Software-Defined Radio (SDR) and Low Noise Amplifiers Working Together

by Simon Ndiritu, Per Vices

A low noise amplifier is an integral part of an RF system, and it is important to choose an LNA with the right specifications LNAs serve a vital role in software defined radio.

LNA Theory and Fundamentals

A low noise amplifier (LNA) is an indispensable component of any RF system. What differentiates the LNA from other amplifiers is that an LNA amplifies the input signal (a very low-power signal) without a significant impact on the signal-to-noise ratio (SNR). Other amplifiers will increase the power of both the signal and the noise present at its input and introduce some additional noise.

The noise figure of an amplifier specifies the noise performance of RF components, including an LNA. The lower the noise figure, the better the amplifier’s performance. One of the major goals in RF circuit design is to minimize the noise of other components before this device to ensure that minimal noise is introduced into the circuit. To make this easier, the LNA is usually placed after the antenna. The noise figure of an RF chain can be computed using the Friis formula, as seen in Figure 1.

Figure 2 highlights the importance of placing an LNA very close to the signal source of the RF circuit. Consider the two RF circuits in Figure 2: they are made up of the same element but have different arrangements. Using the Friis formula, we can observe that placing a component before the LNA will significantly impact the noise performance of the circuit.

How are LNAs Constructed?

There are three leading technologies used in the development of LNAs, pictured in Figure 3: Monolithic microwave integrated circuit (MMIC), Printed circuit board (PCB) design, and Radio Frequency Integrated Circuits (RFIC (Silicon/CMOS)). MMICs are the component of choice for most of today’s high-frequency designs. MMICs offer several advantages over other RF component technologies, such as smaller size, low cost, and high reproducibility, and repeatable performance. The major downside to MMICs is that they are expensive to produce on a small scale.

The other disadvantage of the MMICs is that they are not tunable after fabrication, unlike PCB-based RF components. The PCB design is a low-cost alternative to MMICs that provides good performance. However, PCB-based RF components are not useful for applications where space is a constraint. LNAs made from PCB design also require octave bandwidth matching.

RFIC (Silicon/CMOS) are not useful for high-frequency applications, as their performance is generally limited to 6GHz. Compared to MMICS, the RFICs have an inferior noise performance; they are expensive and are not tunable after manufacture. Advantages of RFICs include that they are mass-producible, integratable into system on chip (SoC) technology and fairly small size.

LNA Specifications

This section highlights the significant characteristics that specify the performance of an LNA. These parameters are essential requirements that must be considered during the development stage of an LNA.

Frequency Range: The frequency range of an amplifier is an important parameter. All amplifiers have a specific frequency range at which they operate. There is a frequency limit at which signal gain drops below 1, and the device is no longer useful as an amplifier. Most amplifiers experience gain roll-off at higher frequencies; hence using an amplifier outside its frequency range will attenuate and distort the output signal.

Noise Figure: The SNR is the ratio of desired signal power to undesired noise power. The noise figure of an amplifier is the measure of the degradation of the SNR between the input and output of the LNA. When a noisy signal (a mixture of noise and desired signal) is applied to the input of a noiseless (ideal) LNA, the noise component and the desired signal will be attenuated or amplified by the same factor; the SNR will be unchanged. Since the LNA is not noiseless (real), it adds noise to the signal at the output. The noise added is measured by the noise figure—the lower the noise figure, the better the performance of the LNA.

Gain: Amplifiers are nonlinear electronic devices; hence the gain of an amplifier varies with frequency and the power of the input signal; as a result, the gain of amplifiers is specified in the datasheet for a specific frequency range. The amplifier gain is the ratio of the output signal power to the input signal power, and it is often expressed in dB or dBm. The gain is usually characterized when the input and output are matched to a 50-ohm source and load.

Input Return Loss: This is a parameter that measures the amount of power that is delivered to the amplifier by the source. It is the ratio of reflected power to incident power at the input port of the LNA. It is mathematically expressed as RL = -20 log |r| (dB), where r is the voltage reflection coefficient; the ratio of the reflected voltage to the incident voltage.

Output Return Loss: This parameter measures the amount of power that is delivered to the load by the amplifier. It is the ratio of reflected power to incident power at the output port of the LNA.

Output IP3: The output IP3 is a measure of the linearity of the amplifier. The amplifier third-order intercept point is commonly measured using the signal generators and spectrum analyzers. When an amplifier is operated in the non-linear region, harmonics are generated; as the input power increases, the power level of the fundamental and the harmonics also increases at the output. In reality, the compression point is reached, and the output flattens out. The output IP3 is the theoretical point where the power of the fundamental signal becomes equal to the power of the third-order harmonic signal.

LNA Design Objectives

From the previous section, the design goals of a LNA are apparent. The design goals are highlighted below.

Low Noise Figure: The noise figure should be as low as possible. An LNA with a low noise figure will have a less significant impact on the SNR of the signal.

Matched Input and Output: Maximum power transfer can occur only when the amplifier is matched at both the input and the output. The load will reflect power to the source when there is a mismatch.

High Gain and Linearity: An amplifier with a high gain is desirable. Achieving high linearity is also essential because it affects the blocker performance of the receiver.

Low Power Consumption: Modern RF circuits are focused on battery or limited power applications; hence power consumption must be considered in LNA development.

What is a Software-Defined Radio (SDR)?

There are many definitions for “software-based radio” (SDR), but an SDR can be generally defined as a radio that uses software-based digital signal processing techniques for receiving/transmitting radio signals. The fundamental goal of the SDR paradigm is to adopt software applications to perform the radio functions on a computing platform, a shift from the traditional hardware-focused, application-specific approach to radio implementation.

An SDR is made up of an RF frontend and a digital backend. The analog to digital/digital to analog converter block sits between the RF frontend and the digital backend. An SDRs radio front-end contains the receive (Rx) and transmit (Tx) block that can down- or up- convert radio signals. The subsystem of the RF frontend depends on the receiver architecture, but the LNA is an indispensable component of the RF subsystem.

SDRs use reprogrammable DSP segments (field-programmable gate arrays (FPGAs), reconfigurable computing (RC), general purpose processors, etc.) for modulation, demodulation, upconverting, downconverting, and other digital signal processing operations. The user can also upgrade or reconfigure the SDR’s capabilities to, say, the latest radio protocols and DSP algorithms; a big difference between traditional and SDR radios. Figure 4 shows the architecture of Per Vices’ Cyan SDR radio.

Importance of LNA in an SDR

LNAs are vital components in the radio front end (RFE) Rx channel of an SDR. In fact, given the current state of radio technology, an SDR is incomplete without an LNA. As RF signals travel from the transmitter to the receiver, the channel introduces noise into the signals. RF signals also experience attenuation, distortion, reflection, and other forms of interference. All these degrade the signal quality at the receiver, hence the reason for amplification. The RF engineer must optimize the sensitivity of the LNA to improve the sensitivity and performance of an RF circuit (including the SDR).

Using the Friis formula, it’s easy to deduce the impact of a slight variation in the noise figure of the LNA on the noise figure of the two circuits. From the formula, we can conclude that the LNA is the predominant noise source in the RF Rx chain; this also highlights the importance of placing the LNA at the receiver’s front end (or close to the signal source). It is the only arrangement that guarantees optimal performance of the RF circuit and low noise figure of the entire RF chain. The LNA is the primary element that determines the performance of the SDR; once the LNA can effectively amplify the incoming signal above the noise floor, the DSP components can efficiently perform further processing.

Only on certain occasions, such as Military and Space applications, are SDR designed explicitly for a particular purpose. SDR used in Military and Space applications has to be radiation resistant and fine-tuned for optimal performance because the cost of replacement will be high. Other SDRs are typically designed to be general-purpose, and therefore, the noise figure is not optimized for any specific application. With an LNA and appropriate filter, performance can be improved by a range of frequencies.

What and Why our Specific LNA Was Chosen

All Per Vices’ SDR products are matched with a 50Ω SMA input, and they use the MMIC GaAs type LNA for several reasons. GaAs MMIC technology offers high operational frequency, low noise, excellent reproducibility, and high integration density. These LNAs come in a tiny, low-profile package (3 x 3 x 0.89mm), making them suitable for our dense circuit board layouts. Moreover, MMICs are the only LNA technology that can operate in the Per Vice’s frequency ranges, as they are capable of amplifiying frequencies as high as 25GHz because of fewer parasitics. Other advantages of the GaAs monolithic LNA include:

• High gain and linearity of the gain
• Ability to work in harsh and rugged environments
• High operating and storage temperature
• Reliability

SDR Applications Where the LNA Plays a Significant Role

LNAs find applications in critical systems such as radars, satellite communication, and wireless medical devices. Radar is a mission-critical system with both military and civilian applications requiring high sensitivity. One of the challenges with achieving this is that at microwave radar frequencies, internally generated noise (thermal noise generated by the receiver) impacts radar system sensitivity significantly. A LNA will improve the system’s performance along with a RF front end architecture incorporating sensitivity time control (STC) before the LNA to prevent amplifier saturation.

In satellite and aerospace systems, the atmosphere and other sources of interference would have severely degraded satellite signals before arriving at the ground station; this is why all forms of satellite communication use an LNA medical devices and IoT devices also need LNAs to transmit and receive signals. Since LNAs are known to be power-hungry devices, design engineers must optimize them to run off of a battery for an extended period.

Per Vices is a leading RF and digital systems innovator supplying multiple industries with wireless communication solutions. They are an industry leader in developing & deploying COTS high-performance software-defined radio platforms with the highest bandwidth, highest throughput, and lowest latency. Per Vices’ area of expertise covers low latency networks, global positioning (GNSS/GPS), radar, test & measurement, medical, spectrum monitoring, and broadcasting & wireless management industries.

Author Bio:

Simon received his BEng in Electrical and Electronics Engineering in 2013 and MSc in Signal Processing in 2019. He has a wealth of experience in designing harware, firmware and software solutions. His areas of expertise include analogue electronics, digital electronics, mixed-signal systems and digital signal processing. In addition, he has a passion for researching and technical writing. 

(1655)