64 Prof. J. Perry on a Steam-Engine 



represents pressure and prevent error due to change of tem- 

 perature of the spring. 



In these days of high-speed machinery the most important 

 defect of an indicator is its slow natural period of vibration. 

 If the natural period of vibration is from ^ to ^ of the 

 time of revolution of an engine, it is found that the diagram 

 is not deformed by waviness, ordinary fluid friction destroy- 

 ing the natural vibration ; if it is as much as y 1 ^, a con- 

 siderable amount of pencil pressure must be employed to 

 obtain a frictional stilling of vibration ; but if it is as low as 

 ^o it is almost impossible, even with great pencil friction, 

 to obtain a decent diagram. To make an indicator have a 

 quicker natural vibration it is necessary to use a stiffer spring, 

 and this means that pressures are indicated to a very small 

 scale. Introducing seven per cent, of inaccuracy in area of 

 diagram, either by friction or smallness of scale, may possibly 

 allow the very best existing indicator to be used on engines 

 of as high a speed as 400 revolutions per minute ; but this is 

 a matter on which assertions are made of higher and very 

 much lower limits of speed than what I have stated. I think 

 that users of indicators will generally agree that I have not 

 exaggerated this defect ; but unfortunately there is scarcely 

 ever any kind of check which can be applied to a measure- 

 ment, of indicated horse-power, so that mere assertions are of 

 very little value. 



Indicated horse-power is now the sole standard of the 

 values of engines, so that its accurate measurement is very 

 important; but for the inventor and improver of engines, 

 and for the physicist, that there shall be no local deformations 

 in the shape of the indicator-diagram is of greater importance 

 than that its area should be correct. Now the natural period 

 of vibration of the indicator before you is about -gfa of a 

 second; allowing twenty periods to one revolution of the 

 engine, I find that this indicator will give diagrams with no 

 wave-deformations until the speed of the engine exceeds 1500 

 revolutions per minute. I can, however, by changing the 

 disk make the natural period of vibration y^oo or 2~so °f 

 a second, or, indeed, what I please, so that there is practically 

 no speed at which this indicator will not give a true diagram. 



Specimens of the indicator are before you, but I will throw 

 upon the screen drawings of it. Fig. 1 is a section, and 

 fig. 2 an elevation. It consists of a very shallow circular 

 box, E, of metal (in some of my specimens it is of cast iron, 

 in others of gun-metal) closed by a thin disk of steel, D. 

 Here, for example, is one an inch and a quarter in dia- 

 meter, and about -^ of an inch thick, which I use for 



