Sec. 1-7] 



MECHANICAL INPUT TRANSDUCERS 



147 



frequency of a cavity is measured. The resonance frequency varies 



with the dielectric constant of the material contained in the cavity. 



If / is the resonance frequency of the cavity when it is filled with a 



reference gas (dry gas, dielectric constant e ) and/j the resonance 



frequency when the cavity is filled with the gas to be examined 



(dielectric constant ej, then 



2 



(£) 



The quantity e x is the complex dielectric constant; however, pre- 

 liminary experiments have shown that the loss factor for gases and 

 water vapor can be neglected in the region of 10 10 cps. 



A schematic diagram of the system is shown in Fig. (1-7)12. Two 

 resonant cavities are excited by a microwave source (klystron) K. 



K 



Fig. (1-7)12. Microwave refractometer for humidity 

 determination in gases; schematic diagram. 



The resonance in each cavity is indicated by the output from a crystal 

 diode D. A frequency modulator S shifts the klystron frequency in a 

 saw-tooth fashion. If the two gases in the cavities have different 

 dielectric constants, resonance will first occur in one cavity and after- 

 wards in the other cavity, as indicated by the curves on the right side 

 of Fig. (1-7)12. The resonance maximum appearing at the crystal 

 detectors will be displaced in time by an amount which is propor- 

 tional to the difference between the resonance frequencies of the 

 cavities. At an operating frequency of 10 10 cps the difference in 

 resonance frequency, if one cavity is filled with dry air and the 

 other cavity with air of a vapor pressure of 100 mb, is 0.27 Mc. The 

 time difference is measured electronically. In the final form of the 

 instrument a null-balance technique is used; the output is applied to 

 a servosystem which mechanically changes the resonance frequency 

 (volume) of the reference cavity. 



The gas temperature and the pressure in either cavity are kept 

 constant; therefore, the output scale can be calibrated directly in 

 vapor pressure. 



Tests of the instrument at high vapor pressure, i.e., in the range 

 from 4 to 16 mb vapor pressure, revealed an error in terms of vapor 



