Antenna Measurement Setup Question

[ferdem]

Hi friends!
Assume we have two dipole antennas located in an anechoic chamber, first antenna is connected to first port of a network analyser and the second is to the second port, the analyser is out of the chamber and the distance between the antennas is ok for the measurement. In this simple setup, can s21 measurement tell us something remarkable? Is not there a mistake? I think there is...
Before the NA measurement, I used SG-SA pair to predict s21 values, the two setup gave almost 25db different s21 values. NA says antennas couple 25 db(!) better comparing to SG-SA measurement. I ended up with following reasoning:

If I use NA, one pole of the antennas are electrically connected through the network analyser ground improving the coupling between the antennas! I think 1:1 transformer isolating the antennas and the analyser should be used for a legitimate measurement. Otherwise, the measurement is not meaningful. Am I right? Comments will be welcome, thanks.

 

[E Kafeman]

Correct antenna setup in an anechoic chamber must include ferrite tubes and/or quarter-wave sleeves for antennas causing reflective current outside of coaxial cable. A ideal 1:1 transformer do not block reflections. Coaxial cable should be placed such that it not can cause any secondary reflections.
The antennas must be placed such that it only can see the planar wave from the other antenna, known as the quite zone as most chambers not can absorb reflections from all directions.
Whole measurement setup should be properly grounded.
Crosstalk between coaxial cables can affect measurement but that is mainly a result of broken cables or faulty connectors.
If system isolation is good enough can easily be decided by replacing both antennas with a termination impedance as a part of system calibration.
If the antenna distance is too short (near field) must that be taken in account for calculation of S21.