Radio Astronomy SA

Radio signals arriving from directions close to the axis of the parabolic antenna are focused by reflection to an antenna feedhorn optimised for operation at 1420MHz.  Antenna gain is assumed to be approx. +30dBi.  Signals then pass through two L band probes phased 90 degrees apart and coupled to a pair of low noise amplifiers (LNA) with noise figures of ≤ 0.2dB at 1420MHz.

The LNAs were originally built, adjusted and tested as a matched pair by Dominique Faessler, HB9BBD located in Switzerland.  Each LNA provides an input 40MHz tuned cavity bandpass filter and two amplification stages producing 40dB of gain. The bandpass filters prevent out of band signals such as those generated by geostationary satellites and other man made interference sources from producing intermodulation products in the SDR's image rejection mixers.

In their original configuration, the two signals are then combined with a 90° quadrature hybrid coupler and fed via low loss cable to a single SDR.  In order to minimise cable losses, the SDR was housed directly underneath the antenna.

Without a low noise amplifier (LNA) the smallest solar radio burst that may be detected is a little more than 9400 solar flux units.  With an LNA, this falls to about 2500 solar flux units.  Experience has shown that the Callisto receiver is capable of detecting less powerful bursts.  It could be that our assumptions on detection threshold and antenna gains are too restrictive.
In professional radio observatories, the LNAs generally are cooled to lower their noise figures.  However, in the VHF range there is no need to cool down the LNA to get lower noise figure because the sky noise is always higher.  For receivers that operate above 1 GHz, where the sky noise is only a few kelvins, it makes sense to cool down the LNA.  For solar burst observations within the Callisto receiver band (45MHz to 870MHz), an LNA is a must; otherwise the science output will be very low.

The Mini-Kits PGA-103 KIT amplifier is a very low noise, high dynamic range receive amplifier that has been designed for operation between 50MHz to 1.3GHz, for this application the upper frequency is 870MHz .  An onboard bias tee allows the LNA to be antenna mounted and powered via 50Ω coaxial cable.  As a driver amplifier, the PGA-103 produces an output between +20dBm & +22dBm @ 1dB compression.  The Output IP3 is +39dBm @ 400MHz, and it has a gain of 20dB and 0.4dB noise figure @ 400MHz.  Both LNAs were fitted with 2dB attenuators to provide some isolation and reduce the combined gain to 36dB.

Power Supply:  The Mini-Circuits PGA-103 MMIC has internal self biasing and can be used with a supply voltage between +3VDC to +5VDC.  A 10Ω resistor is used to reduce the +5VDC supply voltage to the PGA-103 MMIC down to +4VDC for a bias current of 90mA.  The board has an onboard 78M05 voltage regulator which allows a supply voltage between +7VDC to +15VDC to be applied to the RF output jack from a suitable receiver or another bias tee.  We chose an external Mini-Kits EME181A70 Bias tee so that both LNAs could be powered from the same DC supply.
The PGA-103 MMIC is susceptible to being damaged by static including lightning discharges, so the LNAs were ordered from Mini-Kits with cases.  A large stainless steel enclosure has been used to house the two LNAs and a Huber+Suhner lightning protector with a 90V gas cell has been fitted.  Provision was made within the stainless enclosure to add a second, identical LNA pair for another linear polarised LPDA but this never eventuated.

Two x 1MΩ 1206 resistors have been added to the Mini-Kits board directly across the RF SMA input jack, to help dissipate any static buildup on the antenna.

The members of the Radio Astronomy Special Interest Group of the Astronomical Society of South Australia (ASSA) produced our first (Version 1) FEE in 2020.  The objective was to essentially clone the Commercial Off The Shelf (COTS) FEE design, primarily due to the US$500 price tag (plus handling & freight) for a pair of loaded/tested PCBs.  We opted to introduce some minor modifications considered necessary to overcome some issues with parts availability at the time and our modified design has proved to be reasonably successful.  While the spectrometers worked quite well, there were some improvements which we felt would be simple enough to incorporate for future use.
Version 1 used an LM317T device as a replacement for the LM1085-12 voltage regulator on the COTS FEE.  The output voltage was reduced from +12VDC to about +6.20VDC to help set the bias arrangements to the GALI devices.  When the sun was producing lots of power over the summer months and the battery was fully charged, the LM317T tended to run hotter than we would have liked.
The COTS FEE used a TeleTech HX62A, 180° Hybrid Junction to combine the U1 and U2 signal outputs.  The HX62A price back in 2020 was around US$40 per device with an MOQ of 10 (without freight or handling costs).  Version 1 eventually used a Mini-Circuits TC1-1TG2+ Transformer as a replacement for the HX62A.  After a couple of years in daily service, this has proved to be an effective alternate part.

The Version 2 PHI was built in late 2023 for a parts cost of approx. US$75 per board (Qty 2 per LWA antenna).   Minimum order quantities may complicate this estimate but potential users should at least consider a pair and a spare, or collaborate with another Callisto station to mitigate the costs of holding excess stock unlikely to ever be used.
Note that our Version 2 PHI each draw ≈190mA @ +14VDC (2.6W) .  If DC-DC converters are used, the voltage outputs are set for a +9.0VDC (1.8W) input at the Version 2 PHI SMA connector (this includes DC-DC converter losses).

DC Voltage Regulator :  Version 2 replaces the LM317T with an AMS1117 LDO device, set for an output of +7.00VDC at +20°C.  Reliable regulation of the AMS1117 requires the  input voltage at the SMA jack to be at least +8.75VDC, accounting for voltage drops across L3, D1 and track resistance.  While the AMS1117 is more than capable of providing the 230mA required for a COTS FEE configuration, the lower bias currents we use should help provide better stability and reliability at high operating temperatures.  Without consideration for DC power constraints, the Version 2 PHI should operate between an +8.75VDC input (1.75W) and +15.0V input (3.0W).  The optimal input for the Version 2 PHI is +9.0VDC.

GALI amplifier bias :  The bias conditions for Version 2 are considerably different from Version 1.  All four 4.7μH inductors have been replaced with 5.6μH devices  to decrease the low frequency response below 15MHz.  The 4.7μH inductors on the original FEE have a 90MHz series resonance limit, which is close to the 88MHz upper scan frequency for our LWA spectrometers at Sunnydale.  They also have a DC resistance of about 4Ω, which forms a significant percentage of the total bias path resistance.  Our selection of a shielded 5.6μH inductor for both the DC input and GALI bias inductors solves several problems and simplifies the component list, however it does increase the cost of the board assembly.
Additional resistors/pads have been added to the PCB to assist users set up their choice of bias currents and measure each to ensure optimal operation.  Inclusion of the 10Ω shunt resistors (R4, R8 & R12) comes at a cost, reducing the overall gain of the Version 2 PHI to about +34dB, compared to the nominal +36dB of the COTS FEE.

Output attenuator:  To provide some gain control, without the need for 15dB coaxial attenuators inserted between the COTS FEE output and a receiver/upconverter input, the Version 2 PHI has provision for a custom attenuator (R18, R19 & R20) to compensate for coaxial losses and the gain differences mentioned above.

The MWA LNA is connected to the dipole antenna using a twin lead balanced transmission line.  The LNA has a balanced input and an un-balanced 50Ω output.  The DC bias to the LNA is injected from it's output connector using a bias tee.  The design is a single stage matched amplifier stages made of the Avago ATF35143 enhancement mode low noise HEMT.
The transistor is a surface mount component; the board is made on FR4 material.  The dual LNA has 2 differential LNAs, each one of them has its own ground plane and the 2 LNAs are separated by a metal layer.  The circuit topology used for the MWA LNA is: two matched single ended amplifier stages combined in and out of phase using a transformer into a single ended output.  Each amplifier has a 50Ω SMA jack as RF output/ DC input.  The input consisted of microstrip lines on FR4 board where the two twin leads connecting to the antenna are soldered.

The Mini-Kits PGA103-UHF-R2 amplifier is a very low noise, high dynamic range receive amplifier that has been designed for the 70cm band.  It can also be tuned to cover a narrow band use between 375MHz to 500MHz, for this application the frequency is 401.5MHz .  An onboard bias tee allows the LNA to be antenna mounted and powered via 50Ω coaxial cable.  As a driver amplifier, the PGA-103 produces an output between +20dBm & +22dBm @ 1dB compression.  The Output IP3 is +39dBm @ 400MHz, and it has a gain of 20dB and 0.55dB noise figure @ 400MHz.

Power Supply:  The Mini-Circuits PGA-103 MMIC has internal self biasing and can be used with a supply voltage between +3VDC to +5VDC.  A 10Ω resistor is used to reduce the +5VDC supply voltage to the PGA-103 MMIC down to +4VDC for a bias current of 90mA.  The board has an onboard 78M05 voltage regulator which allows a supply voltage between +7VDC to +15VDC to be applied to either the 2.1mm connector or the RF output jack from a suitable receiver or another bias tee.
The PGA-103 MMIC is susceptible to being damaged by static including lightning discharges, so this LNA was ordered from Mini-Kits without a case.  A larger diecast enclosure has been used to house the LNA and a Huber+Suhner lightning protector with a 90V gas cell has been fitted.

Two x 1MΩ 1206 resistors have been added to the Mini-Kits board directly across the RF SMA input jack, to help dissipate any static buildup on the antenna.

Following a couple of months of tracking radiosondes with some success, Peter has decided to introduce some changes/improvements to the setup.  A problem  has been identified when simultaneous balloon launches can occur occasionally from Adelaide and Mt Gambier and the receiver/software locks onto one or both transmissions.  By adding another, more directional antenna option to allow some isolation from the different airborne signal sources when they occur, may be a viable solution.  It also allows the option to change antennas for long range tracking, when appropriate weather conditions prevail.

Mini-Kits have recently discovered a SAW filter version of their PGA103_UHF range of LNAs which didn't make quite make it to their website prior to now, one of these has replaced the version above.  The original LNA won't go to waste, it has now become a key component for another radiosonde tracker to be located onsite at Sunnydale.