Designers have many choices when it comes to wireless connectivity in applications ranging from human interface devices (HIDs) to remote sensors for the Internet of things (IoT). One of the more fundamental decisions that need to be made, and one that many designers still wrestle with, is whether to go with a standards based RF interface such as Wi-Fi, Bluetooth, or ZigBee, or a proprietary RF physical layer (PHY) design and protocol.
The reasons for choosing one over the other are many, but so too are the relative tradeoffs in terms of cost, security, power consumption, interoperability, design time, robustness in the face of interference, coexistence, latency, and certification requirements. Many of these tradeoffs are interrelated so designers must first determine the design requirements, then optimize accordingly.
This article will discuss the factors to be considered when choosing between a standard Bluetooth interface and a proprietary RF protocol. It will then introduce a Bluetooth 5 module, followed by a silicon solution upon which a proprietary protocol can be implemented, with appropriate guidelines for each on how to quickly get up and running.
Proprietary RF pros and cons
The case for proprietary PHY and protocol is strong if a design requires optimization in the direction of security, low power, small footprint, and performance.
Security is critical for many applications, from garage door openers to IoT devices. With proprietary radios, it is addressed in a number of ways. To start, proprietary designs ensure “security-through-obscurity,” in that an RF interface that isn’t well known is harder to hack. There’s also the tendency for proprietary interfaces to be point-to-point, or to operate in closed systems that don’t connect to wider networks, and so stay hidden. Finally, designers of proprietary interfaces are free to develop their own advanced encryption algorithms or tweak established ones, without having to be interoperable with security algorithms from other manufacturers. Just being different is, in itself, a security advantage.
Proprietary radio designs can be advantageous when it comes to ensuring a robust connection in the face of interference from Wi-Fi networks, microwave ovens, cordless phones, and other low-power wireless networks. Without being tied to a standard, designers have the flexibility to make better use of the spectrum using techniques such as direct-sequence spread spectrum (DSSS) and frequency hopping spread spectrum (FHSS). In addition, they can adopt their own preferred coding scheme based on their expected link budget to get higher throughput or lower power consumption.
This flexibility also applies to the data packet structure. Without the packet overhead required to ensure interoperability with standards based wireless devices, the packet structure can be streamlined according to the needs of the application.
From a hardware design point of view, well-understood performance requirements and assurance that those requirements will not change at a later stage, allows designers of a proprietary RF interface to be optimized for space, power, and performance. They can do so by again including only those functions required to meet the application’s needs.
While proprietary RF has many advantages, there are a number of factors that must be taken into consideration. The first is cost: to justify the non-recurring engineering (NRE) cost of a custom RF IC design and associated software, especially for low-cost devices, the expected volume should be >100,000.
Tightly coupled with cost is the design time, especially given the vagaries of RF design and the well-documented scarcity of RF expertise, as well as the time taken to develop the firmware and software required for a successful design.
Bluetooth widely adopted, always adapting
On the other extreme is Bluetooth. Originally designed as a straightforward point-to-point cable replacement technology for HIDs and other devices that were entangling users, it soon became a wireless audio and device-to-device connectivity solution. Benefitting from tight control by the Bluetooth Special Interest Group (SIG), Bluetooth is well understood and designers can be confident their devices will connect and be interoperable with other Bluetooth enabled devices, regardless of the hardware source.
Wide adoption and interoperable devices has resulted in prolific hardware and software, bringing with it lower cost and fast time to market for a design requiring a wireless interface. In addition, Bluetooth has evolved over the years.
It has always operated in the 2.4 GHz industrial, scientific, and medical (ISM) band, starting with GFSK modulation of its seventy-nine 1 MHz carriers, giving a throughput of 1 Mbit/s. This is called Bluetooth Basic Rate (BR). Its adaptive FHSS encoding scheme allows it to continue to remain robust in the face of interferers, even as the IoT brings about more wirelessly connected devices. To get to higher data rates, Bluetooth 2.0+Enhanced Data Rate (EDR) uses π/4-DQPSK (differential quadrature phase shift keying) and 8DPSK modulation, to get to rates of 2 and 3 Mbits/s, respectively.
While Bluetooth is tightly controlled by the SIG, designers need to study closely the changes that came about with the introduction of Bluetooth 4.0 Core Specification in 2010. This introduced Bluetooth low energy (BLE), formerly marketed as Bluetooth Smart. BLE is not backward compatible with Bluetooth Classic, so designers need to be careful here.
The primary goal of BLE is low power. It accomplishes this by moving from Bluetooth Classic’s connection-oriented approach where devices are always connected, to an unconnected approach where they only connect when they need to for short intervals. The applications are wearables like smart watches and sensors for the IoT.
The Bluetooth SIG has continued to improve the specification to meet the various needs of its membership and their applications. For more on how it has evolved, see, “Bluetooth 4.1, 4.2 and 5 Compatible Bluetooth Low Energy SoCs and Tools Meet IoT Challenges (Part 1).”
The latest version, Bluetooth 5, doubles the BLE data rate to 2 Mbits/s from 1 Mbit/s, and increases the range of a 128 kbit/s connection by 4x to up to 50 m by using stronger forward error correction (FEC). The higher data rate allows more packets to be transmitted for a given time slot, so power consumption is reduced as the device can stay in low-power or idle mode for extended periods.
The longer range gives designers more flexibility to trade-off data rate for distance for any Bluetooth device, including Beacons. Beacons are battery driven BLE devices that broadcast their identifier to nearby mobile devices so those devices can perform certain actions when close to the beacon. Popular with advertisers, they also enable precise indoor and outdoor tracking.
However, the SIG implemented another interesting tweak that proprietary RF interface designers can also do: they lowered the overhead-to-payload ratio, requiring fewer transmissions to send a given amount of “real” data, to further reduce power consumption.
What started as a simple cable replacement technology has morphed into something much more useful. As a result, designers are now more apt to look for a quick and easy Bluetooth solution rather than go through the cost and expense of designing their own RF interface.
Getting up and running on Bluetooth
This inclination to opt for a Bluetooth interface is turning into a necessity as time-to-market windows narrow and design budgets shrink. Fortunately, for many designs there’s enough space to accommodate an off-the-shelf Bluetooth module, which will allow the design team to focus on their end application and differentiation.
One such module is the BMD-330 Bluetooth 5 module from Rigado (Figure 1). While there are many modules for Bluetooth, this one is particularly interesting and useful as it has an integrated antenna on the board. Antenna matching and placement is one of the finer arts of RF design, so off-loading it from the designer saves time and helps ensure optimal signal coupling.
Figure 1: The BMD-330 Bluetooth 5 module comes with the antenna and matching circuits included to simplify and speed implementation. (Image source: Rigado)
The module is a complete solution with regulatory approvals, its own on-board DC-DC converter, intelligent power control, and measures 9.8 x 14.0 x 1.9 mm. While the antenna is included, it does need a suitable ground plane to radiate effectively. Also, the area extending out from the antenna portion of the module should be kept clear of copper and other metal, and the module should be placed at the edge of the pc board, with the antenna facing outward.
When mounting the module in an enclosure, make sure there’s no metal near the antenna or it will impact performance. As it is designed and tuned for free-air operation, potting, epoxy, over-molding, or conformal coatings can affect performance, requiring additional measurements after application to ensure the link budget is within specification.
The module is based on a nRF52810 system-on-chip (SoC) from Nordic Semiconductor (Figure 2). This uses an Arm® Cortex®-M4 CPU clocked at 64 MHz, has 192 Kbytes of flash and 24 Kbytes of RAM.
Figure 2: The BMD-330 module is built around the nRF52810 SoC from Nordic Semiconductor, which includes an Arm® Cortex®-M4 CPU and a 2.4 GHz radio. (Image source: Rigado)
This is not much flash space, so Rigado has not supplied any factory firmware on the module. As there’s no bootloader, any firmware needs to be loaded using the serial wire debug (SWD) interface. Once that is done, however, Nordic provides a wide array of protocol stacks called SoftDevices. These are pre-compiled, pre-linked binary files that are downloadable from the Nordic website. The BMD-330 with the nRF52810 SoC supports the S132 (BLE Central & Peripheral) SoftDevice, and the memory-optimized S112 (BLE Peripheral) SoftDevice.
Key specifications for the BMD-330 module include a transmit power of +4 dBm and a receiver sensitivity of -96 dBm (BLE mode). It runs off a 3 volt supply and draws 7.0 milliamps (mA) at +4 dBm and 4.6 mA at 0 dBm in transmit mode. In receive mode, it draws 4.6 mA at 1 Mbit/s and 5.8 mA at 2 Mbits/s. Both transmit and receive specs assume the DC-DC converter is enabled: the current increases when it’s disabled.
Proprietary vs. Bluetooth sweet spot
Between a full custom proprietary radio design and standard Bluetooth, there is another option: an off-the-shelf radio transceiver around which designers can develop their own protocol and coding schemes, or adopt off-the-shelf versions such as Ant, Thread, or ZigBee. With the falling cost of available silicon and a wide range of software support, this may be the “sweet spot” for designers looking for differentiation, some latitude for optimization, and the option to enhance security, all while keeping costs to a minimum and design schedules intact.
A good option for designers interested in this route is Silicon Labs’ EFR32FG14 Flex Gecko proprietary protocol family SoC (Figure 3).
Figure 3: Silicon Labs’ EFR32FG14 Flex Gecko provides a solid hardware platform around which proprietary software can be added or developed. (Image source: Silicon Labs)
Like the BMD-330, the EFR32FG14 also uses an Arm® Cortex®-M4 core, but running at a maximum of 40 MHz instead of 64 MHz, as the chip is very much targeted at low-power IoT applications. It has up to 256 Kbytes of flash and 32 Kbytes of RAM. Note that the chip supports both 2.4 GHz and sub-GHz (915 MHz) operation, and guides to antenna network matching are provided. The chip also includes support for antenna diversity, to mitigate the effects of frequency-selective fading.
A number of flexible I/O and security features are also built in, including: a 12-channel Peripheral Reflex System that allows for autonomous interaction of MCU peripherals; up to 32 GPIOs; and an Autonomous Hardware Crypto Accelerator and True Random Number Generator. The power amplifiers for both 2.4 and sub-GHz operation are also integrated on chip.
To help in the development process, there is the Silicon Labs SLWRB4250A board for the EFR32FG line (Figure 4). It includes the SoC, headers, crystals, and antenna matching circuits, as well as software.
Figure 4: The SLWRB4250A Flex Gecko radio board provides the hardware necessary to experiment with a proprietary, low-power wireless interface. (Image source: Silicon Labs)
There are many reasons to choose either a full proprietary RF design route or a standard Bluetooth radio. Each has its place when it comes to meeting the design and application requirements in terms of cost, time, performance, size, security, and many other factors. However, for designers who want many of the cost and time-saving benefits of off-the-shelf silicon, as well as the flexibility to add some level of proprietary differentiation, vendors are now also providing solid hardware platforms to build upon.