The wiregrid is the main optical element defining the polarization of the source signal. It is expected that the output signal is 99.9% polarized, with a polarization angle defined by the orientation of the wires. The correspondence between the wire orientation and the camera coordinate system can be determined using a laser projecting a diffraction pattern on a flat screen, serving to connect to the metrology system.
The signal polarization observed by a telescope will depend on the relative angle of the wiregrid with the line of sight. The full derivation of this was reported in the following document: Coordinate System . This derivation is based on the assumption that the source's emitting electric field can be represented as a vector laying flat against the wiregrid plane, perpendicular to the wires. The observed radiated electric field will thus correspond to the projection of this vector on a plane orthogonal to the line of sight, as seen from a given telescope. This implies that different telescopes will observe different polarization angles depending on their location on the ground. Moreover, the effective rotation of the polarization angle becomes more significant as the elevation grows.
The following plots show simulations of the expected polarization angle for a perfectly horizontal (Roll0) and vertical(Roll90) electric field, as a function of the offset angle between the source boresight and the telescope line of sight. The simulation is done at different drone elevations.
To have a better understanding of the real effect of the wiregrid, including its frame, we performed simulations using GRASP. The simulation provides the electric fields projected to a sphere at infinity.Â
We observe that the polarization angle field is formed by two components: a softly varying angle variation, very similar to the one expected due to the projection effect, plus a more rapidly varying field, which is due to the frame
When the wiregrid is removed, the slow varying field goes away and we are left only with the fast field, which now appears a lot like an interference pattern.
We conclude that the frame is causing an interference that rotates the polarization angle by factors of a lot more than the acceptable limit of 0.1 degrees. This means that we need to fully understand this effect and improve the wiregrid design to minimize it.