View source for Antenna Cross-Coupling Simulation (David Kwak @ Timbie Group)

I have been looking at the correlated noise of Tianlai dish pathfinder array, a radio-interferometer consisting of 16 radio telescopes (antennae). It was built for Hydrogen (HI/21cm) intensity mapping, and observed at 700-800 MHz band for the first observing run.

The correlated noise is caused by the signal reradiating out of the electric dipole at the feed of each antenna in the array. This signal is picked up by the nearby antennae if they are sufficiently close to each other, and add to the instrument noise. I have been simulating this effect by defining waveguide ports at the feed and looking at the S-parameters between the ports that is computed. I have tried using the Integral Equation Solver (IES), Time Domain Solver (TDS), and Frequency Domain Solver (FDS) for this project, and was only able to obtain sensible results using IES. The use of TDS and FDS seems difficult for this project because the array is too large for the frequency of interests, and the meshing requires prohibitively large RAM for the simulation to run without errors.

According to CST manual, it is advisable to use waveguide ports for this calculation as it extends through to the infinity of the direction the port is facing at, and thus significantly reduces the reflection off of the other structures involved in the simulation. The port was defined as advised in CST guide (cylindrical coaxial tube is used for the model; for coaxial tube, the manual suggest that one should set up the port such that the entire tube is covered). However, we are considering changing the port definition because our collaborator in China used different port definition in their earlier simulation.