[mellanuy] efficiencies OF MULTIPLE-EXPANSION ENGINES 133 



and a blank flange put in its place. A small pipe with a valve led from 

 this flange, the end was perforated and passed into a vessel containing 

 cold water. The steam piping was arranged so that steam could be let 

 directly from the boiler into one of the other cylinders. This cylinder 

 was used to drive the valves of the cylinder upon which the leak test was 

 being made. Whilst the valves were moving, steam was admitted into 

 the steam chest at the required pressure. Any steam that leaked past the 

 valves into the exhaust would pass down the small pipe and be condensed 

 in cold water. The amount could be ascertained by weighing the vessel 

 before and after the experiment. Owing to lack of time, tests were made 

 only upon cylinders 1 and 4. In cylinder No. 1 steam was admitted into 

 the steam chest at a pressure of 100 lbs. (above atmosphere). Two half- 

 hour experiments were made, one at 90 revolutions and the other at 150 

 revolutions per minute. In both cases the leak past amounted to 19 lbs. 

 in the half hour, or 38 11)S. per hour. An experiment was commenced 

 with the pressure at 150 lbs., but unfortunately the lead plug blew in. 

 With cylinder No. 4 experiments were made at pressures 34 and 21 lbs. 

 above the atmosphere ; in these cases the leak amounted to 41 and 29 

 lbs. per hour respectively. The amount of leak past the other valves 

 had to be inferred from these experiments. Unless there is a special od- 

 ing arrangement for directly lubricating the face of the valve, such 

 experiments are dikely to give erroneous results owing to the rubbing 

 away of the oil film between the valve and steam chest faces. In this 

 case it will be necessary to start the experiment almost immediately the 

 engine commences to run, and to take measurements at short intervals of 

 time. A gradient can then be consti'ucted and it will be readily seen 

 * whether the rate of leak increases. It seems only fair that the slowest 

 rate should be taken. In these experiments oil was let directly to the 

 valve face of cyhnder No. 1. Eeadings were taken at intervals of 5 

 minutes and were found to be almost constant. Cylinder No. 4 had no 

 such oiling arrangement. Before the trial the face of the steam chest 

 was well oiled. Readings taken every 3 minutes were found to show 

 practically equal amounts of leak. In cylinder No. 4 steam was also let 

 in whilst the valves were stationary. It was allowed to remain on with 

 a pressure of 30 lbs. above atmosphere for 15 minutes during which time 

 there was not the slightest indication of any leak. This aflbrds satis- 

 factory evidence that the valve may be perfectly steam tight when the 

 engine is standing, but will allow steam to pass by it into the exhaust 

 when the engine is running. 



Method of Working out the Trials. 



Before any of the indicator cards were worked out it has been 

 the custom to put in a graphical form the readings given by the various 

 observers. The steam pipe pressures, revolutions per minute, condenser 



