188 NEW DEVELOPMENTS IN HIGH VACUUM APPARATUS. 



that the results obtained, both in economy and elasticity, have been in some respects quite 

 unexpected and exceed those reported for other types of ejectors. 



Plate 69, as referred to^ on page 176, shows the well-known curve illustrating Dalton's 

 law, but these curves are very misleading in their application to the steam ejector. The 

 curves give the weight of air and steam present in a mixture, but say nothing with regard 

 to tlie specific gravity and volume, which are the important factors in the capacity of the 

 steam ejector. The total weight, as obtained from these curves, will give the actual work 

 performed and will show, in the case of any steam ejector, a ridiculously low mechanical 

 efficiency, but it does not measure the capacity or steam consumption, of the ejector, which 

 is absolutely a volumetric machine. For example, at 90 degrees and 2 inches, the curve 

 shows 1.57 pounds of water per pound of air, or a total of 3.57 pounds; at 80 degrees sat- 

 uration, 0.67 pound of water per pound of air, or a total of 1.67 pounds; at 70 degrees 37 

 pounds of water per pound of air, or a total of 1.37 pounds; so that at 90 degrees the total 

 weight per pound of air is 54 per cent greater than at 80 degrees, and 87^ per cent greater 

 than at 70 degrees, but the ratio of the volumes is entirely different. Those depend upon the 

 partial air pressure; at 90 degrees the volume is 61)4 per cent greater than at 80 degrees, 

 and 116 per cent greater than at 70 degrees, so that, in handling questions of capacity and 

 steam consumption, of the ejector, it is the relative volume and not weight that must be con- 

 sidered. The paper in an indirect way refers to this at the bottom of page 180, where the 

 statement is made that careful experiments have shown the amount of air handled actually 

 less than the calculated theoretical volume based upon weights alone. 



I will not offer any mathematical discussion of the ejector or the varying conditions in 

 the different stages, but would call attention to the likelihood of confusion in interpreting 

 Plate 71. This would forni a very unfair basis of comparison with any other type of pump: 

 for instance, an hydraulic vacuum pimip to rotative dry vacuum pump. Also, while a dry 

 air orifice test in, the shop* is practically the only means of obtaining the capacity of the steam 

 ejector, it must be considerably corrected before it can be applied to actual service. If in 

 service we removed the mixed vapors from a condenser at 2 inches absolute pressure (cor- 

 responding to 101.25 degrees), and the vapors are cooled in leaving the condenser, say, ap- 

 proximately 10 degrees or 91.83 degrees, this will mean, according to Dalton's law, that the 

 steam pressure will be 1}4 inches, but the partial air pressure only ^ inch, and the partial 

 air pressure determines the volume of the vapors to be handled by the ejector. In the shop 

 test, with dry air only at 2 inches total pressure, the ejectors will then evidently handle four 

 times the air that it will in service. In other words, we should divide the quantity of air 

 obtained in the shop' by four to obtain the amount of air that will be handled in actual ser- 

 vice as it comes saturated from the condenser. The actual amount of cooling of non-con- 

 densible vapors will vary with the quantities of steam, air and water as well as the temper- 

 ature of the water. 



Worthington has found a decided advantage in, having a slight degree of superheat in 

 the steam operating the ejector. 



It is unfortunate the author did not give actual figures or calculations in connection 

 with any of the statements made, so there is no opportunity for checking up. The state- 

 ment, however, is made that a steam ejector with a turbine-driven condensate pump will re- 

 quire less steam than a twin beam air pump or even a twinplex pump with the wet and dry 

 cylinder. This is absolutely incorrect, and from actual tests we find that the twinplex pump 

 will take only about 40 to 45 per cent of the steam required by the ejector and condensate 



