Wideband Single Pixel Feeds (WBSPF) is an Advanced Instrumentation Programme aiming to provide a new technical solution for SKA Phase 2. Four of the single-pixel feeds from the current baseline design will be replaced with feeds for two wide bands, Bands A and B, which will extend frequency coverage up to 24 GHz.
We are currently considering re-defining the current Band B (5:1) to 3:1 bandwidth covering Band 5b and Band 5c (8 –24 GHz) and beginning a new study.
The WBSPF consortium is a collaboration of Chalmers University of Technology (Sweden), Low Noise Factory (Sweden), ASTRON, MPIfR, IAF, JLRAT and the Laboratory of Astrophysics (LAB) at the University of Bordeaux. At Chalmers, work is being carried out at Onsala Space Observatory and the Microwave Electronics Laboratory. The next WBSPF Board meeting will be scheduled in May.
During the northern autumn and winter 2016-7 we have been finalising the first prototypes of the project’s sub-elements – feeds and LNAs for each of the two bands. In China (TIPC and JLRAT), work on the cryostat continued. The detailed design of the cryostat’s main body has been selected, and now the design team is focusing on concepts for mounting and supporting the feed horns in the cryostat. Since the design shares elements with the proposal for the cryostat for the Next-Generation VLA (ngVLA), the group is working closely with Sander Weinreb’s group at Caltech.
The primary choice for both bands are feed horns with a quadridge design, similar to the feed for Band 1. These are being developed at CETC54/JLRAT (Band A) and at Onsala Space Observatory (Band B). Design work aimed at improving the optimisation for Band A is underway at CETC54/JLRAT. Currently the feed meets the efficiency specification (78%) over almost 80% of the band.
A prototype Band B feed has been assembled and tested at Yebes Observatory (Figure 1) with very satisfactory results. The current design meets the efficiency specifications (Figure 2). The sensitivity performance is not as good as we had hoped. Testing remains to be done, but our analysis so far shows that with the current technology limitations, the very ambitious specification (similar to the SKADC octave feed specification) is probably out of reach. While the Band B feed and LNA achieves 6.1 m2/K up to 10 GHz and 4.7 m2/K up to 15.5 GHz, the sensitivity is below the 3.5 m2/K specification above 20 GHz (here we are including the latest Aeff/Tsys-analysis atmosphere model).
Figure 1: Band B feed mounted on the scanner at Yebes observatory.
Low noise amplifiers for Band B are being assembled (mHEMT with components from Fraunhofer and InP from Low Noise Factory). Measurements for the InP wafers show very good equivalent noise temperature and flat gain of about 30 dB over the band (Figure 3).
The University of Bordeaux is working on the SKA1-mid band 5 and WBSPF band B receivers. These bands are split into 3 sub-bands (5a, 5b and 5c) for the first prototype because of no ADC are available on the market for sampling the entire band B or band 5 bandwidth. A block diagram of the receiver for the SKA-mid band 5 is given below:
The RF chain (RXS45.RF) of the receiver is described in figure 5.
Concerning the digitiser board (RXS45.DG), the first prototype is based on the 3-bit ADC from Hittite with a 20 GHz bandwidth. This board will be used for bands 4, 5a, 5b and 5c sampling. It will be full tested in October 2017.
A future digitiser board is already under studied for replacing the Hittite board by a low power and larger bandwidth solution. This future board will be based on a 4-bit, 25 GSps ADC developed in 28 nm CMOS technology.
We are currently considering re-defining the current Band B (5:1) to 3:1 bandwidth covering Band 5b and Band 5c ( 8 –24 GHz) and beginning a new study.
The next WBSPF Board meeting will be scheduled in May.