Chapter XXXVII: The Steam Engine (2)

Then from V to V´ draw the line P, and from this line to the centre C of the crank shaft is the amount of steam lap necessary for the valve, while from this line (P) to D is the width of the steam port.

The proof of the diagram is as follows:

When the crank is on the dead centre, the centre of the eccentric is at V, its throw line being represented by the line from V to C, and the valve is about to open the port as shown in the figure.

While the eccentric is moving from V to D, the valve will move in the direction of the arrow and will fully open the port, while the crank pin will move from B to F.

Then, while the crank moves from F to H, the eccentric will have moved the valve to the position it occupies in the figure, having closed the port and effected the cut off.

We have here found the amount of lap and the position of the eccentric necessary for a given point of cut off when the latter is given in terms of the piston stroke. If, however, the point of cut off had been given in terms of the crank pin position, we might find the required amount of lap at once, by simply drawing a line from the centre B, the point to H where the crank pin is to be when the cut off occurs.

From this line we could then draw the dotted circle G, and just meeting the line P, which would give the eccentric position.

To find the piston position, the arc Y would require to be drawn by the same means as before.

If the valve is to have lead, the diagram may be constructed as in Fig. 3313, in which the circle has a diameter equal to the travel of the valve and the cut off is to occur when the piston is at R and the crank at H.

When the valve is at the end of its travel and has fully opened the port, the eccentric will be at D, hence from D we mark an arc G distant from D to an amount equal to the width of the steam port, drop the vertical _m_ from G, and at its lower end V´ is the position of the eccentric centre at the point of cut off. Then draw a line P, distant from _m_ equal to the lead, which will give at V the position of the eccentric when the crank is on the dead centre, and the valve is open to the amount of the lead. The lap is obviously the distance from the centre C of the crank shaft to the arc G.

We have here found all the points necessary except the point at which the valve will open the port for the lead, and this we may find by setting a pair of compasses to the radius B H (or to radius V V´, as both these radii are equal), and from V as a centre, mark at A an arc, which will give the crank pin position at the time the port first opens for the lead, or in other words it will give the position. The proof of the construction is, that if we set the compasses to the distance between the crank pin position on the dead centre and the point of cut off (or from B to H), we may apply the compasses to the points V, V´, which represent the eccentric position when the port is opened to the amount of the lead, and when the cut off occurs.

If the point of cut off only is to be found, we mark from C, Fig. 3314, an arc G representing the amount of valve lap and arc S representing the lead. A vertical P gives the eccentric position V when the crank is on the dead centre at B, and a vertical _m_ from G gives at V´ the eccentric position at the point of cut off. Then with the compasses set to the points V V´, we may mark from B an arc, locating at H the position of the crank at the point of cut off, and from this with compasses set to represent the length of the connecting rod on the same scale as the circle represents the path of the crank, we may, from a point on the line of centres, mark an arc Y giving at R the piston position at the point of cut off.

When, therefore, the lap is given, we mark it from the center C of the crank shaft, and find the other elements from it, whereas, when the lap is to be found, we mark the width of the port from the end D of the valve travel, and find the other elements from that.

A proof of all the constructions is given in Fig. 3314, in which the letters of reference correspond to those in the previous figures, and the positions of the parts are marked in degrees of angle.

To find the piston position at the point of cut off, measured in inches, of the piston stroke it must be borne in mind that as the circle B D represents the full travel of the valve, the diagram gives all the positions of the eccentric and valve full size, but that as it represents the crank path on a reduced scale, therefore we must multiply the measurement on the diagram by that scale.

Suppose, for example, that the piston stroke is 10 inches, and the valve travel 2-1/2 inches, and the circle being 2-1/2 inches in diameter, is, when considered with relation to the eccentric motion, full size, but when considered with relation to the piston or crank motion, it is only 1/4 the size, hence to find the piston position at the time of cut off, we must multiply the distance from B to R by 4.

LINK MOTION FOR STATIONARY ENGINES.

The ordinary mechanism employed to enable a stationary engine to be reversed or run in either direction is the Stephenson link motion. Other forms of link motion have been devised, but the Stephenson form has become almost universal.

Fig. 3315 represents this link motion or reversing gear with the parts in position for the full gear of the forward motion, and Fig. 3316 represents it in full gear for the backward motion.

The meaning of the term full gear is that the parts are in the position in which the steam follows the piston throughout the longest or greatest part of the stroke. When in full gear the link motion operates the valve almost precisely the same as if the eccentric rod was attached direct to the valve spindle and no link motion was used.

Besides enabling the engine to run in both directions, however, the link motion provides a means of reducing the amount of valve travel and thus causes the live steam to be cut off earlier in the piston stroke, thus using the steam more expansively. This is done by moving the reversing lever more upright, the earliest point of cut off being obtained when it is upright and the latch is in the notch marked O on the sector in Fig. 3315. If with the engine standing still we move the link motion from full gear forward to full gear backward and watch the valve, we shall find that the valve lead increases as the reversing lever approaches the upright position, or mid gear as it is termed, and that after passing that point it gradually diminishes again, the valve being so set that the lead is the same for full gear forward as it is for full gear backward.

The reversing lever is used to move the link into the required position and to hold it there (the end of the latch fitting into the notches in the sector being the detaining or locking device); as the link is suspended by its saddle pin S and the link hanger, therefore its motion is to swing or partly rotate on the pin S, and at the same time ending in the arc of a circle whose centre of motion is in the pin at the upper end of the link hanger which is pivoted to the lower arm of the lifting shaft (which is sometimes termed the tumbling shaft). It will clearly be seen that with the position the parts occupy in Fig. 3315, and the crank motion being in the direction of the arrow, the forward eccentric will move the top of the link to the right and therefore the valve will move to the right, while the backward eccentric will move the bottom end of the link to the left.

In full gear, however, the bottom eccentric rod has but a very slight effect indeed on the motion of the valve because both the link hanger and the link block will permit the link to swing on centre of the link block pin as a pivot. If now we turn to Fig. 3316 for the full gear backward, we shall see that these conditions are reversed and the backward eccentric becomes the effective one, being in line with the valve spindle. By shifting the link from one gear to the other, therefore, we have merely changed the direction in which the link will move the valve, and, therefore, the direction in which the engine would run.

In Fig. 3315 for the full gear the parts are shown in position, with the piston at the crank end of the cylinder, and the crank pin on the dead centre, and the eccentrics must be set as shown in the cut, the eccentric rods being open and not crossed. When, however, the crank is on the other dead centre and the piston at the head end of the cylinder, the rods will cross each other, and it is necessary to remember that the rods should be open when the piston is at the crank end of the cylinder. If, however, the running gear contains a rock shaft, or rocker (as is the case in American locomotives), then these conditions are reversed, and the eccentric rods will cross when the piston is at the crank end of the cylinder.

In setting the slide valve of an engine having a link motion, there are two distinct operations. First, to put the crank on the respective dead centres, which will be fully described on page 394 and need not be repeated; and second, to set the eccentrics in their proper positions on the shaft, and correct, if necessary, the lengths of the eccentric rods. The crank being on the dead centre, with the piston at crank end of the cylinder, the eccentric should be moved around on the shaft by hand until there is the desired amount of lead at the crank end port, and temporarily fastened there, a set screw usually being provided (in the eccentric) for this purpose. The lead is best measured with a wedge, W, Fig. 3315. The crank is then put on its other dead centre, and the lead for the head end port is measured. If the lead is to be made equal for the two ports (as is usually the case in horizontal engines) and it is found to come so, the valve setting for the forward gear is complete. If the lead is not equal, the forward eccentric rod or else the valve spindle must be altered so as to make the lead equal. In some engines adjusting screws are provided for the purpose of regulating the length of either the eccentric rod or else of the slide spindle; it does not matter which is altered. The link motion is then put in full gear for the backward motion, and, with the crank on the dead centre (it does not matter which dead centre), the eccentric is moved by hand upon the crank shaft until there is the required amount of valve lead. The eccentric is then fastened on the shaft and the crank put on the other dead centre, and the lead tried for the other port, and made equal by lengthening or shortening the backward eccentric rod. It is to be noted that altering the length of the eccentric rod or of the valve spindle makes it necessary to reset the eccentric, as it affects the amount of lead at both ports; hence, if any alteration of rod length is made, the whole process here described must be repeated after each alteration of rod length.

FLY BALL OR THROTTLING GOVERNORS.

An isochronal governor is one in which the two opposing forces are equal throughout the whole range of governor action, or, in other words, equal, let the vertical height of the plane in which the balls revolve or swing be what it may.

A dancing governor is one that acts spasmodically. Such an action may occur from undue friction in the parts of the governor or of its throttle valve.

The friction offers a greater resistance to starting the parts in motion than it does to keep them in motion after being started; hence, the parts are apt to remain at rest too long, and to move too far after being put in motion.

Rule to find the number of revolutions a governor should make. Divide the constant number 375.36 by twice the square root of the height of the cone in inches. The quotient is the proper number of revolutions per minute.

_Example._--A governor with arms 30-1/2 inches long, measuring from the centre of suspension to the centre of the ball, revolves, in the mean position of the arms, at an angle of about thirty degrees with a vertical spindle forming a cone of about 26-1/2 inches high. At what number of revolutions per minute should this governor be driven? Here the height of the cone being 26.5 inches, the square root of which is 5.14 and twice the square root 10.28, we divide 375.36 by 10.28, which give us 36.5 as the proper number of revolutions per minute at which the governor should be driven.

The construction of the Pickering governor is as follows:

In Fig. 3317 it is shown in its simplest form, and in Fig. 3318 with the driving pulley and speeder (or engine speed regulating device) attached. This speeder consists of a spiral spring whose tension may be adjusted to more or less resist the rise of the governor balls, and thus enable the engine to run at a greater speed for a given amount of rise of the governor balls, hence by increasing the tension the engine speed is increased.

THE SPRING ADJUSTMENT.

The adjustment of the spring tension is made by a worm actuating a worm wheel on a rod passing through the spring, and to which one end of the spring is attached, the other acting on an arm that projects into a slot in the governor spindle. It is obvious that the speeder can be adjusted while the engine is running.

In Fig. 3319 the governor is shown with the speeder and Sawyer's valve, the latter enabling the governor valve to be opened or closed without affecting the rise and fall of the governor balls, which is done by operating the arm shown on the right, whose ends are provided with loops, so that a cord may be attached, enabling the engineer to operate the governor from a distance.

The safety stop or stop motion is shown on the right, Fig. 3320.

It acts to close the governor valve and stop the engine in case the belt that drives the governor should get off the pulley or break. This stop motion consists of a pulley suspended by a rod, and riding on the belt which supports its weight. Should the governor belt break, this pulley falls and severs the connection between the valve and the governor, closing the valve, and holding it closed. Fig. 3321 shows the governor in section to expose the construction of the valve. The valve V is what is termed a poppet or poppet valve, which is balanced, because the steam entering at I, and taking the course denoted by the arrows, acts equally on both ends of the valve and does not press it in either direction, while as the steam surrounds the valve it is not pressed sideways.

At B is a gland or stuffing box to keep the spindle or rod steam-tight. At A is the slot for receiving the arm from the speeder and from the stop motion.

P is obviously the driving pulley, imparting motion to the bevel wheels G, which drive the outer spindle S, the inner spindle _s´_ being connected to A. The balls are upon ribbon springs D, which are secured at their lower ends to a link fast to the spindle S.

The centrifugal force generated by the balls causes them to move outwards, their upper ends pulling down the cap to which they are secured, and this cap operates the valve.

Governors of this class are sometimes termed _fly-ball_ governors.

STARTING A PLAIN SLIDE-VALVE ENGINE.

The method to be pursued before starting a plain slide-valve engine depends upon what the engineer knows about the condition of the engine.

If he knows the engine is in proper running order, all that is necessary is to first attend to the oil cups and start them feeding.

Then, if it is necessary, move the crank into the required position to start it easily; open the waste water cocks to relieve the cylinder of the water that will be condensed from the steam when it enters a cool cylinder, and turn on the steam; giving the throttle valve enough opening to start the engine slowly.

The best position for the crank pin to be in to enable its starting easily is midway between the horizontal and vertical position (or, in other words, at an angle of 45° to the line of centres) and inclining toward the cylinder, so that when the engine moves the piston will travel toward the crank shaft.

There are two reasons why this is the best position for starting. The first applies to all engines because there is a greater piston area for the steam to act on when the piston is moving toward the crank than there is when it is moving away from it. This occurs because the piston rod excludes the steam from a part of the face of the piston. The second applies to all plain slide-valve engines whose slide valves have equal laps and both steam ports of equal widths, because the live steam follows further on the stroke when the piston is moving toward the crank than it does when it is moving away from it, and it follows that more piston power is developed, and the engine is less likely to stop when passing the dead centre.

When first taking charge of an engine, it is proper, before starting it, to ascertain that it is in fair working order.

A complete examination of an engine should include a test of the fit of the piston to the cylinder bore, of the cross head to the guide bars, of the connecting rod brasses to the crank pin and cross head journals, and of the crank shaft to its bearings. It would also include a testing of the alignment of the crank shaft and of the guide bars, as well as the set of the valves and the adjustment of the governor.

The least examination permissible with a due regard to safety would be to move the engine throughout at least one full revolution by hand, and to see that the connecting rod brasses and the main bearings do not fit too tight to their respective journals, and to then start the engine slowly by giving it only enough steam to move it, keeping the hand on the throttle valve so as to be able to shut off steam instantly should it become necessary.

A thorough examination should be made in the following order:

First, slightly loosen the nuts on the crank shaft bearings and also the connecting rod keys.

Then move the fly wheel around until the crank points straight to the cylinder, which will bring the piston up to the outer end of the cylinder bore.

Take off the cylinder cover and also the follower from the piston head, and see that the piston rings are set out to fit the cylinder bore but not to bind it tight. Then bolt the follower up firmly in place again.

Take off the connecting rod and move the piston until it touches the cylinder cover at the other or crank end of the cylinder, and then draw a line across the side face of the cross head guide and on the guide itself.

Put on the cylinder cover and push the piston back until it abuts against it, and then make another line on the cross head guide and the guide bar, and these two lines will show the extreme positions to which the piston can be moved when the connecting rod is disconnected.

Next put on the connecting rod, carefully adjust the keys or wedges, so that the bores of the brasses fit easily to the crank pin and cross head pin, seeing that the oil holes are clear, and that oil will feed properly to the journals.

In making this adjustment it is a good plan, if there is any end play of the brasses on the crank pin, to set up the key or wedge until the rod can just be moved by hand on the pin, by first pulling the rod to one end of the pin, and then pushing it to the other.

In putting on the rod, it will be necessary to move the piston a trifle towards the crank.

In making the adjustment of the crank pin fit to the rod brasses, it is a good plan to drive the key home until the brasses are known to bind the crank pin, and then mark a line across the side face of the key and fair with the top face of the connecting rod strap, to then slacken back the key enough to ease back the brasses to a proper fit, and then mark another line on the key.

The first line will form a guide as to how much to slacken back the brasses to adjust the fit, and the second one will form a guide as to how much the key is moved when making a second adjustment, if one should be found necessary after the engine has been running.

Similarly in adjusting the main bearing boxes to the crank shaft, either the nuts, or what are called leads, may be taken to adjust the fit. Leads are necessary when the joint faces of the brasses do not meet, but are left open so that the wear can be taken up while the engine is running.

It is better, however, to let the brasses abut together, so that it may be known that the fit is correct when the nut is screwed firmly home.

The method of taking a lead is as follows: The top brass is loosened, and between the joint faces of the brasses or boxes on each side of the shaft a piece of lead wire is inserted. For a shaft of, say, four inches in diameter, the lead wire will be about 7/16 inch in diameter, or for a 10 inch shaft the wire should be 1/8 inch in diameter, and should be as long as the brass. The nuts are then screwed firmly home, and the wire will be squeezed between the brasses and thus flattened on two opposite sides, the thickness showing how far the joint faces of the brasses are apart when the bore grips the journal.

A liner, fit strip, distance piece, or shim (all these names meaning the same thing) is a strip of metal placed between the joint faces of the brasses to hold them the proper distance apart to make a working fit of the journal and brasses, when the latter are firmly bolted up.

The fit of the top brass therefore depends upon the fit strip being of the proper thickness from end to end.

Now the lead wire is the gauge for the thickness of the fit strip, the latter being made a trifle thicker than the flattened sides of the lead.

If the lead is thicker one end than the other, or if one lead is thicker than the other, the fit strips must be made so, and the leads must be marked so that it may be known which way they were placed between the brasses so that the proper fit strip may be on the proper side of the brass, and the proper end towards the crank.

Another method that is adopted in the case of large brasses is to screw down the nuts until the brasses bind the journal, and then make a mark on the nut and on the bolt thread. The nut is then slackened back as much as the judgment dictates, and a note made of how much this is, the marks forming a guide.

As the wear takes place, and the nuts screw farther down, a new mark is made on the nut, so that it may always be known how much to screw up or unscrew the nut, to make a light adjustment.

To avoid heating, it is a good plan to press some tallow into the bottom or in one corner of the oil cup, and then pour in the oil used for ordinary lubrication. So long as the bearing remains cool, the oil will feed and the tallow remain.

If the bearing heats, the tallow will melt, and, having a heavier body, will give a more suitable lubrication.

To find if the connecting rod is of the right length to give, as it should do, an equal amount of clearance (or space between the piston and the cylinder cover) at each end of the stroke, move the fly wheel a trifle in either direction, and then move it back until the crank is on the dead centre, and draw a line across the cross head guide and guide bar, and the distance between this line and that drawn when the connecting rod was disconnected, shows the amount of clearance at that end of the cylinder. Then move the crank pin over to its other dead centre, and mark a line across the cross head guide and the guide bar, and the distance between this line and that drawn before the connecting rod was put on will show the clearance at this end of the cylinder.

If the clearance is not equal for the two ends, it should be made so by putting liners behind the connecting rod brasses so as to lengthen or shorten the connecting rod (according as the case may require), and equalize the clearance, while at the same time bringing the connecting rod keys up to their proper heights.

To test the set of the valve, the steam-chest cover must be taken off, the crank placed alternately on each dead centre, and the lead measured for each port.

An unequal or an equal degree of valve lead may be given by suitably altering the length of the eccentric rod, but when the lead is equal for the two ports, its amount must be regulated by moving the position of the eccentric upon the crank shaft.

SQUARING A VALVE.--A method not uncommonly pursued in setting a valve is to what is called _square it_ before trying it.

This squaring process consists in so adjusting the length of the eccentric rod that the valve travels an equal distance over or past the steam edge of each steam port; but since the valve does not, when set to give equal lead, travel equally past each port, therefore the work done in squaring a valve is all thrown away, and may result in altering the eccentric rod from its proper length to an improper one, necessitating that it be altered back again in order to set the lead right.

The proper method is to adjust both the length of the rod and the position of the eccentric, by testing the lead at once, lengthening the eccentric rod to increase the lead at the crank end, or vice versa.

Each alteration of eccentric position may render necessary an alteration of rod length, or vice versa, each alteration of rod length may render it necessary to alter the eccentric position, hence the lead should be tried at both ends of the cylinder after each alteration of either rod length or eccentric position.

In vertical engines the weight of the crank shaft causes it to wear the bottom brass or part of the bearing box the most, thus lowering its position, while the eccentric straps and pins wear most in the same direction; hence the wear increases the lead at the head end of the cylinder when the latter is above the crank, and at the crank end when the crank is above the cylinder.

When the cylinder is above the crank, the weight of the piston, cross head and connecting rod is counterbalanced at the end of the downward piston stroke by giving the crank end port more lead; but when the cylinder is below the crank, it is the head end port that must be given increased lead to prevent a pound or knock, or to allow for the wear downwards of the parts.

After an engine is started, the pet cocks should (if they are not automatic) be closed as soon as dry steam issues, and if this cannot be seen, it may be assumed to occur after the engine has made about 20 revolutions.

The parts that will then require particular attention are the crank pin, main bearings, cross head guides and the pump, if there is one. The former must be kept properly lubricated, so that they may not get hot and the cylinder lubricator (which is usually placed on the steam pipe) must be set to self feed properly.

If the crank shaft bearings should begin to heat, loosen the cap bolts and lubricate more freely, or, if it is at hand, some melted tallow may be applied with the oil, as a heavier lubricant may stop the heating.

The crank pin requires the most attention and is the most difficult to keep cool and to examine, because of its circular path rendering it difficult to feel it. This may be done, however, in two ways, first by standing at the end of the engine bed and gradually extending the hand, until the end of the rod meets it as it passes, and, second, by placing the hand on the connecting rod as near to the end of the guide bar as possible where its motion is diminished and moving the hand towards the crank pin, by which means the end of the crank pin may be approached gradually.

If the end of the rod is hot, the engine speed should be reduced or the engine should be stopped so that the connecting rod key or wedge may be eased back and the oil feed made more copious. Then, after the engine has been stopped for the night, the brasses should be taken out and any rough surface, either on the brasses or on the pin, smoothed down with a file.

Hot crank pins may occur from several causes, but by far the most common ones are from improper oiling, or from the engine being out of line.

A heavier oil will often stop, or at least modify, the heating, but its cause should always be discovered and remedied.

Engines that are used out of doors or are exposed to temperatures below the freezing point must be left so that steam leaks may not condense in any of the parts or pipes and burst them.

Leaky throttle valves may, for example, cause water to accumulate in the steam chest and freeze, perhaps bursting the steam-chest cover.

To prevent this let the engine stand with the crank just past the dead centre, so that the steam port will be open, and open the waste water cocks on the cylinder, and also on the steam chest if there is any.

If the cylinder is jacketed all the drain cocks for the jacket should also be opened.

A leaky check valve may cause the steam to condense in the pump and freeze it up solid or burst it or the pipes. To avoid this, open the pump pet cock.

Open all the drain cocks on the heater and water pipes.

If the water is left in the boiler all night it is liable to freeze.

To prevent this leave a well banked fire.

In extreme weather remember that on exposed engines the oil, if of such quality as sperm or lard oil, may freeze and prevent feeding until the bearings get hot and melt the oil.

To prevent this use a lighter oil, as, for example, a mineral oil. Or, in case of freezing, melt the oil in the cups with a piece of wire made red hot while getting up steam in the morning.

A good plan to prevent oil from freezing and yet have a good quality of oil is to mix two parts of lard oil with one part of kerosene.

Portable engines should stand as nearly level as possible, so that the water will stand level above the tubes and crown sheet of the fire box.

When feed water is drawn from a natural supply, as from a stream, the strainer at the end of the suction pipe should be clear of the bottom of the stream, where it is liable to be choked.

When the exhaust steam is used to feed the boiler, do not open the valve that lets the exhaust steam into the feed-water tank until a little while after the engine has started, because the oil fed to the cylinder will otherwise pass into the feed tank and may cause priming.

In engines having plunger pumps for feeding the boiler it is essential to keep the plunger properly packed, as a leak there impairs or stops the pump from acting.

A gauge glass may be cleaned when the engine is cold by shutting off the cocks leading from the boiler and filling the glass with benzine, allowing it to stand two hours; the benzine must be let out at the bottom of the glass tube, and not allowed to enter the boiler.

In starting a new engine be careful to let the bearings be slightly loose.

At first give only enough steam to just keep the engine going, and keep the hand on the throttle valve ready to shut off steam instantly if occasion should require.

PUMPS.

Pumps are divided into the following classes:

Lift pumps, in which the water flows freely away from the pump, which performs lifting duty only.

Force pumps, which deliver the water under pressure.

Plunger pumps, in which a "plunger," or "ram," as it is sometimes termed, is used.

Piston pumps have a piston instead of a plunger.

A double acting pump is one in which water enters into and is delivered from the pump at each stroke of its piston or plunger, or, in other words, one in which, while water is being drawn in at one end of the pump, it is also being forced out at the other.

A single acting pump is one in which the water enters the pump barrel during one piston or plunger stroke, and is expelled from the pump during the next stroke, hence the action of the suction and of the delivery is intermittent, although the pump is in continuous action.

For very heavy pressures plunger pumps are generally used, the plunger being termed a _ram_.

The advantage of the plunger or ram is that it gives a positive displacement, whereas in a piston pump a leaky piston permits the water from the suction side to pass through the leak in the piston, to the delivery side.

Piston pumps possess the advantage that there is less difference between the contents of the pump and the displacement than is the case in plunger pumps.

The displacement of a piston pump is found by multiplying the area of the pump bore by the length of the piston stroke.

The displacement of a plunger pump is less than the above, by reason of there being a certain amount of clearance or space between the circumference of the plunger and that of the cylinder bore.

It is desirable to keep the clearance space in all pumps as small as the conditions will allow, especially if the pump is liable to lose its water.

Losing the water means the falling of the suction water back into the source of supply, which may occur when the engine has to stop temporarily, and there is a leak in the suction valves.

Rotary pumps are those in which the piston revolves, an example of the most successful form of rotary pump being shown in Fig. 3322, which is that used by the Silsby fire engine.

The advantage possessed by a rotary pump is that it keeps the water passing through the suction in a continuous and uniform stream, as it has no valves.

It may therefore be run at a high velocity or attached direct to the engine shaft.

If a rotary pump leaks, the efficiency is not impaired so much as in a piston or plunger pump, all that is necessary being to run the pump at a high speed.

The principles of action of a pump may be understood from Fig. 3323, which represents a single acting plunger pump shown in section, and with the suction pipe in a tank of water, the pump being empty.

The surface of the water in the tank has the pressure of the atmosphere resting upon it, and as the pump is filled with air, the surface of the water within the pipe is also under atmospheric pressure.

Now suppose the plunger to move to the right, and as no more air can get into the pump, that already within it will expand, and will therefore become lighter, hence there will be less pressure on the surface of the water within the suction pipe than there is on the outside of it, and as a result the water will rise up the pipe, not because the plunger draws it, but because the air outside the pipe presses it up within the pipe.

The water inside the pipe will rise above that outside in proportion to the amount to which it is relieved of the pressure of the air, so that if the first outward stroke of the plunger reduces the pressure within the pump from 15 lbs. to 14 lbs. per square inch (15 lbs. per square inch being assumed to be its normal pressure), the water will be forced up the suction pipe to a distance of about 2-1/4 feet, because a column of water an inch square and 2-1/4 feet high is equal to 1 lb. in weight. In Fig. 3324 the pump plunger is shown to have moved enough to have permitted the water to rise above the suction valve, and it will continue to rise and enter the pump barrel as long as the plunger moves to the right.

When the plunger stops, the suction valve will fall back to its seat and enclose the water in the pump; but as soon as the plunger moves back to the left hand and enters the barrel pump further, the delivery valve will rise, and the plunger will expel from the pump a body of air or water equal in volume to the cubical contents of the plunger, or rather of that part of it that is within the barrel, and displaces water.

If the plunger was at the end of its first stroke to the right and the pump half filled with air, then this air will be expelled from the pump before any water is; whereas if the pump was filled with water, the latter only will be delivered.

Now suppose the first plunger stroke reduces the air pressure from 15 to 14 lbs., and that the second drawing stroke of the plunger reduces the air pressure in the pipe to 13 pounds per inch, the water will rise up it another 2-1/4 feet, and so on until such time as the rise of a column of water within the pipe is sufficient to be equal in weight to the pressure of the air upon the surface of the water without; hence it is only necessary to determine the height of a column of water that will weigh 15 lbs. per square inch of area at the base of the column to ascertain how far a suction pump will cause water to rise, and this is found by calculation or measurement to be a column nearly 34 feet high.

It is clear then, that however high the pump may be above the level of the water, the water cannot rise more than 34 feet up the suction pipe, even though all the air be excluded from it and a perfect vacuum formed, because the propelling force, that is, the atmospheric pressure, can only raise a column of water equal in weight to itself, and it is found in practice to be an unusually good pump that will lift water thirty feet.

Fig. 3325 shows the plunger making a delivery stroke, the suction valve being closed, and the delivery valve open where it will remain until the plunger stops.

To regulate the quantity of water the pump will deliver in cases where it is necessary to restrict its capacity, as in the case of maintaining a constant boiler feed without pumping too much water in the boiler, the height to which the suction valves can lift must be restricted, so as to limit the amount of water that can enter the pump at each drawing stroke.

The delivery valve should lift no more than necessary to give a free discharge without causing the valve to seat with a blow; but if the pump has a positive motion, the delivery valve must open wide enough to let the water out, or pressure enough may be got up in the pump to break it.

A check valve is merely a second delivery valve placed close to the boiler and serving to enable the pump to be taken apart if occasion should arise, without letting the water out of the boiler.

The lift and fall of both valves act to impair the capacity of the pump. Thus, while the suction valve is falling to its seat, the water already in the pump passes back into the suction pipe, and similarly, while the delivery valve is closing, the delivery water passes back.

A foot valve is virtually a second suction valve placed at the bottom or foot of the suction pipe.

The capacity of a pump is from 70 to 85 per cent. of the displacement of the plunger or piston, and varies with the speed at which the plunger or piston runs.

If a pump runs too fast, the water has not sufficient time to follow the piston or plunger, especially if the suction pipe has bends in it, as these bends increase the friction of the water against the bore of the pipe.

The speed of the piston or plunger should not exceed such as will require the water to pass through the suction pipe at a speed not greater than 500 feet per minute, and better results will be obtained at 350 feet per minute.

An air chamber placed above the suction pipe of any pump causes a better supply of water to the pump by holding a body of water close to it, and by making the supply of water up the suction pipe more uniform and continuous. Air chambers should be made as long in the neck as convenient, so that the water in passing through the pump barrel to the delivery pipe could not be forced up into the chamber, as, if such be the case, the air in the chamber is soon absorbed by the water.

Belt pumps are more economical than independent steam pumps, because the power they utilize is more nearly the equivalent of the power it takes to drive them, whereas in steam pumps there is a certain amount of steam, and therefore of power, expended in tripping the valves and in filling the clearance spaces in the cylinder. Furthermore, the main engine uses the steam expansively, whereas the steam pump does not.

Fig. 3326.]