reciprocating pump. A strap was carried from the fly-wheel to a pulley on the spindle of the pump, and as the diameters of the flywheel and the pulley were as 2.3 to 1.0, the pump made 1.15 revolutions per stroke of the steam-piston.

The steam siphon-pump is simply one of the applications of the Giffard Injector. It is constructed as follows: Two suction-mains, or pipes, 2 inches in diameter, are screwed at an angle of 42° into a hollow brass sphere 53 inches diameter. Between them is a nozzle for the steam-jet; its upper end is of an inch above the centre of the sphere, while the lower, to which the steam-pipe is attached, passes below it. Above this a nozzle for the discharge-pipe is screwed into the sphere, shaped like an inverted cone, the lower diameter being 14 inch, and the upper 2 inches; it is 4 inches long, and the discharge-pipe, 22 feet long, is attached to it. During the three trials, three different nozzles for the steam-jet were used. All were of the same height, 63 inches, and of the same inner lower diameter of 14 inch, but the upper inner diameters were successively 3,, and of an inch, according to the boilerpressure of 20, 30, or 40 lbs. per square inch. The water-discharge was of course continuous. The diameter of the steam-pipe was 1 inch.

6

In making the experiments, all the pumps were supplied with steam from the same boiler. The feed-water was pumped into the boiler by a small auxiliary steam-pump, after being carefully measured. The exhaust steam from the auxiliary pump was discharged back into the tank supplying it. Thus the measured feedwater correctly represented the weight of steam used in the cylinder for the pumps, although the auxiliary pump was worked from the same boiler.

Before an experiment, all the tanks, pumps, and boiler were thoroughly tested for leakage. The boiler and steam-pipes were felted.

In the supply-tank, the water was admitted in such a way as to keep it at a certain level during the experiments. It was in all cases raised to a height of 17 feet 8 inches; the discharge-pipe was 1 inch in diameter, 28 feet long for the rotary and reciprocal pumps, and 22 feet long for the siphon. The water was delivered into two receiving-tanks, each holding exactly 160 cubic feet; they were placed side by side on an elevated platform, and carefully levelled. As the cost of the experiments was determined by the weight of feed-water pumped, the weight of coal consumed was not measured. The experiments with the reciprocating and rotary pumps lasted seventy-two hours. With the siphon-pump three experiments of twenty-four hours each were made, in order to ascertain the effect of working it with steam of different pressures, as mentioned above. Every hour there were taken the average temperature of the air outside, and within the experimenting shed, of the water in the supply- and receiving-tanks, height of barometer, and steam-pressure in the boiler. operator also recorded the number of strokes or revolutions per

The

hour of the pumps, and number of lbs. of feed-water pumped into the boiler. The exact time at which each receiving-tank was emptied was noted, and at the end of each experiment the water in the boiler was left at the same height, and with the same steampressure as at the beginning.

RESULTS OF THE EXPERIMENTS.

(Height of water, 17 feet 8 inches.)

I. Reciprocating-pump

Duration of trial, seventy-two hours.

Total cubic feet of water lifted, 52,240 feet.

Number of lbs. of water raised per lb. of steam, 135.6 lbs.

II. Rotary pump

Duration of trial, seventy-two hours.

Total cubic feet of water lifted, 45,280 feet.

Number of lbs. of water raised per lb. of steam, 108.6 lbs.

III. Steam siphon-pump

Duration of trial: three trials of twenty-four hours each.

Total cubic feet of water lifted: First trial, 7.840; second trial, 9,760; third trial, 9,440.

Number of lbs. of water raised per lb. of steam: First trial, 29.6 lbs. ; second trial, 37.3 lbs.; third trial, 37·4 lbs.

It must be remembered that the weight of water lifted by the siphon-pump is not only that lifted from the supply-tank, but also the weight of feed-water pumped into the boiler.

In comparing the results, the Author begins with the steam siphon-pump. This pump was found to work very badly with a pressure of 20 lbs., but acted perfectly when pressures of 30 lbs. and 40 lbs. were applied. The duty in the two last cases varied but slightly, and taking their mean as the correct performance, there would result 661 lbs., of water raised 1 foot high, for every lb. weight of steam used. As this duty includes the weight of steam expended, it represents the total dynamic effect produced by the steam.

The duty with the reciprocating-pump was 2,397 lbs. of water, raised 1 foot high by 1 lb. of steam, or 3.6 times greater than that of the siphon, including with the latter the lb. of steam itself. In this case, however, the steam was used without expansion, and with such high back-pressure that only 43 per cent. of the total pressure could be utilized. But if a very small condensing steam cylinder were used, the total IHP. would be obtained for about 40 lbs. of steam per hour. The pressure being 38 lbs. per square inch, 86 per cent. would be utilized, and the duty in water raised per lb. of steam would be in the ratio of 4.8 to that of the siphon. Where the economy of fuel is of importance, it is evident that any competition of the siphon-pump with the reciprocating-pump is hopeless. The former can only be used advantageously on a steamer to pump out its bilge, after the vessel has come to anchor, with steam in the boilers, or when there is no further use for the main engines. It is very cheap, cannot become deranged, requires

no attention, and can be placed wherever there is room for a pipe of a few inches diameter.

If the duty of the reciprocating and rotary pumps be compared, the marked inferiority of the latter will at once be seen. The pressures with the two pumps were not the same, the mean total pressure with the rotary being less than that with the reciprocatingpump, while the back- and friction-pressures were alike in both pumps. Hence less pressure was utilized with the former than with the latter, and the duty of the reciprocating-pump was therefore 1.14 greater than that of the rotary pump. Some allowance must, however, be made for the friction caused by the strap from the engine, by which the rotary-pump was worked, as also the resistance of the valves in the reciprocating-pump. If all accessory disturbing influences are eliminated, the duty of the two pumps will be found to approximate more closely.

The permanent cause of inferiority in the rotary-pump lies chiefly in the greater water-leakage past its pistons. Nor can this pump, with its uniformly revolving pistons, fill with water the. same proportion of space displacement as the reciprocating-pump with its piston coming to a state of rest at the end of its stroke. Hence, while 0.86 of water was displaced by its piston, and discharged by the reciprocating-pump, the rotary pump discharged only 0.57 of the water displaced by its piston.

B. D.

Notes on New Types of Hydraulic Lifts: Samain's System. By G. CERBELAnd.

(Portefeuille Économique des Machines, 1889, p. 34.)

The types of hydraulic lifts in general use in high buildings are often inconvenient, from requiring deep wells to be sunk for the reception of the long cylinders, and they also require complicated and expensive foundations. When of the direct type requiring a cylinder of the same length, or, indeed, rather longer than the height of the lift, their application is limited, and also the consumption of water under pressure is not always proportionate to the work done.

To avoid these defects many improvements have been introduced, amongst which are those of Mr. Samain. With regard to the usual direct type, composed essentially of a long cylinder fitted with a ram, to the upper end of which the cabin is attached, the ram moving vertically in the cylinder, the latter being fixed in a well of suitable depth, Mr. Samain makes the ram hollow; to economise weight, the top of the hollow ram is sealed, but the lower end is left open. The water under pressure passes through a distributor, the use of which is to admit water to the cylinder, to raise the cabin, or stop the flow, and to hold the ram at any height, or lastly, to let the water flow out of the cylinder and let the ram down. Thus

constructed the lift is a very simple machine, but it is always inconvenient, one difficulty being to balance the dead weight of the ram and cabin without loss of water. The solution of this difficult problem is sometimes effected by a counterpoise nearly equal to the dead weight of cabin and ram. But the problem is complex, and it is necessary to remark that, supposing the pressure of the supply-water to be constant, the pressure exerted on the inferior surface of the ram will vary considerably according to the position of the ram itself; at the lowest position (taken at the ground-level), it will be augmented by the depth of the well, and diminished in proportion to the rise of the ram; or otherwise the ram behaves as a float more or less constrained to displace a variable quantity of liquid, and is submitted to a corresponding variation of pressure. Mr. Samain uses a hollow ram open at the lower end, and as the water fills the whole interior of the ram up to the top, the cabin may be said to rest on the column of water, it is obvious that as the ram rises, the effective pressure decreases by an amount corresponding to the rise of the ram. In order to equalize this pressure without increasing the size of the ram, Mr. Samain has devised a regulator or compensator consisting of a cylinder fitted with a piston, from which latter a band composed of metallic leaves passes through a stuffing-box in the cylinder cover, and thence over a large pulley; a counterbalance is attached to the other end of the metallic band. The capacity of the cylinder is slightly in excess of the total displacement of the ram in the lift-cylinder, but is not so long. The pulley is eccentric, and the axle turns in carriages running upon rollers; the eccentric-pulley is constrained to move concentrically with its circumference by means of two rollers. The supply-water under pressure is admitted from the distributor to the under side of the piston, and the top of the cylinder is connected to the ram-cylinder of the lift. The action is as follows: the water under pressure being admitted below the piston, the latter rises and displaces the water above it, which in turn raises the ram of the lift; the eccentric-pulley makes one-half revolution for the total lift, and the axle being eccentric with the rim of the pulley, as the latter turns, the leverage on the counterbalance side increases, and on the piston side decreases, thus a very perfect balance is effected; as the lift descends the opposite action takes place. The consumption of water is reduced to a minimum; it is equal to the volume of the compensator cylinder, and in consequence of the open hollow ram, and the fact that the cabin is supported almost directly on the water, all the moving parts can be made lighter than usual. The description is illustrated by a plate with several figures showing the compensation. The figures also show in detail. the ram of a telescopic hydraulic lift, also of Mr. Samain's design. H. H. P. P.

[THE INST. C.E. VOL. XCIX.]

2 I

Experiments with a Compound Winding-Engine at the Skalley, Shaft I Colliery, Saarbrücken.

(Zeitschrift für das Berg-, Hütten- und Salinen-Wesen, 1889, p. 191, one plate.)

The compound engine under notice was originally a single horizontal winding-engine, with a 43.3-inch cylinder, by 5 feet 1.8 inch stroke, winding two full tubs, each weighing 10 cwts., from a depth of 1,115 feet, with a steam-pressure of 44.12 lbs. per square inch. In 1886 in Dingler's engine-works in Zweibrücken the engine was altered to a compound one, to suit an increased steampressure of 95 lbs. per square inch, by adding a second cylinder 31 inches diameter. Until the new boilers were in working order, the engine had to be so made to act as a double-cylinder engine, with a pressure of 3 to 4 atmospheres, but it was found unnecessary to use the engine except as a compound one. The engine winds four tubs at a time in two decks, each tub weighing when full 10 cwts. nearly, from a depth of 372 yards, assisted by a counterbalance rope under the cage, weighing 8.06 lbs. per yard. In order to show the advantage of the compound system here introduced, the Author gives, together with the experiments of this engine, experiments with two other winding-engines (see next page).

Engines No. II and III were counterbalanced with rope under cage. The consumption of fuel per HP. per hour is 40 per cent. less by the compound than by the double winding-engine at Skalley shaft, No. II. Expansion is obtained in the compoundengine without any special mechanism, and in a certain degree by the unequal cylinders and the reversing lever. The Author points out that with the compound-engine less heating surface is necessary, consequently there is a saving in boilers, stokers and repairs. The plate contains diagrams of the engine. The whole winding is done in forty seconds, the average speed per second being 27 89 feet, and the average piston speed 4.95 feet per second. The engine is fitted with a receiver, from which live steam can be supplied to the low-pressure cylinder if required; this also assists the engineman in lifting the reversing-lever. The engine works observably quietly. A table, showing experiments with seven winding-engines, of which four are compound, all made at the Dingler's engine-works, finishes the Paper.