The dielectric constant of a board simulating the medium in half space is measured in the laboratory using the space wave method including the metal-scaling method and the dual polarization method The experiment is performed in X frequency range

the amplitudes and the phases of the received fields are measured simultaneously The dielectric board consists of paraffin

plaster and graphite First

the scattered fields by different Fresnel zones in the board are calculated by Kirchhoff formula It is concluded that the area larger than 10 Fresnel zones can be approximately regarded as infinitely large The thick of the board is determined by the attenuation coefficient of the elctromagnetic waves After repeatedly measuring the dielectric constants in the waveguide for different prescriptions

the weight ratio in the board is chosen as paraffin:plaster:graphite=2 38:1:1 8

in which case the attenuation is desired The thickness and the area of the board are chosen as 10cm and 135cm×105cm The experiment frequency is 9 965GHz

the 3dB half width of the wave bean is 9°。The microwave receiver（Atlanda 1780） receives the phases and amplitudes of the reflected fields in the specular direction In the metal-scaling method

transmitting and receiving are performed in a same polarization

the reflection coefficient of the medium is obtained by comparing the phases and amplitudes of the fields reflected by the medium board and a metal board

then the permittivity is deduced from the reflection coefficient The incident angle is adopted as 15°in the metal-scaling method

the measurements indicate that the results using parallel polarization is better than that using vertical polarization In the dual polarization method

however

the incident angle is taken to be 60°

a wave with both vertical and parallel polarizations is transmitted by an inclined horn

the amplitude and the phase of the reflected field in the specular direction is received for each polarization respectively A simple formula to deduce the dielectric constant from the amplitude ratio and phase difference of these two polarizations is given in the paper The measurements indicate that the result of dual polarization method agree well with that of the matal-scaling method using parallel polarization The measured result of the long sample method in the waveguide has a little difference（less than 20%）with their results

which originates from the inevitable difference between the sample part and the whole board From the contrast with the long sample method in the waveguide

it is found that our metal-scaling method is intrinsically a long sample method in the free space The paper also gives the corresponding formulas for the waveguide to the free space The paper analyses the errors of the phase measurement in the metal-scaling method which mainly come from the flatness of the metal board and the stability of the instruments In the experiment the real part of the permittivity is not sensitive to the phase error

the imaginary part only have a relative error of the same order with the phase error

thus the results of the experiment are rather stable

which verifies the feasibility of the phase method Both the metal-scaling and the dual polarization methods have the advantages that the measurements and the inverse formulas are simple and the results are stable But for the purpose of on-site measuements

the dual polarization method is better