From correlation receivers to zero-spacing interferometers

The SARAS system develops correlation receivers in which the sky power received by the antenna is split in a power splitter to form two signal paths, and these signals are then amplified, band limited and processed in a cross-correlation digital spectrometer. Once the path delays are equalized by a calibration process, the real component of the complex correlation yields the total power spectrum of the antenna temperature and hence a measure of the sky spectrum.

Such a single-element spectral correlator unfortunately also responds to ohmic losses in the signal path upstream of the splitter, and via second order effects responds to receiver noise.

ZEBRA (ZEro-spacing interferometer measurements of the Background RAdio spectrum) shifts the power splitter upstream to a location ahead of the antenna. The idea is that sky radiation is split by a space beam splitter into two components that are sensed by a pair of antennas, and these coherent signals corresponding to the two antenna temperatures are band limited and amplified in a pair of analog receiver chains before the cross power is computed in a digital cross correlation spectrometer.

The ZEBRA concept

The ZEBRA configuration we develop has a vertical space beam splitter and a pair of antennas on either side; the configuration is shown below:

Sky radiation is incident on both sides of the vertical sheet and these incident radiations are partly transmitted and partly reflected. A pair of antennas receives these radiations, providing coherent signals for the measurement of the cross power spectrum of the sky. For details on beam splitter, have a look at Mahesh et al.

We have shown that for ZEBRA to respond to uniform sky, the space beam splitter needs to be resistive, with sheet resistance equal to half the impedance of free space.

Test and Measurement Results

We have designed, built, and measured the performance – transmission and reflection coefficients – of such a resistive beam splitter. As shown in the photographs below, this is built as a soldered resistive grid: each segment contains a resistor of value 180 Ohms.

The grid size of the screen we have built is 10 cm, giving fairly frequency independent performance below about 375 MHz. The lower frequency performance is limited by the total size of the screen – we built a screen of 3 m x 4 m.

We have tested this concept by making a zero-spacing interferometer by placing a pair of fat dipole antennas on the two sides of the vertical screen and measuring the cross power spectrum.

Resistive Grid
Screen with fat dipole antenna

Comparison of the measured power with the resistive screen, with that measured with the screen removed, showed that the screen does indeed give an additional component in the cross power spectrum corresponding to the uniform sky (apart from the inevitable emission from the resistive screen!).

Designing antennas more appropriate for this configuration is the path to developing this concept.