756 



THE BELL SYSTEM TECHNICAL JOURNAL, MAY 1953 



The actual procedure adopted in calibrating the coupling iris was this: 

 The secondary cavity was set to a fairly high and known value of Q^ , say 

 1500, and the coupling iris adjusted to the smallest width which still 

 resulted in overcoupling, i.e. in Q^k > 1, and thus displayed the de- 

 sired load hysteresis on the oscilloscope screen. By means of a high-Q 

 absorption wavemeter the crossover frequencies were determined and, 

 along with the known value of Q, substituted into equation (4.7) which 

 then yielded the value of A;. As a double check the above procedure was 

 repeated for several values of Qs and in accordance with theory the 

 coupling coefficient, k, found to be a function only of the iris coupling 

 width. A typical set of readings is given in Table II. They show that 

 though Qt was varied by a ratio of nearly 3:1, the resulting values of k 

 only differed by about one-half per cent. Measurements as shown in 

 the table were repeated for various iris widths. They resulted in the 

 curve of Fig. 19. 



To complete the calibration of the secondary cavity it was necessary 

 to extend the relation between Qs and micrometer readings to values 

 lower than could be obtained by the reflected power technique. This 

 was carried out with the aid of Fig. 19 in the following manner: 

 With the coupling iris adjusted to its maximum width of 0.5", the 

 secondary Q was varied until the condition of critical coupling, as ob- 

 served by the formation of a cusp on the oscilloscope pattern (see Fig. 

 16(c)), was reached. Since this condition is characterized by Qsk = 1 

 and since the value of k corresponding to the iris opening could be read 

 off the curve of Fig. 19, the value of Qa then followed simply as the 

 reciprocal of k. This measurement was repeated for successively smaller 

 iris openings and the points shown as triangles in Fig. 18 were obtained. 

 Inspection of this figure also reveals that these values of Qa , ranging 

 from 65 to 1100, overlap and coincide with values of Q, determined 



Table II 



