476 THE BELL SYSTEM TECHXICAL JOURNAL, MARCH 1957 



noise figure of the receiver F. We found that a noise figure of about 



11-14 db is easily attainable with commercial I.F. amplifiers and 



balanced mixer. This would give us a ratio of 



// 

 Xmix-obs ^ - 



-77 — ^ 



Xmin-th 



which is plotted together with the other crystal schemes in Fig. 9. Al- 

 though this system does necessitate 2 stable microwave sources, it is not 

 difficult to operate once the}" are set-up. This was not considered as a 

 major disadvantage at least not at X-band. The phasing problem dis- 

 cussed in connection with the mixer barretter scheme of comparable 

 sensitivity is eliminated. An additional small advantage is the rugged- 

 ness of crj^stals in comparison to barretters and the availabihty of good 

 commercial balanced cr\'stal mixers. There are other double frequency 

 schemes which do not need 2 separate microwave signal generators. 

 The other frec[uenc3' may be obtained b}' amplitude or phase modulating 

 one signal generator b}" an IF frecjuency. The side bands which are pro- 

 duced m this way are displaced by just the IF frequency and may be 

 utilized instead of the second signal generator. Schemes of this sort 

 look particularly promising for frequencies well above X-band in which 

 case it might prove difficult to maintain the difference frequency of two 

 separate microwave generators within the band width of the IF. 



F. Experimental Determination of Sensitivity Limits 



1 . Preparation of samples 



In order to get an experimental check on the previous analysis, 

 samples with a known number of spins had to be prepared. Two sets of 

 samples were made. One consisted of .single CuSOi -51120 crystals of 

 varying sizes hermetically sealed between 2 sheets of pol3'eth3dene. The 

 other set con.sisted of different amounts of diphenyl picryl hydrazj-l* 

 which were similarly sealed up. D.P.H. samples having less than lO'' 

 spins were prepared by dissolving known amounts of the free radical 

 in benzene and putting a drop of this solution on the poh-eth\dene. 

 After the benzene had evaporated, it was sealed up with another sheet 

 of poh^ethylene. The ^-values of CUSO4 -51120 and D.P.H. differ enough 

 so that both samples can be eonvenientl}' run simultaneousl3\ This 

 was done in order to check the self consistency of the two sets of samples. 

 The measured integrated susceptibility of all the D.P.H. samples 



We are indebted to A. X. Holden for supplying us with this material. 



