Semi-automatic High-Pressure Installation. 161 



The roller R is jonrnallcd in one end of the lever L, -which 

 is so loaded that R always rests in light contact with the 

 undermost of the pile of weights. When the weights fall 

 the lever L is tilted clockwise and a few cubic centimetres of 

 mercury contained in the glass tube T attached to L run 

 down to the right-hand end of the tube and make contact 

 between two iron electrodes E. The tube T is exhausted 

 to a Fleuss vacuum and hermetically sealed, the electrodes 

 being closely fitted through ground conical tubulures, and 

 made tight with sealing-wax. 



It may bo worth while noting that this type of vacuum 

 •contact-maker is capable of carrying currents up to some 

 5 to 7 amperes at 100 volts D.C. without sustaining injury. 



Sensitiveness of the gauges. — It is well known that in this 

 form of gauge the sensitiveness depends largely on the rate 

 at which the pistons are rotated. (Several experiments were 

 made to find out the most suitable rate, i. e. the slowest rate 

 compatible with sufficient sensitiveness. The outcome of 

 these being that G 3 should revolve once in 6 seconds, G 2 once 

 in 26 seconds, and G x once in 45 seconds. At these rates it 

 was found that with the pistons of both G 2 and G 3 floating 

 under a pressure of 800 atmospheres, but with the latter 

 slowly falling while the former is slowly rising, an addition 

 of 20 grams to G 2 (equivalent to a change of pressure of 

 0*17 atmosphere) the movements would reverse, and G 2 

 would fall while G 3 rose. 



Pressure release-valve. — Whilst working with this in- 

 stallation it was found an advantage to be able to release 

 the pressure somewhat rapidly, but at the same time to do 

 so without jerks, An ordinary needle-valve is not suitable 

 because the small opening soon gets clogged by any solid 

 particles there may be in the oil (or even any material of 

 greater viscosity ?), and on enlarging the opening the 

 obstruction flows away, and a rapid variation ensues. 



The following device successfully avoided the trouble, 

 especially at the higher pressures, where it is most needed. 



A cylinder, closed at one end by a needle-valve, was bored 

 out to g- inch and tapped 26 threads to the inch. A plug- 

 carrying the same thread was loosely fitted into the cylinder 

 and the whole joined on to the pressure-system. On opening 

 the needle- valve, the plug travels forward until there is a 

 firm contact between the adjacent thread surfaces, thus 

 forming a narrow channel through which the oil can flow. 

 The length of the channel (some 50 inches) acts as a resistance 

 to the flow, but its depth is such as to avoid clogging. 

 Phil. Mag. S. 6. Vol. 32. No. 187. July 1916. M 



