This paper presents a method of measuring the temperature coefficient of the material permittivity and the coefficient of thermal expansion of dielectric resonators in a purely electrical way. The radial mode matching method has been succesfully applied to the accurate analysis of the test structures containing dielectric resonators.
I NTRODUCTI ONRecently, dielectric resonator materials with very small dielectric losses and very good temperature stability have been developed [1,2]. This prompted a need for precise and simple techniques of the dielectric resonators parameters measurements, as it now appears necessary to certify the value of permittivity er within an error less than .3!.-, quality factor [l0 higher than 30000 and the stability of resonant frequepcy versus temperature Trf with precision better than . lppmrV C. Several excellent papers considered :r and UO measurements have been published [3,4]. Also .lppmnvC accuracy of the Tf measurement has been achieved (5,6], but this only partly solves the frequency stability problem. There are three main effects which cause the resonant frequency variations. The first being temperature dependence of the material permittivity T r" the second being the change in dimensions of the res inator aL and the third responsible for change in shielding structure dimensions. In many cases it is possible to design a resontor structure in a way that the three effects are compensated, however to obtain that goal one should have precise information about each of the effects indepedently. We can assume that it is possible to obtain the parameters characterizing the shielding structure but information about resonator parameters T r and aL is not always provided by the resonator suppliers or s not sufficiently accurate. Catalogue values of Tf give information about the total frequency drift with an assumption that the following equation is valid [2]: Tf -0.5 Ter-c (1) As i s shown i n Tabl e 1 f or mul a (1) i s not al ways val i d. The measurements provide information about the total frequency drift caused by material parameter change and also by varying resonator dimensions. To separate these two effects the manufacturers often make separate mechanical measurements of the temperature dilation of material dimensions. Such measurements are expensive C usually special long samples of
