Efabless Logo
Self-Interference...
public project

Title: Self-Interference Cancellation Low Noise Amplifier for 5G and beyond.                                

Abstract: This is a proposal for IEEE PICO design contest. Conventional radio transceivers operate in either TDD or FDD that is half-duplex. The main hurdle behind the implementation of the full-duplex transceiver is Self-Interference. Self-Interference is the leakage of TX signal towards its RX when uplink and downlink are at the same frequency. This project explores the cancelation of SI in the analog domain at the LNA which is the first active element in the RX chain. Proposed LNA inherently cancels the leakage signal. LNA cancels the SI while providing gain to the actual RX signal. For the proof of concept an anti-SI is modelled as a single component and injected at different additive and subtractive nodes of LNA. The main advantage of incorporating SIC in LNA into the transceiver chain is that it fulfills the requirement of analog cancellation of SI with less hardware. Furthermore, the receiver chain's overall NF will remain the same because of the absence of additional SIC components in the RX chain.

Introduction and Problem Statement: In-Band Full-Duplex (IBFD) radio is the enabling technology for 5G systems and Self-Interference (SI) is the core problem behind the implementation of IBFD radios. Many researchers have proposed several techniques for Self-interference cancellation (SIC). The most popular design proposed by Stanford[1] requires an adder block preceded by the Low Noise Amplifier (LNA). This could results problems and complications: degraded Noise-Figure (NF) and impedance mismatch. In a system of cascaded devices where each device adds its additional noise and noise are amplified successively, it has become evident that the noise and gain of the very first stage dominate the overall noise of the cascaded system, as told by Friss’s Formula. LNA is the first active and vital component of the receiver chain. If any additional component is added before LNA to suppress SI, it is necessary to keep the NF and gain of the device low and high, respectively. The solution opted by the Stanford and Rice designers may degrade the overall NF and gain of receiver chain.

Proposed Idea: The proposed solution considers the above mentioned problem. It is proposed to design a wideband LNA and enhance its capability so that it can suppress SI within it. In this way, the overall NF and gain of receiver chain will remain unchanged. The proposed idea is achieved by using the additive or subtractive nodes in an LNA and injecting anti-SI signal at these nodes in current form. In this way, the SI is subtracted within an LNA in current form using very basic Kirchhoff’s Current Law as mentioned in Equation (1).

iTotal=i(RX+SI)-iSI=iTotal=iRX                                                                                  (1)

Diagram of Proposed LNA:

Figure 1: Block Diagram of Proposed LNA

 

Schematic Diagram of Half Circuit of Differential LNA:

Figure 2: Schematic diagram of Half circuit of proposed Differential LNA

Description of Half Circuit of Differential LNA:

The ultimate design approach for an LNA for 5G are inductor-less design, wide bandwidth, and linearity. As the MOSFETs node scale down, the resistor and inductor do not scale as the transistor scales down. RFIC designers focus more on the latest transistor node for high ft, single-chip integration, and low cost. Therefore, inductor-less designs are preferable. As more and more radios are being integrated on a single device, wideband LNA is a better choice. Conventionally an RF filter preceded by narrow-band LNA blocks the blocker. However, in the case of wideband LNA, the blocker can destroy the biasing of LNA. Therefore, LNA must be highly linear. The proposed LNA is, inspired from [2], comprised of Common-Source (M1) and Cascode (M4) for high gain, current reuse device (M5) for gain with low power, and Common-Drain (M2) for partial noise cancellation. Author in [3] improved the design to cancel the noise of M2 and improved gain by adding additional stage. In this proposal, it is proposed that the two additional devices (M11 and M12) are added at two different nodes, highlighted as “1” and “2”, of LNA to suppress SI. M11 and M12 work as current sources. The anti-SI signal is injected using these devices to suppress the SI towards GND.

Design Goal:

  1. Implementation of SIC capabilities within wideband LNA by exploiting its differential property for 5G.

Target Performance:

Factors:

Value

SIC

>20dB

Bandwidth

Sub-6GHz

Supply Voltage

1.4-2V

Gain

>15dB

S11

<-10 dB

Power

20-25mW

NF

<4dB

Area

0.02mm2

P1dB

>-20dBm

Teamlead: Engr. Hamza Saleem (i201305@nu.edu.pk) NUCES(FAST-NU)

Team Members:

 Hamza Atiq NUCES (FAST-NU)

 Mudassir Ali NUCES (FAST-NU)

 Dr Hassan Saif NUCES (FAST-NU)

 Dr Rashad Ramzan NUCES (FAST-NU)

References:

[1]       D. Bharadia, E. McMilin, and S. Katti, “Full duplex radios,” SIGCOMM 2013 - Proceedings of the ACM SIGCOMM 2013 Conference on Applications, Technologies, Architectures, and Protocols for Computer Communication, pp. 375–386, 2013, doi: 10.1145/2486001.2486033.

[2]       R. Ramzan, S. Andersson, and J. Dabrowski, “A 1.4 V 25 mW inductorless wideband LNA in 0.13 mm CMOS,” in IEEE International Solid-State Circuits Conference, 2007, pp. 424–425.

[3]       S. Jamil, M. Usman, H. Atiq, and R. Ramzan, “28-32 GHz Wideband LNA for 5G Applications,” 2021 1st International Conference on Microwave, Antennas and Circuits, ICMAC 2021, 2021, doi: 10.1109/ICMAC54080.2021.9678221.

 

 

If you are collaborating on this project, please click here to access your collaboration files, and click "Accept Share" in the actions column if you haven't done so already.

Owner

Hamza Saleem

Organization URL

http://isb.nu.edu.pk/rfcs2/

Summary

This is a proposal for IEEE PICO design contest. Conventional radio transceivers operate in either TDD or FDD that is half-duplex. The main hurdle behind the implementation of the full-duplex transceiver is Self-Interference. Self-Interference is the leakage of TX signal towards its RX when uplink and downlink are at the same frequency. This project explores the cancelation of SI in the analog domain at the LNA which is the first active element in the RX chain. Proposed LNA inherently cancels the leakage signal. LNA cancels the SI while providing gain to the actual RX signal. For the proof of concept an anti-SI is modelled as a single component and injected at different additive and subtractive nodes of LNA. The main advantage of incorporating SIC in LNA into the transceiver chain is that it fulfills the requirement of analog cancellation of SI with less hardware. Furthermore, the receiver chain's overall NF will remain the same because of the absence of additional SIC components in the RX chain.

Version

Final

Category

opamp