Cup Anemometer 1 1 



It remains to explain this pure scale effect on h which was found previously by Marvin (cf. 



Page 3). 



The Reynolds Number introduced in the discussion of k cannot enter here because evidently 

 V a cannot have any effect as V becomes very great and h represents V/v as V approaches in- 

 finity. This is substantiated by Harrington's tests (Table II) where it is seen that h remains 

 fairly constant over a great variation of V a (P. 9). 



Neither is it plausible that any turbulence inherent in the wind tunnel flow could account 

 for a scale effect such as this; for at the higher rotational velocities the anemometer creates its 

 own turbulent whirl which would dominate over any initial turbulent character of the airstream. 

 The results of tests made at the Bureau of Standards on Marvin s beaded cup anemometer reveal 

 a very small percentual change of h for turbulence extremes of .7% and 4.6%. Table III. 



The influence of the wind tunnel size which might have introduced such a scale effect has 

 been shown to be negligibly small for the various sizes of wind tunnels which have been used for 

 anemometer tests. 



It is just conceivable that a Reynolds Number incorporating the velocity of sound might 

 account for this scale effect but this would mean that very high velocities would have to be at- 

 tained within the whirl set up by the anemometer. Radial exchange of momentum within this 

 whirl would bring about the condition for the absence of vorticity v r r = const, where v r = velocity of 

 whirl at any radius r, and thus v T would increase inwards until it reached a limiting value. This 

 explanation, however, seems rather inacceptable and it is hoped that experiments in a variable 

 density wind tunnel may reveal the nature of this scale effect. 



THE NUMBER OF CUPS 



The discussion so far has been confined to the four cup anemometer but a certain number of 

 experiments were carried out on anemometers having 3 and 2 cups in order to find out the varia- 

 tion of the various factors with the number of cups used. 



Examination of the test results indicates that the factor h is substantially unaffected by the 

 number of cups, all other dimensions being kept constant, k appears to be less for the 3 cup than 

 for the 4 cup anemometer. Patterson 1 has noted this in his remark that the factor for the 3 cup 

 anemometer appeared to approach the asymptote more quickly than that for the corresponding 

 anemometer having 4 cups. 



The variation is, however, not sufficiently significant to enter into the discussion of the rela- 

 tive merits of 3 or 4 cups. In this connection the point brought out by Fergusson 8 might be men- 

 tioned. Fergusson has repeatedly stressed the importance of sensitivity of the cup anemometer 

 although this question has not been considered by other writers; if, as he states, the 3 cup is more 

 sensitive it should be used in preference to the 4 cup. 



