30 ENGINE TESTS. 



greater simplicity and ease of operation, together with its free- 

 dom from the particular errors noted. 



In calibrating the springs for pressures below the atmosphere, 

 the dead-weight apparatus referred to is not applicable, and 

 resort has been had in these tests to comparison with a mercury 

 gauge, or with a standard vacuum gauge, the former being pre- 

 ferred. In making these comparisons in the shop or laboratory 

 it is necessary to obtain a vacuum by the use of some form of 

 pump or exhauster, and this often proves an inconvenience. 

 For this reason the author has been in the habit of making 

 them in the engine room where the indicators are being used, 

 and where a vacuum is obtained by connecting the testing- 

 apparatus with the condenser. All that is required for ap- 

 paratus is the connection to the condenser, a tee for the 

 attachment of the indicator-cock, and a mercury gauge applied 

 to one end of the tee. With this apparatus the spring can be 

 calibrated down to the lowest pressure to which it is subjected. 

 It is desirable to make the calibration of an indicator spring 

 that is used for pressures below the atmosphere under condi- 

 tions of vacuum as well as under conditions of pressure; for the 

 fact that the spring is correct when subjected to compression, as 

 it is when a pressure is applied to it, furnishes no positive 

 assurance that it is correct under tension, as it is when it is 

 subjected to a vacuum. 



It is of no little importance that the scale of the spring 

 should be known within reasonable limits of error; for upon 

 this knowledge depends the whole accuracy of the indicator 

 work, and consequently of all the results of the tests depending 

 upon it. 



MANNER OF WORKING UP THE TESTS. 



The results of the feed-water tests are computed from the 

 hourly consumption of feed- water corrected for the leakage of 

 the boilers, pipes, and connections, as explained, and the indi- 

 cated horse-power developed. The " steam accounted for by 

 the indicator" is determined from measurements of the dia- 

 grams and computations based thereon. 



