Chapter XXXIV: Wood-Working (2)

"Let us suppose the pattern-maker to be uninformed of the purpose the casting is to serve, or how it is to be treated: in such a case he is guided partly by his knowledge of the use of such patterns, and a consideration of being on the safe side. The form shown in Fig. 2773 would suggest itself as being a very ready method of making the pattern; by coring out the hole, it can be made parallel, which the drawing seems to require. The advantage of leaving the hole parallel is that less metal will require to be left for boring in case it should be necessary; because, if the hole is made taper, the largest end of the bore will require to have the proper amount of allowance to leave metal sufficient to allow the hole to be bored out true, and the smaller end would, therefore, have more than the necessary amount; while just the least taper given to the exterior would enable the moulder to withdraw the pattern from the mould. Made in this way, it would be moulded as shown in Fig. 2774, with the flange uppermost, because almost the whole of the pattern would be imbedded in the lower part of the flask, the top core print being all that would be contained in the cope; and even this may be omitted if the hole requires to be bored, since the lower core print will hold the core sufficiently secure in small work, unless the core is required to be very true. The parting of the mould (at C D, Fig. 2774) being level with the top face of the flange, much taper should be given to the top print (as shown in Fig. 2773), so that the cope may be lifted off easily. Were this, however, the only reason, we might make the top print like the bottom one, providing we left it on loose, or made it part from the pattern and adjust to its place on the pattern by a taper pin; but another advantage is gained by well tapering the top print, in that it necessitates the tapering of the core print at that end; so that, when the two parts of the mould are being put together, that is to say, when the cope is being put in place, if the core has not been placed quite upright, its tapered end may still arrive and adjust itself in the conical impression, and thus correct any slight error of position of the core. The size of the core print should be, at the part next the pattern, the size of the core required; for if the extremities are made of the size of the core, and the taper or draft is in excess, there will be left a useless space around the core print, as shown at A B in Fig. 2774, into which space the metal will flow, producing on the casting, around the hole and projecting from the end face, a useless web, which is called a fin, which will of course require to be dressed off the casting.

"We will now suppose that our piece, when cast, is to be turned under the flange and along the outside of the hub or body, and that the hole also is to be bored. In this case the pattern made as above would still be good, but could be much more easily made and moulded if it has to leave its own core, its shape being as shown in Fig. 2775; because the trouble of making a core is obviated, and the core is sure to be in the centre of the casting, which it seldom is when a core is used. We must, however, allow more taper or draft to a hole in a pattern than is necessary on the outside; about one-sixteenth inch on the diameter for every inch of height on work of moderate size is sufficient. The allowance for boring should be one-sixteenth inch at the large end of the hole, provided the diameter of the hole is not more than five or six inches, slightly exceeding this amount as the diameter increases; whereas, if the pattern had been made with core prints, an allowance of one-eighth inch for small, and three-sixteenths inch for larger work would be required. These are the advantages due to making the pattern leave its own core. We have still to bear in mind, however, that, if the casting require a parallel hole, a core must be used; and furthermore, if the hole is a long one, we have the following considerations: The separate dry sand core is stronger, and therefore better adapted to cases where the length of the hole greatly exceeds the diameter. Then again, if the hole require to be bored parallel, it can be more readily done if the hole is cast parallel, because there will be less metal to cut out. The casting also will be lighter, entailing less cost, provided it has to be paid for by the pound, as is usually the case. The moulder is given more work by making the core; but the saving in metal, and in turning, more than compensates for this, provided the length of the hole is greater than the diameter of the bore.

"Let it now be required that the casting is to be finished all over. It would, in that case, be preferred that if the casting should contain any blow or air holes, they should not be on the outside face of the flange, and this will necessitate that the piece be moulded the reverse way to that shown in Fig. 2773: that is to say, it must be moulded as shown in Fig. 2776, with the flange downwards; for it may be here noted that the soundest part of a casting is always that at the bottom of the mould; and furthermore, the metal there is more dense, heavier, and stronger than it is at the top, for the reason that the air or gas, which does not escape from the mould, leaves holes in the top of the casting or as near to the top as they can, by reason of the shape of the casting, rise. The bottom metal also has the weight of the metal above it, compressing it, and making an appreciable difference in its density. It must, therefore, be remembered that faces requiring to be particularly sound should be cast downwards, or at least as near the bottom of the mould as they conveniently can. Following this principle, our gland will require to be moulded as shown in Fig. 2777, P P representing the line of the parting of the mould; so that, when the cope is lifted off, the loose hub A will rise with it, leaving the flange imbedded in the lower half of the mould. It is evident that in this case the pattern must be made, as shown in Fig. 2776, the body and core prints being in one piece and the flange in another, fitting easily on to a parallel part on one end, and adjoining the core print, as shown at A. For glands of moderate size, this method is usually adopted, and it answers very well for short pieces; but in cases where the length of the body approaches, say, three diameters, the horizontal position is the best, and the pattern should be made as shown in Figs. 2778, 2779, or 2780. Even in short pieces, when the internal diameter approaches that of the external, this plan is the best, because it is difficult for the moulder to tell when his core is accurately set in position.

"For a pattern to be moulded horizontally, Fig. 2780 shows the best style in which it can be made. Its diameters are turned parallel; the required draft is given by making the rim of the flange a little thinner than at the hub, and by making the end faces of the hub and the core prints slightly rounding. If the hub is very small, as, say, a half-inch or less, and the flange does not much exceed it, the pattern may be made solid, as shown in Fig. 2778; but if the hub be small and the flange large, it should be made as shown in Fig. 2776.

Fig. 2773.

Fig. 2774.

Fig. 2775.

Fig. 2776.

Fig. 2777.

Fig. 2778.

Fig. 2779.]

"To construct the pattern shown in Fig. 2773, we proceed as follows: From a piece of plank we saw off a piece of wood a little larger and thicker than the required flange, and turn it up between the lathe centres, using a pattern makers' contraction rule, which has its measurements larger than the actual standard ones in the proportion of one-eighth inch per foot: so that a foot on the contraction rule is 12-1/8 standard inches, and an inch is 1-1/96 standard inches. The reason for this is, that when the metal is poured into the mould, it is expanded by heat; and as it cools it contracts, and a casting is, therefore, when cold, always smaller than the size of the mould in which it was made. Brass castings are generally said to be smaller than the patterns in the proportion of one-eighth inch per foot, and cast-iron castings one-tenth inch per foot; and so, to avoid frequent calculations and possible errors, the contraction rule has the necessary allowance in every division of the foot and of the inch. It is not, however, to be supposed that the possession of such a rule renders it possible for the pattern-maker to discard all further considerations upon the contraction of the casting; because there are others continually stepping in. Such, for example, is the fact that the contraction will not be equal all over, but will be the greatest in those parts where the casting contains the greatest body of metal.

"In the smaller sizes of patterns, such as those of 6 inches and less in diameter, there is another and a more important matter requiring attention, which is, that after a moulder has imbedded the pattern in the sand, and has rammed the sand closely around it, it is held firmly by the sand and must be loosened before it can be extracted from the mould. To loosen it, the moulder drives into the exposed surface of the pattern a pointed piece of steel wire, which he then strikes on all sides, causing the pattern to compress the sand away from the sides of the pattern in all directions; and as a result, the mould is larger than the pattern. In many kinds of work, this fact may be and is disregarded, but where accuracy is concerned, it is of great importance, especially in the matter of our example (brasses for journals), for they can be chipped and filed to fit their places much more rapidly than they can be planed, and it is necessary to have the castings as nearly of the correct conformation as possible. In cases where it is necessary to have the castings of the correct size without any work done to them, the shake of the pattern in the sand is of the utmost importance. If it is required to cast a piece of iron 3 inches long and 1 inch square, supposing the pattern were made to correct measure by the contraction rule, the moulder, by rapping the pattern (as the loosening it in the mould is termed) would, by increasing the size of the mould above that of the pattern, cause the casting to be larger than the pattern; that is to say, it would be longer and broader, and therefore, in those two directions, considerably above the proper size, since even the pattern was too large to the amount allowed for contraction. The depth, however, would be of correct size, because the loosening process or rapping does not drive the pattern any deeper in the mould. It follows that, to obtain a casting of as nearly the correct size as possible, the pattern must be made less in width and in length than the proper size, to the amount of the rapping; and to insure that the moulder shall always put the pattern in the sand with the same side uppermost, the word "top" should be painted on the face intended to lie uppermost in the mould. The amount to be allowed for the rapping depends upon the size of the pattern, and somewhat upon the moulder, since some moulders rap the patterns more than others; hence, where a great number of castings of accurate size are required, it is best to have two or three castings made, and alter the pattern as the average casting indicates. For castings of about 1 inch in size, the patterns may be made 1/32 inch too narrow and the same amount too short; but for sizes above 6 inches, allowance for rapping may be disregarded.

"In patterns for small cast gears, the rapping is of the utmost consequence. Suppose, for instance, we have six rollers of 2 inches diameter requiring to be connected together by pinions, and to have contact one with the other all along the rollers; if we disregard the allowance for rapping, the pinions will be too thick, and we shall require to file them down, entailing a great deal of labor and time, besides the rapid destruction of files.

"Let it be required to cast a pillow block to contain a babbitt-metal bearing. In this case there requires to be a cavity to receive and hold the babbitt metal. This is provided by casting ridges of metal around the edges of the bearing, as shown in Fig. 2781, at D E and on each side at F, the pieces D E may be made solid with the pattern, but those for the sides must be removable, having dovetails as at _c_ _c_ to hold them in position while being moulded, or in place of the dovetails, wires as at F F may be employed, in either case the pattern would be extracted from the mould, leaving the side strips to be removed afterwards. If, instead of a pillow block, a bracket or frame, such as in Fig. 2782, were required, it must be moulded in the direction of the arrow, and in that event it would be desirable to core out the journal bearing. This would be accomplished by providing a core print to block up the whole opening B. A suitable core box for the bearing would be as in Fig. 2783. The core print must project below the casting so as to form in the mould a core print for the core, and it is obvious that the core itself must be made of increased depth to the amount allowed for core print; hence the end piece B, Fig. 2783, is increased in thickness to the amount allowed for core print."

Patterns for cylindrical bodies, especially those that are hollow and thin, are constructed in pieces by a process termed "building up." The pieces are usually segments of circles, and the manner of marking them is as follows:--

Let it be required to make a pattern for a flanged pulley, such as shown in section in Fig. 2784. It would be constructed in two halves composed of a number of courses as from 1 to 8, and each course would be composed of segments of the form shown in Fig. 2785. The length of the arc of these segments must be such that it will require a certain number of these to complete the circle of that part of the cylinder which the segment is to form; and the manner of accomplishing this is shown in Fig. 2786, in which the circle C is of the diameter of the outside, while circle D is that of the outside of the pulley proper, circle E is of the diameter of the inside of the pulley rim. These circles are divided into as many equal divisions as there are to be segments in the circumference; hence the number of divisions determines the length of arc of the segments. Thus A would be a segment for the body of the pulley, and F a segment for the rim. A template is then made of each one of these segments, as at A and F. This template must be made slightly larger in every direction than the respective divisions, to allow for the stuff that will be turned off in truing the pattern in the lathe and in jointing the segments to one another during the building. The templates are employed to mark out on the board which should first be planed to the required thickness. This will be a trifle thicker than the course so as to allow for truing the surface of each finished course in the lathe. The courses are best built up on the chuck of the lathe on which they are to be turned, and a saving in time will be effected if there are two chucks, so that a course on one half of the pattern may be built up while the glue of another course on the other half is drying. On the lathe chuck, and directly beneath, where the joints of the segments will come, pieces of paper as at _a_, _c_, _e_, _g_, Fig. 2787, and if the segments are long ones, intermediate pieces of paper, as _b_, _d_, _f_, _h_, will be necessary. The radial edges of the segments are trimmed on what is termed a shooting board, which is a device such as shown in Fig. 2788, in which A is a piece of board on which is fastened the piece B. S is a piece projecting above B, and is provided to rest the segment S´ against, the flat surface of the latter lying on the board B. It is thus held in a fixed position, ready to have its edges E planed, the whole being laid upon the bench against the bench stop G. If, however, it is more convenient to rest the shooting board across the bench, a piece C may be fastened beneath A, so as to come against the edge of the bench as in Fig. 2789, in which T is the bench. The plane is laid with its side on A as in Fig. 2790, so that the surface of A acts as a guide, keeping the edge of the plane vertical, and thus planing the edges of the segment square. The plane is operated by hand in the usual manner (save that it lies on its side), taking its cut most off the outside or inside of the edge of the segment S´, according as the position of the latter is varied. In some of the shooting boards manufactured by tool makers, the height of B from A is adjustable, so that all parts of the plane blade edge may be used, which saves grinding, since only that part of the edge that is used dulls. Also there is provided means whereby the required lateral position of the segment may be adjusted; such a device is shown at P, Fig. 2788, which is a plate having a slot through it, through which passes the thumb screw V, which screws into S. Hence the plate may be adjusted so that when one end of the segment rests against the end of S, and the other against the end of P, its edge E will be in the proper position to be planed to correct angle by the plane, whose line of action is in this case rendered positive by means of a slide on the plane, acting in a groove in the base A.

The first segment is glued to the pieces of paper on the chuck, as shown in Fig. 2787, S´ representing the segment. A second segment is then added, being set fair to the pencil circle O, and jointed and glued both to the chuck and to the ends of the first segments. Successive segments are added until the whole circle or course is completed, and when dry the radial face of this course is turned in the lathe so as to be true, flat, and of the required thickness, and the diameter is trued. The second course may then be added, but the joints at the ends of the segments should not come over those of the first course, but in the middle as shown by the dotted line. The ends of the segments should be made to bed properly against each other, and glue should be applied to the joint between the two courses and at the ends. By adding the successive courses the whole may be built up on the chuck ready to receive the arms. As each segment is added it should be clamped or weighted to press it firmly to its seat and press out the excess of glue.

If the pattern consists of two, or say three, courses, the glue will be sufficient to hold it to the chuck while turning, but if there are more courses a screw should be inserted through the chuck and into each segment of the first course. The cylinder must then be turned inside and out ready to receive the spokes. These are made of pieces equal in length to the internal diameter of the rim, or a trifle longer, so that the ends may be let into the rim. A line is then marked along the edge of the rim, dividing its thickness into two divisions, and in the centre of the length a recess should be cut out from the face to the line, the width of the recess equalling the width of the arm, so that one arm will let into the other, forming a cross, of which the flat surfaces lie in the same plane. This cross is let into the rim of the wheel and fixed temporarily with brads. The lathe may then be started and the centre of the arms (and therefore that of the cylinder or pulley) be found by a pencil point moved until it marks a point and not a circle when the lathe revolves. The arms may then be marked to shape and a recess turned at their centre to receive the hub. The arms being marked to their respective places and their outside faces being marked with a pencil so that they may be replaced in the same position in the wheel, they may be removed and shaped to the required dimensions and form, and then replaced and glued to the rim.

If the wheel is to have six arms they may be constructed as follows:--

Instead of taking two pieces of the diameter of the rim, as in the case of four arms, three pieces are necessary, and in this case the thickness of the edge of each piece is divided by two marked lines which will divide the thickness of the edge into three equal divisions, as shown by the dotted lines 1 and 2 in Fig. 2791, which will divide the thickness of the edge into three equal divisions of thickness. From the centre of the lengths of each of the three pieces we mark on the flat face a circle whose diameter will equal the width on the flat face of the pieces themselves.

With an angle square having its adjustable blade set to an angle of 60°, and set so that the back is fair with the edge of the piece, and one edge coincident with the perimeter of the circle, lines tangent to the circle and crossing each other are drawn on the pieces A C. On the piece B four of such tangent lines (two on each side) must be drawn. The piece A is recessed between one pair of tangent lines to the depth of the second lines on its edge, or, in other words, to a depth of two-thirds its thickness, and between the other pair to a depth of one-third, as shown, the two-thirds at D, the one-third at E. The piece D must be recessed between its tangents on each side to a depth of one-third its thickness, as denoted at F F, while on C the whole space between the tangent lines must be recessed to a depth equal to two-thirds its thickness, as shown at G. The pieces may then be put together so that the two diametrically opposite arms will be in one piece. If an odd number of arms is employed this form of construction cannot be followed; hence each spoke will be a separate piece, extending from the rim to the centre and jointed at the latter, as in Fig. 2792, which is for five arms.

For this construction draw a circle _c_, Fig. 2792, and divide it into as many equal points of division as there are to be arms in the wheel. From these points of division draw lines to the centre, and these lines will show the required bevel at the end of each spoke, as shown in the figure. The ends should be verified for bevel by striking from the common centre a second circle, as D; and measuring if the arms are equidistant, measured at the circle and from the edge of the arm to that of the next, finished along the full length. When fitted, corrected, glued and dry, the spokes may be let into the wheel and a recess turned into the centre to receive the hub.

The rim and all parts that can be got at may then be turned in the lathe, the pattern then being reversed in the lathe to turn the inside of the rim, or the other side of the spokes, when the job will be complete. When, however, the rim is to be a very thin one, it may be necessary to fasten the segments together at the ends by other means as well as glue, hence a saw-cut may be made in each end, and a tongue inserted.

It is obvious that each half of the pattern is constructed by similar segments, the line of parting being through the centre of the arms, as at A B, in Fig. 2791. To keep the two halves coincident when in the mould, pins are inserted in the rim and arms of one half, fitting closely into holes provided in the other half.

To construct a pattern for a pipe, the pattern would be made in two halves, and constructed of what are termed staves, that is, pieces of wood running lengthways of the pipe. The number of these staves is optional, save that it must be even, so that each half pattern will contain an equal number.

Let it be required to make a pattern for a pipe 18 inches in diameter, and to be 1 inch thick. Draw the line A B, Fig. 2793, and from a point on it, as C, draw a semicircle A B, equal in diameter to the diameter of the outside of the pipe. Also the circle D E F, equal to the diameter of the inside of the pipe, and these will represent an end view of the pipe. Divide these semicircles into as many equal divisions as it is decided to have staves in the half pattern--as 1, 2, 3, 4, 5, 6; and from one of these divisions make a template as denoted by the oblique lines at 2, leaving it slightly larger than the division, to allow stuff to work on in fitting the staves, &c.

Now, when the staves are cut out it is necessary to have some kind of a frame or support to hold them while jointing them; hence, draw also from the points of division, as D, E, F, the lines _a_, _b_, _c_, _d_, _e_, _f_, and these will form the sides of a half-disk polygon, whose diameter is from D to F. A sufficient number of these polygonal half-disks are cut out to stand about two feet apart along the whole length of the pipe, as in Fig. 2794, and on these, temporarily fastened to the board B, the staves are jointed and fastened together by glue while each stave is held to its place on each half-disk by a screw. The top stave may be put on first, as it will act as a stay to the half-disks. If the pipe is so long that it is composed of more than two pieces, the end pieces should be put on first, and the intervening space filled up last, which enables the ends to abut firmly. The second half may be added to the first one, putting a piece of paper between the edges of the two to prevent their sticking together.

If the pipe has a bend, it is built up separately, instead of being formed of staves, the process being as follows:--

In Fig. 2795 let B represent the centre of the bend curve, the line C representing one end, G the other end, H the inner and J the outer arc of the bend. Let it be determined to build up the bend in five pieces, as shown at 1, 2, 3, 4, 5, which represents an end view of the half pattern. Templates are then made for each of the pieces 1, 2, &c., being formed as denoted by the oblique lines, whose dimensions slightly exceed the half circle E of the pattern, to allow wood for dressing up. To find the curve for these pieces, set the compasses to a radius from B to the outer corner of piece 1, and draw the arc K. Set the compasses to the radius from B to the inner corner of piece 1, and draw the arc L, and the space between these two arcs, which space is marked 1 T, is a template for the curve of piece 1. By a similar process applied to pieces 2, 3, 4 and 5 similar templates for their respective curves are obtained; and selecting timber of a proper thickness, we mark out the respective curves from these templates, which may be of thin board or of stiff paper. In putting these pieces together the lower ones are set to lines forming a plan of the bend, being set a little outside the lines to allow wood for truing the pieces to shape after they are put together. The lower pieces are temporarily fixed to the board on which the plan is marked, and the upper ones fastened to the lower by glue, the joint surfaces of each line being planed true previous to being glued. It is a great assistance, however, to cut out two half circles, representing the ends of the pipe, and to place them on the board to build upon. When a bend of this kind occurs in a covering for a pipe that is exposed to view, it is necessary, for the sake of appearance, to have the pieces composing the bend to correspond with those on the straight part of the pipe, as shown in Fig. 2796. The part A would be got out in staves, as described for the pattern of a pipe. The bend B would be also got out as described for that figure for a bend, save that the number of staves for the bend would equal the number on the pipe. But in this case each stave should be fitted to its fellow by pins, or its edge fitting into dowels on the edge of its fellow; thus one edge of a stave would have the dowels and the other the pins; the whole, when finished, being bound together by metal bands, as shown in the figure.

The patterns for a globe valve, such as shown in section in Fig. 2797, would be made as follows (which is taken from "The Pattern Makers' Assistant"):--

"The flanges vary in shape; but as a rule small valves are provided with hexagons and large ones with round flanges suitable for bolting to similar flanges to make joints. For small valves, say up to 2 inches, the pattern is usually made with the hexagons cut out of the solid, but for sizes above that, they should be made in separate pieces, as shown in Fig. 2798, and screwed to the pattern, so that in case of necessity they may be removed, and flanges substituted in their stead. In Fig. 2799, we have a perspective view of the finished pattern; and Fig. 2800 represents the pattern as prepared, ready to receive a flange or hexagon as may be required. A globe valve pattern should be made in halves, as shown in Fig. 2801, the parting line of the two halves being denoted by A B. To make this pattern, we first prepare two pieces of wood so large that, when pegged together, the ball or body of the pattern can be turned out of them, and long enough not only to reach from P to P, in Fig. 2799, but also to allow an excess by means of which the two pieces may be glued or otherwise fixed together. These two pieces we plane to an equal thickness, and then peg them to retain them in a fixed position, taking care, however, that the pegs do not occur where the screws to hold the flanges will require to be. We also place two pegs within a short distance of what will be the ends of the pattern when the excess in length referred to is turned off. We next prepare, in the same way, two more pieces, to form the two halves of the branch, shown at B, in Fig. 2801, for which, however, one peg only will be necessary. These pieces must be somewhat wider than the size of the required hexagon across the corners, that is, supposing the hexagon is to be solid with the branch; otherwise we must make them a little wider than the diameter of the hub of the flange, or of the round part of the hexagonal pieces. Their lengths must be such as to afford a good portion to be let into the ball or body of the pattern (as shown by the dotted lines in Fig. 2800), which is necessary to give sufficient strength. The two pieces must be firmly fixed together, and then turned in the lathe.

"During the early stages of the turning, or, in other words, during the roughing out, we must occasionally stop the lathe and examine the flat places on the body; for unless these places disappear evenly, the work is not true, and one half will be thicker than the other, so that the joint of the pattern will not be in the middle. It was to insure this that the pieces were directed to be planed of equal thickness, since, if such is the case, and the flat sides disappear equally and simultaneously during the turning, the joint or parting of the pattern is sure to be central. If the lathe centres are not exactly true in the joint of the two pieces, they may be made so by tapping the work on the side having the narrowest flat place, the process being continued and the work being trued with the turning tool at each trial until the flat places become equal. By this means, we insure, without much trouble, two exact halves in the pattern, which is very important in a globe valve pattern on account of the branch and other parts, not to mention the moulding. Having turned the body of the pattern to the requisite outline, and made, while in the lathe, a fine line around the centre of the ball where the centre of the branch is to come, as shown in Fig. 2800 by the line A, we make a prick point (with a scriber) at each crossing of the line A and the joint or parting of the pattern. We then mount the body upon a lathe chuck, in the manner shown in Fig. 2802. A point centre should be placed in the lathe and should come exactly even with the line A. In Fig. 2802, V V are two [V]-blocks made to receive the core prints. These [V]s are screwed to the lathe chuck, and the pattern is held to them by two thin straps of iron, placed over the core prints and fastened to the [V]s by screws. If the chuck and centre point run true, the [V]-blocks are of equal height, and the core prints are equal in diameter, the prick point opposite to the one placed to the centre point will run quite true; and we may face off the ball or body to the required diameter of branch, and bore the recess to receive the same. We make the holes in the flanges of the same size as the core prints; but we should not check in the print, because, if a flange with a different length of hub were substituted, it would be a disadvantage. To obtain the half flanges, we take a chuck and face it off true in the lathe; then, with a fine scriber point, we mark the centre while the chuck is revolving. We then stop the lathe, and, placing a straight-edge to intersect the chuck centre, we draw a straight line across the chuck face. We then take two pieces suitable for the half flanges, and plane up one flat side and one edge of each piece. If the flanges are not large ones, they may be planed all at once in a long strip. We place the pieces in pairs, and mark on each pair a circle a little larger than the required finished size of flange. We then fix each pair to the chuck, with the planed faces against the chuck, and the planed edges placed in contact, their joint coming exactly even with the straight line marked on the chuck face, and we may then turn them as though they were made in one piece and to the requisite size.

"In Fig. 2803 we have a representation of one half of a suitable core box, the other half being exactly the same, with the exception that the position of the internal partition is reversed. To get out this core box, we plane up two pieces of exactly the same size and length as the pattern, and of such width and thickness as will give sufficient strength around the sphere, allowing space for the third opening. After pegging these two pieces together, we gauge, on the joint face of each, lines representing the centres of the openings and the centre of the sphere. We then chuck them (separately) in the lathe, and turn out the half sphere. We next place the two halves together, and chuck the block so formed in the three positions necessary to bore out the openings; or if preferred, we may pare them out. The partition (A, in Fig. 2803) follows the roundness of the centre hole, and is on that account more difficult to extract from the core than if it were straight and vertical. When, however, the partitions are of this curved form, the pieces of which they are formed are composed of metal, brass being generally preferred. Patterns have in this case to be made wherefrom to cast these pieces, and they may be made as follows: First, two half pieces are turned; each is then cut away so as to leave the shape as shown at A in the same figure, and is then fitted into the spherical recess in the core box, letting each down until both are nearly but not quite level. The two wing pieces are then fastened on, and this pattern is complete. When the pieces are cast, they must be filed to fit the core box, and finished off level with its joint face, a small hole being drilled in the centre, and a pin being driven through the piece and into the box to steady the corners. We then saw the pieces in halves with a very fine saw.

"If the partition, instead of following the roundness of the valve seat, is made straight, the construction of the core box is much more simple. In this case, a zigzag mortice is made clear through each half of the box, its size and shape being that of the required partition. Fig. 2804 represents a half-core box of this kind. A piece of wood A is fixed, as shown, to the partition, to enable the core maker to draw it out before removing the core from the box. The mortice for the partition should be turned out before the half-spherical recess, the mortice being temporarily plugged with wood to render easy the operation of turning.

"In very large valves (say 10 or 12 inches) a half-core box is generally made to serve by fitting the two half partitions, shown at A, in Fig. 2803, to a half-core box, and keeping them in position by means of pegs, a half-core being made first with one and then one with the other in the core box. It is often necessary to form a raised seat in the body of an angle valve, such as shown in Fig. 2805, which represents a section of such a body. It is shown with flanged openings, though in small valves hexagons to receive a wrench would be substituted.

"Fig. 2806 is a plan of half the core box necessary for forming the raised seat. From this construction, it will be seen that the large core, though solid with the branch core, is not solid with that forming the hole in the seat and the part below it; therefore the core prints on the body pattern must be left extra long to give sufficient support in the mould for the overhanging cores. The loose round plug P, is made of the size of the outside of the seat and fitted to the box. The part outside the box is a roughly shaped handle to draw it out by. The diminished part D is a print, and into the impression left by it is inserted the core made in box shown in Fig. 2807. The print D is of the same diameter as the hole in the seat; and the print on the pattern is of the size of the increased diameter below the seat. Large angle valves are made with half a core box by making a branch opening in the box right and left, a semicircular plug being provided. Two half-cores are made with the plug, first in one and then in the other branch opening. The plug P should be in this case only half round."

For finding the lengths of the sides of regular polygons, scales, such as shown in Figs. 2808 and 2809, may be used, the construction being as follows:--

Draw a horizontal line O P, Fig. 2809, and at a right angle to it the line O B. Divide these two into inches and eighths of an inch, and draw lines meeting the corresponding divisions on O P, O B. From the point O draw the following lines: A line at 55-1/2 degrees from line O P, which is to serve for polygons having 9 sides; a line at 52-1/2 degrees to serve for polygons having 8 sides; a line at 49 degrees for polygons having 7 sides; a line at 45 degrees for 6 sides; a line at 40 degrees for polygons having 5 sides. It may be added, however, that additional lines may be drawn at the requisite angle for any other number of sides.

The application of the scale is as follows:--

The point O represents the centre of the polygon; hence from O to the requisite line of division on O B represents the radius of the work. From the line O B to the diagonal line (measured along the necessary horizontal line of division) is shown the length of a side of the polygon. From the point O, measured along the line having the requisite degrees of angle, to the horizontal line denoting the radius of the work, gives the diameter across corners of the polygon. The diameter across the flats of a square being given, its diameter across corners will be represented by the length of a line drawn from the necessary line of division on O B to the corresponding line of division on O P. A cylindrical body is to have six sides, its diameter being 2 inches, what will be the length of each side? Now, the radius of the 2-inch circle of the body is 1 inch; hence, find the figure 1 on line O B and measure along the corresponding horizontal line the distance from the 1 to the line of 45 degrees, as denoted by the thickened line.

A body has six sides, each side measuring an inch in length, what is its diameter across corners? Find a horizontal line that measures an inch from its intersection of the line O B to the line of 45 degrees, and along this latter to the point O is one-half the diameter across corners.

_Example 3._--It is desired to find the diameter across corners of a square whose side is to measure 3 inches. Measure the distance from the 3 on line O P to the 3 on line O B, which will give the required diameter across corners.

This scale lacks, however, one element, in that the diameter across the flats of a regular polygon being given, it will not give the diameter across the corners. This, however, we may obtain by a somewhat similar construction. Thus, in Fig. 2808, draw the line O B, and divide it into inches and parts of an inch. From these points of division draw horizontal lines; from the point O draw the following lines and at the following angles from the horizontal line O P:--

A line at 75° for polygons having 12 sides.
" 72° " 10 "
" 67-1/2° " 8 "
" 60° " 6 "

From the point O to the numerals denoting the radius of the polygon is the radius across the flats, while from point O to the horizontal line drawn from those numerals is the radius across corners of the polygon.

A hexagon measures 2 inches across the flats, what is its diameter measured across the corners? Now, from point O to the horizontal line marked 1 inch, measured along the line of 60 degrees, is 1-5/32 inches; hence the hexagon measures twice that, or 2-5/16 inches across corners. The proof of the construction is shown in the figure, the hexagon and other polygons being marked for clearness of illustration.

Let it be required to make a pattern for a section of pipe such as shown in section and in plan in Fig. 2810, which is from "The Pattern Maker's Assistant." This pattern would be made to mould, as shown in the section, lying horizontally, and must therefore be made in two halves, the line of joint for the two halves being along A B in Fig. 2811.

"The body A and the branch B would be made separate from the flanges, and would be reduced in diameter at the ends to receive them. To form A, take two pieces of timber, say three inches longer than the length of A, including the core prints, and measuring a little more than half the diameter of the pipe one way, and a little larger than the full diameter of the pipe the other way, and glue them together at the ends for a distance of 1-1/2 inches, which will serve to hold them while turning them in the lathe.

"The pieces may then be turned in the lathe to the required diameter. During this turning, however, it is essential to insure that the joint of the two pieces be exactly in the centre, otherwise one half of the pattern will be (when the halves are separated) thicker than the other.

"The ends are then turned down to receive the flanges, the reduced diameter being necessary so as to leave a shoulder for the flanges to abut against to keep them true, or at a right angle to the axial line of the body. The branch is turned up in the same way, and the flanges are then turned and put on.

"The end of the branch may be cut to fit the circumference of the body as follows:--

"Set a bevel square to an angle of 45°. Take the halves of the branch apart, and rest the stock or back of the bevel against the end face, and let the blade lie on the joint face, and mark two lines A B in Fig. 2812, which lines must just meet in the centre of the branch at the end. Cut away the angular pieces C and D down to the lines A B. This performed on each half will leave them when given a quarter turn as shown in Fig. 2812, and the curve shown by the junction of the horizontal with the vertical shading lines is the curve for the end; hence the surface covered with the horizontal lines requires to be cut away.

"When this is done on both halves the branch will fit to the body, as shown in Fig. 2813, in which A is the body and B C the two half branches. For a temporary pattern the branch may be fastened to the body with a few screws; but for a permanent pattern it should be glued also, which is done as follows:--

"Lay one half of the body A, Fig. 2813, on a board, with the flange overhanging to be out of the way, and clamp it there; lay the branch also on the board, and draw it firmly up to the body by clamps, while also clamping it flat down to the board, as shown in Fig. 2814. This will insure that the joint faces are true with one another, that is, lie in the same plane. Paper should, however, be placed between the joint faces and the board to prevent them from becoming glued to the board, and the edges, therefore, from breaking away. The second half can be put together as the first one, paper being put between the two to prevent them from being glued together; and to further strengthen the joint, let into each half a piece of hard wood P, Fig. 2815, and put in the screw shown at A.

"Suppose now that the diameter of the branch had been smaller than that of the body of the pattern, then the length of curve necessary on the branch end to let it abut fairly against the cylindrical pattern body may be found as follows:--

"Draw on a piece of board the line A B, Fig. 2816, and from any point C mark a semicircle equal in radius to that of the radius of the body of the pattern, draw the line E parallel to A B, and distant from it to an amount equal to the radius of the branch, then from the junction of E with the semicircle as at D, mark the line F at a right angle to A B. Let it now be noted that the semicircle A G represents half the pattern body, and E D F B the branch; hence from F to G is the length of the branch end that will require to be curved to fit the circumference of the body, while it is also the length to be added to the distance the branch requires to stand out from the body. To draw the curve on the end D F G of the branch the gauge or marking instrument, shown in Fig. 2817, is employed. The branch P is placed in [V]-blocks (Fig. 2818), resting upon a plane surface. The gauge consists of a stand E carrying a vertical bar A; upon A is the closely fitting cross-tube carrying the arm C, which in turn carries the marking pointer D, which is set distant from the centre of the bar A to the amount of the radius of the piece of work or the cylinder is to fit against.

"If the branch required to stand at an angle to the body, as in Fig. 2819, the marking may be performed by the same gauge and in the same manner, but the axial line of the branch must be set, when marking one side, at an acute angle to the axial line of A, and at an obtuse angle to A when turned over to mark the other side, which may be done in each case by raising one of the [V]-blocks until the branch lies in either case at the same angle to A as it will require to stand to the body on which it is to fit.

"When the body is much larger in diameter than the branch, a hole may be bored in the former to receive the end of the latter, by giving to the branch end a stem, as in Fig. 2820, and then cutting in the body a recess for the branch end and its additional stem. This recess may be cut out in the lathe, chucking the body as in Fig. 2821.

"Should it occur that one end of the [T] is of larger diameter than the other, one chucking [V] must be deeper than the other, and we may find their respective depths by the following process:--