122 INSTRUMENTATION IN SCIENTIFIC RESEARCH [Chap. 1 



sensitivity can be reduced by a factor of less than 0.5 or can even be 

 negative. Even at axially symmetric flow, the sensitivity is 7.4 per 

 cent lower than that computed from Eqs. (1) and (2). In narrow 

 rectangular channels the sensitivity is (according to Shercliff) inde- 

 pendent of the velocity profile. In such conduits the electrodes 

 should extend over the entire width of the channel. 



The output voltage is in general in the microvolt range. A par- 

 ticular sensitive and accurate instrument is described by James; 1 it 

 has an input range from zero to about 40 mm/min and furnishes, at 

 the transducer output, an a-c voltage (11.4 cps) of 0.5 / ttV/(ml)(min) 

 ■i 1 per cent for liquids of an electric conductivity of at least 

 3.5 x 10 -6 ohm _1 -cm -1 . The electric conductivity of the liquid 

 determines the output impedance which may vary from several 

 ohms, for mercury, up to several megohms. 



The output is proportional to the flow velocity; the deviation from 

 linearity can be kept small and corresponds, in the above-mentioned 

 instrument of W. G. James, to about 0.25 ml/min. The output 

 follows the variations of the flow velocity without inertia. Flow 

 transients of 1 msec duration have been measured by Arnold. 2 



Errors may arise from parasitic voltages between the electrodes. 

 Such voltages are independent of the flow and may stem from gal- 

 vanic potentials between the electrodes and other metallic parts in 

 the system 3 or from polarization of the electrodes at d-c operation. 

 Parasitic d-c voltages due to flow (with the magnetic field removed) 

 have also been observed. They can be eliminated by the use of 

 graphite electrodes. The use of nonpolarizable electrodes is generally 

 required at d-c operation. The parasitic a-c voltages mentioned 

 above can be canceled by an auxiliary compensating voltage of 

 appropriate magnitude and phase. The amount of random noise 

 originating in the transducer and the influence of stray electric and 

 magnetic fields increases, of course, with increased resistivity of 

 the liquid. 



The advantages of the method are reliability, simplicity, and 

 ruggedness. There are no moving parts or constrictions in the tube. 

 The response is fast, it is linear, and the output is independent of the 

 physical properties of the liquid, with the exception of the electrical 

 conductivity. Disadvantages of the method are the small output 

 voltage, which requires amplification, and in some cases the high 



1 W. G. James, Rev. Sci. Instr., 22, 989 (1951). 



2 F. S. Arnold, Rev. Sci. Instr., 22, 43 (1951). 



3 A. J. Morris and J. H. Chadwick, Trans. AIEE, Conf. Paper T-l-58, 346 

 (1951). 



