MICKOWAVE PARAMAGNETIC RESONANCE ABSORPTION 475 



Putting in the numerical values for the 1N23C as quoted previously 

 we get that B = lA X 10~^ /Prf . Equation (56) becomes with the opti- 

 mum bucking parameter 



Xmin-obs 



77 

 Xmin-th 



2F 



AMP 



(58) 



For the case under discussion this ratio turns out to be c^50 independent 

 of Po . Fig. 9 shows a plot of (58) . The values in parenthesis indicate the 

 degree of bucking necessary to accomplish this ratio as determined from 

 (57). It should be noted that the negative bucking can be easily accom- 

 plished by means of the slide screw tuner in the magic T arm (see Fig. 6) 

 whereas for large positive buckings a scheme like in Fig. 8 has to be 

 used. (Some positive bucking can of course be also accomplished by 

 means of the slide screw tuner) . 



c. The siiperheterodyne scheme 



The RF bucking system just described bears a certain resemblance 

 to the balanced mixer barretter scheme. In both cases additional micro- 

 wave power was added to the detector in order to increase the con- 

 version gain. However in the crystal scheme this resulted in an increase 

 in noise power whereas this should not be the case with barretters. The 

 question arises whether a decent conversion gain in crystals has to be 

 always accompanied by a large noise power. An inspection of (49) 

 shows that around frequencies of tens of megacycles* or higher the 

 noise output of the crystal becomes negligible. As pointed out earlier 

 such high magnetic field modulation frequencies are not feasible. How- 

 ever in a superheterodyne system the crystal outputs will be at an 

 intermediate frequency of 30 or 60 mc. This will make the flicker noise 

 components negligible even at high powers where the conversion gain 

 is good. The conventional way to obtain the intermediate frequency is 

 to beat the reflected signal from the cavity with a local oscillator (see 

 Fig. 10) which is removed from the signal generator by the I.F. fre- 

 quency. In order to eliminate the noise from the local oscillator a bal- 

 anced mixer should be employed. The ratio 



/xMiN^sN ^^^^^^^^^ ^hen from equ. 52 f^i^-ti-IliY (59) 

 Vxmin-th / \ ^ / 



The expression in the brackets is called in radar work'^ the overall 



* It was shown by G. R. NicolP that this equation holds up to this frequency 

 range. 



