NEW DEVELOPMENTS IN HIGH VACUUM APPARATUS. 189 



I'unip, operating at the same capacities. There is considerable misinformation, in various 

 circles with reference to the twin pump. There have been papers extensively circulated 

 abroad, comparing twin pumps unfavorably with other types of vacuum apparatus, but, 

 where definite observations were given, they are ridiculous, in comparison with the known 

 performances as readily tested out. For one thing, dry air was used in their shop test, the 

 air being drawn over water at a given temperature, and the assumption made that the air 

 l>ecame 100 per cent saturated at the temperature of the water, which is incorrect, as is per- 

 fectly apparent. Any pump must reduce in efficiency as the temperature of saturation nears 

 the temperature corresponding to the total pressure, and in the papers referred to a zero ef- 

 ficiency is given at the point where an actual efficiency of between 30 and 35 per cent is read- 

 ily maintained. 



Another important point is that the builder is seldom asked to specify the size or guar- 

 antee the performance of a twin pump. The size has been specified and definitely asked for 

 and is ordinarily much larger than necessary. Doubtless those of you who are interested have 

 noticed, in the various tests of the vessels of the U. S. Navy as reported through the Jour- 

 nal of the Society of Naval Engineers, that the speed of the twin pump' is seldom more than 

 one-tliird to one-half of its normal rated speed, and in many of those trials where the manu- 

 facturer's representative has desired to reduce the speed of the pump very materially, in 

 order to obtain a better economy of the pump, he has been instructed to keep the speed up 

 in order to obtain the amount of exhaust steam necessary to heat the feed to the proper 

 temperature. 



The efficiency of the twinplex pump is much greater than that of the twin pump, and 

 it covers the requirements of the separate wet and dry vacuum system. 



The steam ejector has not been adopted in the high grade stationary plants for various 

 reasons, one of the principal beifig the fact that perhaps the gi-eatest economical develop- 

 ment of turbine-condensing plants in recent years has been due to the adoption of the wet 

 and dry system alluded to by the author. In that system the non-condensible vapors are 

 handled by a separate pump and are thrown away while the feed is pumped into a closed 

 heater (or open heater with cover) and recirculated without this air; consequently, in a closed 

 system, a very small amount of air is present. Tests reported some years since (American 

 Society of Mechanical Engineers, Transactions 1912, page 737) upon a plant at the Boston 

 Elevated Station showed that a dry vacuum pump, which would normally operate at from 

 100 to 135 revolutions, showed no drop' in vacuum if operated at around 45 to 50, which 

 was as slow as this pump would operate without centering. The speed was reduced by hand 

 control to 8 or 9 revolutions per minute without drop in vacuum, and the pump was finally 

 closed down entirely; in one-half hour's operation of the turbine with the vacuum pump 

 stopped, the drop in vacuum was only 0.3 inch. 



The use oi the steam ejector can only be countenanced where the heat in the steam can 

 be reclaimed for heating the feed, but in order to do this the steam must be discharged 

 through the feed water, which is thus fully recharged with air. I have alluded earlier to 

 the effect of air in reducing the efficiency of the condenser. 



There is another most important feature already referred tO' in connection with Plate 

 69. When the non-condensible vapors leave the condenser to enter the steam ejector, they 

 are seldom cooled more than 10 — perhaps 15 — degrees below the temperature correspond- 

 ing to the vacuum. But where an hydraulic vacuum pump is used (which uses hurling water 

 of the temperature corresponding to the circulating- water), the vapors are reduced tO' or 



