Our goal is to not only make an inexpensive antenna array, but also a very sensitivity one. To get an idea of the sensitivity of our receivers, we measure the galactic hydrogen line at 21 cm wavelength. The tile was mounted on the roof of my mother-in-laws’ house, which is located in the Vredefort dome, a fairly radio quite enviroment. The reference transmitter was set to 1419.8 MHz as can be seen by the large spike right in the middel of the figure below. To normalise the receiver signal, the signal is compared to the expected HI signal. To calculate the expected signal, HI profiles was downloaded from here. Astropy was then used to calculate the movement of the earth and to calculate a sky image at the measurement time and location. Two simple beam patterns was then assumed, sin(alt)^2 and sin(alt)^3. Multiplying the beam patterns with the sky image then give the expected signal. The expected signal fit the measured signal quite well if a system noise of about 90 K is assumed. This is higher than the simulated receiver noise temperature of about 30 K (see here), probably due to side lobes that picks up the ground and RFI. The large RFI noise on the right are probably signals from the tile itself (rtl-sdr dongles, Odroid or switch mode power supply) or from close-by computers.
6 pm measurement
7 pm measurement
Our expected model can be improved by using a more realistic antenna beam pattern. But as a first estimate, a 90 K system noise is not too bad for a broadband PCB antenna that is not very directional. It is comparable to the first EMBRACE system noise temperature measurement of about 100 K at 1.2 GHz. This proof that our receivers are sensitive enough to do even some astronomy.