The following text may have been generated by Optical Character Recognition, with varying degrees of accuracy. Reader beware!

March, 1910 HIG. 3. ed and started on Sept. 7, 1909, ex- tensive preparations had been made to determine the efficiency of the apparatus on the assumption that there might .be a transmission loss of at least 5 per cent. After calculating the results of a few preliminary trials, it was found that the apparent efficiency was over 98 per cent, and this raised the perplexing question as to how to determine this unexpectedly small transmission loss with a satisfying degree of exactitude. If the efficiency were only from 90 to 95 per cent, its measurement, within reasonable limits of accuracy, would have 'been an easy matter. Rising as it un- questionably does, above 984 per cent, it will be readily understood that the problem requires a refinement of meth- od, and a degree of painstaking care far beyond that which is usual in or- dinary engineering investigations. For the reason that the efficiency is so remarkable, it has been thought ad- visable to describe the methods em- ployed and to tabulate the principal observations with a minuteness of de- tail that might, in many cases, be con- sidered superfluous, if not actually wearying. Naturally the first trials of the gear were made at very moderate speeds and loads, with frequent stops to- examine the condition of the teeth and 'bearings. During the first few days the speed was gradually raised to 1,500 R. P. M. of the turbine shaft, and 300 revolutions of the gear shaft, and a load of over 6,000 B. H. P»was carried: The load was limited only by the power of the turbine, as there were no_ indications that the ultimate safe capacity of the gear had from any standpoint been nearly approached. On Saturday, Oct. 16, 1909, an end- urance test at maximum load and speed was started at 3:15 p. m., and continued until 7:15 a. m. on the following Mon- day, or a total of 40 hours. This test was witnessed by Commander R. S. Griffin, U. S. N., and Lieut.-Commander U. T. Holmes, U. S. N., specially dele- gated for this purpose by the Bureau of THE Marine REVIEW 107 Steam Engineering of the United States instant may be determined with greater Navy Department. The load was applied by the hydraulic dynamometer, described on the preced- ing pages. The radius arin -of this dynamometer is 6% ft. and the pres- sure exerted is transmitted to the plat- form scale through a lever having a ratio of approximately 1 to 4, as shown in Figs. 1 and 2 in the detailed descrip- tion of this piece of apparatus. The precise distances between the _ knife edges on the lever being, however, 18 in. and 72 5-32 in. respectively, the ef- fective length of the radius arm of the dynamometer is 6.5 X 72.15625 18 The constant for 1 lb. pressure on the scale, and 1 R. P. M. is consequently 26.056 X 2 X 3.1416 33000 = 26.056 feet. = 0.0049612 horse- power. Bigs 4, During this trial the following average conditions were maintained: Gross, weight 'on 'scale, Ibs... 3.23 32. 4443.4 Dead weighton 'scale lbs. 2 es 388.0 Net. weight on scale, lbs; 0.00.0. 4055.4 Speed Re es Pa Me sera ceo es ee 300.6 BH Pee a ates ek Ssh espa ee elec 6048.0 With the dynamometer the measurement of the power output of the gear becomes - a very simple matter. But to determine at the same time the effective power input, is an engineering problem that is more than a little out of the ordinary. If. the gear were driven with a reci- procating engine, the natural procedure would be to measure the indicated horse- power, and either estimate or measure the friction of the engine, and thus ob- tain a correction which, when applied to the indicated horsepower, would give the effective power delivered to the gear. There being no way in which to meas- ure the indicated horsepower of a steam turbine, it became necessary to establish the exact brake horsepower in some other way. Happily, there is one char- acteristic of the steam turbine that makes it possible to calibrate any par- ticular machine, in such a way that its output in effective horsepower at any accuracy than it is possible to determine even the indicated horsepower of a reci- procating engine. As long as the speed and exhaust pressure are maintained constant, the absolute inlet pressure of commercially dry steam, at any instant, is a very accurate measure of the brake horsepower the turbine is developing. By substituting for the reduction gear, a dynamometer connected directly to the turbine shaft, and operating the turbine at a fixed speed and' with a constant vacuum in the exhaust pipe, we may determine the inlet pressures cor- responding to different loads at this speed. Plotting these observations on a dia- gram in which the ordinates represent absolute pressures per sq. in. and the abscissae represent brake horsepowers, and drawing a line through the several points, we get a scale by means of which ~the horsepower corresponding to any in- let pressure may easily be read. This line is practically straight, so that read- ings made between the points represent- ing actual observations, are quite de- -pendable. -- The diagram in Fig. 9 shows the cal- ibration of the turbine used on these tests at 1,500 R. P. M. with a vacuum of 22.78 in., corrected to a barometric pressure of 30 in. The radius arm of the high speed hy- draulic dynamometer used in calibrating the turbine is 48.375 in. The constant 'per pound pressure on the scale is therefore ; 48.375 X 2 X 3.1416 X 1500 = 1.1513 12 X 33000 TABLE I. 2 a » zi = bh 2 5 os z 6 oS 2 é i a » oY 28 ee 22 32 #5 aa OS Be ae esta 109.6 3,585 385 3,200 3,675 . 114.8 3,755 385 3,370 3,880 119.3 3,925 385 3,540 4,076 124.3 4,095 385 3,710 4,271 128.8 4,265 385 3,880 4,467 133.3 4,435 385 4,050 4,663 139.3 4,690 385 4,305 4,956 143.8 4,860 BRS) 4 as 5,152 147.3 5,030 385 4,645 5,348 151.8 5,200 385 4,815 377) 55548 157.8 5,455 385 5,070 S50aa Table I gives the observed absolute inlet pressures, loads on scale, and cal- culated brake horsepowers over a con- siderable range above and below 5,000 H. P. From these observations and cal- culations, a calibration curve similar to that shown in Fig. 9, was drawn to a large scale so that small differences of -- loads and pressure could be read ac- carately. In these tests, as well as in the tur- bine calibrations, at every load before reading the inlet pressure, the speed of the turbine was brought exactly to 1,500 R. P. M., as indicated by a vibration