10 INSTRUMENTATION IN SCIENTIFIC RESEARCH [Chap. 1 



flow pattern approximates that of three-dimensional flow in a thick 

 sheet, i.e., the potential difference rises rapidly when s/t becomes less 

 than unity. 



A set of "contact squares" is used. Each contact square contains in 

 an insulator four resiliently mounted contacts (phonograph needles), 

 as shown in Fig. (1-1)5. A set consists of squares with dimensions 

 decreasing by a factor of 2; i.e., s = 4, 2, 1, ^, %, | in. First, a large 

 contact square is pressed against the plate to be examined; the cur- 

 rent is adjusted to a convenient 

 value and maintained constant 

 thereafter. The potential differ- 

 ence is noted. Then the next 

 smaller contact square is used. If 

 the size of the large contact square 

 is considerably larger than the 

 plate thickness, the potential dif- 

 ference for the same current will 

 be only slightly higher with the 

 smaller contact square. However, 

 if the size of the larger contact 

 square approaches the plate thick- 

 ness, then the ratio of the poten- 



3.0 



2.8 



2.6 



^2.4 



S2.2 



2.0 



o 

 > 



1.4 

 1.2 



tial difference e x obtained with a 

 (smaller) square of the size s to e 2 , 

 the potential difference obtained 

 with a (larger) square of the size 

 2s increases rapidly. The ratio of 

 is noted, and the thickness of the 



1.2 1.3 1.4 1.5 1.6 1.7 1.8 

 Potentiol difference read with smoller contact square 

 Potential difference read with larger contact square 



Fig. (1-1)6. Function f(e l je 2 ) versus 

 ratio of measured potential differences 

 [from A. O. Warren, J. Inst. Elec. Engrs. 

 (London), 84,91 (1939); by permission of 

 The Institution of Electrical Engineers, 

 London}. 



these potential differences eje 2 

 plate is determined from 



t = 



where f{e 1 /e 2 ) is to be found from Fig. (1-1)6, and s is the side of the 

 smaller of two successive contact squares used. 



The error of the method is less than 3 per cent. The presence of 

 paint or rust does not influence the result. The method can be used 

 to measure the thickness of restricted plates and to detect corrosion 

 pits. The method does not require any knowledge of the electrical 

 conductivity of the plate material. 1 



A similar method has been described by B. M. Thornton and W. M. Thornton, 

 Proc. Inst. Mech. Engrs. (London), 140, 349 (1938), and B. M. Thornton, 

 Engineering, 159, 81 (1945). 



1 For underlying theory, see Warren, ibid. 



