VI. PERFORMANCE 



The tunnel has operated over wide ranges of amplitude and period in a 

 special series of tests and in the completed permeability experiments 

 (Lofquist, 1975). The special tests investigated the constancy of the 

 rate of rotation of the driving arms throughout a cycle and the leakage 

 past the piston seals. In these tests, made before the tunnel was modi- 

 fied, the depth of flow in the test section was only 15.6 centimeters and 

 the amplitude of the horizontal velocity was given by 



„ 1505 /S \ . 



rather than by equation (17) . These tests extended to values of U able 

 to sweep away sand and were performed with wooden boards in place of a 

 sand bed surface. Also, no screens were mounted in the spools. 



The tests for constancy of speed were made with the electronic timing 

 apparatus in the ranges: S/Sjjj between 0.095 and 0.64, T between 3.0 and 

 25 seconds, and U between 30 and 150 centimeters per second. Generally, 

 variations in speed from average values increased with U, but all varia- 

 tions were small. For U less than 50 centimeters per second, the varia- 

 tions remained less than ± 0.25 percent; for U less than 100 centimeters 

 per second, less than ± 0.5 percent. The largest observed variation, at 

 U«150 centimeters per second, was around ± 0.8 percent. The smallness 

 of these variations shows that the steepness of the motor torque-speed 

 curve, the inertia of the system, and the active counterweight combine to 

 offset the effects of friction in the flow and between cylinders and pis- 

 tons. For given S/Sjji and T, variations in speed were observed to increase 

 as the active counterweight was moved from its proper position given by 

 equation (16) . 



Ideally, a constant rate of rotation of the driving arms would ensure 

 pure simple harmonic motion of the pistons. However, the elasticity of 

 the scotch yoke and the clearance between the ball-bearing pins and the 

 scotch yoke slots permit the friction between the piston seals and the 

 cylinder walls to hold the pistons at rest for a finite time interval, 

 6t, at their extreme upper and lower positions. This interval persists 

 while the pins move through some small vertical distance and the driving 

 arms turn through some small angle equal to 2tt 6t/T, so that 6t/T in- 

 creases as S/S]j, is reduced. At small strokes this effect can be detected 

 as a slight jerk in the flow just following flow reversal. 



An average rate of leakage past the piston seals was obtained by 

 measuring the changes in depth of shallow layers of water over the pis- 

 tons after an operation of around 100 cycles. No attempt was made to 

 determine the instantaneous rate of leakage as function of phase. The 

 changes in depth were hard to detect, and did not exceed a few tenths of 

 a millimeter. This represents a volume leakage per half cycle of not 



24 



