RF front end design is deceptively simple once you know which specifications are important. The overall signal chain design tasks are roughly the same from system to system, it’s individual component selection and layout that gets complicated quickly. Luckily there are some basic specifications you can follow in RF front end design to help prevent signal degradation.
Getting started in RF front end design is all about determining the level of integration you need in your signal chain. If you’re designing a small IoT product with relatively short range over a standard wireless protocol (e.g., WiFi or Bluetooth), you won’t need much more than a typical MCU SoC and an antenna. For higher power products, anything that needs to sweep across broad bandwidth, or other applications, you’ll need to carefully consider which components you’re using and the important specifications in your signal chain. Balancing these against your system’s functionality requirements can be difficult, but hopefully this guide will help reveal what your RF front end might look like in a newer wireless product.
RF Front End Design Fundamentals
The RF front end consists of all circuitry needed to interface between antennas and the digital section in an RF system. The digital block contains the processor that sends and receives data, which then needs to interface with the RF source and supporting circuitry in the signal chain. The RF front end contains a number of components that work together to ensure signal integrity throughout the signal’s bandwidth. This includes preparing a signal for transmission, as well as receiving and demodulating a signal received on the Rx side.
The functional block diagram below shows the general topology of an RF front end. Here, we’ve lumped the Tx and Rx sides together using an antenna switch, which allows an incoming/outgoing signals to be routed on the Rx/Tx sides of the signal chain, respectively.
This block diagram shows signal flows on the Rx and Tx side with an antenna switch.
In this topology, the antenna switch forms the I/O, and it controls whether the transceiver circuitry at the I/O is transmitting on the Tx side or receiving on the Rx side. The Rx side can also include an ADC to convert the demodulated analog signal into a digital signal. There are other inputs as well, depending on the modulation/demodulation method; typically a mixer will be used as part of a supeheterodyne receiver on the Rx side to pull the modulation signal off of the received signal.
Since this topology can appear as a single component or spread across many components, you’ll need to decide which type of RF front end design is best for your needs. If you use an integrated transceiver module, you’ll basically have an entire front-end solution in a single component. Gain and output frequency might be controllable over a defined bandwidth via standard digital interfaces (SPI, UART, etc.). You can get the same features and controllability if you use separate ICs and other components, but your PCB layout will be more complicated.
Selecting Components
The components you select for an RF front end design can vary widely. Some SoCs and transceivers integrate the entire front-end into the chip, and you only need to worry about impedance matching the antenna to the RF output. In other cases, such as when you need wideband operation and/or high power operation, everything needs to be designed from separate components and laid out on the PCB.
In designing the signal path shown above and selecting components, there are some important design goals that need to be satisfied. These goals revolve around high frequency signal integrity, preventing crosstalk between different circuit blocks, and ensuring the received signal can be properly demodulated and information recovered:
- Prevent distortion on the Tx side. The power amplifier on the Tx side normally runs near saturation. The input signal should not be so large that it causes compression distortion. In addition, load-pull techniques are normally used to determine appropriate impedance matching to maximize power transfer to the antenna.
- Remove noise on the Rx side. This is why an LNA is normally used on the Rx side as noise needs to be minimized.
- Ensure flat phase and gain throughout the bandwidth. Ensuring flat phase and gain within the relevant signal bandwidth prevents signal distortion.
- Prevent crosstalk between different Rx and Tx channels. In devices with MIMO, different channels can interfere with each other if not isolated properly. In addition, crosstalk needs to be prevented between the RF front end and the digital section in your board.
If you’re looking for components and preparing a layout for your next design, here’s what you’ll need to watch for in your next PCB.
3OIP point and 1 dB Compression
These values are particular to your RF power amplifier on the Tx side and will effectively limit the amount of useful output power your amplifier can provide. The inherent nonlinearity in the amplifier will cause the input-output power curve to saturate. The third-order intercept point (3OIP) and the 1 dB compression point are closely related, and either can be used to set an upper limit on the input power in the RF amplifier. Beyond the 1 dB compression point, the RF amplifier will cause greater saturation and more intense intermodulation products on the output.
Preventing Crosstalk Between Board Regions
This is basically a problem of isolation. The analog section with the RF front end needs to be given its own region in the board, and return paths need to be carefully planned to prevent interference from the digital region into the analog region. The simplest method simply involves placing guard traces along microstrip lines, but high power and high frequency signals need greater isolation to keep noise within desired limits. This is where you need to use an alternative routing scheme like coplanar waveguide routing or substrate integrated waveguides. Once you get to mmWave frequencies, you may need even greater isolation through the use of multiple ground planes, shielding, or electronic bandgap structures.
Relationships between 1 dB compression and intercept points for various intermodulation products. The 3rd order intermodulation products are the most important as they lie closest to the desired bandwidth. The point where the 3rd order products and the desired tone intercept (3OIP) is usually about 10 dB above the 1 dB compresion point. The latter of these tells you the signal level that experiences 1 dB of compression compared to the ideal output signal without saturation.
Noise Figure, Phase Flatness, and Gain Flatness
These values are functions of the components you choose. Different components will provide different levels of gain flatness and phase flatness throughout the relevant bandwidth. Wideband amplifiers with adjustable output can have highly variable gain throughout the bandwidth and phase flatness. Noise at different frequencies will also be amplified by different levels due to gain dispersion. Ensuring desired functionality means you need to carefully match your component bandwidths to your required gain, power output, signal bandwidth, and power consumption.