NEW DEVELOPMENTS IN HIGH VACUUM APPARATUS. 195 



Mr. Kotuny: — I cannot Ijut feel gratified to see that this paper has ehcited such an 

 interesting discussion. The subject is obviously one which interests marine engineers, and 

 one of the reasons for submitting this paper to this Society is that I have been personally 

 associated for the last ten years with the development previously mentioned. It is undoubt- 

 edly of furtiier interest, in view of the revolutionary nature of the development. 



When I prepared this paper I had in mind just to record this development, but not to 

 talk about the advantages of the different makes of air ejectors and the different types of 

 air pumps. 



Mr. Katzenstein, in his discussion, first criticised Plate 68, showing the amount of per- 

 missible air leakage. He quoted some figures from a British paper, which showed 41 pounds 

 of air leakage per hour for 50,000 pounds of steam. The curve illustrated on Plate 68 shows 

 only 20 pounds per hour air leakage for 50,000 pounds of steam. Immediately afterwards 

 he stated that, on a turbine consuming 350,000 pounds of steam per hour, the maximum pei"- 

 missible air leakage was only 16 to 33 pounds; in other words, considerably less than the 

 amount taken from the curve on Plate 68. 



From this it is not quite clear what Mr. Katzenstein wishes to convey — whether the 

 limits of air leaks given on Plate 68 are too small or too large. Mr. Smith correctly stated 

 that the air leakage is an arbitrary value, and you will agree that there should be certain 

 limits for this arbitrary value. The curves published in Plate 68 should not be taken as 

 absolute in every sense — they were published more with the intention of suggesting a stand- 

 ard, so that marine engineers interested along this line could get together and establish a 

 standard limit for this arbitrary value, as has been done in many other branches of engi- 

 neering. 



Mr. Smith also stated that on the ships on which they used the steam-air ejector the 

 vacua obtained Avere generally from 0.35 to 0.40 of an inch higher than those guaranteed. 

 This is interesting, and I am glad to hear it because the sizes of the air pumps selected for 

 these installations were based on the air leaks shown in Plate 68. This would also indicate 

 that the values given in the curve on Plate 68 are amply large, and not too small, as Mr. 

 Katzenstein stated. 



Mr. Katzenstein also criticised Plate 69, showing the relation between vapor and air. I 

 do not think I need comment on this, because this curve is based on physical laws, and I do 

 not pretend to be author of the curve — it is a curve which may be found in many text-books 

 and which have been previously published, and I have no doubt the curve is correct. 



Mr. Katzenstein called attention to the likelihood of confusion in interpreting Plate 71, 

 illustrating the characteristics of an air ejector, and said that it would fomi: a very unfair 

 basis for comparison with any other type of pump. However, these statements are at vari- 

 ance with his explanations which followed them. On page 177 of the paper it is clearly 

 stated that the curves illustrated on Plate 68 only give the amount of free air entering by 

 leaks, but not the air-handling capacity of the pump. The latter has to be calculated by using 

 Plates 68 and 69. Consideration should be also- given to the fact that the work of compres- 

 sion of a mixture consisting of air and vapors is somewhat different from that of dry air, 

 This, however, applies to any type of dry-air pump, whether air ejector or rotative dry 

 vacuumi pump. 



Mr. Katzenstein's statement that the twin-beam air pump will take only 40 to 45 per 

 cent of the steam required by the ejector is incorrect. The ejector plus condensate pump 

 takes less steam than the twinbeam or twinplex air pump, when handling the same amount 

 of air and condensate. Comparative tests based on the same air and condensate handling 



