Ask any Wi-Fi developer what’s the hardest part of their job and you’ll likely hear something along the lines of, “designing an antenna and managing worldwide RF regulatory certifications.” Silicon Labs SiWx917Y ultra-low-power Wi-Fi 6 modules address this with an integrated antenna and worldwide RF-certifications, removing the biggest development headache and saving development time and costs. To lower the bar for exploring the benefits of SiWx917Y Wi-Fi 6 modules, we are excited to introduce a new, simplified Wi-Fi development kit, SiWG917Y Module Wi-Fi 6 and Bluetooth LE Explorer Kit.Wi-Fi Development Kit Overview

The total number of IoT devices is expected to reach 30-40 billion units globally by 2030. Historically, the IoT has been dominated by low-power wireless technologies including Zigbee, Thread, and Bluetooth Low Energy (LE). These are light-weight protocols that offer the level of energy-efficiency required by small, resource-constrained, and battery-operated IoT devices.

However, the rapid evolution of Wi-Fi 6, the most energy-efficient generation of the protocol, is quickly closing the power-consumption gap with the traditional low-power IoT protocols and becoming a viable option for the IoT. Silicon Labs’ SiWx917 ultra-low-power Wi-Fi 6 wireless MCU has been tested to enable multi-year battery life on IoT devices, further proving the case for Wi-Fi’s growth in the IoT space. However, IoT is more than just low power consumption. This blog explains the typical characteristics of the IoT-optimized Wi-Fi and how device makers can optimize Wi-Fi for their IoT devices.

Requirements for Modern Wi-Fi IoT Devices

For IoT device makers, power consumption has been the most complicated design challenge in Wi-Fi. Modern IoT devices and smart appliances are tasked to run increasingly sophisticated applications in smart homes, building automation, hospitals, enterprises, and industrial environments. They’re collecting data from many types of sensors, executing advanced computation locally, connecting to the cloud over multiple protocols, and communicating fluently with various ecosystems. IoT devices need to run advanced security algorithms to keep users safe while also inferencing vast amounts of data through machine learning algorithms (ML) to detect patterns and anomalies. During their multi-year operational lifetime, IoT devices also require ample memory capacity to accommodate frequent software and security over-the-air (OTA) updates.

Consequently, when designing Wi-Fi IoT devices, manufacturers need to recalibrate their perspective on optimizing Wi-Fi for IoT. It’s not only about minimizing power consumption, but also about implementing all the advanced capabilities such as compute, memory, security, AI/ML, ecosystems, and peripherals on compact form-factors and scarce resources of IoT devices while adhering to tight energy regulations.

SiWx917 is the Most Optimized IoT Wi-Fi Connectivity Solution

Silicon Labs’ SiWx917 was built for IoT from the ground up. It’s the most IoT-optimized Wi-Fi solution in the market, comprising a comprehensive ultra-low-power wireless MCU with all the crucial elements of IoT within a footprint of 7×7 mm.

At Silicon Labs, we are dedicated to advancing the Wi-SUN standard and ensuring that our solutions meet the highest industry standards. Silicon Labs has been at the forefront of developing Wi-SUN solutions for over a decade, and this continues with the certification of our EFR32FG25 for FAN 1.1 router and border router applications. Our focus on Wi-SUN FAN 1.1 certification is a testament to our commitment to providing reliable, efficient, and future-proof connectivity solutions, exemplified by being one of the first 3 companies in the world to achieve this certification. Our extensive experience and expertise have enabled us to create a comprehensive portfolio that supports both OFDM and FSK technologies. Our Wi-SUN certified ICs, reference designs, and development tools are designed to simplify the certification process and accelerate time-to-market for our customers.

As the IoT continues to transform industries, design decisions around wireless connectivity components become increasingly complex. Engineers often face the dilemma of choosing between ICs and wireless modules for their IoT applications. Both options offer unique benefits and trade-offs, impacting factors such as design complexity, cost, scalability, and time to market. This blog explores the technical considerations and challenges associated with each, drawing insights from Silicon Labs’ expertise in wireless solutions. We’ll provide a technical comparison between ICs and modules, highlighting the key differences, benefits, and hidden costs associated with each approach.

Understanding ICs and Modules

An IC (SoC) is a semiconductor device that contains a processor, memory, and other functional elements on a single chip. Engineers designing with ICs must develop a supporting PCB, including antenna design, RF matching, power management, and other components.

A module, on the other hand, is a pre-integrated system that includes an IC along with all necessary supporting components, such as RF-pin/built-in antenna, EMC shielding, power supply filtering, RF matching components, antenna components, and worldwide regulatory certifications. Modules are designed to minimize the hardware environment required to implement a complete solution, reducing design complexity and certification efforts.

Design and Development Considerations for ICs and Modules

RF Design Complexity

• IC: RF engineers must meticulously design and optimize antenna layouts, PCB trace lengths, and matching networks. Even slight variations in layout can degrade signal performance, requiring extensive debugging.
• Module: Pre-optimized RF design eliminates the need for complex antenna placement and matching, reducing development time and effort.

Certification and Compliance

• IC: Products using SoCs require separate regulatory approvals for each target market (FCC, CE, etc.), which can be costly and time-consuming.
• Module: Modules come with pre-certified regulatory approvals, significantly reducing certification costs and risks.

Time to Market

• IC: The complexity of RF design and certification can delay product launches by 3-6 months.
• Module: Faster development cycles enable quicker market entry, a crucial advantage for competitive industries.

Cost Analysis

• IC: Lower initial component costs, but higher design and certification expenses. Suitable for large-scale production where economies of scale justify the investment.
• Module: Higher per-unit costs, but reduced design and development expenses. Ideal for low to medium production volumes.

Hidden Costs of IC Design

As highlighted in the Silicon Labs whitepaper, “Six Hidden Costs in a Wireless SoC Design,” designing with SoCs introduces hidden costs that are often overlooked:

• RF Expertise: Hiring specialized RF engineers can cost $100,000 to $200,000 annually.
• Lab Equipment: Spectrum analyzers, anechoic chambers, and other RF test equipment can cost up to $50,000.
• PCB Layout: Achieving optimal antenna performance requires iterative PCB design and manufacturing cycles.
• Certification Costs: Regulatory testing for SoCs can exceed $50,000 over a five-year period.

SiP (System-in-Package) vs PCB Modules

When considering ICs and modules, another factor is the form factor used.

SiP Modules

System-in-Package (SiP) modules integrate multiple components into a single package, offering a compact solution with optimized RF performance with smaller size (<12mm x 12mm), pre-integrated components, advanced packaging techniques (e.g., stacked memory), optimized for high-performance applications and requires careful thermal and mechanical design.

PCB Modules

PCB modules consist of a carrier board with components mounted separately. They offer easier development and in-house prototyping, more flexible design changes, easier second sourcing, and generally larger size (>10mm x 10mm).

SiP Modules for Miniaturized IoT Designs

Silicon Labs’ System-in-Package (SiP) modules, such as the BGM220S, exemplify the benefits of module-based design. These modules feature integrated RF components and shielding, offering superior size and performance optimization for space-constrained applications like wearables and smart sensors. With a compact 6.5 mm x 6.5 mm footprint, SiP modules reduce PCB real estate while ensuring robust RF performance.

When to Choose IC vs Module

ICs are ideal when:

• High production volumes justify the upfront design and certification costs.
• The design team has access to RF expertise and advanced lab facilities.
• Custom RF optimization is critical for the application.

Modules are preferable when:

• Rapid time to market is essential.
• Regulatory certification costs need to be minimized.
• The application requires a compact and standardized design.

A Balanced Approach

The decision between IC and module depends on specific project requirements, resources, and business goals. Silicon Labs’ ability to provide both solutions ensures that customers can seamlessly transition from modules to SoCs as their production volumes and design capabilities evolve. By partnering with a single supplier, companies can protect their software investment and optimize their IoT designs for long-term success.

Silicon Labs, a leader in IoT connectivity, has consistently set new benchmarks in MCU performance, capabilities, and applications. Today, we’re announcing the general availability of the PG26 – a general-purpose microcontroller with a large combination of flash and RAM that features a dedicated matrix vector processor to enhance AI/ML hardware accelerator speeds. This addition to our portfolio is equipped to address a wide array of challenging applications, making it an essential tool for AI/ML-driven applications.

A 32-bit Microcontroller Specifically Designed for Energy-Efficient Embedded Development

The PG26 offers an impressive 3200 kB of flash and 512 kB of RAM, both of which lead Silicon Labs’ Series 2 platform. This combination allows for more complex AI/ML models, larger customer applications, and bigger ML models for more accuracy. It also offers a rich peripheral set with a substantial number of GPIOs which enables better system integration. Its 64 GPIOs and four additional analog-only pins reduce the need for external components, which is also facilitated by an integrated LCD driver that can support up to 288 segments. The high performance, multicore compute platform (80MHz Cortex-M33 Core) is also equipped with Secure Vault™ High to improve efficiency and security. Like other Series 2 MCUs, the PG26 is firmware-compatible across the portfolio, making it easier for developers to work across multiple devices. The PG26 maintains the low-power performance Silicon Labs developers are used to, making it an ideal main microcontroller for sensors and home automation devices.

The PG26 MCU Brings Advanced AI/ML Functionality, More Memory, and Additional GPIO

The PG26 provides developers with a wide variety of options to meet diverse application needs. Its low power consumption and small form factor are ideal for energy-sensitive applications. Its precision analog capabilities and integrated LCD are ideal for applications requiring accurate measurements and display functionalities while maintaining low-power performance. The PG26, with its enhanced AI/ML capabilities, more memory, and additional GPIO, is designed for advanced applications that require high processing power and flexibility. Lastly, the latest PG26 offers the most memory, a comprehensive peripheral set, and extensive GPIO options, making it well-suited for complex and memory-intensive applications like industrial automation, predictive maintenance, and monitoring.

The PG26 Microcontroller Makes it Easy to Transition to Embedded IoT Designs

The xG26 platform (BG26, PG26 and the most recently announced MG26) supports a dual strategy, allowing developers to choose between connected and non-connected configurations based on their specific needs. This flexibility ensures that the xG26 can be used in a wide range of applications, from standalone devices to complex IoT systems. One of the key advantages of the xG26 platform is its firmware compatibility and common toolchain across the entire Series 2 portfolio. This ensures that developers can migrate their applications quickly and easily to take advantage of the latest features and improvements without extensive rework. The xG26 platform also offers footprint compatibility, making it easier for developers to upgrade their designs and integrate new features without significant changes to the hardware layout.

Silicon Labs has a rich history of developing cutting-edge microcontrollers, consistently pushing the boundaries of innovation and reliability. The PG26 builds on decades of expertise and technological advancements, bringing Silicon Labs’ track record of integrating advanced connectivity features to enable seamless communication with various devices and networks. This combination of robust performance, advanced security, and wireless capabilities makes the xG26 a standout choice for modern applications.

We recently had a Matter roundtable at Silicon Labs Works With 2024. It was a fantastic chance for ecosystems, device makers, development partners, and others to learn more about the current state of Matter, its challenges and limitations, what is being done to address its shortcomings, new device types coming in the future, and new markets or use cases that are being worked on.

The goal of the roundtable was to allow people to have a more open and honest conversation, so we did not stream or record the event as has been done in previous Works With events. This allowed for attendees to share their own experiences for bad or good, confusion with the goals of the standard, or just general concerns. It is never easy to hear this but is important to get the open feedback.

One of the pieces of confusion that we heard had to do with the need for multiple apps on your mobile device. Shouldn’t Matter mean I can get rid of all my other smart home apps? I never need to even download another smart home app again right? Well, almost.

Matter as a standard has done great things for providing a baseline of interoperability to greatly simplify the commissioning and standardize IP based IoT communications. This helps so many new and existing device makers avoid inventing ways of doing the same basic commissioning that Matter already provides. It doesn’t mean there isn’t room for improvement in the Matter device setup, but the goal is to ensure that proprietary improvements will be really innovative and not just mildly improved.

When it comes to smart home apps, I see them falling into three categories:

• Generalist smart home apps
• Device feature apps
• Services apps

Generalist Smart Home Apps

There is a lot of simple functionality that can be accessed via Matter, and that means users can do most all of it with just a single app. Matter already has support for most of the popular device types. Put more succinctly, I don’t need multiple lighting apps on my phone just because the smart bulbs happen to be from different manufacturers.

Apps that will handle most of the basic home automation, I call generalist smart home apps. These will handle associating devices with different rooms, scenes, routines, and labelling them to make logical sense with the way we naturally live in our smart homes. Users only need one of these generalist apps on their phone. However, it’s important that we allow users the option to switch between generalist smart home apps if they want. One of the goals of Matter is to avoid manufacturer lock-in by ensuring devices can be paired to multiple smart home apps simultaneously, even if the user will only use one.

From a developer perspective, creating your own mobile app is a high barrier to entry and could mean a sub-par experience for something basic. I recently bought a temperature sensor system that had a hub and a number of remote sensors so that I could monitor the temperature in remote locations in my house (the garage, the attic, and a couple of rooms where the thermostat wasn’t located). My goal was to get a better sense of the temperature swings during the cold winters and the warm summers here in the Boston area. These temperature sensors required me to download the mobile app, create an account on the manufacturer’s website, commission the hub, commission each sensor individually, label the sensors, and only then could I finally achieve what I wanted. Monitoring the system required me to re-launch the app regularly, which on occasion made me log back in just to see the temperature. The worst part is that this particular system had extremely poor range and I couldn’t find the right place to install the hub to make all the sensors happy. I eventually gave up and returned the system.

This entire experience has a lot of overhead just to get a temperature sensor up and running. This really shouldn’t be the case, for either a user or developer. It is easy to say that a temperature sensor doesn’t need its own Mobile App. This is an obvious case where a generalist smart home app could easily provide support for temperature sensors along with its other features and functionality. Apple, Google, Samsung, and Amazon all provide this kind of experience through their current hardware and software offering.

Smart Home Device Feature Apps

But now let’s take something more sophisticated like a robot vacuum. This is a more complex device type that can be setup with a schedule, maps of rooms, excluded vacuum areas, control for mopping vs vacuuming, sensors for indicating battery charge, the level of the vacuum bag, and more. Although Matter does have support for robot vacuums, you can easily see why you may want to have a device feature app on your phone for your a robot vacuum and its more complex management. Now certain simpler functionality could be handled, and I would argue should be handled, via a generalist smart home app means. Start, stop, or notifications could easily be handled by a generalist app. If your robot vacuum starts running while you are in the middle of having a kids party, you may want to very quickly use a voice assistant to tell the device to stop or delay its routine. Perhaps the device gets stuck, and you would like to see a notification pop up on a more convenient screen such as a Matter-enabled TV.

From a user perspective, it should be intuitive when you need a dedicated app.

From a developer perspective, Matter enables simpler device integrations but also provides flexibility for a more complex interface experience via a dedicated device feature app. Building a smart home mobile app is now a choice based on the user experience and the device capabilities, rather than just a necessity in all cases.

For some device classes, Matter will grow in its feature set and the major ecosystem will evolve to provide a “good enough” experience for your device that you can phase out your own specialized app. For smaller companies, this means they can focus on adding new hardware features rather than being required to hire mobile developers to re-create yet another a smart home mobile app.

Smart Home Device Services Apps

Now, for a users there will be situations where a generalist smart home app will be too simple in what it can do but where a device feature app will be too limiting. A device feature app is focused in its usage to interact with a single device or single device type. What a user may want is to integrate with a feature across multiple devices. This is where a services app comes in.

Take for example the case of energy management. Matter has enabled devices to report their energy usage and provide the means to aggregate that data in a single unified place. A services app can interact with multiple devices in the home all in support of tracking energy usage. An energy management app needs to understand big energy consumers, like thermostats connected to HVAC, laundry dryers, laundry washers, and a few other related devices like ceiling fans. It doesn’t need to know about lights or sensors or many other Matter enabled nodes that are just not relevant to that goal. It doesn’t need to provide home automation routines or integrate with a voice assistant. This app may have knowledge of your local utility’s cost for electricity or that you have solar panels on your roof, or a generator hooked up to the house. An energy management app may have dedicated screens showing your energy usage over time, correlating this to the past weather conditions. It could have various modes focused on how you can fine tune energy in your home or business to achieve lower costs or just going more green and comparing to your neighbor’s usage. This kind of app will need access to your Matter network but is not meant to be a substitute for the other kinds of apps, and in particular will not be a generalist app.

This kind of specialized service offering is common on the internet. Take for example Slack. Slack is still hugely important for business messaging. It integrates with a number of other tools to provide a means of getting updates and even interacting all within Slack as a means to provide a better user experience. However, it is not a substitute for those other tools. You can integrate Slack with Jira to get updates and perform some simple operations, but it doesn’t replace Jira. I foresee the same kind of thing will occur with user’s Smart Homes.

Smart Home App Conclusion

Ultimately it’s up to the user to decide how they want to manage their network. If they want only a single generalist smart home apps because their usage is simple enough that’s their choice. But if they need a bit more functionality there is option for a blend of smart home apps based on what they do or which one the user likes best.

Having the option for multiple apps is a good thing because it will allow companies to specialize and innovate, and of course compete. Device makers can focus on building the best-in-class device and how to highlight those unique elements that users desire. Services apps can build a service offering based around a theme or specialized use case without having to support all the Matter enabled device types. For all the typical automation use cases, you have the generalist apps.

To put it more succinctly, Chris La Pre Head of Technology at CSA had this great way to summarize it: “You’ll intuitively know when you need a separate app for your smart home device”.

As wireless communications continue to evolve, the demand for devices that can support multiple protocols simultaneously has grown significantly. This capability, known as concurrent multiprotocol (CMP), allows devices to operate across different wireless standards at the same time, increasing their versatility and adaptability. CMP is a benefit because while devices can traditionally handle multiple protocols, this often requires incorporating multiple radio ICs. CMP simplifies this by enabling devices to achieve the same functionality with just a single radio IC, making it more cost-effective for developers. In this blog, we will dive into the concept of CMP and examine the distinctions between CMP – single channel, CMP with concurrent listening, and CMP with Bluetooth Low Energy (LE) Dynamic Multiprotocol (DMP).

Concurrent Multiprotocol CMP: Single Channel

CMP allows a device to support multiple wireless protocols simultaneously that are based on the same IEEE 802.15.4 standard, such as Zigbee and Thread.

By sharing the 802.15.4 PHY and MAC layers, and with Zigbee and Thread operating on the same channel, this feature enables the device to concurrently receive Zigbee or Thread packets using a single radio (and no time slicing). It achieves the differentiation between the protocols through the unique PANID associated with each protocol stack, a functionality referred to as the Multi-PAN feature.

The key advantage of CMP lies in its simplicity and ability to operate on multiple networks with no decrease in performance other than medium congestion.

Works With 2024 has come to an end, but you can catch every minute of this year’s Virtual event on-demand any time here.

It’s hard to believe that this was our fifth Works With event, which we created specifically to provide a platform for IoT developers and manufacturers to learn, network, and exchange ideas. For just a bit of perspective, the first Works With took place at the height of the pandemic, and the world was changing before our eyes. The way our industry – and the world – has adapted and evolved in that time is hard to describe, but the role connectivity has played in that transformation is everywhere you look.

Creating the Works With agenda and content is a long process of assessing customers’ technical market training needs and providing industry-leading content to satisfy this need. We take excruciating care to plan topics, sessions, and speakers that provide practical value for IoT developers, bring together the growing technology ecosystems, and foster collaboration and interoperability. This year’s programming carried that mission forward with technology-specific tracks focusing on Wi-Fi, Bluetooth, LPWAN, Matter, Compute, and IoT Trends. But one of the most significant changes this year was that we hosted four live, in-person events in Austin, San Jose, Hyderabad, and Shanghai. Taking Works With on the road gave us the opportunity to bring these topics to our customers and partners and connect with the global community of innovators who share our passion for connecting devices and improving lives.

We’re extremely proud that today, Works With is considered a can’t-miss industry event and is still the only place wireless developers can access this level of hands-on training from the best and brightest in the industry at no cost and without traveling. We’re also happy to provide every session and keynote from Works With 2024 on-demand. Check out the highlights:

Day 1 Keynote: Connecting the Future: How AI is Accelerating the Next Wave of IoT Evolution

Silicon Labs CEO Matt Johnson and CTO Daniel Cooley, along with David Niewolny, director of business development for healthcare/medical at NVIDIA, discuss the relationship between AI and IoT.

Day 2 Keynote: Works With 2024: IoT Challenges & Solutions Around the World

We announce the Matter Challenge winner, which was a joint program with Arduino and Mouser, followed by Adam Scraba, director of product marketing at NVIDIA, Niranjan Patil from Samsung Research India, and a panel with Jamie Gomez, the CEO of Habitat for Humanity of Metro Denver, about how IoT is delivering on its promise to transform lives by tackling pressing challenges around the world.

Bluetooth Developer Training Sessions

• BLE201: Silicon Labs: Accelerating End Device Development
• BLE203: Enable Accurate Distance Estimation Using Channel Sounding
• BLE204: Harnessing Ambient IoT: A Leap Towards Sustainable Connectivity

LPWAN Developer Training Sessions

• LPW101: Sub-GHz IoT: An overview of protocols, applications, and key considerations for long-range IoT
• LPW102: In-Depth Analysis of LPWAN Protocols: Technical Comparison of Wi-SUN, Z-Wave, and Amazon Sidewalk for Smart City and IoT Applications
• LPW103: Z-Wave Long Range: Bringing Beyond-the-Home Connectivity

Matter Developer Training Sessions

• MAT101: Introduction to Matter: Why, What, and How
• MAT102: Matter: Is it the Right Technology for my Products?
• MAT103: Matter: Technology and Adoption from a User’s and Ecosystem Perspective

IoT Trends Sessions

• IOT101: Biggest Security Trends and What to Expect
• IOT102: Revolutionizing large scale IoT deployment with Open CSMP: A Cisco and Silicon Labs Partnership
• IOT103: Wired to Wireless for the Industrial IoT
• IOT104: Top Five Medical IoT Trends for Product Designers – Specialist Panel
• IOT105: Simplifying IoT Development with an RTOS
• IOT106: Bluetooth Trends of Today and Tomorrow

Wi-Fi Developer Training Sessions

• WF101: Evolution of Wi-Fi 4 to Wi-Fi 6 and Wi-Fi 7 and How it Affects Your IoT Strategy and Roadmap
• WF201: The Wi-Fi IoT Developer Journey to Achieve Ultra-Low-Power for your IoT Device
• WF302: Build a Wi-Fi Tracking Platform with Silicon Labs SiWx917

IoT Partners and Ecosystems Sessions

• PAR101: Hands-on with Arduino Nano Matter
• PAR202: Edge Impulse Tutorial for Motion Classification and Anomaly Detection
• PAR203: Transforming Smart Home Gateways From Wi-Fi to Multiprotocol IoT and Matter

IoT Edge Computing Sessions

• CMP102: Introduction to Simplicity Studio 6
• CMP101: Defining the Ideal MCU Platform for IoT Development
• CMP103: Edge Intelligence: How to leverage Silicon Labs AI/ML to improve efficiency and performance

Now that Works With 2024 is in the books, it’s time to turn our attention to next year. We’ll never have more time to prepare than right now, and with the recent release of Matter 1.4, continued innovation in Bluetooth, Wi-Fi, and other foundational technologies that are driving our industry forward, we better get started.