The Coaxial L-P Cryogenic Receiver of the Sardinia Radio Telescope

The design and characterization of the coaxial dual-band L-P radio astronomical receiver for the prime focus of the Sardinia radio telescope are presented. The main feature of this receiver is to allow simultaneous radio astronomical observations in the P (305 – 410 MHz) and L (1.3 – 1.8 GHz) frequency bands. This functionality, which has been requested by the Pulsar research group at the National Institute for Astrophysics to estimate, among the others, the ionospheric dispersion in Pulsar observation, is currently missing in any other radio astronomical facility throughout the world. Also, single band operation is ensured by the proposed design both in linear and circular polarization, making this L-P receiver an ideal instrument for a wide range of radio astronomical and space applications. Some components of the receiver chain have been housed inside a cryostat and refrigerated at 20 K to reduce the noise temperature, resulting in a good performance compared to the receivers of other large radio telescopes. Several challenging issues have been faced in the design, mainly due to the large dimension and weight of the overall structure to be mounted in the prime focus position. Moreover, the design of the cryostat was constrained by the limited space available in the direction of the optical axis inside the focal cabin of the radio telescope, requiring a compact and light realization of the components of the receiver chain. This called for a home-made design of several devices, requiring a strong collaborative effort by researchers, engineers, and astronomers.


I. INTRODUCTION
The Sardinia Radio Telescope (SRT) is a general-purpose fully steerable 64-meter diameter radio telescope designed to operate with high efficiency across the 0.3-116 GHz frequency range [1]- [4]. The radio observatory is the result of the scientific and technical collaboration among three separate organizations of the Italian National Institute for Astrophysics (INAF): the Institute of Radio Astronomy (IRA), the Cagliari Astronomical Observatory (OAC), and the Arcetri Astrophysical Observatory (OAA). The main funding agencies are the Italian Ministry of Education, University and Research (MIUR), the Sardinia Regional Government (RAS), The associate editor coordinating the review of this manuscript and approving it for publication was Yiming Huo .
the Italian Space Agency (ASI), and INAF itself. The SRT is designed to be used for astronomy, geodesy, and space science, both as a single dish and as part of European and International Very Large Baseline Interferometry (VLBI) networks. The SRT operates as an international facility, with regular worldwide distributed calls for proposal, and no a priori limitation based on the affiliation of the proposers. A large fraction of the observing time (of the order of 80%) is devoted to radio astronomy applications, while 20% of the time is allocated to activities of interest to ASI, i.e. space applications and the follow-up of space science missions [5]. Recently, the telescope has also been used for Space Surveillance and Tracking (SST) [6]- [8].
The telescope is located 35 km North of Cagliari (Italy) at about 600 m above the sea level. To minimize spillover and VOLUME 10, 2022 This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ standing waves, the SRT optical design is based on a quasi-Gregorian configuration (Fig. 1) with a shaped 64-meter diameter primary reflector (M1) and a 7.9-meter diameter secondary reflector (M2). The primary reflector is equipped with an active surface, consisting of 1008 aluminum panels (with a panel manufacturing root-mean-square error (RMSE) less than 70 µm) and of 1116 electromechanical actuators able to compensate the gravitational deformation of the backup structure. The primary reflector has been aligned with a RMSE of 290 µm using photogrammetry. Currently, the metrology system is being upgraded to improve the RMSE to a level of 150 µm and enable high antenna efficiency observations up to the 3 mm band (with 116 GHz maximum frequency). SRT has been designed to host up to twenty receivers, which can be installed in six different focal positions: i. Primary focus (F1) with focal length to diameter ratio (F/D) equal to 0.33, ideal for receivers operating in the frequency range 0.3 GHz -20 GHz. ii. Gregorian focus (F2) with F/D equal to 2.34, frequency range 7.5 GHz -116 GHz. iii. Beam-wave guide foci (F3&F4 and F5&F6) with F/D 1.38 and 2.81, respectively, frequency range 1.4 GHz -35 GHz. In this work we will describe the development, design, and realization of the cryogenic coaxial dual-frequency L-P band receiver of the SRT, which has been installed on the primary focus (F1) of the antenna and allows concurrent observations in the P and L frequency bands, as well as observations in a single band (either P-band or L-band). This receiver operates in both linear and circular polarization, allowing observation of a wide range of radio astronomical phenomena.
Among the largest radio telescopes in the world (see table 1), to the best knowledge of the authors the SRT is the  only one equipped with a dual-frequency receiver that covers the P and L frequency bands (Fig. 2), and this makes it the ideal instrument for several radio astronomical applications. Among them we can mention the study of the mechanism that generates eclipses in eclipsing binaries [9]. Concurrent observations in different frequency bands can also be exploited to compensate for the ionosphere dispersion in Pulsar survey [3]. In fact, when performing high precision timing it is important to account for the variable effect of the interstellar medium along the line-of-sight to the source (on time scales longer than a few days). To properly remove this contribution, it is necessary to accurately measure the delay between the times of arrival of the same pulse at two well-separated frequencies. In this context, the dual-band operation of the L-P receiver allows halving the time required to obtain a high precision timing of a target, which in turn, implies the possibility to double the number of useful observations for the given target (in an assigned telescope time).
The coaxial L-P receiver of SRT can also be used to observe a radio source with a single band. For example, the VLBI observations require L-band operation [10]. Recently, from 20 to 24 February 2020, the P-band receiver has been used to perform the lowest-frequency detection to date of three radio bursts, observed at 328 MHz from the periodic repeater FRB 180916.J0158 + 65 [11]. Simultaneous observations with the L-band receiver at 1548 MHz were also performed but did not detect any bursts. A further application of the SRT in the P band channel (and linear polarization) is space debris detection, as the receiving part of the BIRALET bistatic radar [6]- [8].
The bandwidth of the L-P receiver has been chosen after astronomical considerations but also primarily after Radio Frequencies Interference (RFI) measurement around the SRT site. Indeed, at the start of the project the P-band was supposed to cover the frequency range 305-425 MHz, but after several RFI campaigns, it was decided to narrow down the band to the present frequency range, i.e. 305-410 MHz. Based on the same considerations, the operating frequency range of the L-Band has been selected between 1.3 GHz -1.8 GHz to achieve the widest radio astronomical performance, while avoiding adjacent-channel interference. The RFI monitoring is performed periodically to identify any possible new source of interference [12].

II. ARCHITECTURE OF THE L-P RECEIVER
The L-P receiver is installed on the primary focal position F1. For this reason, it must meet strong weight and dimension constraints: it has a volume of about 1.5 m × 1.5 m × 1.5 m and a weight of 700 Kg.
Because of its relatively low operating frequency range, this receiver is characterized by the absence of the mixer section. Indeed, the two RF bands (P-band and L-band) are inside the intermediate frequency (IF) of the SRT, which covers the frequency range 0.1 -2.1 GHz [13]. As a consequence, the down-conversion of the RF bands is not required.
The main features of the receiver are listed in the following: • Frequency range: 305-410 MHz (P-band) and 1.3-1.8 GHz (L-band) simultaneously.
• Linear and circular polarization for each frequency band.
• System Temperature (T REC ) ∼ 20 K in the P-band and ∼10 K in the L-band. • Selectable linear or circular polarization for specific observations.
• Selectable narrow band for VLBI observations. The basic architecture of the receiver is shown in Fig. 3(a), wherein we report the main blocks and components that form the front-end path. Starting from the left side of the figure we can find: a) The coaxial feed converting the received (P-band or L-band) free-space electromagnetic wave into a guided wave, operating at environmental temperature (300 K).
b) The coaxial orthomode junction (OMJ) of the P-band receiver, which is integrated in the coaxial feed and works at room temperature (300 K). c) The cryostat (or Dewar), wherein some of the components of the P-band and L-band receiver paths are housed, operating at the temperature of 20 K. d) The noise calibration unit, which provides the receiver calibration for both L and P frequency bands by injecting a noise source in the RF paths. The block diagram is described in Fig. 3(a). e) The P-band linear to circular polarizer ( Fig. 3(b)). f) The L-band linear to circular polarizer ( Fig. 3(b)). g) The P-band filter selector, which allows selecting narrow-bands within the P-band for specific radio astronomy observations. The block diagram is shown in Fig. 3(c). h) The L-band filter selector, which allows selecting narrow-bands within the L-band for specific radio astronomical observations ( Fig. 3(d)). In the following subsections we will describe in detail the design and the architecture of the blocks a), b), and c) listed above.

A. ARCHITECTURE OF THE COAXIAL FEED
The feed of the dual-band L-P receiver is realized using a coaxial configuration, which includes also the coaxial orthomode junction (OMJ) of the P-band. The latter operate at room temperature (300 K), whereas the OMJ of the L-band works at cryogenic temperature (20 K) (see Fig. 3(a)). The design of the P-and L-band feeds and of the P-band OMJ have been made using the commercial software CST Studio Suite. The section view and 3D view of these components are shown in Fig. 4.
The P-band OMJ and the P-band feed consist of a coaxial waveguide with an outer diameter equal to D P = 650 mm (which is the P-band feed aperture, as well), and a variable inner diameter, which corresponds to the L-Band waveguide and is composed of three sections: a first section, starting from the back-short, of length L 1 = 575 mm and diameter D L1 = 196 mm; a matching conical section of length L 2 = 120 mm; a third section, terminating at the feed aperture, with length L 3 = 220 mm and diameter D L2 = 171 mm (see Fig. 4).
Four matching cylindrical metallic irises are arranged along the optical axis of the P-band feed. Two of them, I1 and I2, are attached to the inner L-band waveguide near to the planar back-short (see Fig. 4). The other two metallic irises, I3 and I4, are located near the aperture of the feed: one is attached to the external P-band waveguide, the other is attached to the internal L-band waveguide (Fig. 4). The dimensions of these irises are reported in Table 2, wherein L Ii is the distance of the iris Ii from the backshort, and T Ii is the thickness of the iris Ii (Fig. 4).
The sketch of the P-band orthomode transducer (OMT), which includes the coaxial OMJ and two 180 • Hybrid power dividers, is shown in Fig. 5 (only one polarization). Two linear polarized signals (Pol. 1 and Pol. 2) are extracted through four identical metallic coaxial probes connected to the central pin of commercial 50 7/16-type coaxial connectors ( Fig. 6(b)). The probes are located at L I 1 = 177.5 mm from the back-short of the coaxial waveguide and each one consists of four cylinders of different diameters with the same axis of the coaxial connectors. The geometry, the position and the dimensions of the probes have been optimized for a better matching and are reported in Fig. 6(c). To avoid mechanical microphonics problems due both to the radio telescope movement and to the vibration of the cryogenic pump, the inner part of the probes has been  hollowed and fixed to one of the irises (I1) through robust Teflon (ε r = 2.01) supports. This solution allows reducing the weight of the probes as well.
The fundamental mode of the coaxial waveguide described so far is the TEM mode and the cut-off frequency of the first higher order mode, the TE 11 , is 232 MHz. However, using the four-probes OMJ feeding configuration, the fundamental TEM mode does not propagate. Under these conditions, in the operating P-band of the receiver (305 MHz -410 MHz), only the TE 11 mode propagates and its wavelength at the center frequency (357.5 MHz) is about λ g = 1100 mm.  Finally, to improve the radiation pattern of the P-band feed external corrugations have been placed on the aperture flange. They are composed of three rings with a maximum diameter of 1148 mm and a thickness of 2 mm each. The distances of the throats from the optical axis are: R C1 = 574 mm, R C2 = 510 mm, and R C3 = 428 mm, with a depth of D C = 320 mm (see Fig. 4).
The L-band feed is a simple truncated circular waveguide, which is inside the circular waveguide of the P-band feed (see Fig. 4). As pointed out before, this waveguide is made of three sections (Fig. 4): a section ''A'' of length L 1 and constant diameter, a matching conical section ''B'' of length L 2 , and section ''C'' of length L 3 and constant diameter. The waveguide wall thickness is 3 mm. Therefore, the inner diameter of sections ''A'' and ''C'' is DI L1 = 190 mm and DI L2 = 165 mm, respectively. The structure has one physical port but two electrical ports because two degenerate TE 11 modes are excited inside the circular waveguide [14].
Two irises, I5 and I6, located near the aperture of the feed (Fig. 4), have been used to match the free space impedance to the impedance of the circular waveguide. The dimensions and position of the irises are listed in Table 1, whereas the diameter reduction (from DI L1 to DI L2 ) is aimed to improve the return loss of the OMJ [14].
The photo of the L-P feed installed on the primary focus of SRT is shown in Fig. 7.

B. P-BAND RECEIVER CRYOGENIC RF PATH
As discussed in Section II.A, the coaxial orthomode junction of the P-band receiver is integrated into the coaxial feed and operates at room temperature (300 K). On the other hand, the components of the P-band receiver, cascaded as shown in Fig. 3(a) and operating at a cryogenic temperature of 20 K, are: • A cryogenic 180 • hybrid with integrated directional coupler [15]. A planar fractal 180 • hybrid configuration has been used to cover the operating bandwidth of the receiver with a significant size reduction, which is mandatory to save space (and refrigerating power) inside the cryostat. The return loss is larger than 20 dB over the operating bandwidth, whereas amplitude and phase imbalance are 0.5 dB and ±10 • , respectively. A coupled lines directional coupler with weak coupling (−26 dB) and high isolation (below −45 dB), used to calibrate the receiver chain, is cascaded to the 180 • hybrid and realized in the same layout. The planar size of the resulting six-port component is only 108 mm × 160 mm. This component has been successfully tested at the cryogenic operating temperature (20 K), with a percentage power loss of less than 1.5% over the operating frequency band.
• A commercial cryogenic coaxial mechanical switch (model Agilent 8761B). It is used to isolate the P-band low noise amplifier (LNA) should the SRT be equipped in the future with a transmitter for deep space and nearearth applications. The insertion loss of this component is estimated at 0.01 dB at 20 K and the equivalent noise temperature is negligible.
• A cryogenic bandpass filter, which is proposed in [16]. Typically, a filter is not used before the LNA in the RF path since its insertion loss worsens the system temperature of the whole receiver. However, in our case, the band pass filter has been used to reject a strong radio frequency interference near 420 MHz, which is slightly outside the receiver bandwidth. This filter has been designed with a high-temperature superconducting (HTS) technology to reduce the losses before the LNA. A magnesium oxide substrate with a thickness of 0.503 mm and a dielectric permittivity of 9.65 has been used. The superconducting transmission lines are made by yttrium barium copper oxide (YBCO). The metallic box has been realized in (gold-plated) titanium to match the thermal expansion of the substrate. The insertion loss of this device is about 0.09 dB (plus a ripple of 0.12 dB) at 20 K and the equivalent noise temperature is less than 1 K.
• A cryogenic LNA. The LNA designed in [17] has been used for the P-band receiver. The input and output matching are, respectively, better than 10 dB and 20 dB in the operating bandwidth. The measured gain is 27 dB at the center frequency and the noise temperature at 20 K is less than 5 K in the operating bandwidth.

C. L BAND RECEIVER CRYOGENIC RF PATH
A 125µm thick Kapton vacuum barrier (with dielectric permittivity 3.5 and tanδ = 0.001) and suitable thermal gaps [14] separate the L-band feed, at environmental temperature (300 K), from the cryogenic part of the receiver. The components of the L-band receiver operating at a cryogenic temperature of 20 K, cascaded as shown in Fig. 3(a), are: • The orthomode transducer [14], [18], which consists of a cylindrical orthomode junction (OMJ), and two identical 180 • hybrid power combiners in a double ridged waveguide (DRWG). This component provides isolation of 40 dB, cross-polarization less than −35 dB, and an estimated noise temperature less than 1.3 K at the cryogenic temperature of 20 K.
• A commercial cryogenic switch (model Agilent 8761B) used to isolate the L-band LNA in case the SRT should be equipped in the future with a transmitter for deep space and near-earth applications.
• A cryogenic microstrip directional coupler with high directivity, low insertion loss (less than 0.16 dB) and weak coupling (−26 dB at the center frequency), employed to calibrate the L-band receiver chain by injecting a noise source and a weak coherent comb signal in the RF path [19].
• A cryogenic LNA. The LNA designed in [20] has been used for the L-band receiver. This LNA is composed of three cascaded high electron mobility transistors. The input and output matching are, respectively, better than 10 dB and 15 dB in the operating bandwidth. The measured gain is 38 dB at the center frequency and the noise temperature at 20 K is less than 4 K in the operating bandwidth.

III. RADIATION PERFORMANCE OF THE COAXIAL L-P FEED
In this section we analyze and discuss the performance of the L-P band feed. All the simulations presented here have been performed using CST studio Suite. Photos of the receiver installed on the SRT and the CST model of the feed are reported in Fig. 8. The effect of the metallic box containing the receiver has been considered in the CST simulations, and a metallic frame has been added to the model to account for the presence of the prime focus positioner (Fig. 8).
In Fig. 9 the simulated radiation pattern in the principal planes (E-plane, 45 • , and H-plane cuts) and the crosspolarization (45 • cut) is reported at the center frequency (357 MHz). Table 3 summarizes the feed gain, the crosspolarization level, and the taper at the edges of the SRT aperture (which corresponds to 74 • ) at 357 MHz and at the ends of the operating P-band.
The optimum theoretical value of edge taper for the SRT is about −12 dB at 74 degrees. However, we have designed the feed with a value of the edge taper slightly lower with respect to the theoretical value of −12 dB to reduce the effect of the spill over. The consequence is a slight under-illumination of the primary mirror.
Finally, the improvement of the P-band feed radiation pattern due to the external corrugations (see Fig. 4 and Section II.B) has been evaluated by CST simulations. In Fig. 10, we compare the radiation pattern of the feed at the center frequency of 357 MHz with or without these corrugations. The position of the corrugations has been optimized using the CST model reported in Fig. 4, i.e. without the metallic box containing the receiver. The simulations show that a reduction larger than 10 dB of both the backradiation and the cross-polarization can be observed thanks to the choking effect of the corrugations.
We conclude the discussion on the P-band feed with an experimental assessment of the OMJ, which is included in the same guiding structure of the feed, as reported in Section II.A. This component has been characterized by means of the output return loss measurement, whereas the insertion loss and the cross-polarization have been estimated using CST simulations and theoretical considerations.
The experimental setup for the measurement of the output return loss is shown in Fig. 11. The reflection from the coaxial output of the OMJ has been obtained by terminating the coaxial waveguide input with an Eccosorb R load. The return loss, measured using a Vector Network Analyzer connected at the output of a commercial 180 • power splitter, is reported in Fig. 12 and compared with simulated results. The agreement between simulation and measurement is excellent and the reflection is below −20 dB in the operating bandwidth.
The OMJ of the L-band receiver has been extensively tested both with simulations and measurements in [14] and VOLUME 10, 2022  its performance has been summarized in section II.C. Therefore, we focus here on the far-field pattern of the feed in the L-band. In Fig. 13 the simulated radiation pattern in the principal planes (E-plane, 45 • , and H-plane cuts) and the cross-polarization (45 • cut) is reported at the center frequency   (1550 MHz), whereas table 4 summarizes the feed gain, the cross-polarization, and the edge taper at 74 • at 1550 MHz and at the ends of the operating L-band. As made for the P-band, the L-band feed has also been designed with an average value of the edge taper slightly lower than the theoretical value of −12 dB to reduce the effect of the spill over.

IV. THERMAL DESIGN OF THE CRYOSTAT
To reduce the system noise temperature some devices of the receiver chain have been housed inside a cryostat, a vacuum container with inner coaxial radiation shields, to operate at cryogenic temperature (see Fig. 3(a)). The cryogenic system consists of a two-stage commercial cryocooler (model 350C Cryodyne R [21]). We have set the upper nominal temperatures of the two stages at 20 K and 77 K, respectively. As shown in Fig. 14, the desired physical temperatures are obtained when the heat power dissipation is less than 20 W on the first stage and less than 4 W on the second stage.
To obtain an estimate of the cooling power required by the cylindrical Dewar fabricated for the L-P receiver, a complete cryostat was thermally modeled. We divided the thermal model of the Dewar into three stages: ''stage 0'' (at 300 K), ''stage 1'' (at 77 K), and ''stage 2'' (at 20 K). Basically, three physical phenomena contribute to the thermal load: convection, radiation, and conduction.
Since the typical pressure inside the Dewar is below 10 −7 mBar, the contribution of the convection has been neglected [22].
The thermal radiation [22], [23] depends mainly on the radiating surface of the Dewar and on the emissivity of the material used to fabricate it (aluminum in our case) [22]. The thermal radiation has been reduced using additional ''thermal shields'' made of a cylindrical 2mm-thick structure in aluminum coated with a sheet of polyester superinsulation, composed of 10 layers (e.g. type COOLCAT 2 by RUAG). We have employed two thermal shields, the first between ''stage 0'' (at 300 K) and ''stage 1'' (at 77 K), the second between ''stage 1'' and ''stage 2'' at 20 K (see Fig. 15).
The major thermal contribution due to radiation comes from the L-band OMJ [14] (Fig. 16), which operates at 20 K. Indeed, the backshort of the OMJ ''sees'' the conically tapered section of Styrodur R 3035 CS at higher temperature (∼300 K) as shown in Fig. 16(c). Clearly, thermal shields cannot be used in the internal circular waveguide. Using the emission coefficients of the Styrodur R 3035 CS (estimated equal to 0.08-0.1) and of the aluminum (about 0.05), Then, we need to account for the thermal conduction [23], [24]. Four main contributions should be considered: i) The thermal conduction across the four cylindrical columns of G10 material, used to sustain the mechanical parts of ''stage 0'' and ''stage 1,'' which are separated by a small air-gap ( Fig. 16(a)) [14]. ii) The thermal conduction across the four (flat and holed) supports of G10 employed as a support for the air-gap between ''stage 1'' and ''stage 2'' ( Fig. 16(a)) [14]. iii) The thermal conduction across the four cylindrical columns of G10 material used to sustain the copper plates at 20 K ( Fig. 16(b)) wherein the cryogenically cooled devices are housed (Fig. 15) [14]. iv) The thermal conduction due to the coaxial cables employed in the P-band and L-band RF paths. v) The thermal conduction due to the electrical connections. The contribution of the G10 supports has been estimated by thermal simulations of the OMJ performed with the commercial software Solidworks, whereas the contribution of the coaxial cables and electrical connections has been computed through analytical expressions derived by [22]- [25].
Finally, the thermal load due to the power dissipation of the cryogenic LNAs should be considered.
The results of the thermal analysis are summarized in table 5, showing a total thermal load of 3.8 W and 2.5 W, respectively for stage 1 and stage 2, well below the heat power dissipation limits of the cryocooler (Fig. 14). VOLUME 10, 2022  We have placed three thermal sensors into the cryostat to monitor the temperatures of the inner stages. A sensor monitors the temperature of the first stage (nominal at 77 K) and it is placed on the aluminum plate at 77 K (see Figure 16(a)), two other sensors monitor the temperature of the second stage (nominal at 20 K) and are installed on the P-band LNA and on the L-band LNA. In Fig. 17 the cooldown curve is reported. It shows that the two sensors near the LNAs measure a temperature below 20 K, and the sensor on the aluminum plate measures a temperature around 70 K, as required by the thermal design and predicted by our thermal analysis.

V. EVALUATION OF THE RECEIVER SYSTEM TEMPERATURE
The overall noise temperature of the receiver (T REC ) is one of the key parameters of the front-end. Its value impacts the system noise temperature, which includes the atmospheric contribution and the spillover, and consequently the sensitivity achieved during radio astronomy observations [26]. Each component of the P-band and L-band RF paths described in section II has been characterized with measurement, simulations, and/or theoretical considerations to estimate the final noise temperature of the P-band and L-band receivers. For each passive component of the receiver chain, the equivalent noise temperature has been computed by wherein T REFi is the physical temperature of the component and NF i its noise figure. Then, the final noise temperature of the receiver T REC has been computed using the well-known Friis formula for n cascaded blocks [26], [27]: where T ei and G i are the noise and the available power gain of the i th stage. The predicted values of the T REC are reported in Table 6 for the P-band receiver and in Table 7 for the L-band receiver.
The estimated values of the system noise temperature of the receiver reported in Table 6 and 7 have been assessed by measurements, for both receivers (P-band and L-band) and both polarization channels (Horizontal and Vertical). the DRWG Hybrids and the backshort of the OMJ are connected to copper plates @ 20K (not shown in (a)), the DRWG hybrids pass through holes in the aluminium plate at 77 K (b), without any contact with this plate; (c) cross-section of the OMJ; for more details and geometrical dimensions of the L-band OMJ see [14].
The Y-factor method [28] has been used to perform these measurements. The Y-factor is the ratio of two noise power levels, measured by switching on/off a suitable noise source. These two measurement points are employed to draw a line representing the linear noise power as a function of the noise temperature. Finally, the Y-intercept at the temperature of the device under test (DUT) indicates the noise added by the DUT.
In our case, the required power levels are generated by switching between a load at room temperature (300 K) and the Maury Microwave MT7118A Cryogenic Termination  Chamber, which provides a load at 77 K. Then, the system noise temperature of the receiver at 20 K has been extrapolated, and the results are reported in Table 8 and 9. An excellent agreement with the T REC estimated in Table 6 and 7 can be observed for both polarizations and at different frequencies.
The noise temperature of the L-P receiver of the SRT has been assessed against the noise performance of the receivers of other large radio telescopes for radio astronomy (Table 1) in the frequency range of interest. The results reported in Fig. 18 show the good performance of the proposed coaxial dual-frequency receiver in comparison with the receivers available in the main radio astronomical facilities.

VI. COUPLING BETWEEN THE COAXIAL FEED AND THE SARDINIA RADIO TELESCOPE
To compute the reflector antenna gain, efficiency, and crosspolarization, the coupling between the coaxial L-P feed and  Fig. 3). F = free-space; DW = Dewar; POL = Polarizer; FS = Filter selector.  the SRT has been evaluated using the commercial software GRASP by Ticra. GRASP is a reliable tool for the analysis and design of reflector antenna systems, which employs an advanced Physical optics (PO) algorithm as the baseline analysis method, supplemented by Geometrical Theory of Diffraction (GTD) and Moment Method (MoM) solvers for advanced applications. In the GRASP model of the SRT we VOLUME 10, 2022 TABLE 9. Measured T REC of the L-band receiver (Y-factor method). have considered the blocking effect of both the sub-reflector (M2) and the quadrupods (see Fig. 1). At present, in the absence of measurements from a calibrated celestial radio source, the software GRASP, whose results can be considered equivalent to experiment [29], has been also used to compute the SRT beam pattern. In Figs. 19 and 20 we show the simulated radiation pattern of the SRT coupled with the P-band and L-band feeds, respectively.
The radiation features are summarized in table 10 for the P-band and in table 11 for the L-band. The cross-polarization is lower than −30 dB in the P-band and has an average value of about −25 dB across the L-band, with a worst value of −21.5 dB at 1.55 GHz. The antenna Gain reported in tables 10 and 11 is: 2642 VOLUME 10, 2022  Let D the actual antenna directivity and D M the maximum directivity of the aperture, defined as: wherein R is the radius of the reflector aperture and λ is the free space wavelength. Then, η A is the aperture efficiency, defined as D = η A D M ; η S is the spillover efficiency; η B is the blockage efficiency; η is the overall efficiency. The values of η A , η S , and η B are reported in tables 10 and 11, whereas the return loss efficiency η RL has been estimated as 97% in both L-band and P-band. The Ruze RMS efficiency, the ohmic efficiency, and the cross-polarization efficiency have been neglected since they are larger than 99%. Finally, the Half Power Beamwidth (HPBW) is 0.96 deg and 0.216 deg at 357 MHz and 1.55 GHz, respectively.

VII. CONCLUSION
We have described the design of the cryogenic dual-band L-P receiver of the Sardinia Radio Telescope. This receiver is a unique instrument in the panorama of the radio receivers installed on the radio telescopes available in the world. Thanks to this equipment, the SRT is the only large radio telescope able to observe the sky concurrently in the P-band and in the L-band, but also, separately, in each of these single bands both in linear and circular polarization. The coaxial L-P feed of the receiver has been designed using the commercial software CST Studio Suite. The measurement of the feed radiation pattern was not possible due to the difficulty of finding an appropriate anechoic environment fitting the large dimension of the antenna. However, simulations of the far-field have been performed using CST, which can be considered equivalent to experiment for a wide range of applications [30], [31]. The coupling of the coaxial feed with the SRT has been evaluated using GRASP by Ticra, a highly reliable software for the analysis and design of reflector antenna systems.
Both home-made components and commercial devices have been employed for the realization of the receiver chain. All of them have been accurately characterized by simulations and measurements. Hybrid couplers, coaxial switches, noise injection directional couplers, and low-noise amplification stages are placed inside a compact vacuum vessel and cryogenically cooled at a physical temperature of 20 K by a commercial cryocooler. Two HTS Band Pass Filters in the P-band and an OrthoMode Junction in the L-band are also cryogenically cooled inside the cryostat, allowing to achieve low noise performance. The receiver noise temperature, measured in the laboratory prior to installation of the instrument on the SRT, was in the range 17-21 K in P-band and 10-13 K in L-band. These values match closely the predicted performance.
The successful deployment of the L-P band coaxial receiver has enabled the telescope to operate in the framework of the Very Large Baseline Interferometry and Pulsar Timing networks. Furthermore, L-P band radio astronomical observations are regularly performed in single-dish mode (continuum, spectral lines, spectro-polarimetry and Pulsars), driving numerous important scientific discoveries.
This receiver, in the P-band operation has also been employed for space applications, as space debris detection and tracking, and in the near future will operate within the program Sardinia Deep Space Antenna by the Italian Space Agency.