Both 1-port and 2-port measuring methods have been demonstrated and compared to determine the real part of the frequency-dependent effective relative permittivity,
re
(f
), of microstrip lines fabricated on an alumina substrate. Signal flow diagram analysis has been used to derive equations from which
re
(f
) was calculated at discrete frequencies for the substrate with a known dielectric constant,
r
= 9.5. Both methods gave similar results for
re
(f
) from 45 MHz to 50 GHz. In addition, the measured frequency dependence of
re
(f
) was in agreement with the Kirschning-Jansen dispersion model within better than 2%. From the measured S
-parameter data it was also possible to determine approximately the attenuation coefficient of the microstrips. Measured values obtained by the 2-port method were about 1.5-2 times the calculated values whereas the 1-port method suffered from radiation loss at the open end of the microstrips in the millimetre wave region. Both 1-port and 2-port methods can be used, for example, for quality checking purposes to verify how well the
r
value of the substrate material employed is within the specified range in a broad frequency band. However, the 1-port method can be more easily used to determine
r
of the substrate material at low microwave frequencies because it is faster than the 2-port method, the test structure is simpler and the calibration routine is easier to perform.