Nowackt and Sharma 
was mounted at the appropriate propeller clearance (45 mm from 
wheel center to the vertical stem profile). Every measurement was 
repeated at least once. 
It has been noted elsewhere that the measured wake in both 
forward and reverse motion showed somewhat erratic undulations at 
the outer radii (see Fig. 6 and 7). This could possibly be blamed on 
the method of measurement. We had no way of establishing just how 
accurately the uniform flow calibrations could be relied upon for de- 
termining the circumferential averages of a varying axial velocity in 
a complex nonuniform flow involving significant circumferential and 
radial components. 
A.5. Wave Measurement 
A stationary wave probe was mounted at a point about midway 
along the length of the towing tank and at a fixed transverse distance 
Yo = 605 mm from the center plane of the model (which coincided 
nearly with the center plane of the tank itself), Hence, a time record 
of local wave height at the probe, while the model passed by, was ob- 
viously equivalent to a longitudinal cut z = {(x,y,) through the steady 
wave pattern of the model ina coordinate system Oxyz moving with 
the model. 
A thin light beam was set up across the tank at a known fixed 
distance (x; = 336 mm) upstream of the probe. During the run a shutter 
affixed to the model at a known fixed distance (x, = 2933 mm) forward 
of the midship section interrupted the light beam and generated an 
event signal marking the point x = x,-x, on the wave record, thus 
defining the coordinate origin. 
The wave probe itself was of the conductance wire type adapt- 
ed from the HSVA design of Luft (1968) to match the available Sanborn 
carrier frequency preamplifiers. The circuit output was fed into one 
channel of a Sanborn strip chart recorder. The overall sensitivity was 
set at 6 to 9 mm deflection per inch of wave height so as to produce 
full scale deflection at the measured wave peaks. Wave height records 
were manually read off at about 350 points at equal time intervals 
At = 0.03 sec, thatis ata step size Ax = -VAt, and key-punched on 
IBM cards. All further analysis was done by computer programs, 
It should be noted that at the highest Froude number investi- 
gated the length of useful record (taken before running into tank wall 
reflection) was not really adequate to establish with confidence the 
asymptotic character of the wave profile behind the model which is 
needed for the application of a truncation correction (see Fig. 16). 
However, this was due to a purely geometrical constraint resulting 
1886 
