If your design involves a wireless (RF) link, sooner or later you'll have to consider your antenna selection. This applies to “small” RF situations such as IoT nodes, and goes up from there to conventional radios, base stations, and much more. In some cases, the antenna choice is defined by or largely constrained by the application; in others, the designer has many degrees of freedom in making the choice.
There's some irony in the diversity of antenna options: after all, an antenna is a “just” passive component formed by a carefully shaped conductive material (wire or metal, in most cases). Still, that hasn’t stopped creative minds who understand the almost magical aspects of these conceptually simple transducers between electrical and electromagnetic energy (and the reverse) from devising with all sorts of creative approaches. As with all design decisions, there is no single “best” approach, and every option is part of the tradeoff among factors such as efficiency, size, cost, radiation pattern (do you want wide or narrow?), bandwidth (same question), and other more.
[Note that “antenna diversity” is not the same as a “diversity antenna.” The latter is a set of two or more independent but linked antennas, where the one with the best received signal strength and SNR is dynamically selected to feed the input amplifier; a complementary situation is used with the transmit-side signal chain and output amplifier to provide the best radiated coverage.]
The antenna choice is especially critical since in nearly all situations, the operating characteristics of the overall RF product must meet various regulatory mandates, and a poorly designed or chosen antenna can make an otherwise good RF chain look bad and even fail. In fact, an antenna chosen to highlight one characteristic such as radiation pattern may be technically wonderful but cause regulatory-approval problems. These technical challenges are also growing, with standards such as 5G adding to the multiband requirements.
For these reasons, and because antennas are an excellent example of a “could be simple but is not” component, I try to keep up with antenna developments. Just trying to develop a family tree of these components is daunting; I usually start by looking if the antenna is a discrete component or part of the PC board layout itself. Each group has unique features: discrete antennas can be located away from the PC board, if needed, and so offer packaging and installation flexibility, but have a BOM cost; PC board antennas can take on some very intriguing and complex forms, and have zero BOM cost, but also require precious PCB real estate, and their location affects – and is affected by – the PC board.
Among the many articles I have seen and read in just the last few months are these:
The Challenge of Mobile Phone and IoT Antennas (Design News)
Impact of Antenna Design, Tune and Match on Wireless Range (High Frequency Electronics)
Antenna Performance Soars into the Stratosphere (Microwaves and RF)
Flexible Antennas Look to the Sky (Microwaves and RF)
A Hybrid Hexaband Cellular Antenna (Microwave Journal)
This list does not include the many very practical antenna-related features in QST, the publication of the Amateur Radio Relay League (ARRL). Nor does it include a hard-to-believe but real antenna using a seawater fountain, see here and here.
Of course, there's a closely related subject that also needs serious attention: matching the antenna to the source or load for efficiency and minimum VSWR. Many antennas of not have a simple, purely resistive, 50- or 75-ohm impedance, so there's expertise, skill, and art to both matching it and also deciding how to implement the matching circuit (discrete components, or PC board traces), all helped by the venerable Smith Chart, nearly 100 years old, see “The Smith chart: more vital after all these Years” (EDN).
If you really want to dive deep into a world of rigorous numerical analysis combined with what seems like magic, you can look into the EM-field simulation programs which let you analyze antenna performance across multiple parameters and evaluate the tradeoffs. One thing is sure with respect to antennas: we've come a long way from the classic “long wire” used by early wireless pioneers, as well as teenagers building crystal radios.
What's your experience with various antenna approaches and their actual implementations? Do you keep it simple, or go for a sophisticated approach?