With Gartner predicting 26 billion connected devices by 2020, it is not surprising that the Internet of Things (IoT) formed a backdrop for this year’s Embedded World Exhibition and Conference. Many already see it as the single most significant trend ever to emerge in the semiconductor industry, and as it continues to transition from concept to commerce, it is likely to broaden its reach.
Many of the most significant announcements made at this year’s event had the IoT at their heart; and while it may still be emerging, the IoT as an industry seems to have really arrived. The potential it offers is now driving developments that will shape the world we live in. Connected devices will be prevalent for decades to come, a real paradigm shift.
While the IoT will undoubtedly play a key role in the near future, the electronics industry has many established sectors that continue to be served. Here is a roundup of the major announcements made at Embedded World 2014 and how they fit into the industry.
With so many intelligent nodes expected to emerge over the next few years, demand for even smaller microcontrollers is also building, and manufacturers are meeting that demand by packing ever-more functionality in to ever-smaller packages. Freescale
used Embedded World to announce the smallest yet: measuring just 1.6 by 2.0 mm, the Kinetis KL03. That makes it more than 15% smaller than its predecessor and the company sees the miniaturization of MCUs as a key driver of IoT evolution. The device also has the ARM® Cortex™-M0+ core at its heart, held up as the industry’s most power-efficient 32-bit ARM core. It also features a high-speed 12-bit ADC, a ROM-based boot loader (which allows for factory-programming and on-line system firmware upgrades), and an internal voltage reference source for the ADC. The inclusion of an ADC will allow the device to interface to analog-based sensors, which will be prevalent in IoT sensor nodes. As Freescale puts it, “when size is no longer a barrier to incorporating microcontrollers into edge-node devices, we can start to redefine what is possible for the Internet of Things”.
Figure 1: Freescale’s Kinetis KL03 is supplied in the smallest package yet.
Also based on the Cortex-M0+ and targeting similar applications, STMicroelectronics
was promoting their recently-announced STM32L0x family, which at 3 mm by 3 mm is not quite as small as Freescale’s KL03, but still offers significant functionality in a tiny package. It has a slightly-lower top frequency than the KL03 (32 MHz as opposed to 48 MHz) but both operate down to 1.71 V, and ST says the L0x continues to operate at full speed across the entire voltage range. Perhaps more significantly, it also continues to operate at the same low supply current throughout its extended operating temperature range, up to 125°C, which ST says makes it the ‘premium’ choice for industrial applications. It also boasts the lowest current-consuming ADC in the industry, according to ST; 48 µA at 100 ksamples/s with 12-bit resolution, while a hardware-based oversampling feature allows it to sample at 16-bits with no increase in conversion current (although it does take longer to perform the conversion).
For IoT applications that make use of a USB connection, some devices in the L0 range offer a crystal-less USB 2.0 FS interface thanks to an integrated 48 MHz oscillator, while also supporting the Battery Charging Detection feature.
While ARM’s Cortex-M family of cores has become the ‘de facto’ architecture for MCUs, there are still many proprietary solutions available and being actively developed. A case-in-point is the brand new Application Oriented Controller MCU families from FTDI
, which as the name implies are aimed at defined application areas. The initial device within the family is the FT51, which is designed to address control systems and is based on the venerable 8051 core. Announced at the same time (during the Exhibition) is the FT900, which uses a proprietary 32-bit RISC core and targets high-performance systems such as video imaging, multimedia and other demanding, cost-sensitive activities involving large amounts of data.
The FT51 reaches 48 MIPS performance and integrates a USB hub (USB technology being the ‘point of origin’ for FTDI), which enables sub-systems to be cascaded. It also features a range of ADC/DAC peripherals that support sensor (IoT) interfacing. The FT900 achieves 293 DMIPS at 100 MHz, which FTDI says is achieved by using shadow RAM to store the program memory, thereby allowing the core to run with zero wait-states. The addition of the 8- and 32-bit devices rounds out FTDI’s MCU family, which already featured the 16-bit Vinculum devices.
Figure 2: Application-Oriented Controller MCU families from FTDI.
Part of the challenge of using low-power MCUs in IoT applications is accommodating the communication protocols needed in devices that, typically, do not always integrate large amounts of RAM or Flash memory. Of course, the amount of memory varies within the family in order to hit different price-points, but sometimes there is no substitute for size and the IoT may be a case-in-point. Addressing this dilemma, Renesas
used Embedded World to announce its new RX64M group of MCUs, based on its new RXv2 32-bit core.
The new devices will be the first products in the RX family to be fabricated on a 40 nm process and use high-speed embedded Flash to also achieve zero wait-state execution at speeds of up to 120 MHz. Renesas says that the expansion of the connected society and IoT markets means there has been a rapid growth in diverse IoT-related products, such as connected network and industrial equipment used to enable ‘smart factories and smart buildings’. This is what is driving the demand for increased memory: to accommodate the wide range of communication functions, adding that increased memory capacities will also play a key role in supporting the further miniaturization of devices while providing security.
To meet this growing need, the RX64M group of MCUs offer up to 4 Mbyte of embedded Flash and 512 kbyte of SRAM, but at the same time offer a 40% reduction in power over previous RX MCUs, while performance is also up; to 504 Coremark, which is 1.6 times greater than older devices. These improvements are attributed to the move to a 40 nm process.
Despite being in long-term decline, the market for 8-bit MCUs remains strong and, in some cases, is even showing some growth. One particular area where Microchip
is seeing demand is in Class B safety-critical applications using 8-bit devices, and to meet this demand it announced the PIC16(L)F161x family. The new family features what Microchip calls Core Independent Peripherals, as well as fault-detection hardware designed for safety-critical applications. This illustrates the company’s strategy to develop devices that use the core less and it includes a Cyclic Redundancy Check with Memory Scan unit that can run without stalling the program, which it intends to add to all new devices. Microchip stated that implementing safety and control in an embedded system often requires extensive code and additional components; the PIC16F161X family is intended to reduce this complexity through the integration of its specialized hardware features.
While the industry is generally in agreement that the IoT will, for the foreseeable future, rely heavily on battery power (as opposed to energy harvesting in the long term), the need for in-the-field charging is not going away, and one of the most significant developments in this area recently has been the advent of wireless charging.
As well as showcasing a new range of motor control MCUs at the exhibition, NXP
was championing a new DSP-based Qi wireless charger device, the NXQ1TXA5, which integrates all the critical circuits needed in a single, 5 mm by 5 mm package. Aimed specifically at mobile phones, it requires fewer than ten external components to build a complete low-power 5 V Qi A5/A11 wireless charging transmitter, alongside the Qi coil and resonant capacitors.
Figure 3: NXP’s new DSP-based Qi wireless charger device.
However, as with many new technologies, more than one solution exists for wireless charging and, recognizing this, IDT
presented what it claims is the industry’s first dual-mode wireless power receiver compatible with both the Wireless Power Consortium (WPC) 1.1 Qi standard, and the Power Matter Alliance (PMA) 1.1 standard. As such, the IDTP9023 can be used to make mobile devices fully compatible with both Qi and PMA charging bases.
The device features a high-efficiency synchronous buck converter combined with an embedded MCU, allowing it to support all currently available WPC and PMA receiver coils, as well as proprietary and PCB-based coils. IDT claims the MCU provides functions that exceed the Qi and PMA specifications, allowing even more sophisticated power delivery systems to be developed. It also uses a proprietary transmitter/receiver back-channel communication protocol to provide compatibility with IDT’s wireless power transmitters.
Figure 4: The IDTP9023 can be used to make mobile devices fully compatible with both Qi and PMA charging bases.
A key element of the IoT is connectivity and, unsurprisingly, wireless connectivity is emerging as the prominent medium. In particular, the advent of Bluetooth Low Energy (BLE), incorporating Smart and Smart Ready, will be an enabling technology.
Digi-Key has been actively expanding its linecard over the last 12 months to include more Wi-Fi, Bluetooth, and BLE SoCs and modules, and its latest signing, Nordic Semiconductor
, was announced at Embedded World.
Nordic’s Bluetooth low energy solutions include the nRF51822
SoC (based on the ARM Cortex-M0) and the nRF51422
SoC, which uniquely supports both BLE and ANT+ wireless technology in a single chip.
Mark Zack, Digi-Key’s Vice President for Global Semiconductor Product, said, “Bluetooth low energy has become the gold standard in low-power connectivity implementations, and the solutions offered by Nordic Semiconductor address this area and provide a welcome addition to our Bluetooth low energy product lineup”.
Thanks to its impressive power savings and compatibility, BLE is expected to play a key role in the IoT and, in particular, in wearable technologies. It means the opportunities for new solutions are growing as rapidly as the desire to connect ‘things’ and, answering this growing need, Toshiba Electronics
Europe was present at Embedded World 2014 to talk about its latest offering in this space, the TC35667FTG. Based on an, albeit, older ARM core (the ARM7), the device integrates the MCU alongside EEPROM and can support server and client functions defined by the Generic Attribute Profile (GATT).
Figure 5: Toshiba’s ARM7-based TC35667FTG can support server and client functions defined by the Generic Attribute Profile (GATT).
Embedded World represents the electronics industry’s largest single event; over 850 international exhibitors and 26,600 visitors (up 18% on last year’s event) flocked to Nuremberg in order to discover where the industry is going and how the IoT will be realized. As usual, they were not disappointed.