TM No. 377 



The instability between 70 and 90 is probably caused by the gross drag— 

 the boundary effect around the cylinder that is normal to the flow or towing 

 direction. The horizontal projected area of" the cylinder is about 180 cm . 

 'This drag inducing area is, of course, absent in the single meter system. 

 As with the single meter, at 6 = 90° no output was sensed except for a slight 

 wobble occurring at speeds greater than l80 cm sec -1 . 



These calibration tests showed that the average error of measurement of 

 a sinusoidally oscillating vector, using the OMDUM III device in a two- 

 dimensional flow field (in the XZ plane), is of the order of 7-12 percent. 



Error Correction for OMDUM III. Since the ducted meter system produces 

 a bias in the measurement of current speeds at various angles with respect 

 to the meter axes, an attempt was made to devise a method of correction. 

 The approximate error in the velocity components was determined to be a 

 function of the angle 9. Now, if 9 is known, the value of R can be applied 

 as a correction coefficient to the velocity value obtained from the zero 

 angle calibration curve. This is, of course, predicated on the assumption 

 that, for a particular ratio of the outputs of the u and w meters, there is 

 a unique value of 9°, i.e., that the relative magnitudes of the two orthogonal 

 velocity components determine a unique 9. 



For example, examine the relationship of tne ratio of the detected hori- 

 zontal velocity zo the detected vertical velocity as a function of the angle 

 9 (defined as the angle made by the axis of the w meter with respect to the 

 flow). The relationship is shown in figure II-2U (dashed curve). The 

 numbers and the vertical bars adjacent to the points refer to the number 

 of data points and to the two sigma spread, respectively. These data 

 points are the result of tests at steady tow speeds and at various values 

 of Pw. The curve is very similar to the tangent 9 curve (dotted line). 

 The greatest deviation occurs beyond 80°, which is, of course, where the 

 greatest error occurs in the determination of the w component. (This 

 corresponds to the unstable region of the error curve shown in figure 11-22). 

 The similarity of the ratio curve to the tangent curve is easily explained 

 if the two speed components are considered as approximately following the 

 cosine lav. In this case: 



w = |V| cos 9 , (11=20) 



and u = | V| cos (90=9) = |V| sin 9. ( 11-21 ) 



Thus, u/w a tan 9. ( 11-22) 



The question arises about the possible dependence of the ratio u/w upon 

 the flow speed; also about the possible effects of accelerative flow on 

 this relation. Figure 11-25 shows a plot of the ratio u/w as a function 

 of the angle 9 and of the tow speed. The abcissa is the tow speed in 

 cm sec" 1 , and the ordinate is the numerical value of the ratio u/w plotted 

 on a logarithmic scale for convenience. The crosses along the ordinate 

 axis indicate the ratios determined under steady towing conditions. These 



38 



