances may be used as binders a complete list of which the reader will find in books on founding. The ones most easily available are given in Table II, together with the amounts to be used.

When the cores are thin and fragile, they are frequently reinforced with iron wires. If the cores are very bulky, provision should be made for conducting away the gases that will be formed in them and that will blow the cast apart if allowed to remain unvented. Such provision is made by laying strips of wax in the sand as the cores are being made. When the cores are baked, the wax is driven off, leaving holes for the gases to escape through.

The completed cores are placed in the molds, and the flask is reassembled. The halves–the cope and drag– should be securely clamped together, for otherwise the metal may actually float the cope off and let out the metal as shown in Fig. 21(t). The mold is now ready to be filled as illustrated in Figs. 21(r) and 21(s).

For production work, when numbers of identical casts are required, match boards are used. These are boards having pins and sockets that correspond to those in the cope and drag, one to fit the cope and one to fit the drag. The two halves of the pattern are permanently fixed to the boards, properly placed to insure correct matching of the two. The halves of the mold are then made separately. They may even be made by separate workmen, and are not assembled until the time for the pouring. In the case of metal patterns several are placed on each board so that a number may be cast at a time.

The molten metal should be poured gently into the mold. It is well to skim back the slag and scum that form on the top of the metal in the crucible or ladle. For the soft metals, which can be handled in iron containers, a kettle having a spout that communicates with the bottom may be used. The floating slag and oxides will be left in the kettle. In molds involving large masses of metal it is common for the workmen to stir the metal in the mold. They churn it up and down in the riser with an iron rod.

This action prevents the metal in the riser from solidifying until the outer shell of the cast has set. As the center of the cast sets, the metal in the riser descends and prevents shrinkage pockets from being left in the top of the cast. The escaping gases usually burn briskly, but if they burn too much at the division of the mold, the flames should be doused with water to save the flask from being badly charred.

Large casts are usually left overnight to cool. Smaller ones may be immediately dug out of the mold. In fact, it is the practice in a certain instrument company to remove the small casts immediately from the sand and throw them red-hot into the water. The steam generated blows away the sand and even blows out the cores, leaving the casts virtually clean.

When the sprues and gates and risers have been cut off, the cast is ready to be machined or otherwise finished as shown in Fig. 21(u).

In having the casts made it is often desirable to estimate the weight of the finished casts before they are ordered. Such an estimation is easily made: Weigh the pattern, allow for the cores, and multiply the result by a coefficient which is the ratio of the weight of the pattern material to the weight of the metal of the cast. A short list of these coefficients is given in Table III.

If one is making one's own casts, enough metal must be figured to include the sprues and risers. Foundries do not charge for this metal, since they cut it off and use it over again.

Babbitt metal is essentially a mixture of lead and tin with the addition of enough antimony to cause the metal to expand slightly when it freezes. Many variations of this alloy are commercially available having different properties, some being suitable for high-speed bearings and others for bearings which must work under heavy loads. Bearings of babbitt are usually cast in some sort of carrier, so that the babbitt forms a bushing. These bearings are sometimes cast as solid plugs, which are then bored and reamed to size. More frequently they are cast in two halves with a dummy shaft in place. For many purposes such bearings are good enough just as they are, but for precision bearings they should be scraped to fit the spindle. A dummy shaft should always be used, because the hot babbitt metal is apt to warp the shaft. Fig. 23 shows the method of casting a split bearing.

Cuttlebone casting. There remains one other method of casting which might be found useful in the laboratory. This is cuttlebone casting. The desirable properties of the method are the ease and rapidity with which a cast can be produced: A mold can easily be made and filled in half an hour. The objects to be cast should not exceed the dimensions of 1/4 inch in thickness, 1-1/2 inches in width, and 3 inches in length. The patterns should be of metal, since they must be subjected to pressure. Draft angles can be very slight, or they may be ignored altogether.

There can be no cores. Fig. 25 shows the method quite clearly. Cuttlebones can be procured from a pet shop or drug store for a few cents. The soft, calcareous face is easily crushed and takes a very firm imprint of any object that is pressed into it. Difficulty of pressing thick patterns into the cuttlebone can be overcome by repeatedly pressing the pattern into the bone, the crushed material being brushed out of the imprint after each operation. The cuttlebone will stand quite high temperatures and is sufficiently porous to allow the air in the mold to escape.