Chapter XXIX: Erecting Engines and Machinery (2)

It may not happen that a crank pin is out of truth in a direction in which the error will show plainest when the crank is on its dead centres, or at half-stroke; but if a crank pin, tested in those four positions, fails to show any error when tested by the connecting rod, it will be true enough for all practical purposes, and true enough to avoid heating and pounding, both of which evils accompany an untrue crank pin. Suppose, now, that a crank pin stands out of true in the direction shown in Fig. 2545, in which A A represents the axial line of the cylinder bore prolonged, and B B the axial line of the crank shaft (the two being parallel or in proper line). Let E E represent the centre line of the connecting rod when the crank is on one dead centre, the axial line of the crank pin being at C C. Then the brasses being keyed up to fit the crank pin, the centre or axial line of the connecting rod would stand as denoted by E E. But the brasses at the other end of the rod being keyed up to fit the cross-head journal, and their lines being at a right angle to the line A A, we have that the rod is at that end endeavored to be held parallel to A A; hence, keying up the connecting-rod brasses on the crank pin would tend to bind the rod, one end standing parallel to A A, and the other parallel to E E.

This would place great strain on the outer radial face of the cross-head journal, as well as on the cylindrical body of the journal.

When, however, the crank pin arrives at the opposite dead centre, as denoted by the dotted lines in Fig. 2545 (G G representing its axial line, and F F the centre line of the connecting rod at a right angle to G G), the want of truth in the pin throws the cross-head end of the connecting rod against the inside face of the cross-head journal. Hence, twice in each revolution is the connecting rod bent, and twice does it jam from side to side of the cross-head journal.

It may now be pointed out that if we take either dead centre singly, and connecting the rod at the crank-pin end, try it at the cross-head end, and it will be a difficult matter to determine whether any want of truth at the latter end is caused by the crank pin being out of axial truth, or whether it is the crank shaft itself that is out of line. But there is this difference between the two cases. When the error is due to want of alignment in the crank shaft, the connecting rod will show the error _on the same side_ of the cross head, no matter on which dead centre the crank pin stands; but when it is due to the crank pin, the rod will fall inside the cross head on one dead centre, and outside when tried on the other dead centre, as is shown by the respective lines E and F, in Fig. 2545; E being at a right angle to C, and F at a right angle to G.

Again, it has been shown that when the shaft was out of line, a point on the crank-pin journal passed outside of the cylinder centre line at one dead centre and inside at the other; but when the pin is axially out of parallel, the path of a point on its journal will remain in the true plane, as is shown in Fig. 2546, the point being taken at the intersection of E and C C. A A represents the path of rotation of the same, which is parallel to the true face B of the crank.

From the angle of the axial line of the pin being in opposite directions, when on opposite dead centres to the axial line of the crank shaft, the bore of the brasses cannot wear to suit the error, which, therefore, only diminishes by the wear of the crank pin. Suppose the error to be 1/64 inch in a crank-pin journal 3 inches long, and that the connecting rod is 6 feet long, the error at the cross-head end of the rod will amount to 3/8 inch.

In Fig. 2546 the error is shown to exist in an opposite direction, throwing the rod to the other side of the cross-head journal. But, in this case, the crank, when on the dead centre nearest to the engine cylinder, throws the connecting-rod end against the inside face of the cross-head journal, as denoted by the line E, which is on the opposite side of A A to what it is in Fig. 2545. Again, when on the other dead centre, the line F F, in Fig. 2546, falls _outside_, while F F, in Fig. 2545, falls _inside_ of A A, and it is by this difference that we are enabled to know in which direction the crank pin is out of true. To find the amount to which it is out of true in the length of its journal, place the crank on one dead centre, and with the connecting-rod brasses keyed up firmly home on the crank pin, and the other end of the connecting rod entirely disconnected from the cross head, mark on the latter a line coincident with the side face of the rod end, as at D, Fig. 2547. Then, with the crank pin placed on the other dead centre, mark another line on the cross head, coincident with the other side face of the rod, at C, Fig. 2547. Now, suppose that the line D shows the rod to fall 3/8 too much on that side, and line C shows it to fall (when on the other dead centre) 3/8 too much on the other side of the journal, and that the length of the rod is 6 feet, while that of the crank-pin journal is 3 inches, then the latter, divided into the former, gives 24, and this sum divided into the 3/8, the rod end falling out of true at C and D, Fig. 2547, gives us 1/64-inch as the amount the crank pin stands out of true in its length; hence, to correct the error, we may file on the crank pin a flat place at each end, as shown in Fig. 2548 by the lines C D, and then file on the top and the bottom of the crank pin a flat place B, 1/128-inch deep, and of equal depth all along the journal; by then filing the crank pin round and bringing the flat places just up to a circle, we shall have reduced the diameter of the crank pin by 1/64 inch, and have made it axially true with the cross-head journal. It is important, however, to bear in mind that, in this case, the crank pin is supposed to be out of true in the direction shown in Fig. 2545, and to stand axially true with the cross-head journal, when the crank is placed at half stroke, top and bottom, the crank shaft being in proper line.

If the axial line of the cross-head journal stands truly horizontal, the flat places on the crank pin may be filed horizontally level, with the crank placed on the corresponding and respective dead centres. But as the length of the cross-head journal is so short, it is difficult to gauge, if it does stand axially exactly horizontal, hence it is better to try the rod, or follow the above directions; especially as the cross-head journal and crank shaft may be in line without being axially horizontal.

Suppose now that the axial line of the crank pin stands true with that of the cross-head journal when the crank is on either dead centre, but out of true when at the top and bottom half stroke. The connecting rod, connected as before, and tried with the cross head, will fall first to one and then to the other side of the cross-head journal, and the direction in which the crank is out of true may be known from the position of the crank pin when the error shows itself.

If the error exists to an extent that is practically measurable, a pound in the journals, as well as their heating, is the inevitable result. In Fig. 2549, for example, the rod end is shown in section, and it will be noted that the error being in the direction there shown, and the crank pin in the respective positions there shown, the brass bore only contacts with the journal at each end, and that the diameter of the bore of the brasses is greater than the diameter of the crank pin journal to _twice_ the amount the crank pin is out of line. Now let us place the crank at the top of its revolution, as in Fig. 2550, and as its axial line then stands parallel to that of the cross-head journal, the brass bore is too large to fit the crank pin journal and there is lost motion.

From the time the crank pin passes the dead centre this lost motion increases in amount until it becomes sufficiently great to slam the rod over against the side of the cross-head journal, while at the same instant the crank pin pounds in the connecting-rod brasses. At what precise part of each quarter crank revolution this action will occur, depends upon the amount the crank pin is out of line; but the more it is out the nearer to the dead centre it will be, and, conversely, the nearer true it is the nearer the crank will approach its highest and lowest positions before the pound takes place. If it is attempted to key up the brasses so as to spring the rod and let them close along the journal, the brasses will heat in proportion to the amount of error; hence when the crank pin pounds with the brass properly adjusted, and heats while keyed up enough to stop the pound, the crank pin is out of true.

To test the alignment of an engine with stretched lines take out the piston and rod, and take off the connecting rod, then fasten a piece of iron at the open end of the cylinder so that it will hold a stretched line true with the axis of the cylinder bore. Provide at the crank end of the engine bed a fixed piece of wood to hold the other end of the line, and then with a piece of wire as a gauge set this line (tightly stretched) true with the cylinder bore. Then place the crank pin at the top of its path of rotation and drop a plumb line from the centre of its journal length, and this line should, if the crank shaft is horizontally level, just meet the stretched line. If it does not do so place a spirit level on a parallel part of the crank shaft, and if the shaft is not level it should be made so, and so adjusted that the line from the centre of the length of the crank pin journal just meets the stretched line from the cylinder bore.

To test if the axial line of the crank shaft is at a right angle to the cylinder bore axis move the crank pin nearly to its dead centre, and measure the distance from the middle of its length to the stretched line. Then move the crank pin over to nearly the opposite dead centre, and (by means of the plumb line) measure the distance of the plumb line from the stretched line. To be correct the plumb line from the crank pin will during this movement just touch the stretched line.

To test if the stretched line is fair with the centre of the crank shaft place a square on the end of that shaft and even with its centre, and the blade should then just meet the stretched line.

The edges of the guide bars may also be tested with the stretched line, and the top and bottom of the guide-bar flanges may be tested to prove if the bars are of the correct height.

To further test the bars place a spirit-level across them and lengthwise on them.

If the piston rod and connecting rod are in place the alignment may be tested as follows; Let the piston rod be as far out of the cylinder as possible, and stretch a line to one side of it, just far enough off to clear the guide bars, &c. Set this line as follows: Let it be in line with the rod as sighted by the eye when standing some few feet away from it but horizontally level with the centre of the rod, set it parallel to the rod with a rule or its equivalent. Then the centre of the crank-pin journal should measure from the stretched line, the distance of the line from the piston rod added to half the diameter of that rod. This test, however, is not very accurate on account of the difficulty in setting the line, and because the piston rod may not have worn equally on each side.

SETTING SLIDE-VALVES--An engine slide-valve may be so set as to accomplish either one of three objects. First, to give equal lead for each stroke; second, to cause the live steam to be cut off and expansion to begin at an equal point in each stroke; and third, for the exhaust to begin at an equal point in each stroke.

If we, set the eccentric so that the exhaust will begin at corresponding points for the two strokes, the valve lead will not be equal, and the exhaust opening will be greater when the piston is at one end of the cylinder than it will be when the piston is at the other end.

If the eccentric be set to cut off the steam at corresponding points for the two strokes, then the lead, the admission, and the exhaust of the steam at one port will differ (with relation to the piston movement) from that at the other. It is generally preferred to set the eccentric so as to give equal lead for the two ports when the piston is at the respective ends of its stroke, which gives an equal amount of exhaust opening when the piston is at the respective ends of its stroke.

The only operations properly belonging to the setting of a slide-valve are those of finding the true dead centres of the crank pin, and setting the eccentric to give the valve the desired amount of lead. It is generally found, however, that the length of the eccentric rod requires a little correction, and as this must be done before the eccentric can be set, the setting operations should be conducted with a view to making the correction as early as possible.

In many of the instructions given by various writers it is directed to first square the valve, which is to attach the parts and move the engine crank, or fly-wheel, through one revolution, to ascertain if the valve moves an equal distance on each side of the centre of the cylinder ports, correcting the length of the eccentric rod until this is the case. This is an error, because on account of the angle of the eccentric rod the valve does not, when set to have equal lead at each end of the stroke, move an equal distance on each side of the cylinder ports, but travels farther over the port nearest than it does over that farthest from the crank.

When the travel of the valve is equal to twice the width of the steam port, added to twice the amount of steam lap, the valve does not fully open the farthest port from the crank. When the valve-travel is more than this amount both ports may open fully, but the error due to the unequal valve-travel from the angularity of the eccentric rod is increased. That the amount of error induced by squaring the valve is appreciable, may be seen from the fact that with 1-1/4 inch steam ports, 3/4 inch steam lap, and 4-1/2 inches of valve-travel, it amounts to about 1/8 inch with an eccentric rod 4 feet long. As the eccentric rod has (if a solid one, as in the case of a locomotive) to be operated upon by the blacksmith to alter its length, and requires some accurate setting for alignment after having its length corrected, it is obviously preferable to obtain its exact length at once. This may be done with less work than by the squaring process, which is entirely superfluous.

Assuming, then, that all the parts are properly connected and oiled, the valve is set as follows: Upon the face or edge of the fly-wheel an arc, true with the centre of the wheel, should be drawn, as at A B, in Fig. 2551, marking it on opposite sides of its diameter and opposite to the crank pin P. The engine should then be moved in the opposite direction to that in which it is to run, until the guide block I is very near its full travel. A straight-edge must then be placed to bear against, or be coincident with, the end face of block I, and held firmly while a line is drawn across the edge of the guide bars, as shown at C. There should then be fastened to the floor (which must be firm, and not give under the engineer's weight), a piece of iron W, having a deep centre-punch mark, or its equivalent. A steel tram-rod T, pointed at each end, is then set in the centre-punch mark at W, and with the upper end D a line made across the wheel edge or face. The fly-wheel must then be moved so that the crank passes the dead centre, the guide block moves back and away from the line C, and then approaches it again. When the end of the guide block is again coincident with the line C, the tram should be set as before and a second line, F, marked on the fly-wheel rim, and from these two lines, D and F, the crank may be placed upon its true dead centre as follows:--

In Fig. 2552 a section of the fly-wheel rim is shown (enlarged for clearness of illustration); from the lines D, F the centre E is found, and marked with a centre punch dot to define it. It will be obvious, then, that if the fly-wheel be moved until this line and dot come fair with the upper edge of the tram T, the guide block will be at the exact end of its travel, and the crank, therefore, on its dead centre. By a similar operation performed with the guide block at the other end of the guide bars, and with lines on the other side of the wheel rim (as shown at B, J, K), the other centre L may be found. In obtaining these centres, however, a question arises as to the direction in which the wheel should be moved for bringing the guide block up to the lines at C, and for marking the lines D F and J K, or for bringing E or L true with the tram point. If the fly-wheel be moved in the opposite direction to that in which the engine is to run, the cross-head journal and crank pin will bear against the boxes of their brasses in the direction in which they will have contact when the engine is running. Suppose, for example, that the top of the fly-wheel when the engine is in motion moves from the cylinder, then the cross-head and crank-pin journals, driven by the piston, will bear against the half-brass nearest to the cylinder, which, _when the force-producing motion is applied to the fly-wheel instead of to the piston_ will be the case when the fly-wheel is moved in the opposite direction. By moving the fly-wheel in an opposite direction to that in which the engine is to run, the lost motion in the journals and bearings is therefore taken up in the proper direction so far as the connecting-rod brasses are concerned, and any lost motion between them and their journals will not impair the set of the valve, as would be the case were the fly-wheel moved in the direction in which it is to run.

But by moving the fly-wheel backwards the play in the eccentric and in all the joints between it and the valve spindle is up in the wrong direction, because the power to move the rods is being applied in the opposite direction to that in which it will be applied when the engine is running, and, therefore, the play motion of the jointed or working parts will cause a lost motion impairing the set of the valve.

Now there are generally more working parts between the eccentric and the valve than between the crank pin and the piston, and hence more liability for lost motion to exist, and it follows that in such case it is better to move the engine in the direction in which it is to run.

It may be remarked, however, that the play may be taken up in the proper direction in both cases, and the engine be brought upon its dead centre, by moving it in the opposite direction to that in which it is to run, and that in setting the eccentrics they be moved on the shaft in the direction in which the engine is to run, as forward for the forward eccentric, and backward for the backward one (assuming the engine to have a link motion, and, therefore, two eccentrics).

It is obvious that any other resting place may be used instead of the floor for the tram; thus in a locomotive the wheel guard may be used, the tram T being used to mark lines on the upper part of the wheel rim, instead of opposite the crank. To set the valve, place the fly-wheel on its dead centre, moving the fly-wheel as directed until one of the points (E or L, say E) comes fair with the point of the tram; then move the eccentric on the shaft until the steam port is open to the required amount of lead, and fasten the eccentric to the main shaft. Next move the fly-wheel around until on the opposite dead centre, and if the lead is the same in amount for both ports the valve is set. Suppose, however, that in this last case the lead is too great; then it shows that the eccentric rod is too long, and it must be shortened to an amount equal to half the difference in the lead. Or suppose that the lead when the wheel was tried on the last dead centre L, was less than for the other port; then the eccentric rod must be lengthened to half the amount of the difference. Assuming that the rod was too long by 1/32 of an inch, then it may very often be shortened by simply heating about six inches of its length to a low red heat, and quenching it in water. If the rod has a foot which bolts on a corresponding foot on the eccentric, then to lengthen it a liner of the requisite thickness may be placed between the two feet.

Suppose there is an equal amount of lead at each end but the amount is not sufficient or is too great: then the eccentric must be moved on the shaft until the proper amount of lead appears at the port. The lead must then be again tried at the other dead centre. In moving the eccentric, however, it must, under all conditions, be moved in the direction in which it will rotate, for reasons already given. The best method of measuring the lead where the lines on a rule cannot be seen is with a lead wedge P, as shown in Fig. 2553; this, if slightly forced in, will mark itself, showing how far it entered.

In some practice the position of the valve is transferred to the valve stem outside of the stuffing box or gland, as shown in Fig. 2554, sectional view. The valve stem being disconnected from the rod or arm that drives it, the valve is moved by hand to have the proper lead, as at A; a centre-punch mark is then made outside the stuffing box and a tram B rested thereon; with the other end of the tram a mark C is made on the valve stem. A similar mark is made on the stem when the crank is on the other dead centre, and the tram and marks, applied as shown, are employed instead of measuring the lead at the ports themselves. This involves extra work, but gives no more correct results. It involves marking lines on the valve stem, which is objectionable. If several trials have to be made there is a confusion of lines on the valve stem, and the wrong one is apt to be taken. On the other hand it affords a facility for setting the valve without having the steam chest open, which may in some cases be desirable. If this plan be adopted the lines on the valve rod should not be defined by centre-punch marks, for they will cut the packing in the stuffing box.

When the eccentrics are secured to the shaft by a set-screw only, and not by a feather, it is an excellent plan, after they are finally set, to mark their positions on the shaft, so that if they should move they may be set to these marks without moving the engine around.

For this purpose take a chisel with the cutting end ground to the form of a fiddle drill, one cutting edge being at a right angle to the other. The chisel must be held so that while one edge rests upon the axle, the other edge will bear against the radial face of the eccentric. A sharp blow with a hammer upon the chisel head will make a clean indented cut upon the axle and the eccentric, the two cuts exactly meeting in a point where the eccentric bore meets the axle circumference, so that when they coincide the eccentric is in its proper position.

If the eccentrics of a locomotive should slip when the engine is upon the road, and there are no marks whereby to readjust them, it may be done approximately as follows:--Put the reverse lever in the end notch of the forward gear, then place the crank as nearly on a dead centre as the eye will direct, and open both the cylinder cocks, then disconnect the slide-valve spindle from the rocker arm, and move the valve spindle until the opening of the port corresponding to the dead centre on which the crank stands will be shown by steam blowing through the cylinder cock, the throttle valve being opened a trifle. The position of the valve being thus determined, the eccentric must be moved upon the shaft until the valve spindle will connect with the rocker arm without being moved at all. The throttle valve should be very slightly opened, otherwise so much steam will be admitted into the cylinder that it will pass through any leak in the piston and blow through both cylinder cocks before there is time to ascertain which cock first gives exit to the steam.

Instead of finding when the crank pin is on the dead centre by means of the process shown in Fig. 2551, it may be found as in Fig. 2555, which is for a vertical engine. On the face of the crank and from the centre of the crank shaft as a centre, draw a circle B equal in diameter to the diameter of the crank pin. Then take a spirit-level C and apply it to the cylinder bore and note where its bubble stands. Then apply the spirit-level to the perimeter of the crank pin A and circle B, and move the crank until the spirit-level bubble stands in the same position as it occupies when applied to the cylinder bore. If the cylinder bore stands truly vertical the bubble will in both cases stand in the middle of the spirit tube; but in any event, the bubble must stand in the same position when applied to the crank as when applied to the cylinder bore, in which case the crank will be on its dead centre whether the cylinder bore be horizontal, vertical, or at an angle, the dotted line E passing through the centre of the crank and the axis of the cylinder bore.

When an engine has two eccentrics, so as to enable the engine to run in either direction, as in the case of a locomotive, it is necessary to consider which eccentric is to be set for the forward, and which for the backward motion. In American locomotive practice it is usual to let the eccentric nearest to the wheel, and, therefore, the most difficult to get at, be for the backward motion, which is the least used, and therefore the least liable to get loose upon the axle.

The eccentric that connects to the top of the link is usually that for the forward motion, and hence that which connects with the eccentric farthest from the wheel.

In testing the lengths of the eccentric rods, work may be saved after the engine is first placed on its dead centre by putting the reverse-lever in the forward notch of the link, and adjusting the forward eccentric until the valve has the proper lead. Then set the reverse-lever in the back notch and move the backing eccentric (in both cases moving them in the direction in which they will run), until the proper amount of lead appears. The engine may then be placed on the other dead centre, and the lead both for forward and backward gear measured, so that if there are any errors both the rods may be corrected for length; but for the final trial the crank pin must be set on its dead centre for each direction of motion separately, so as to take up any lost motion in the connecting-rod brasses.

In the case of large marine engines it is not practicable to move or rotate the engines to set the valves, and the eccentrics are therefore adjusted to their positions on the crank shaft by lines before the crank shaft is put into its place or bearings. First, the throw of the crank is set to stand horizontally true by the following method: From the centre of the crank shaft strike a circle of the diameter of the crank pin, as shown in Fig. 2556, at A, and draw upon the face of the crank a line that shall just meet the two circles as denoted by the line B, using a straight-edge, one end of which rests upon the crank pin, while the other end is coincident with the perimeter of the circle A.

By means of the wedges shown at C D adjust the crank until the line B stands horizontally level, tested by a spirit-level. A straight-edge having straight and parallel edges is set horizontally level, beneath the eccentric, so that its edges will stand parallel with the throw line of the crank. On this straight-edge, and parallel to the edges, is marked the line A A, Fig. 2557. The first process is to mark on A A the centre of the crank shaft K, which is done as follows: Over K is placed the fine line B B, suspending the weights or plumb bobs at B B; coincident with this line and across A A, are marked two lines C D; midway between C D is marked E, which therefore stands directly beneath the shaft centre. From E the line F is drawn distant from E to the amount of lap added to the lead the valve is to have. From F as a centre two lines are drawn across A, their distance apart equalling the full diameter of the eccentric; the plumb line is then placed over the eccentric, and the latter is rotated on the shaft until the plumb lines come exactly fair with the lines G H.

It is obvious that instead of using plumb lines a square may be employed to mark the lines C D, and to set the eccentric to the lines G H, the square being applied as at S and S´, in Fig. 2558.

In this example it has been assumed that the direction of crank rotation was to be as denoted by the arrow; but, suppose the crank rotation required to be in the opposite direction, then the marks on the straight-edge would require to be located precisely the same, but the position of the eccentric throw-line would require to be as in Fig. 2559, the perimeter of the eccentric being set to the lines G H as before. The eccentric rod being supposed to connect direct to the valve spindle, without the intervention of a rock shaft, for if there is no rock shaft the eccentric leads in the direction of rotation, while if the engine has a rock shaft the eccentric follows the crank-pin in the direction of rotation, and F must be marked on the crank-pin side of E, as in Fig. 2560.

If two eccentrics are used, as in a link motion, the lines for setting one eccentric are equally applicable to both; the lap and lead line F being located on the crank-pin side of E when there is a rock shaft, as is supposed to be the case in Fig. 2561; and on the other side of E when there is no rock shaft; and in this case the eccentric that is to operate the valve to make the engine run forward must have its throw-line following the crank pin, as at J, in Fig. 2561; the eccentric K operating the valve for running backward. Conversely, in the absence of a rock shaft, the throw-line of the forward eccentric leads, while that of the backward eccentric follows the crank pin.

When the line of connection of the eccentric rod is not parallel to the axial line of the cylinder bore, the crank must be placed horizontally level (or if it be a vertical engine, on the dead centre), but instead of the straight-edge being placed parallel to the throw-line of the crank, it must be placed at a right angle to the line of connection of the eccentric rod.

Thus in Fig. 2562 the engine is supposed to be a vertical one, and the crank is, therefore, placed on its dead centre, its throw-line being vertical instead of horizontal as in our previous examples (which were supposed to be for a horizontal engine). It is also supposed to have a rock shaft A; hence the straight-edge is set at a right angle to the line of connection of the eccentric rod which is denoted by B.

It is obvious that to set the crank throw-line vertical the circle B in Fig. 2509 may be used, the spirit-level being resorted to to discover when the crank stands vertical.

An example in the erection or setting of framed work is shown in Fig. 2563, which represents a side elevation of a frame put together in four parts, two side and two end frames. A and B are journal bearings requiring to stand parallel and true one to the other, B being capable for adjustment in distance from A by means of the adjusting screws G, H. The bearings C, D, E, F, are to be parallel one to the other and to A, B. Their proper relative distances apart, and the axes of all the shafts, are to stand at a right angle to the side frames.

Fig. 2564 represents an end view of the frame, the ends T being bolted to the side frames S and S´ at I, J, K, and L.

Now it is obvious that the ways for the bearings A, B, C, &c., may be trued out, ready to have the brasses fitted before the framework is put together, and that from their positions they would have to be planed out at separate chuckings; supposing, of course, the frame to be too large to be within the capacity of the machine table. It would be difficult to cut all the surfaces of the bearing ways to stand in the same plane, unless there were some true plane to which all might be made common for parallelism.

Furthermore, unless the surfaces where T is fastened to S and S´ are properly bedded to fit each other, bolting them up would spring and bend the frames out of their normal planes. To meet these requirements, there are given to the side frames a slightly projecting surface where the feet of T meet them, and furthermore, the feet of T themselves project beyond the sides of T as shown. These projecting pieces may therefore be planed to a common plane without planing the sides of the respective frames; and this plane should be as nearly as can be parallel with the body of each frame surface. The surfaces of the bearing ways may then be planed parallel to those of the projections, and the jaw surfaces true to the side surfaces, and all the bearing ways will stand true if the frames be properly set--when put together with the bolts. But unless the bedding surfaces at I, J, K, L, be made to bed and fit properly, the whole truth of the bearing ways and their distances apart across the framework may be altered. Thus, supposing the feet of T at I and J to meet S as denoted by the dotted lines O R, and whether the fault lie with the feet of T or with the projections on S the result will be that the pressure of the bolts holding I J to S will bend S so that its plane will be a curve as denoted by the dotted line P P, and the distances apart of the journal ways B B and D D respectively will be wrong, being too wide on account of the bend outward of S.

But the feet may touch on the opposite corners, the surfaces of S´ or of T being out of true or out of full contact, as denoted by the dotted lines V W on K, L; in this case the frame S´ would be bent to the curve Q Q, and the journal ways would be too close together.

On the other hand, the want of fit between these surfaces may be in the direction of the length of the frame instead of the direction of its height, as has been supposed; or it may be in one direction on one foot and in another direction on another foot. But in whatever direction it does exist, it will inevitably bend and twist the frame.

It must not be taken for granted, that because these surfaces have been planed or milled, that therefore they are true; because frames of this class cannot, if large, be held without springing them to some extent from the pressure of the bolts or other devices necessary to hold them to be cut.

It is not uncommon to plane the surfaces as true as may be, and put the frames together, bolting them up tight, and then applying the straight-edge trammel and rule to test the truth, correcting any error that may be found by inserting pieces of paper, sheet tin or material of requisite thickness on one side of the surfaces, so as to offset the error in their fit and bring the framing true; but this is not the proper way, because it reduces the area of contact, and furthermore renders a new testing and adjustment necessary whenever the frames are taken apart. It is better therefore to apply a straight-edge to the surfaces and true them to it, testing them vertically as by placing the straight-edge across K L, and longitudinally across S´, at K and the corresponding projection at the other end of the frame, filing them until they appear true.

The holes through the frame may be drilled before filing these surfaces, so as to reduce the area to be filed. Since the end frames T do not in this example carry any journals or mechanism, the position of T is not so particular as it otherwise would be; hence, the holes in its feet may be marked off and drilled independently of the frame, the holes being drilled a little too small to allow for reaming with the holes in the frame. The framing will then be ready to put together (all machine work upon them being supposed to be done). The feet of all such frames should be planed true, so that the frame, when put together, may stand true and steady when placed upon a level floor or foundation, and in this case the distance and parallelism of the feet surfaces will be true with the ways or bearings, affording much assistance in holding the frame while putting it together. The height of the holes may be measured and marked from the feet surface, thus insuring truth as far as height is concerned. Lines may be drawn or marked on each side frame, at the proper distance from and parallel with the jaws of the ways A, B, thus completing on the side frames the marking of the location of the centres of the holes for bolting the end frames on.

If the frames were of a size to be sufficiently easily handled, the end frames might be put in their places, and the whole framework set true, so as to mark the holes in the end frames from those already drilled in the side frame. But if the use of a crane were necessary to lift them, it would be better to mark the holes on the end frames, and drill them before putting the framework together at all, leaving sufficient to ream out of the holes to bring them fair, notwithstanding any slight error in drilling them. In this case, a line denoted by the dotted line X in Fig. 2564, should be drawn across the frame, and the holes at I and J be made equidistant on each side of it, as well as the proper distance apart. X must be at a right angle to the trued foot surfaces at I J, so as to cause the side frames to stand vertical while their feet are horizontal.

Supposing now the holes to be drilled and the frames are to be bolted together, the whole frame may be held temporarily together by bolts passing through the side frames at each end, or a bolt may be passed through the holes F to steady it. Indeed, if these holes F have been accurately bored, a neatly fitting mandrel passed through them should hold the side frames true. The end frames T having been set to stand at a right angle to the side frames, and with their holes at I J, &c., as near fair as may be with the holes in the side frames, two feet, as I J, may have their holes reamed fair with the holes in the side frames, and tightly fitting bolts be driven in and screwed firmly home. Before reaming the other holes (as K L) of each end frame, the jaws to receive the bearing boxes should be tested for alignment one with the other. Truth, in this respect, being of the utmost consequence for the following reasons:

Suppose the bearing ways on one side frame to stand higher than those on the other, then, the shafts will not stand level in the frame unless (except in the case of the brasses or boxes in B) the lower brasses are made of unequal thickness through the crown, to an amount equal to that of the error. In the case of the brasses in A, C, D, E, the joint faces of all the brasses of one side frame would require to be made thinner beneath the journal than above it on the high frame, and thicker beneath than above on the low frame, This would entail much extra work in planing, marking, and boring the bearing boxes or brasses, and be an inferior job when done.

Again, the bores of all the brasses would not be parallel to the crown or bedding faces, and this error would entail the following extra work: 1. Ascertaining the amount of the error, and allowing for it in marking the brasses; 2. The setting of the ways of the brasses out of true with the ways when clinking them for boring; and, finally, extra fitting or filing the brass bores when fitted with the shafts in place. This extra fitting would be necessary for the following reasons:

When the surfaces of work are to be parallel, they can be measured with calipers. Surfaces to be at a right angle can be tested with a square; those to be in line can be tried with a straight-edge, and in each case the truth or alignment of the surfaces is tested by contact of the testing tool. But in the cases where surfaces at an angle are tested or measured the tools must be set to a line or lines, and the work must be measured or cut to lines, thus: Suppose it were found that the bedding surface of the brass B was a certain amount out of alignment with the corresponding bedding surface on the other side frame, and, by measurement, this amount determined to be 1/64 inch, then there is a liability to error in measuring this 1/64. The brasses must be marked (for boring the same 1/64 out of square, inducing another liability of error in marking that amount); this marking being done by lines, there is a liability to error in setting the work to the lines. From these liabilities to error, it is generally found that work not true in alignment requires, when it comes to be put together, to have each piece fitted to its place and corrected for alignment.

But, suppose the ways are made true and in proper alignment, then the brass bores are simply made of equal thickness at the crown, and on the sides at a right angle to the inside faces of the ways; and truth, in these respects, may be measured by actual contact, with the square or calipers, eliminating the chance of error.

In repairing the machine, or putting in new bearings or brasses; the measurement and transferring of the error in the ways to the brasses has all to be gone through with again, and the parts fitted for alignment; whereas, if the ways are true, the brasses can be made true, and to go together, with but little, if any, adjustment when tried in their places.

The most accurate method of testing the adjustment of the ways is as follows: Fig. 2565 represents a plan view of the frame; N represents a straight-edge applied to the surfaces of the jaws _a_ _b_. The method of applying this straight-edge is to place one end across a jaw, as _a_, while the other end is elevated above _b_; then, while pressing the end firmly against _a_, lower the other end to the face of _b_; if its edge at that end falls fair with _b_, so as just to touch it, the process may be reversed--one end being pressed to _b_, and the other lowered upon _a_. By this means, it will not only be discovered whether the jaws _a_ and _b_ stand square across the frame, but also whether the frame on either side is sprung. A square _c_ may also be rested against N, and its blade _d_ tested with the side face of the way, as shown. The same process of testing should be applied to the other jaw faces _e_, _f_.

Suppose, however, that the width between the jaws _a_, _f_ was less than that between _e_, _b_, then the straight-edge, when pressed to _a_, would show a space between its edge and _b_; and also a space between its edge and _e_, when its other end was pressed to _f_; and, when these spaces were equal in amount, the frames would be set true in one direction. To test the truth in the other direction, the straight-edge should be applied after the same manner to the bottom surfaces _g_, _h_.

It will not answer to rest the straight-edge against the two surfaces and observe their coincidence with its edge, because any error cannot be sufficiently, readily, or accurately tested by this means. Nor will it answer to test by the bearing marks of a straight-edge applied with marking, unless the coat of marking be very fine and the straight-edge be moved without any vertical pressure on it; because, under such pressure, the straight-edge will bend.

The ways for all the bearings should be tested in this manner; so that, if from any error in the machine work, some of them will not come fair, the frames may be set to align those that it is of most importance to align truly; or if there is no choice in this respect, then those carrying the largest bearing should be set true; because, if it be decided to correct the error on the other bearing or bearings, there will be less area to file or operate upon. The setting being complete the holes may be reamed and the remaining bolts put in, the testing being repeated after the frame is finally bolted together. If this final test shows that bolting the frame up has altered the alignment by springing the frame, the bolts in one foot, as say I, Fig. 2564, may be slackened and the test repeated; and, if the frame is then found true, it is the bolting at I that causes the spring, on account of the bedding surfaces not fitting properly. If I is not found to be at fault, it may be bolted up again and J tested by loosening its bolts, and so on, until the location of the error is detected. Furthermore, when the frame is bolted up, the width of the bearings, as from _a_ to _b_, should be tested; for in a job of this kind, it will pay to have the framework so true to the drawing that, if the other parts, as the shafts, bearing parts, &c., be also made to the drawings, the parts will go together, thus avoiding the necessity of varying all the other parts from the drawing to accommodate errors in the framework.

Among the jobs that the erector is often called upon to perform is that of patching or repairing pieces that have cracked or broken. Fig. 2566 represents a case of this kind, the fracture being at D. The principle to be observed in work of this kind is to cause the bolts to force the fractured pieces together, so that the irregularity or crookedness of the crack, as at D in the figure, may serve to lock the pieces together.

Suppose, for example, we were to put on a patch P, Fig. 2567, and there would be but little to prevent the crack from opening under severe strain, and the patch would stretch, permitting the crack to open and finally causing the bolts to break or sheer off. A preferable plan, therefore, is to put two patches on the sides in the following manner:--

The holes should be drilled through the beam and the plates held against the beam so that their holes may be marked by a scriber passed through the holes in the beam. The holes in the plates should be drilled closer together than those in the beam, so that when driven in they will serve as keys to close the two sides of the crack together, as shown in Fig. 2568, where it is seen that one side of the bolt bears against the holes in the patch and the other against the holes in the beam. To facilitate getting the bolts in place the plates may be heated so as to expand them.