We present an end-to-end model of the signal-chain of the correlator and beamformer of the Square Kilometre Array Phase 1 Mid telescope. The objective of this model is to determine whether the proposed signal processing architecture can meet the SKA’s stringent requirements on signal quality. The model consists of two parts: 1) the “Reference” model, and 2) the “Realizable” model. Both are implemented in MATLAB. Both the reference and realizable models are implemented with double-precision floating point arithmetic, however, the realizable model considers quantized data and coefficients for the signal processing algorithms. The time-varying delay due to the rotation of the Earth and the synthesized celestial source is modeled in the signal generator. The signal generator also supports the modeling of signals sampled according to the Sample Clock Frequency Offset method. Both reference and realizable models utilize a two stage delay correction process with integer delay correction and resampling with a fractional-delay filter-bank. Combinations of oversampled polyphase filter-banks, critically-sampled polyphase filter-banks, digital down-convertors, threshold-based RFI detectors, phasedelay beamformers and complex cross-multiplier accumulators are used to model the continuum/spectral-line imaging, zoom-imaging, pulsar-search-beamforming, pulsar-timing-beamforming, VLBI-beamforming and VLBI visibilities to parameter evaluation for calibration. This model has been used to verify that the proposed delay correction method is sufficient to achieve the required sensitivity. Also, this provided evidence that the phase-delay beamforming method can be successfully used for pulsar-timing beamforming. The study of the degradation of the signal quality in response to various RFI scenarios, which are expected on the telescope site, has also conducted and published.
A three dimensional (3-D) spatio-temporal analog signal processing scheme is presented for the selective removal
of off-dish interference and noise from focal plane array (FPA) received signals. The method exploits specific
geometrical properties of the 3-D spatio-temporal frequency spectrum of FPA signals to perform the filtering
operation. A 3-D infinite impulse response (IIR) filter having a cone-shaped filter passband in the 3-D spatio-temporal frequency space is employed to extract the spectra of desired FPA signals while rejecting the spectral
components from undesired off-dish interference and coupled noise from front-end electronics. A proof-of-concept
example is provided by considering the filtering operation in 3-D spatio-temporal frequency domain.