In previous columns, we covered several aspects of IoT design, including sensors, connectivity, and software. In this part, we will explore the advantages of designing with single-chip IoT architectures, present and in the future.

The Definite Integral

As we’ve explored in the previous columns of this series, IoT nodes are tiny eyes, ears, and brains that speak a common language to communicate with each other and with higher capability systems. Therefore, they require multiple integrated devices to perform their functions. As the number of devices used in a system increases, so does size and cost. IoT devices are projected to far outnumber PCs and smartphones. This creates, yes, you guessed it – a very competitive market. To play in this fast growing and competitive IoT market, embedded devices need to be able to embrace a faster rate of innovation in a key area – integration.

There is a clear need to densely pack electronics so that they occupy the minimum real estate while preserving full functionality. Therefore, designers seek maximum integration of functionality in a single chip. MCUs are a good example of continued integration in the embedded market.  MCUs pushed standard peripherals, as well as volatile and non-volatile memories, into the otherwise standalone MPU. SoCs pushed even more peripherals, regulators, and clocks into the device. For the rich user interfaces the market has come to love, integration of display driving and touch sensing circuitry is required. We’re not done yet, since there are many more features that we can pack into a single device, starting with programmable wireless, high complexity analog sensor interfaces, and voice commands for connected user interfaces. Another important prerequisite of an IoT device is security - billions of interconnected devices bring significant security challenges. Security done right is security done in both hardware and firmware. Such a hybrid security approach combines the robustness of custom hardware with the upgradability of firmware.

Today there are MCUs that integrate an entire programmable system into a single chip. Such an integrated embedded system makes it fast, easy and cost-effective to implement an IoT system using the powerful, flexible programmable system-on-chip architecture. With the ability to create a variety of sensor interfaces with programmable analog blocks and custom display and communication interfaces with programmable digital blocks, these devices can be adapted to suit a myriad of IoT applications.

Continuing advancements in manufacturing technology have led to increasingly smaller processing nodes.  Smaller processing nodes, in most cases, reduce the overall cost of the chip, as they allow more devices to be manufactured on the same piece of silicon wafer. Performance and power are also directly impacted; smaller processing nodes have lower power consumption and/or increased frequency headroom for higher performance. However, in the future, shrinking nodes raises major challenges for the manufacturing process of IoT devices. This is because in addition to digital capabilities, future integrated IoT devices require high performance analog with low leakage, and faster and lower power radios. There’s no fit-for-all manufacturing node that can optimally fit analog, digital, and radio. This is where the advances in packaging and multi-chip modules (MCM) come in.

Analog loves high supply voltages and big transistors, features that power supply blocks such as LDOs and switching regulators also enjoy. Moreover, analog blocks don’t shrink quite as much as digital blocks do with a smaller processing node. After all, they must worry about current carrying capacities, integrated passive components and transistors that need to have large active regions instead of simple on-off switches that are enough for digital circuitry. This causes analog to become costly at small processing nodes. Analog and power would fit on a 180nm – 90nm die. Digital, on the other hand shrinks quite nicely with a smaller processing node. This is where the smallest processing node that can be afforded for a given chip complexity and integrated memories make sense. Sub 40nm works well for digital chips with MCUs, peripherals, and memories. RF subsystems prefer an optimized bulk process or the fully depleted silicon-on-insulator (FDSOI) concept custom tailored for integrated IoT radio that hold huge promise.

While MCMs can be a “best of both worlds” solution, the choice between MCMs and monolithic IoT devices comes down to the required performance and cost. But one thing is certain, integration will push IoT into applications never deemed possible before!

In the next column, we will be exploring how to remove the remaining two wires from your IoT – the power supply ones! Stay Tuned!

Jaya Kathuria Bindra works as an Applications Manager at Cypress Semiconductor Corporation where she is managing the Embedded Applications Group and Solutions Development using the PSoC platform. She has 14+ years of experience in the Semiconductor Industry. She earned an executive management credential from IIM, Bangalore and holds a BS in Electronics Engineering from the Kurukshetra University.

Nidhin MS works as a Staff Applications Engineer at Cypress Semiconductor Corporation. He has seven years of technical experience with analog, power electronics, touch sensing, embedded computing and connectivity and holds a bachelor’s degree in Electronics and Communication Engineering.