The following text may have been generated by Optical Character Recognition, with varying degrees of accuracy. Reader beware!

~~ that is: the very low efficiency which he has found for fans, an efficiency which he states for the conditions of ship ventilation come to about 50 percent, in some cases _ the efficiency rising to 60 percent or more but usually for steady flow; higher pressures than allowed. This is a very different result from what is usually accepted among many who have made fan tests, and the question is Of great interest where the difference comes in. Mr. Taylor relies altoge- ther on the Pitot tube, which he says on page 2 that he is convinced its accuracy is well established by independent investigations. That is a point on which there is a very great deal of difference of 'opinion. Mr. Taylor mentions an experiment on page 4 wherein, with an axial flow, he gradually drew back the Pitot tube from the mouth of the pipe delivering air, and found that the Pitot tube also exerted a pressure higher than the atmosphere, He says: "If the Pitot tube had been an error, as used to be sup- posed, the pressure recorded in the center of the free jet would have been materially below that of the atmsophere, while as a matter - fact it never quite reached the at- mospheric. pressure." : Mr, Taylor: pressure. Mr. Macalpine: I don't think that that experiment proves very much for us: On pagé 27 we have, in the appendix, the well known equation given for steady flow of air, but in developing the theory of the Pitot tube theré are cer- tain assumptions made. They are not entirely independent. They may be stated as three: first, that there should be secondly, that there should be axial flow: thirdly, that. the air in the central stream should be brought to rest. Mr. Taylor's experiments go up to 22 pounds per square foot, which, with a squaré delivery, would extend to 140 feet per second. With such velocity That is the side préssure, not the impact even approximately steady flow is an impossibility, arid in fact there is a proof of-that in the sho of the paper 'it+ self. Mr. Taylor: That 22 pounds per square foot is) not all velocity. 2 Mr. Macalpine: No, but it runs up to that pressure. Mr. Taylor: To the pressure, but not velocity, Mr, Macalpine: Well, there was a velocity of 100 to 200 feet per second spoken of in the paper. Now, as to the necessary assumption that the center of the stream line-is brought to rest. Of course that would be. so where the stream line is perfectly axial. If it was one deg-ee from being axial it would not be true. If it were 5 degrees it would be very far indeed from being true. In fact; on page 4 Mr. Taylor says that in his experi- ments it was thought necessary to use ten or more. tubes in? large pipes. in order to reduce 'to a negligible qauntity the. error due to the velocity of the air. That ts a well- known fact; and if he had given us the pressures at the different parts of thé pipe, we would have found also that the. pressures were very far from uniform, I méan a con- siderable percentage of the trouble. If we refer to a test of thesé p-essures with Mr. Taylor's mode of experiment- ing it might appear that it was effected by the instrument, but if we refer to a test by Commanders Carnegie, Dinger, and Johnson, at the Society of American Engineers, in February 1905, they also used as one of their modes of measuring, Pitot tubes. The nozzle experimented with was 24 inches diameter, in which twelve measurements were taken in positions shown. Taking one of the measure- ments, while one of the tubes: showed 7% 'inches pressure, another tube showed 6144, or 16 per cent over the lower. Now I,am not taking up one of the experiments alone,. nor. were these pressures, these large gradients, transverse gradients, at the 24 inches apart. In two of the experi- 29 "THE MARINE REVIEW - deal of vortex motion. by Messrs. Carriegie, Dinger and Johnson. ments we have the Pitot tubes within about six inches of one another, registering 74 inch difference of pressure. This is not a perfectly constant feature, showing that it is not an experimental error that has been introduced; but is so far constant as to show that it is a real fact that 'high and low pressures occurred at these points near one an- other. Now, if we have such heavy pressure gradients transverse to the tube, it naturally means that the stream lines are very much curved, that is that there is-a great The Pitot tube reer no note, no proper note, of this vortex motion. Before passing on, I might refer to another | iresalt got They have a box; they deliver into a box which is only a smaller en- largement of the pipe, and they meastire the pressure by putting in three tubes--three I think--transverse to the stream. Now it is well known that a tube transverse to the stream will give a very defective measurement of pressure; as a matter of fact, taking the averages, while -the Pitot tubes gave--the pressure in the box of course should have rather higher value than that at the nozzle, since the velocity ® a good déal lower, but instead of hav- ing a higher value where the Pitot tubes gave an average of 6.64 in the experiment made, the pressure in the box was 4.6. That is to say, the Pitot tubes registered 44 per- cent more pressure in the tube with air blowing across dt,. It.is not, I think, open to doubt, that a similar error of lesser degrée will exist if the pressure is absolutely axial, which-cannot be realized. Now, as further evidence regarding the umsteadiness, the: defective ieasurements due to Pitot tubes, I might say that on July 16, 1904, I saw two fans tested. They delivered into a box cut of which the air flowed through an 18-inch nozzle. The press- ure here was pretty high, 21 inches of water. The Pitot tube was very small, but the pressure measured at differ- ent parts of the opening of the nozzle was very different indeed, and at one part, over one area, it was exceedingly unsteady, the water jumping eight or ten inches or more continuously. Here we had what, if it had been a' steady flow, would have been an axial flow, and the tube so small that it would have no sensible effect on the current. Now the question is; is there another way of testing? A very frequent way of testing--there are two others to which I referred--the first, a very frequent way, is by delivering the air into a very large box so that the air currents, in- stead of being exceedingly swift, are very much reduced, not by any means seriously. There is still kinetic energy, which will register itself in the pressure gauge, but they are very much reduced, so that the pressure gauge has not a very powerful current blowing across it. A large kinetic energy is transferred into potential energy. . Referring again to that test on July 16, both methods were used. By the Pitot tubes, taking an average as near as could be, from the excessively unsteady measurements, _ there was a calculation of 55 percent efficiency. "The ob- servations from the gauge in the box gave 69 percent, a difference of 14 percent. Now, as I have pointed out, there were still powerful currents in the box, and con- sequently axial efficiency would. usually be. considerably sensibly higher than that often shown by the box. In most experiments the results which I have seen--and I have seen quite a number--whether there is a difference be- tween the Pitot tube results and the pressure gatige re- sults, I think in all cases the Pitot tube result was the low- er. Now there have been a great many fan tests made, in which-the efficiency has-been believed to be consider- ably over 60 percent in efficiency. I have seen fan results where, even with Pitot tubes, it, was calculated from the observations that the efficiency was as high as 70 percent. That at Trunham which I have referred to was 69% per-