Available via license: CC BY
Content may be subject to copyright.
Received: 5 May 2023 |Revised: 15 June 2023 |Accepted: 29 June 2023
DOI: 10.37917/ijeee.19.2.13 Early View |December 2023
Open Access
Iraqi Journal for Electrical and Electronic Engineering
Original Article
Power Efficient LNA for Satellite Communications
Haidar N. Al-Anbagi*1,2, Abdulghafor A. Abdulhameed1,3 , Ahmed M. Jasim4,5, Maryam Jahanbakhshi1, Abdulhameed Al
Obaid6,7
1Department of Electronics and Information Technology, University of West Bohemia, Pilsen, Czech Republic
2Department of Communications Engineering, University of Diyala, Baqubah, Diyala, Iraq
3Department of Electrical Techniques, Qurna Technique Institue, Southern Technical University, Basrah, Iraq
4College of Engineering, Design and Physical Sciences, Brunel University London, Uxbridge, London, UK
5Department of Computer Engineering, University of Diyala, Baqubah, Diyala, Iraq
6RIV Lab, Department of Computer Engineering, Bu-Ali Sina University, Hamedan, Iran
7Thi Qar Governorate Council, Thi Qar, Iraq
Correspondance
*Haidar N. Al-Anbagi
Univerzitn´ı 2732, 301 00 Plzeˇn 3, Czech Republic
Email: alanbagi@fel.zcu.cz
Abstract
This article presents a power-efficient low noise amplifier (LNA) with high gain and low noise figure (NF) dedicated
to satellite communications at a frequency of 435 MHz. LNAs’ gain and NF play a significant role in the designs for
satellite ground terminals seeking high amplification and maintaining a high signal-to-noise ratio (SNR). The proposed
design utilized the transistor (BFP840ESD) to achieve a low NF of 0.459 dB and a high-power gain of 26.149 dB. The
study carries out the LNA design procedure, from biasing the transistor, testing its stability at the operation frequency,
and finally terminating the appropriate matching networks. In addition to the achieved high gain and low NF, the
proposed LNA consumes as low power as only 2 mW.
Keywords
Ground station front end, LNA, Noise figure, Satellite communications, Satellite IoT.
I. INTRODUCTION
Currently, existing Internet of Things (IoT) networks offer the
interconnection for hundreds of millions of devices worldwide,
providing an endless list of convenient daily life applications
such as healthcare [1, 2], agriculture [3], and danger detection
and alarming [4, 5]. However, these services are not yet in-
clusive to isolated areas such as oceans, wildwood, deserts,
and south and north poles [6]. Thus, future 6G networks seek
the engagement of satellites into conventional IoT network
towards the aimed global coverage with no topographical lim-
itations creating an emerging new field named satellite IoT
(SIoT) [6]. Out of multiple satellite categories, small satellites
have gained research attention for such engagement because
of their low orbit altitude, the affordable cost to manufacture
and launch, and low path degradation losses [7]. In addition,
most recently launched satellites are small ones, reflecting an
even better opportunity for realistic ubiquitous coverage.
Nevertheless, small satellites are always designed with
size and resource restrictions implying low gain transmitting
antennas and low power transmitted signal [8]. At the re-
ceiver side, the receiving ground terminal will have difficulty
retrieving the original data from such received weakened and
noisy signals. The ground terminal’s operator must then uti-
lize an expensive steerable high gain antenna along with a
well-designed low noise amplifier (LNA) [9] . LNAs are vital
components of such systems to amplify the received weak
signals without adding extra noise [10]. Moreover, achieving
signal amplification with very low noise levels allows more
accurate and reliable signal detection in successive procedures.
Furthermore, the performed high gain LNA helps widen the
margin of the link budget calculation at the receiver side.
In the last decade, many designs of LNA structures have
been proposed for different applications like, IoT applications
[11], magnetic probes [12], and 5G [13]. More relevant LNAs
This is an open-access article under the terms of the Creative Commons Attribution License,
which permits use, distribution, and reproduction in any medium, provided the original work is properly cited.
©2023 The Authors.
Published by Iraqi Journal for Electrical and Electronic Engineering
|
College of Engineering, University of Basrah.
https://doi.org/10.37917/ijeee.19.2.13 https://www.ijeee.edu.iq |110
111 |Al-Anbagi et al.
for satellite communications applications were reported in
[14
–
19] and should be examined in deep in the following
paragraph.
In [14], a 3-stage LNA was designed to serve GPS ap-
plications at a frequency of 1.57 GHz, achieving a gain of
23.89 dB, a noise figure (NF) of 1.77 dB, and a power con-
sumption of 6.54 mW. 2-stage LNA was reported in [15] for
satellite communications at 401.635 MHz resulting in a gain
of 28 dB and a relatively high NF of 3.6 dB. Another 2-stage
LNA for CubeSats at 29.15 GHz was presented in [16], where
the achieved high gain was 39 dB but on the price of high-
power consumption of 420 mW and high NF of 2.8 dB. The
study conducted in [17] was to design a LNA for CubeSats at
13-14 GHz consuming 162mW of power to generate a gain
of 15.5 dB and a NF of 2.4 dB. Serving the same previous
application and at the same operation frequency, the study
in [18] depicted an LNA design that consumes a higher power
of 3.2 W to produce a high gain of 54 dB while maintaining
almost the same NF of 2.3 dB. Lastly, the findings of LNA de-
signed for radar, nanosatellites, and GPS applications in [19]
revealed a wide-band operation frequency of 0.1-2 GHz with
a power gain of 11.3 dB and a high NF of 2.9 dB.
Enlightened by the surveyed LNAs in the literature, this
work presents an optimal LNA design for satellite communi-
cations ground terminals whose novelty is to achieve a high
gain of amplification and yet maintain very low NF. More-
over, the depicted LNA takes into consideration the power
consumption efficiency. As for this aimed LNA, the design
and simulation were conducted in the environment of AWR
Microwave Studio [20]. The rest of this article is organized
as follows: Section II describes the LNA design procedure,
including the biasing circuit, transistor stability verification,
and matching networks. Section III presents the study results
and compares findings versus previous designs. Section IV
briefly concludes the presented LNA achievements.
II. LNA DESIGN
After selecting (BFP840ESD) transistor for the intended
LNA design, the following parts describe in detail the de-
sign procedure, including the biasing, stability, and matching
networks.
A. Transistor biasing circuit
The biasing circuit of a transistor is necessary to ensure
the operation in the active region. Figure 1 demonstrates the
biasing network for the transistor after connecting the biasing
elements.
For this NPN transistor to function as an amplifier, its
emitter-base and collector-base junctions must be forward
and reverse-biased, respectively. The datasheet of the selected
transistor (BFP840ESD) provides the following specifications:
Fig. 1. The biasing network for the transistor (BFP840ESD)
to work as an amplifier
The collector current is (Ic= 2 mA), collector-emitter voltage
(VCE=1 V), forward base-emitter voltage (VBE=0.8 V), and
DC current gain (hFE=250). Assuming the VCC voltage is
1.2 V, the other elements of the biasing circuit are calculated
using the well-known biasing equations and listed in Table I.
TABLE I.
RESULTING BIASING CIRCUIT PARAMETERS
Parameter IBRB1RB2Rc
Value 8µA 10KΩ4.54KΩ96KΩ
B. Transistor stability
The transistor can be either unconditionally stable or po-
tentially unstable based on the magnitude values of input and
output reflection coefficients,
|Γin|
and
|Γout |
. If
|Γin|<1
and
|Γout |<1
, the transistor is unconditionally stable. Otherwise,
the transistor’s stability may oscillate for different loads. The
unconditional stability factor (Rollet stability factor or K fac-
tor) is a measure of the LNA’s stability. It is calculated using
the transistor’s scattering parameters as in equation (1).
K=(1− |S11|2+|∆S|2)
(|S12|2− |S21 |2)(1)
where,
S11
is the input reflection coefficient,
S12
is the reverse
transfer parameter,
S21
is the forward transfer parameter, and
112 |Al-Anbagi et al.
∆S
is the determinant of the reflection coefficients matrix.
The stability of the transistor should be examined at the aimed
frequency of 435 MHz. The AWR Microwave studio offers
two metrics to identify whether the transistor is stable or not.
These metrics are the Rollet factor (K factor) and Stability
Circle Impedance Ratio (SCIR).
In the Rollet factor linear plot diagram, the transistor must
have a value slightly more than 1 at the operation frequency
to be stable. On the other hand, the SCIR is calculated by
measuring the distance from the center of the Smith chart to
the point where the stability circle intersects the unity gain
circle. The SCIR is a useful parameter to evaluate the stability
of an amplifier design because it provides information about
the impedance match between the source and the load. For
the transistor to be stable, the curve line must be outside the
unity circle in the SCIR polar plot.
After terminating the transistor to the input and output
port in AWR Software, as shown in Figure 2 The evaluated
K factor and the SCIR were as depicted in Figures 3 and 4,
respectively.
Fig. 2. Initially generated schematic diagram
Fig. 3. LNA Rollet factor versus frequency
Fig. 4. SCIR polar plot for stability test at 435 MHz
Looking at the simulated results presented in Figures 3 and
4, the Rollet factor has a value of 0.05, while the SCIR curve
lies inside the unity circle, implying an unstable transistor at
this operation frequency. Thus, a stabilizing method should
be involved. The simulation was performed on a range of
(400-470) MHz with a resolution of 1 MHz. Thus, the LNA
operation frequency of 435 MHz is located at the step (36) as
indicated in the legend of Figure 4.
One common method that can be used as a stabilizer is
to connect a series or shunt resistor to the transistor collector.
In this design, a shunt resistor is utilized with a slider to
precisely determine the required value for the transistor to be
stable. The slider is slightly increased and decreased while
observing the Rollet factor and the SCIR plots. Following this
mechanism, the optimal resistor value is 60 Ω.
The stabilizer resistor shunt termination is shown in Fig-
ure 5 and the resulting modified Rollet factor and SCIR dia-
grams are presented in Figures6 and 7, respectively. Figures
6 and 7 show a stable transistor with Rollet factor of slightly
more than one and a curve outside the unity circle in the SCIR
plot.
Fig. 5. Modified schematic diagram after terminating the
transistor to a shunt stablizer
113 |Al-Anbagi et al.
Fig. 6. Rollet factor versus frequency for the stable transisor
Fig. 7. SCIR polar plot for the stable transistor
C. Matching networks
The next step is to evaluate the losses caused by the reflec-
tion of the power at the input and output ports. The amount
of the reflected power at the input and output ports is relevant
to how well-matched these two ports are to the source and
the load. The impedance mismatch between the input port
and the source results in more reflection of the input power.
Similarly, the return losses at the output port are increased
when that port is mismatched to the load. Consequently, the
impedance mismatching causes the efficiency of the amplifier
to reduce.
S11
and
S22
parameters indicate how these two
ports are well-matched to the source and the load.
Impedance matching is crucial to improve the perfor-
mance of an LNA by maximizing the power transfer and min-
imizing the NF through diminishing the losses of impedance
mismatch. Furthermore, selecting the right and suitable match-
ing method depends on several criteria. For instance, a simple
design with a minimum number of lumped elements is pre-
ferred because it is affordable to implement and introduces
less losses. Figure 8 demonstrates the matching performance
of the transistor at the operation frequency of 435 MHz.
Fig. 8. S-parameters of the stable unmatched LNA
Observing the performance of the transistor in the
S11
and
S22
graphs, neither the input nor the output ports are
perfectly matched to the source and load terminals. The
S11
curve indicated a magnitude value of -0.22 dB which implies
a complete reflection of the input power at port 1. On the other
hand, the
S22
parameter curve showed a low reflection value
of -19.92 dB at port 2. Despite the low value of the reflected
power at port 2, the resulting flat
S22
response also inspires the
importance of the matching network for this transistor since
such a wide band response may cause more vulnerability
to electromagnetic interferences. As a result, the transistor
should be terminated to input and output matching networks
to provide the aimed performance at the specified frequency.
The input and output matching networks design starts
with the evaluation of the current values of the input and out-
put impedance values from the Normalized Frequency Circle
Impedance Ratio (NFCIR) and
S22
polar plots, as presented
in Figures 9 and 10.
The NFCIR plot helps visualize and analyze the input
impedance characteristics of a circuit over a range of frequen-
cies. By plotting the normalized impedance on a polar graph,
it provides insights into the impedance variation and the cor-
responding stability of the circuit. To evaluate the output
impedance matching, the output reflection coefficient
S22
is
used in its polar form. The input impedance from the NFCIR
1.805 +2.660i
is taken into Smith chart software after nor-
malized to
50 Ω
and conjugated. The resulting impedance
of (90-133i) is used in a Smith chart software to evaluate
114 |Al-Anbagi et al.
Fig. 9. NFCIR polar plot for evaluating input impedance
the required lumped elements at the input matching network.
Similarly, from
S22
polar plot, we get the output impedance
of
1.251 −0.161i
, which becomes
62.55 −8.05i
after normal-
ized to
50 Ω
. The schematic diagram in Figure 11 shows
the input and output matched transistor with the calculated
lumped elements.
Fig. 10. S22 polar plot for evaluating output impedance
The performance of the transistor after the matching net-
works design is again observed through the
S11
and
S22
pa-
rameters and presented in Figure 12.
Fig. 11. Modified schematic of the matched stable transistor
The
S11
and
S22
curves in Figure 12 are not yet promis-
ing enough for the transistor to meet the requirements of the
LNAs for small satellites. Figure 12 shows an input reflection
coefficient of -0.32 dB which emphasis the almost total reflec-
tion of the input power at the input port. On the other hand,
the resulting reflection coefficient at the output port indicated
the perfect matching of that port to the terminated load.
Thus, the lumped elements were optimized using the AWR
optimizer algorithm of Particle Swarm. Particle Swarm is a
well-known optimization algorithm that can be used to find
optimal values of several variables after setting certain goals
to meet. This algorithm is a built-in option in AWR and was
used to evaluate the optimal values of the lumped elements
to achieve the requires
S11
and
S22
goals of (
<−10
) and
(
<−40
), respectively. These introduced goals must be met
within the frequency range of (430-440) MHz which repre-
sents a ±5 MHz around the operation frequency of 435 MHz.
Fig. 12. S-parameters after terminating matching networks
The optimized lumped elements are terminated to the tran-
sistor as in Figure 13. The optimized lumped elements were
115 |Al-Anbagi et al.
standardized to realizable values according to the market avail-
ability. The depicted values of the lumped elements in Fig-
ure 13 achieved the given gaols of
S22 <−40
and
S11 <−10
for the ±5 MHz around the operation frequency of 435 MHz.
Fig. 13. LNA schematic diagram with optimized lumped
elements
III. RESULTS
The input and output-matched transistor performance re-
sults are presented in terms of the S-parameters, the power
gain, and the NF. In Figure 14, the
S11
and
S22
parameters
show good magnitude values that satisfy the satellite com-
munications design requirements. Furthermore, the designed
LNA offers a power gain of about 26.149 dB at 435 MHz, as
demonstrated in Figure 15. The proposed LNA achieved a
low NF value of 0.459 at the targeted operation frequency,as
presented in Figure 16. The achieved high gain along with
the very low NF were not at the price of design complexity
since the depicted LNA design is a single stage amplifier with
proper biasing and matching networks.
Fig. 14. S-parameters of the optimized LNA
Fig. 15. Optimized LNA’s Power gain
Fig. 16. NF versus frequency for the optimized LNA
Further insight into other achievements from the literature
versus the findings of the presented work are compared in
Table II. Throughout this comparison table, we can clearly see
how the listed previous designs were subject to the tradeoff
to either achieve good gain at the price of poor NF values
or vice versa. Even when the previous LNAs achieved high
gain and an acceptable noise figure, they lacked consideration
regarding power consumption. This study offered an LNA
design capable of achieving high gain at a very low NF and
consuming considerably low power.
These achievements emphasis the novelty of this design to
not only fulfill the aspirations of the satellite communications
requirements but also meet the aim of constructing affordable
and simple amateur ground stations for small satellites.
116 |Al-Anbagi et al.
TABLE II.
COMPARISON OF THE PROPOSED WORK WITH PREVIOUS LITERATURE
Specifications [15] [19] [21] [22] [23] [24] [25] This work
Frequency (GHz) 0.401 0.1−2 0.1−1.45 0.2−1.6 0.13 −0.93 0.1−1 0.434 0.435
Supply voltage (V) 1.08 9 1.8 1.2 1.8 1.2 3.3 1.2
Consumed power (mW) 6.5 - 9.5 0.7-3.1 3 2.7 - 2
Gain (dB) 28 11.3 16.9 6.2-8.8 16.6-19.6 14 16.5 26.149
Noise figure (dB) 3.6 2.8 2.5 2.31-3.68 3.6-5e 4 0.63 0.459
Input return loss (dB) -16.4 -8.9 -17 -20 <−10 <−10 -2.6 -14.06
Output return loss (dB) -18.7 - -15 - - - -16.4 -46.84
Applications
Sat.
Comm. GPS
Wideband
receiver - IoT
Wireless
communication
Sat.
Comm.
Sat.
Comm.
IV. CONCLUSIONS
In summary, the proposed LNA design can serve satellite
communications applications at the amateur’s frequency of
435 MHz. The findings of the simulated LNA revealed a
high gain of 26.149 dB while maintaining a very low NF
of 0.459 dB. Besides the high gain and low NF, the pre-
sented LNA only consumes 2 mW of power when supplied
by 1.2 V. These promising findings come along with 14.06
and -46.84 dBs for the input and output return losses, re-
spectively. Such LNA enables the ground station terminal to
receive, process, and decode the weak signals transmitted by
small satellites when engaged in future SIoT networks. More
promising results at the ground station terminal are assured
if this LNA is confronted by a well-designed band-pass fil-
ter, which prevents the LNA from being overdriven by the
received broad power spectrum.
CON FLICT O F INTEREST
The authors of this article hereby declare no conflict of interest.
They have read and reviewed the manuscript and agreed on
the publication as open access material.
REFERENCES
[1]
A. M. Jasim and H. Al-Raweshidy, “Towards a coopera-
tive hierarchical healthcare architecture using the hman
offloading scenarios and srt calculation algorithm,” IET
Networks, vol. 12, no. 1, pp. 9–26, 2023.
[2]
M. M. Saleh, “Wsns and iot their challenges and appli-
cations for healthcare and agriculture: A survey.,” Iraqi
Journal for Electrical Electronic Engineering, pp. 37–
43, 2020.
[3]
B. N. Alhasnawi, B. H. Jasim, and B. A. Issa, “Internet
of things (iot) for smart precision agriculture,” Iraqi
Journal for Electrical Electronic Engineering, vol. 16,
pp. 28–38, 2020.
[4]
N. K. Jumaa, Y. M. Abdulkhaleq, M. A. Nadhim, and
T. A. Abbas, “Iot based gas leakage detection and alarm-
ing system using blynk platforms,” Iraqi Journal for
Electrical Electronic Engineering, vol. 18, pp. 64–70,
2022.
[5]
J. A. AL-Hammoudi and B. H. Jasim, “Design and im-
plementation of monitoring and warning (iot) system for
electricity poles,” Iraqi Journal for Electrical Electronic
Engineering, pp. 105–111, 2020.
[6]
D. C. Nguyen, M. Ding, P. N. Pathirana, A. Seneviratne,
J. Li, D. Niyato, O. Dobre, and H. V. Poor, “6g internet
of things: A comprehensive survey,” IEEE Internet of
Things Journal, vol. 9, no. 1, pp. 359–383, 2021.
[7]
H. N. Al-Anbagi and I. Vertat, “Collaborative network
of ground stations with a virtual platform to perform
diversity combining,” in 2022 International Conference
on Applied Electronics (AE), pp. 1–6, IEEE, 2022.
[8]
H. N. Al-Anbagi and I. Vertat, “Cooperative reception
of multiple satellite downlinks,” Sensors, vol. 22, no. 8,
p. 2856, 2022.
[9]
H. N. Al-Anbagi and I. Vertat, “Pre-detection combining
of small satellite downlink’s replicas,” in 2022 Interna-
tional Conference on Electrical, Computer and Energy
Technologies (ICECET), pp. 1–6, IEEE, 2022.
[10]
A. S. Abdullah, S. A. Hbeeb, H. N. Al-Anbagi, and
A. A. Abdulhameed, “Power efficient wideband power
lna for wsn,” in IOP Conference Series: Materials Sci-
ence and Engineering, vol. 1076, p. 012011, IOP Pub-
lishing, 2021.
[11]
D. Kim and D. Im, “A reconfigurable balun-lna and tun-
able filter with frequency-optimized harmonic rejection
for sub-ghz and 2.4 ghz iot receivers,” IEEE Transac-
tions on Circuits and Systems I: Regular Papers, vol. 69,
no. 8, pp. 3164–3176, 2022.
117 |Al-Anbagi et al.
[12]
A. A. Abdulhameed and Z. Kub
´
ık, “Switchable
broadband-to-tunable narrowband magnetic probe for
near-field measurements,” Sensors, vol. 22, no. 19,
p. 7601, 2022.
[13]
M. M. Joshi, R. Mathew, P. Sarkar, A. Dutt, S. Tiwari,
and P. Nigam, “Performance analysis of radio frequency
(rf) low noise amplifier (lna) with various transistor con-
figurations,” in 2020 5th International Conference on De-
vices, Circuits and Systems (ICDCS), pp. 88–91, IEEE,
2020.
[14]
A. Siddiqui, A. S. Rajawat, A. Rathi, and G. Mehra, “3-
stage lna for high gain l1 band applications using 180nm
cmos technology,” in 2022 4th International Confer-
ence on Inventive Research in Computing Applications
(ICIRCA), pp. 137–140, IEEE, 2022.
[15]
R. M. Timbo, H. Klimach, and E. Fabris, “A 130 nm
cmos lna for satellite application,” in 2019 IEEE 10th
Latin American Symposium on Circuits & Systems (LAS-
CAS), pp. 253–256, IEEE, 2019.
[16]
F. Alimenti, P. Mezzanotte, G. Simoncini, V. Palazzi,
R. Salvati, G. Cicioni, L. Roselli, F. Dogo, S. Pauletto,
M. Fragiacomo, et al., “A ka-band receiver front-end
with noise injection calibration circuit for cubesats inter-
satellite links,” IEEE Access, vol. 8, pp. 106785–106798,
2020.
[17]
G. Orecchini, G. Schiavolini, P. Mezzanotte, S. Pauletto,
A. Loppi, A. Beltramello, F. Dogo, D. Mani
´
a, V. Palazzi,
G. Simoncini, et al., “Low-noise ku-band receiver fron-
tend with switchable siw filters for cubesat applications,”
in 2022 29th IEEE International Conference on Elec-
tronics, Circuits and Systems (ICECS), pp. 1–4, IEEE,
2022.
[18]
G. Orecchini, G. Schiavolini, P. Mezzanotte, S. Pauletto,
A. Loppi, A. Beltramello, F. Dogo, D. Mani
`
a, V. Palazzi,
G. Simoncini, et al., “Low-noise block downconverter
based on cots and siw filters for ku-band cubesat
transponders,” in 2023 IEEE Space Hardware and Radio
Conference, pp. 1–4, IEEE, 2023.
[19]
S. Zafar, S. Osmanoglu, B. Cankaya, A. Kashif, and
E. Ozbay, “Gan-on-sic lna for uhf and l-band,” in 2019
European Microwave Conference in Central Europe
(EuMCE), pp. 95–98, IEEE, 2019.
[20] M. Office, “In awr microsoft studio,” 2014.
[21]
L. Ma, Z.-G. Wang, J. Xu, and N. M. Amin, “A high-
linearity wideband common-gate lna with a differential
active inductor,” IEEE Transactions on Circuits and
Systems II: Express Briefs, vol. 64, no. 4, pp. 402–406,
2016.
[22]
M. Lee, X. Liu, Z. Yang, J. Jin, and L.-S. Wu, “A com-
pact 0.2-1.6 ghz 20 mhz-bandwidth passive-lna exploit-
ing an n-path 1: 3 transformer,” IEEE Transactions on
Circuits and Systems II: Express Briefs, 2023.
[23]
S. Tiwari and J. Mukherjee, “An inductorless wideband
gm-boosted balun lna with nmos-pmos configuration
and capacitively coupled loads for sub-ghz iot applica-
tions,” IEEE Transactions on Circuits and Systems II:
Express Briefs, vol. 68, no. 10, pp. 3204–3208, 2021.
[24]
D. Kim, S. Jang, J. Lee, and D. Im, “A broadband pvt-
insensitive all-nmos noise-canceling balun-lna for sub-
gigahertz wireless communication applications,” IEEE
Microwave and Wireless Components Letters, vol. 31,
no. 2, pp. 165–168, 2020.
[25]
A. Bouyedda, B. Barelaud, and L. Gineste, “Design
and realization of a compact size active antenna for
uhf satellite communication,” in 2021 97th ARFTG Mi-
crowave Measurement Conference (ARFTG), pp. 1–4,
IEEE, 2021.
