average size cylinder used on the majority of roads did not exceed 18 or 19 inches and the boiler pressure averaged 160 pounds, the balance D-valve was satisfactory. The notches in the quadrant were usually made to engage the latch of the reverse lever in the 6, 9, 12, 15, 18 and 21-inch cut-offs. In order to give the engineer an opportunity to obtain a finer adjustment of the cut-off, the fine notched quadrant was applied to many of the simple engines, as with the old style quadrant the engine was often worked in the 9 or 12-inch notch because she would not make the time working the 6 or 9-inch, and the 9 and 12-inch was as much too heavy as the others were too light. The notches allowed the engineer to obtain the desired cut-off and regulate it to effect economy. This was appreciated by enginemen until the cylinders increased in size from 19 to 22 inches in diameter, which required larger admission and exhaust ports, also a larger valve. This with the increased boiler pressure to 190 or 200 pounds made the valve hard to handle when the throttle was well opened and necessitated the partly closing of the throttle when it was desired to change the position of the lever on the quadrant. When the valve was running rather light on the lubrication, if the latch was disengaged sometimes the engineer and the lever would both go into the corner. Then the fine notches were not used as intended and many a fireman can testify that he has shoveled extra coal into the firebox for the reason that the throttle would be eased off and longer cut-off used than necessary because it was too hard work to hook the lever back and the speed and power were regulated by the throttle. To overcome this and make the engine easy to handle, the piston valve was applied to the large modern engines, and was expected to enable the engineer to manipulate the lever with ease as well as to reduce the cost of maintenance or repairs. The fact that they will run from one to two years and then show very little wear on the rings and bushings, that there are no leaky steam chest joints for steam to escape and obscure the vision of the engineer, no valve seats to plane, or balance plates, strips, or rings, for adjustment and very little trouble with the valve rod packing, as it is exposed only to the pressure of the exhaust steam, and the drops of valve oil that are fed through the sight feed glasses of the lubricator seem to be more effective with the piston valve than with the slide valve in case either becomes dry, are all in favor of the

piston valve. The report of face valves is not entered on the work book as frequently as before the introduction of the piston valve. We are all familiar with the manner of obtaining the marks on the valve stem as shown in Fig. 1. By placing a piece of tin in the port and pushing the edge of the valve up to it then scribing the mark on the stem, with a piston valve internal admission it would be impossible to locate the position of the valve in that manner. Therefore they are usually provided with a hole drilled into the valve chamber at each end that registers with the admission ports and plugs are screwed into the peep holes to prevent steam escaping. When it is desired to mark the stems the plugs are removed and a piece of small copper pipe with a piece of wick inserted in the end that forms a very convenient torch for this purpose is inserted in the hole and the edges of the packing rings may be easily seen and the position of the valve located when the edges of the rings are opposite the edges of the ports, and the stems are marked accordingly with the same accuracy as they were with the strip of tin and the cover removed on the D-valve. In case of disconnecting on one side, the ports may be covered by placing the valve central on its seat, which may be determined by the steam ceasing to flow from the cylinder cocks on the disabled side when the throttle is opened or by placing the rocker arm in a vertical position. As for the engine tearing herself to pieces when descending a grade with the throttle closed and the lever full gear, it is not considered good practice to drift at high speed with the throttle closed and lever in full gear forward. It is better to leave the throttle slightly opened or cracked enough to admit sufficient steam to the cylinders to hold the air valves shut, thus preventing the pistons from forming a partial vacuum in the cylinders that will cause the hot gases and cinders to be drawn into them. This is the reason for not tightly closing the throttle. The reason that the lever should be notched well up in the quadrant will appear clear from the following example: Take an engine with a 60-inch driving wheel including tire. This sized wheel will make 336 revolutions per mile. The valve will weigh possibly 175 pounds. If the engine is equipped with an intermediate rocker or arm, we have another hundred pounds, neglecting the weight of the rocker and friction we have 275 pounds and a valve travel of 5% inches. If we multiply the travel of the

rings of the air piston are practically never perfectly air tight.

While it is too much to expect such tightness at the discharge valves as would permit of pouring oil through the cup with the pump stopped and pressure in the main reservoir, yet there is no excuse for more than just perceptible leakage when holding a finger or loose strip of paper just above the cup opening.

Neither is it practicable, as previously implied, to prevent all leakage by the rings, nor does an absence of any oil cup blow on the down stroke indicate the absence of leakage, as might at first thought be supposed. When the piston is descending it is creating a space to be filled to atmospheric pressure. Therefore, any air escaping at the oil cup on the down stroke is only a portion of the leakage over the amount required, with that drawn in from the outside, to fill the space above the piston to atmospheric pressure.

By supposing a loosely fitting piston wtihout rings, an air-tight cylinder and the rod moving slowly up and down it will be seen that the air would be churned from one side of the piston to the other yet at no time would there be any material pressure on the compression side or vacuum on the other. Rare cases are met with where piston rings are so worn and closed in by overheating that a very slow speed after pressure is fully pumped up will, with gummy receiving valves, actually bring about the operation just described. At such times the steam pressure required to drive the piston is so low that the exhaust is very weak, the pump working with little sound. A good pump when working against the usual pressure carried will have a distinct chug at each reversal. It is not a metallic blow and does no harm.

With the pressure and speed given, no oil cup blow will be felt on the down stroke when the air cylinder is in good order. A blow over the last one-fourth of the stroke warrants immediate repairs and one for half of the stroke is about as bad as is possible.

Many times the report, "Pump air valves sticking; clean them," should advise instead of leakage past the air piston packing rings or back into the cylinder from the main reservoir. Receiving valves are opened by suction or, in other words, the partial vacuum created by the air piston. Back leakage or leakage past the piston packing rings so nearly supplies the partial vacuum as created as to allow a slight amount of gummy oil to stick receiving valves which would

otherwise open. When receiving valves stick, remember this and test for the faults described. But first, by cleaning, get the upper receiving valve to work freely, as such a fault, this valve being put in with no lift or bad leakage at the air stuffing box are the only causes which will prevent the test described from indicating air piston packing ring leakage or back leakage from the main reservoir.

After pumps are overhauled in the repair room they should, in addition to the oil cup test, be made to pass a more severe one, an actual efficiency test.

The first essential with a locomotive is that she haul the proper tonnage satisfactorily and no matter how much expense has been put on her in repairing the results are not considered satisfactory if this tonnage is not hauled on time. The air pump is not put to any similar service test; hence, one on the lines described is necessary to guard against those repaired going into service with a lower efficiency than warranted. There is no better teacher for the observing repairman than a good efficiency test and some, at least, who have thought they were doing good work would sometimes be surprised at the results of such a test.

Broken air valves or seriously defective seats, permitting strong leakage, are not only indicated by the tests described, but will cause unequal pump strokes, as will also a stuck valve or one with insufficient lift. These faults are easily detected and located with a little reasoning after making the tests described, carefully noting the results and bearing in mind the ordinary operation of the pump.

The efficiency test should not develop the greatest capacity of the pump, but, instead, should be one requiring a certain minimum and moderate delivery or output of air at a fixed pressure and certain speed. It is very desirable that the pressure worked against be near that ordinarily met with in service, or about 90 pounds, but where this is impracticable a lower pressure will answer, though it should not be less than 60 pounds. The writer considers a speed of about 60 single strokes or exhausts per minute to be better than a higher one and that even less affords a more severe and accurate test. The slower the speed of the pistons the more opportunity there is for leakage, the common and most serious fault, to reduce the efficiency. By providing a series of leakage openings, to be used one at a time, each having a slightly greater area than the other, the efficiency of each pump can be meas

ured comparatively by the opening it can supply when working at, say, 60 strokes per minute and against 90 pounds pressure. These two factors can, within the limits implied, be made to suit the facilities, mainly steam pressure, available, but should be the same for all pumps tested.

A less accurate test, yet far better than none, is to use a fixed opening and pressure and note the number of strokes per minute required to maintain the pressure. Its weakest feature lays in the fact that, within reasonable limits, the air delivered per stroke increases with the speed, there being less time for leakage to have its effect.

This is illustrated by a leaky bicycle pump which will only pump up the tires when worked rapidly.

How many repairmen have a provision by which the pump exhaust can be discharged for a time at a point where its character can be accurately noted? Yet this test of the steam cylinder is easily provided for and is well worth making. Its absence results in serious leakage being overlooked. This test should be made with high air pressure, preferably 90 pounds, and rather slow speed as these afford the most accurate observations. At the same time, the condition of the air cylinder should be good, as the absence of the resistance this will afford would result in a low steam cylinder pressure which would not plainly indicate steam leakage existing.

A repairman should never await the complete assembling of pump to test the discharge valves, the cap, seat and cage if air pressure is at all times available, as it should be. With the air piston out and the discharge valves fitted and in place the air pressure should be applied by a connection at the discharge opening and each outside joint, as well as the air ports, inspected for leakage, repairing at once any found.

With the pump completed this can readily be extended to the cylinder heads, receiving valves, the cap and cage, by removing a discharge valve before turning air pressure into the discharge opening. F. B. FARMER.

Combustion.

Forney gives the amount of carbon in soft coal, as a general average, at 80 per cent., hydrogen 5 per cent., and the rest as incombustible or waste material, and the amount of air required for the com

bustion of each pound of coal at 12 pounds, equal to 150 cubic feet. The rate of this air admission must be proportionate to the rapidity of the coal consumption. Too much, especially over the fire, cools, and too little does not supply sufficient quantity for combustion, and in either case there is a loss of heat.

If dampers and ash pan were perfectly tight and the openings above the fire were adjustable to the demand for air, then one could control the air admission to suit the need of the hour, but in locomotive practice it is easily seen that this would be almost impossible of accomplishment and, therefore, there will always be a loss of more or less heat from this cause, and also from the ease with which the gases escape if not quickly consumed, and the condition and quality of the coal a locomotive is liable to receive on a single trip over a division of a railroad, and also the condition of the firebox in regard to leaking must be taken into consideration.

While the amount of carbon and hydrogen in soft coals on a general average is placed at 85 per cent., the amount in any certain kind of soft coal can only be determined by the test; the comparison of the performance sheets of parallel lines of railroads is no practical standard by which to determine the ability of the enginemen as economical coal consumers unless the same grade of fuel is used, the locomotives of the same class and condition, and the physical characteristics of the roads the same.

Hydrogen is the lightest gas, and, as the tendency of any gas is to expand and fill the surrounding space the hydrogen rises as soon as released from the fuel and, as the interior of the firebox when a locomotive is working is in miniature a storm center, it is immediately consumed or escapes through the flues, and in its combustion a certain amount of moisture is consumed. After the hydrogen is consumed the carbon is also burned, and this process is so rapid as to be practically continuous. There is nothing destroyed in the economy of nature. The coal is changed into heat and waste material; the oxygen of the air changes its form; the water is converted into steam, yet while the form of the element is changed it still exists in another form.

In addition to the loss of heat by the gases escaping unburned, or a part of the solid part of the carbon being unconsumed, there is a loss of heat which goes out with the escaping unconsumed gases and air, which, if they do not burn and give off heat, absorb it, and by radiation

or the giving off of heat to the surrounding atmosphere or adjacent bodies, by a heated one.

When the firebox and boiler are not properly lagged with a good nonconductor of heat there is a great loss of heat by radiation. Owing to the necessity of frequent stay-bolt inspection and replacement it is general custom to leave the outside of the firebox unlagged. If one desires to prove that a large amount of heat escapes unabsorbed up the smokestack, stand on the smoke arch and one can feel the escaping heat radiate as it comes out from the top of the smokestack. The loss of heat, while unmeasurable, amounts to quite a per cent. of the real heat value of the fuel consumed.

The arch is made of firebrick extending in the form of an arch across the firebox from side sheet to side sheet, from a point below the flues in the flue sheet and sloping up and extending back to such a distance as desired. The number of bricks in the arch or, more properly speaking, the size of the arch, being varied to meet the master mechanic's idea of each particular road's requirements. Sometimes an engine will do better with a larger than a smaller arch, or vice versa, and this also should enter into the consideration. The bricks next to the tube sheet are hollowed out so as to permit of a certain amount of draft in front of the arch for the fuel next to the flue sheet and which otherwise, on account of the arch lying so close to it, would burn too slowly for satisfactory results, but the intention in having an arch is that by throwing the gases and the air back and over the arch to the tubes there will be more time for their mixing and burning than where they go directly into the flues.

The brick arch serves also as a protection to the flues by preventing the cold air that is admitted above the fire from coming in direct contact with them and cooling them down, and thus causing them to leak. Where air is admitted above the fire the use of an arch is very necessary. As the arch prevents the sudden cooling of the tubes it also causes their rise in temperature to be gradual, as it heats and cools slowly. Long after the fire is out of the engine the arch will be found to be glowing hot, and this has led many roads to avoid its use, particularly in freight service, where the demand for power to handle the business being acute and, therefore, there being not sufficient time to wait for the arch to cool before the boilermaker could do his work, and many of the engines having flues that re

quired calking every trip, rather than be to the expense of removing and replacing the arch every trip the use of the arch has been abandoned, temporarily at least. The arch being very hot and retaining its heat it aids in bringing the gases rising from the coal to an igniting temperature more quickly than could otherwise be done, and in this respect it also aids combustion.

Some roads have tried the use of the water table instead of the brick arch as it is more durable. This consists of two flat sheets of boiler iron about four or five inches apart, attached to the side sheets and filled with water. Its use, like that of the hood over the door to deflect the air from the flues, is not so general as that of the brick arch.

If gases enter the flues unburned there is no chance, practically, of their being burned there owing to the low temperature in the tubes and the amount of heavy gas in them. A light lowered in a well where there is carbonic acid gas will immediately be extinguished, and the knowledge of this fact is used by well diggers to protect them before descending into old wells for the same reason that when a blaze enters the flues charged with gas at a low temperature it dies at once. The flues being surrounded by water and being in direct contact with the fire only at the firebox end where the blaze enters them for a short distance are always at a low temperature compared with the adjacent parts of the firebox, and this temperature diminishes toward the front end of the boiler. Flues seldom leak at the front tube sheet from the fact that there they are not subjected to the radical changes in temperature that they are at the firebox end, therefore, at the front tube sheet end the contraction and expansion is not so great. A small tube is used in locomotives for two reasons: First, in a given boiler shell more heating surface can be obtained by the use of small flues, and a thinner one that will transfer the heat passing through it more readily to the water can be used. This can not be carried to an extreme, as too thin a flue would collapse and too small a flue would stop up quickly, yet a 2-inch tube can be made much thinner than a 4-inch tube and stand the same pressure. As the same amount of heat that raises one pound of water one degree will raise nine pounds of iron one degree the fact that iron is a good heat conductor is fully demonstrated. The same amount of heat will raise thirty pounds of mercury or gold one degree, and from this it is plain that water is not as