One interesting innovation is our compact orthogonal mode transducer (or OMT) which utilises two crossed dipoles in the waveguide environment, which then feed the orthogonal linear receive signals to the LNAs through a nominally –35 dB integrated directional coupler (for test noise injection). It is electrically much smaller than a traditional quad-ridge structure and introduces less than 2 K noise when cooled to around 100 K physically (corresponding to a signal loss of less than 0.06 dB). This compact and remarkably efficient OMT resulted in a compact cryostat and thus less loading on the Gifford-McMahon (GM) system – which is significant given the low efficiency of GM cooling and the cost of power. In addition, it allows a very compact solution when adding lower frequency bands to old systems that do not have enough physical space for traditional OMTs.
We have also been careful with important secondary characteristics such as receiver signal path stability. Here we employed thermally-stabilised platforms for the primary signal-path components, optimised our cabling choices, ensured substantial RFI isolation of power supply noise and used custom-developed low-speed control and monitoring circuitry and protocols.
We required a passband that is as flat as possible over the entire frequency band, and given the typically modest input impedance match of noise-optimised LNAs, we spent many design hours on matching the OMT to the horn, and on matching the combination to the reflector – so that reflection at the LNA input will reradiate without causing standing waves and a resultant passband ripple.
Our cryogenic receiver systems rely on Gifford McMahon (GM) cooling, and has been optimised for low additive noise without compromising the effective usage of the reflector surface area – thereby maximising the prime metric in radio astronomy – sensitivity. On MeerKAT, which has an offset Gregorian reflector system with conical surfaces, we have achieved a measured normalised noise temperature (full system temperature / reflector efficiency) of 22.5 K in L-band. In the case of SKA, which has lesser sidelobe constraints and lightly-shaped optics, we predict with high certainty that the normalised noise temperature will be better than 18 K over the entire
950 MHz to 1760 MHz frequency band and all elevations down to 30° above the horizon. Our good grasp of the electromagnetic behaviour of electrically medium-sized reflector-feed systems helped us to reach these performance levels.
The RFI requirements imposed by the MeerKAT and SKA telescopes are severe – especially in the case of any support equipment that live out in the open on the telescope structures. We have successfully developed essentially RFI-free helium compressor and vacuum pump controller circuitry, and use layered RFI reduction techniques integrated into the packaging of our control and monitoring subcomponents.
Since our solutions meet the stringent MeerKAT and SKA requirements, they may be of use in your system – should you also be prohibited from generating any measurable electrical emissions.