While the Internet of Things is a worldwide wonder that promises to connect everything, it’s probably worth remembering that the study of electromagnetism started nearly 200 years ago and was, in a very real way, the birth of radio. It’s also fair to say that the first widespread use of wireless technology was for data, in the form of Morse Code, rather than voice. Today, voice and data often coexist and share bandwidth that is expanding at a phenomenal rate.
A lot of emphasis has been put on the underlying technologies enabling the IoT, such as IPv6, which will enable the packet-switched network to extend its reach far beyond comprehensible limits. As an indication, consider that the size of just one standard subnet in IPv6 is the square of all the addresses available within IPv4. If the entire meta-net was made up only of IPv6-enabled devices, it’s realistic to say we would still never run out of unique addresses.
And this picture doesn’t begin to factor in those devices that are not directly addressable over an IP network. Gateways will exist that will coordinate smaller groups of devices connected in a way we may not even consider to be a network. This will include direct connections between two devices that create and consume data or services that have no place on the wider web. In reality, this conglomeration of disparate networks was the landscape before the IoT and it will still have its examples within the IoT, particularly as we hurtle towards an era of edge networking and distributed AI capabilities. We may have conceived the IoT as a way of connecting every bit of data to every other bit, but the expectation that all data generated will flow over a packet network unprocessed is fast evaporating. Instead, the IP backbone will be reserved for aggregated and pre-processed data, while the raw processing takes place on the periphery.
This creates an environment where using IP-addressable, standards-based communication will not be the only option for manufacturers looking to connect their ‘things’. Over recent times the term ‘wireless communications’ has become synonymous with technologies such as Bluetooth, Wi-Fi, ZigBee and Thread, and in the IoT where endpoints need to be directly addressable over the Internet these technologies will likely be dominant, but for those devices and applications that will be indirectly addressable there are many alternatives to PANs, WLAN, and WWANs, that can offer significant benefits in terms of cost, power, performance, security and functionality.
Manufacturers of proprietary wireless communications offer a powerful alternative to going down the standardized route, but it is one that requires greater design effort and appreciation for the radio frontend. As data becomes king, the demand for wireless communications that can handle voice as well as greater data bandwidths increases, which is driving innovation at the board level. One technology that is enabling OEMs to develop new solutions that are better aligned to a data-centric world is Direct Conversion. This is the technique that converts an RF signal to a baseband without any intermediate steps, and it is being applied across wireless applications because it offers significant benefits in terms of component count, with respect to the incumbent Superheterodyne approach.
Direct Conversion, which is also known as Zero-IF due to the absence of intermediate frequencies, has two additional variants; Low IF and Near-Zero IF. They are differentiated mainly by the frequency of the Local Oscillator that is used, resulting in the presence of an IF. It is particularly applicable in the use of I/Q modulation, which uses two independently modulated carrier signals that are 90 degrees out of phase. When combined with Direct Conversion, I/Q modulation can enable smaller receivers with lower power consumption than a receiver based on a Superheterodyne topology, as it doesn’t require any bandpass filters. In addition, the requirements on the ADCs used to convert the I and Q waveforms are less strenuous than in a Superheterodyne system, which also helps to lower the BoM further.
Despite its benefits, Direct Conversion can suffer from several features that make its use more challenging, including imbalances between the I and Q signal paths, non linearities that cause harmonics, and DC offsets caused by ‘self-mixing’. In particular, large DC offsets can reduce the dynamic range of the receiver and while using AC coupling can help overcome the DC offset, the capacitor can ultimately impact the low frequency element of the baseband signal and, in turn, the decoding algorithm. For this reason, other ways of avoiding DC offset are preferred.
To compound the challenge, a DC offset can vary with time and temperature, as well as the Local Oscillator. Careful design of the receiver demodulator is an effective way of combating DC offset, which starts with a mixer that has excellent Second Order Intercept Point (also known as Input Intercept Point, or IIP2). The CMX994 family of Direct Conversion Receivers has been designed to provide an IIP2 mixer performance of +79dBm and is extremely effective at attenuating DC intermodulation. The family has been designed to meet the Radio Equipment Directive (RED) requirements for digital PMR (Private Mobile Radio), and is ideal for a number of RF applications including digital/analog multi-mode radio, Software Defined Radio, Data Telemetry Modems, as well as narrowband (50kHz, 12.5kHz, 6.25kHz) and wideband (>1MHz) wireless data systems.
As we move closer towards a world where distributed networks support both voice and data, flowing seamlessly, the type of radio technology we need to enable that world is developing. Direct Conversion is an effective solution to creating smaller and more robust RF receivers that can meet those demands, but can present challenges for designers in terms of DC offset. By developing innovative and high performance solutions with DC offset mitigation, CML Microcircuits is pushing the boundaries and giving developers and manufacturers the technology they need to meet tomorrow’s radio challenges.
