Study of different methodologies to determine relative dielectric constant of given substrate using fabrication, modelling and measurement strategies

Sammanfattning: In this thesis, different methods for measuring the relative dielectric constant is investigated using simulations and experimental approaches. The first method is known as the S-parameter inversion method. In this method formulae is used to calculate the characteristic impedance of the microstrip line from measured S-parameters. To calculate the value of relative dielectric constant, four expressions for the microstrip line that relate the characteristic impedance to the relative dielectric constant were used. Second method uses an implementation of a band-pass filter and is commonly known as the band-pass filter method. In this method, a band-pass microstrip filter was designed using a predicted relative dielectric constant value. Center frequency of the band-pass filter is chosen to be at the frequency of interest, since relative dielectric constant will be determined around this frequency. The designed band-pass filter was manufactured and the frequency response was measured. To determine the true relative dielectric constant one changes the relative dielectric constant parameter used in the simulation until it matches the measured response of the manufactured PCB. Third method is called the quarter wavelength stub method. It uses implementations of a microstrip quarter wavelength stub because it resonance at different frequencies. The relative dielectric constant is determined using the frequencies on which the reflection occurred. In the fourth method, called the two microstrip line method, two lines of different lengths were designed and the phase difference between the propagating waves were measured. The phase difference and difference in length of the two lines is then used to calculate the relative dielectric constant. The thesis shows that a majority of the methods generates a similar result, thus indicates that they are suitable to determine the relative dielectric constant of any given substrate. The two methods that gave the most accurate results are the quarter wavelength stub method and the band-pass filter method. S-parameter inversion method is the method that has high variations in the results. Since the characteristic impedance that is calculated using the S-parameters are sensitive towards any sort of disturbances. The resulting relative dielectric constant aren't within the expected range for FR-4 both higher and lower values were obtained. The band-pass filter method gives the most accurate results of the methods. As the resulting relative dielectric constant are within the expected range for FR-4. Quarter wavelength stub method gives the results of the relative dielectric constant that are within the expected values of FR-4 and the variation is moderate. Two microstrip line method shows deviations in the results and has non-linearity as well. This is probably coming due to resonance of the line that gives rise to a phase change. This method is also acceptable since the results of the relative dielectric constant are within the expected range for FR-4.