MEASURING OF AIR. 453 



111 comparing these measurements, it is to be remembere'd 

 that the reading of the Barometer is the only one common to 

 both sides and that even this enters very ditlcrently in tlie two 

 calculations. 



It is to be remembered that the quantity of air taken in 

 should be greater than the quantity sent out l)y the leakage 

 from the glands of the compressor. Moreover there is a 

 slight uncertainty as to the effect of the butterfly valve on the 

 down stream side of the orifice. On the discharge side of 

 the compressor there is some evidence that the coefficient 0.675 

 used for the plate orifice should be increased by al:)Out Vi per 

 cent., had this been done the agreement would have been even 

 closer than it is now. From this test can be concluded : 



(i) that in a well designed turbo com])ressor the leakage 

 through the labyrinth glands is a negligible cjuantity being less 

 than the deviation of the measured quantities. Xo leakage was 

 observed at any of the glands or flanges. 



(2) that when the tests are properly conducted and ori- 

 fices used with the necessary precautions, the measurement at 

 low (normal atmospheric) pressure and of the same quantity 

 of air at high (deli\ery) pressure agree. 



(3) thati the order of agreement, i.e.. the certaintv with 

 which air can be measured, is as high as with electric measure- 

 ments. 



II. RECIPROCATING COMPRESSOR TEST. 



The test has been made on May nth, 1913, on No. 3 

 Pakorny and W'ittekind Compressor of the Brakpan Com- 

 pressor Station. The compressor delivers 7.500 cu. ft. of free 

 air per minute at a pressure of 90 lbs, gauge when running 

 with 75 revolutions per minute. 



The air discharged by the compressor was measured l>y a 

 Venturi tube, which had been arranged on the downstream 

 side of the large station air receiver. (See fig. No. 5. ) The capa- 

 ci.ty of this receiver is 1080 cu. ft. or 57 times the capacity of 

 the H.P. air cylinder. Its size was sufiicient to stop com- 

 pletely the fluctuations of pressure, due to the intermititent^ 

 discharge of the compressor, so that ]>erfectly steady readings 

 were obtained at the Venturi tube. 



A throttle valve was arranged between recei\er and 

 Venturi tube in order to regulate the pressure at the engine 

 terminus. After passing the Venturi tube the air is throttled 

 down to atmospheric pressure b}- four orifices arranged in 

 series. 



The test was made under ordinary working conditions. 

 No special preparations were made except that the intercooler 

 was boiled out three weeks before the test and slight readjust- 



