In the last section, you learned all about how to create the design files that will bring your project from just a figment of your imagination all the way to a working prototype and then into a design that is ready to be manufactured. In this section, I will give you an overview of the process to take those design files to market.
Step 6: Manufacturing Prototypes to Production
If you have made it this far, you have a design file that represents the PCB artwork to be laid out in copper. This is where a real board is finally created from all of your careful planning and months of researching, breadboarding, and designing. You have many options to turn that artwork into a real, physical board and then assemble that board with real components.
Manufacturing Options By Phase
The first PCB’s that you bring up will not be perfect. You will not go straight to production. Despite all of your hard work in schematic and layout, you will fail to connect everything perfectly. It is highly recommended to build just a few boards in the first batch and then bring up just those few to ensure that the artwork is correct before ordering thousands upon thousands more. Then, a pilot build will help hone your process for the onslaught of thousands or even millions of boards. Do you want to find out that your process is broken after you have developed 100 boards or one million boards?
PCB Manufacture Options
· DIY laser printer & etching
· Circuit Printers
· Quick-turn board shop
· DIY Toaster oven and hand-soldering
· Prototype assembler
· Quick-turn board shop
· Prototype assembler
· Domestic vs. foreign board shops
· Domestic vs. foreign assemblers
· Self-purchase parts, distributor, or assembler purchase
Prototyping Manufacturing Phase – DIY Options
The very first time you build a PCB for a new design, the quickest and cheapest way to get it done in very low volumes is to simply do the work yourself. There is some upfront investment in tools that has to be made, but you can continue to use those tools for future projects. There is also a real risk that you will do things improperly and create your own headaches. It depends on how willing you are to learn a new skill and take a risk that you could end up wasting your time and effort.
Old-School Board Etching
In order to build the PCB yourself, you have some options. The old-school way to build your own board was to purchase copper-clad circuit board blanks, lay plastic traces on the board, and then etch away the exposed copper using a powerful and toxic etching solution. This method is fine as long as you have large pin pitch components. It won’t work for today’s modern 1/2 mm pitch components.
Hand-drawn circuits with marker and etching kit
Laser Printer Board Etching
To reach a finer degree of precision, a laser printer can be used to print the Gerber artwork from a CAD program on special transfer paper, available from PulsarFX as a PCB Fab-in-a-Box kit. A laser printer’s toner is essentially powered plastic that is fused to the paper. By fusing the toner to transfer paper, the idea is that you can reheat the toner and it will release from the transfer paper onto your blank copper circuit board, which is part of the kit. You will need a high-temperature laminator, and it can be tricky to find one that works just right. In my experience with this method, it was a little spotty. Thankfully, you can trace over any incomplete traces with an ordinary marker pen, and that will help resist the etching compound from removing the copper at the gaps in the traces. However, you lose some of the precision that was gained by the laser printer process. In order to create a multilayer board, you must print out each layer and etch it separately, and then glue the layers together. Then, you have to drill through any of your through-holes or vias and make a connection between the layers with solder. With enough practice and experimentation, this method can be successful.
Laster printed etch resist and final product with SMT LEDs
Direct Circuit Printer
A recent invention in circuit board prototyping is the circuit printer. There are many exciting developments in this area that will surely produce useful tools for your prototyping work in the near future. Some notables to check out are: Cartesion Co’s Argentum and the soon-to-be-produced 3D circuit printer from Voltera. The big concern that I have for these devices in the near term are with fine pitch devices.
Argenutm and Voltera Circuit Printers
Quick-turn PCB Services
The best choice for many makers even with these available DIY choices is to simply order a quick-turn board from a board shop. They can be found with very reasonable prices and ship the boards in a few days. You can still assemble the board yourself, which is where you will get more value.
However you acquire your PCB, the next step in the DIY process is to assemble the board. Before you can do that, you have to source the parts. This is a time-consuming activity. It is nice if you can find all of your parts from a single vendor, but often you will have to order from several and coordinate shipping so that you aren’t waiting on a single part to start your build. You can’t start meaningful assembly until all of the parts arrive.
When ordering parts, be prepared to make decisions that you never had any idea that you would need to make. You may know the exact major electronic components you will be ordering, but there are a collection of supporting components that need to surround those major components. You will find that there are many more types of resistors, capacitors, crystals and other miscellaneous parts that have a million variants. You have to specify things like brand, package, size, value, tolerance, material, lead or no-lead, temperature, wattage, operating voltage, etc., when all you want is a 1µF capacitor!
Your parts arrive packaged in little antistatic baggies or boxes. Now you have to match up each one of those components to a refdes on the schematics and formulate a plan. I line all my baggies up in a shoe box in the order that I will apply them to the board and prepare notes for myself so that when the assembly starts, I don’t forget what I am doing.
If you are planning to have your boards assembled by an outside company, you can save money by ordering the parts yourself. Assemblers will charge a markup on any of the parts they order on your behalf. Just be sure to ask your assembler what kinds of machines they will use so that you get the right kind of packaging. An automated “tape and reel” machine requires that your parts are delivered in that fashion. If you don’t order a full reel of parts, can your assembler accept “cut tape” parts that have been cut off of a reel?
Parts ready for assembly
SMT Board Assembly
You can assemble your SMT board yourself with the help of a toaster oven and solderpaste. If you had your PCB developed by a board shop, you have the option to order a stencil for the solderpaste. They are laser cut, and you can use this to spread solderpaste across the board, laying it down only on the areas where a component pin will land on a pad. You can get by without a stencil, but there is a higher chance of solder bridges, which are bits of solder that connect adjacent pads. Solderpaste needs to be kept refrigerated, and you only have an hour or so before the solderpaste becomes a liquid, although it depends on the temperature of the room. I have extended it to a few hours with no issues. The hardest part of placing components by hand is to get them to land on the solderpaste cleanly, and then not bumping them when you are placing nearby components.
Solderpaste on a board and then the parts are placed and baked
Commercial SMT ovens bake the boards using a precise temperature profile to ensure that all of the solder melts and you don’t cook your components too long. You can purchase a home SMT oven for a few hundred dollars, which includes the temperature profile. It’s all automatic, which is great. I have only used a regular toaster oven from a household store. The directions that I found online were to simply insert the board in the oven, crank the temperature to 400 degrees, then watch for the solder to melt. The grey solderpaste becomes shiny silver when it melts. You can watch this process happen through the window of the toaster oven, as the temperature washes over the board, in my case from the corners inward. Once all of the solder turns silver, you wait about 15 seconds more and then remove the board from the oven. There is risk with this method of some cold solder joints, and you will have to touch up any solder bridges you find with flux and a fine-tipped soldering iron. Use a 10x microscope if you can find one, and look for issues. Take your time. It is better to find the problem now than when you power up or your board or debugging a software issue. Use a multimeter to check for shorts between power and ground and between traces that are hidden. I have been able to get several good boards out of this process in the first and only time that I tried it. Your mileage may vary.
Prototype/Pilot Manufacturing Phase – Outsource Vendor Option
The low-volume board development process is not cheap nor fast. If you want it fast, prepare to spend a lot more. Quick-turn PCB builds can be done for a premium in 24 hours, if you are on a tight schedule. Things get a lot more affordable as you approach a 5-day turn, and become most affordable at a 2-week turn. You will have to quote the boards in many different volumes because you may be surprised to find that building 100 boards isn’t ten times as expensive as building 10 boards. It can be just twice as expensive, for example, and may be worth it to get as many prototype boards as you need for your testing purposes.
When you receive your prototype boards, the first thing to do is a visual examination to look for copper traces that may be touching where they are not supposed to be touching. You will need to refer to Gerber files and layout to compare the real PCB to the design files. Measure a few points with a micrometer. Once the board passes visual checks, you can probe the board with a Volt Ohm Meter (VOM) and ensure that there is no short circuit between the power and ground pins of any device. This is heartbreaking if you find it, because it can be hard to trace back where the error occurs. All vias pass through any internal power and ground planes, and if there was not sufficient anti-pad clearance inside those layers around all vias, you could get a power and ground short. You will need to then check every signal versus power/ground to find the offending signal. If you pass that check, you can then measure to ensure that all signal traces don’t short against power or ground. If you find that all of the boards check out, move on to the assembly step.
The assembly of your electronic components onto a low volume of prototype boards can be very expensive if completed by a prototype assembler vendor. Be prepared for some pricey quotes! To make matters worse, the prototype assemblers are not in control of the schedule. They have large repeat customers that will knock your product to the back burner no matter what they may have quoted you regarding turn times. A fast turn time for the assemblers is five days. It takes time for the assembler to examine the BOM, order the parts, receive the parts, configure the tape-and-reel machines (if used for your volume order), find pin #1 on all devices in your silkscreen and layout files, and then do hand soldering touch ups after your boards have been run through the automated assembly process.
You may not have a test procedure for the prototype builds, but it would be helpful to define manufacturing tests by the time the pilot builds occur. The pilot builds are meant to test your process against a larger volume of manufacturing that will follow.
Volume Production Manufacturing Phase
When moving from pilot builds to manufacturing builds, you biggest concern will be cost. You certainly don’t want to pay the pilot build prices or you will go broke. You must choose between domestic and foreign assemblers and board shops, keeping in mind the tradeoffs between cost, language barriers, shipping charges, and travel abroad to fix issues. Some assembly shops have domestic pilot lines with foreign volume lines, and those can be a good choice because their teams already work together. This phase is more about business strategy than engineering, and approaches will vary greatly between developers.
That’s the big picture overview of what lies ahead if you plan to bring your own gadget to market.
Today we announced the latest addition to the Wireless Gecko portfolio with the release of the first dual-band multi-protocol and sub-GHz wireless SoCs. Designed to support proprietary sub-GHz protocols as well as ZigBee, Thread, and Bluetooth standards, the new devices make it possible to use the same multi-protocol device for operation on multiple bands, reducing cost, simplifying design, and decreasing time to market.
The Connect Networking Stack
To address the broad range of applications built on proprietary wireless protocols, we’ve released the new IEEE 802.15.4-based Connect software stack. Connect abstracts the complexity of low-level radio configuration and network formation so developers can focus their time on application development.
Think of it as a template for peer-to-peer, star, and extended star networks.
With the needs of the ever-changing IoT market in mind, the combination of these new SoCs, tools and the Connect stack represent what we think is a major step forward in helping to ease the complicated wireless development process. The low active currents, innovative sleep modes, and superior radio performance of the SoCs, coupled with the customizable nature of Connect serves as a great starting point for IoT application development. Connect supports all device types (end nodes, coordinator, range extender) and is optimal for applications that require low-power consumption. The included sample applications and easy-to-use tools can help accelerate time-to-market. With Connect, everything for network formation and radio configuration is already done- all that is left is adding the application on top.
The Connect stack is built on top of our Radio Abstraction Interface Layer (RAIL). RAIL is essentially a library of functions built on top of the radio itself. Example code for things like collision avoidance, address filtering, and other features is built on top of the RAIL layer and a graphical user interface (GUI) lets you customize RF parameters so you can get what you need from the radio. RAIL brings in the needed configuration that enables you to build your own proprietary protocol on top of it.
Whether you are using sub-GHz or 2.4 GHz bands, Connect simplifies radio configuration and network formation by supporting many combinations of frequencies, data rates, and radio modulations. It is designed for compliance with regulatory specifications across worldwide geographic regions. Connect is ideally suited for applications such as smart lighting, home automation, and wireless sensor networks.
Smart lighting, for example, consists of a network of lights, switches and sensor nodes divided into zones. A smartphone equipped with Bluetooth functionality and a smart lighting application could interface with the Blue Gecko SoCs, for commissioning of the network. The Flex Gecko run the Connect proprietary networking stack for lighting control over sub-GHz network.
Simpler is Better
Just a single set of software tools is required, accessed through Simplicity Studio, to support the protocols. A Wireless Starter Kit is provided and serves as a common hardware platform. Simplicity Studio provides access to a comprehensive set of software tools to aid in development and optimize performance. Simplicity Studio’s Application Builder allows customization of the Connect stack, and the final binary represents the feature set selected by the user. This, along with one-click access to quick start guides and product documentation, enables a user to quickly get started with application development.
From customization to simplification, the Connect networking stack allows developers to ease the software development process and build their own connectivity solution faster.
Get all the details about the Connect networking stack here.
Everyone has teeth. Yet more than one-third of Americans do not have dental health coverage. That’s a big risk for the part of their body responsible for biting, chewing, talking, and more. Alex Frommeyer saw this problem as an opportunity, creating Beam Dental, of which he is co-founder and CEO. His company has changed the way dental insurance policies are defined, helping insurance rates go down while dental health goes up, thanks in part to Beam’s smart toothbrush. Here’s more on how connecting our brushing habits to the Internet is bringing the dental healthcare industry into the future.
How did you get the idea for Beam Dental and the beam brush?
About four years ago, my colleagues and I started an R&D services business, where we had the chance to work on all kinds of medical devices. Around that time, we received a contract from a dental manufacturing company that opened our eyes to the dental industry. We discovered that while there were plenty of startups out there helping build and evolve healthcare, almost nobody was doing the same for the dental industry. So we decided to go after this corner of the healthcare market.
It turns out that only two out of every three Americans have dental insurance coverage. Why is that?
Primarily because it’s commoditized! For any insurance company, it’s always difficult to scope out the right kind of policy rates for the actual need. You try to get as much data and context as possible to make a good match, but there’s always a bit of guess work. This was particularly the case with dental insurance and often meant expensive rates for the kind of coverage people could get, especially in individual or family policies. Those without an employer offering dental benefits are today mostly left out of the market.
Our solution: provide policy holders with a connected toothbrush that operates like any other electric toothbrush, but also collects behavioral data and sends to a mobile device via Bluetooth. This data helps us build a better picture of the likelihood of future claims and volume, helping us adjust policy rates based on good brushing behaviors. A connected toothbrush gives us an advantage when identifying how to insure our customers, especially when compared to our competitors. Over time, we expect to lower dental premiums because we can become more accurate in our predictions. The better you care for your teeth, the lower your premiums.
We created our own smart toothbrush, along with replacement heads, toothpaste, and floss. Every one of our members gets enrolled into our Beam Perks program that ships them the beamâ brush, and replenishes a supply of brush heads, toothpaste, and floss every three months.
The point of this program is to show how much we care about prevention. We don’t just tell people to brush their teeth; we’ve investing in everything needed for our members to practice good dental care. We know that by providing these goods, and encouraging the preventative care, we can change long-term behavior for the better, and ultimately create better dental health over time.
What’s really exciting is that our data is showing us that we ARE affecting behavior in a positive way. Our users are brushing their teeth almost two minutes per brushing on average. Every dentist wishes their patients would brush even just one minute per brushing, and our users are doubling that. They are brushing longer and more frequently by a factor of almost two.
How did you go about designing the beam brush?
First off, we had to figure out how to manufacture this product within our economics of selling insurance plans. The more plans we sold, the more toothbrushes we have to make, leaving an extremely small margin of error for cost of goods.
For our design requirements, I feel like we looked at everything: performance, size, power, cost; everything went through the ringer. Once we decided on Bluetooth for communication, the Silicon Labs BLE112 started to stand out. It met all of our technical requirements and the module was pre-certified, which is a major time and cost saver when trying to bring a product to market in the healthcare industry. And while cost was a big driver for us, so was performance. With a Silicon Labs Bluetooth Smart Module, we could do some of the processing on the module itself, accomplishing tasks at the board level. On top of all that, this had to be a battery powered device, so power consumption management was key.
You’ve certainly found a perfect application for a smart IoT device. In your opinion, what does the future of IoT look like?
There is still so much ground to cover. If IoT is going to be a homerun, we haven't even made it to first base yet. There are exciting things happening: machine learning, quantum computing, wireless, renewables. All these things are converging thanks to small, high performance, low-powered, connected sensors and devices. That opens up tons of new opportunities.
The toothbrush is one of the most common consumer products in existence. When we started a few years ago, something like the beamâ brush was just barely financially possible to produce. Today, there are an increasing number of compelling connected product. And in five years, I’m sure there will be new capabilities we can’t predict yet. There's no telling where that will take us.
I am writing this blog/book to help makers create a real product and introduce it to the market, with a focus on embedded firmware design. You should be able to read these lessons and have a good idea how to get things working in embedded firmware and understand the hardware and firmware development process.
This chapter will introduce you to the process and terminology that you will encounter as you develop your embedded hardware project and bring it to production. It is by no means an exhaustive coverage on this subject and is intended to give you an overview of what is possible. You will find a great deal of information out there on each of these topics if you look for it.
Product Development Process
It all starts with an idea. You found something that the market needs, or is just a super cool idea and it absolutely must be brought to market. I’ve been there. Most people would have no clue what to do next. Hardware engineering is HARD, and don’t let anyone convince you otherwise. But you might have a lot of experience in software, or perhaps you are good with your hands, or like me, you are just crazy enough to dive in and try to build this all by yourself. Good for you! You can do it if you never give up. Here is the general flow of how things should proceed.
Step 1: Hacked Prototype
When creating your own project from scratch, the first thing to do is to look around and see if you can find something similar. Tear it apart and figure out what makes it tick. Modify it however you want to get your point across. It can be ugly, and it’s OK at this stage to be really rough. When pitching your idea to others, it helps to have something that may be clunky but gets the point across. I see this all of the time. You can get an artist to render something awesome looking on paper or even with animations, then bring out your clunky hacked prototype with duct tape and glue and show your audience that your product is already in the process of becoming real. It’s not just an idea any more. This step is optional, but you will gain more insight into the problem that you are attempting to solve and gain feedback from others.
Accelerometer hacked into a pre-existing foam sword
Step 2: Part Selection
When you have shared the idea, gained valuable feedback, and feel that you are ready to move forward to creating a manufacturable product, the next step is to select your electronic components, also known simply as your parts. This is a very big step. This is the sole focus of another chapter; it’s that important. You should try to find support from the part distributors like Arrow and Avnet, as they have application engineers and connections to manufacturing representatives that can help you find the parts that you need for your application. There are millions of parts available from thousands of vendors and suppliers. There are also a million details that you have to get right so that all of your chosen parts will work together. Will they run on a common voltage? Can they talk to each other using the same types of communication interfaces? Can they perform quickly and efficiently enough? This step takes a while, and you might revisit this step after you have already begun your PCB development, because it is very tricky to get this step right the first time. But one way or another, you will need to pick parts and start moving forward with a design.
Step 3: Breadboard Prototyping
My book is aimed squarely at breadboard prototyping. This is the step in which you have already picked your parts, you have decided that you know how to use them, and now it is time to make sure that you really understand all of the implications of your particular part. It starts with dissecting the pinout and making sure that you understand the specs. Sometimes part specs are written in foreign countries and translated to English, making the intentions less than clear. Some part specs are hundreds of pages long. If you miss a footnote on a single page, it can bite you. It is best to learn exactly how to use the part when you are only using a breadboard. Prototype firmware doesn’t need to be pretty and maybe won’t be as well-performing as the production firmware. Your goal at this stage is to get basic communication and function from your chosen parts. Note that you don’t have to use an actual breadboard, either. I sometimes skip the breadboard entirely and make my connections right on the breakout board, then connect the breakout board to the pins on my Starter Kit.
SMT breakout board and jumpers to a Starter Kit
Step 4: PCB Design (Schematics)
Once you have real-world, first-hand experience in physically connecting your parts together on a bread board, the specs will make a lot more sense, and you should be very familiar with the part pinouts. You now have the experience necessary to enter your designs into a schematic capture program. You will most likely need to create your own schematic symbols, which are little more than boxes that represent the part with pin names and numbers on them. Some tools include a library of schematic symbols, which is great if you can find your parts in there, but in my experience, they don’t ever seem to contain the majority of your parts.
Example Schematic symbol and connections for an RF component
You have a lot of options when it comes to schematic capture programs. Unfortunately, there is no industry-standard tool for makers. They are all inferior to the standard that exists in my mind. These are not easy programs to master! There are free tools, some inexpensive tools, and expensive tools. I will outline a few below that are popular, and you can download a demo or trial software until you find one that you like. All of the listed tools have schematic and layout tools integrated into one. Keep in mind that the tool that you choose for your schematics could heavily impact the type of tool that you choose for layout, especially if you plan on contracting the layout out to an expert.
Step 5: PCB Development (Layout)
The process of placing your components on a Printed Circuit Board (PCB) and drawing the copper traces that connect the parts together is not for the faint of heart. Before I can even describe the process, I need to introduce the terminology...
Whew! Now I can finally explain how to design the PCB using a layout editor.
Note: It is perfectly OK to outsource the layout of a PCB to a professional. This step can be very challenging for the novice. You must have an exportable netlist that is compatible to the layout tool that your contractor uses, so figure that out beforehand.
Example Layout Viewer
Most combination schematic and layout tools will integrate the process of laying out the board. However, it starts with the schematic connections. Once a few of those connections are completed, layout can begin. It is often necessary to create the footprint for your components yourself using a footprint editor and the mechanical drawings found in the part specs. The pins in these footprints are then mapped to the schematic symbol pins. Then, the components are placed on a board. Careful study of the pinouts of components are necessary to ensure that signals can breakout from the chip and connect to the other chips without creating impossible routes. This is the ultimate puzzle! If a trace is blocked in by traces in the area that are surrounding it, then you will have to “drop a via” to another layer and continue the routing on that other layer. You can see the potential for a big mess is very real if you fail to plan ahead.
When routing traces, the layout software uses routing rules as a guide to prevent you from inadvertently placing things too close together. If you don’t follow the rules, your board shop will reject your board with DRC checks and you will have to re-layout the board. Note that even with these rules, many layout tools will still violate the rules and either not warn you, or warn you later when you go to generate the Gerber files. Often, the board shop’s Gerber analysis software will find very small rule violations that are rounding errors, and it is up to you to haggle with the vendor to allow it, or go back into your tool and fix the rounding error. It’s a pain, and it feels like everything is stuck in the stone ages.
When routing your traces on the board, the pin-to-pin connections are shown as a rats nest; all you have to do to make the rats go away is provide a copper path between pins of components that are connected in the schematic. When all rats are gone and all DRC checks are clear, you can then export Gerber files and upload them to PCB board shops for quoting and manufacturing.
Schematic and Layout Tools:
In the next section, we will finish the chapter with the overview of how to get that design built in the real world.
I was at Computex a couple of weeks ago when several customers came up and asked me this question: How do I know which products in the market will work with the product I'm developing? Great question indeed. When a customer decides to go the standards route (e.g. ZigBee HA1.2) how do they know which other products will work with theirs? That brings up the topic of ecosystems.
What is an ecosystem and how does it relate to IoT? The word "ecosystem" can mean many different things to different people. There are hardware, software, and cloud ecosystems that are all related but different. In the end, a few things remain the same:
Diversity is a great benefit of any ecosystem. Multiple players in the same ecosystem allows for more things to be connected in the same setup. It also means the suppliers can specialize and work together to build out a larger community. A bigger community of things means a bigger reach in the market.
Interoperability is closely related to diversity. Interoperability means different "things" from different suppliers can all work together, hence creating a more homogeneous environment. In order to achieve interoperability, a common set of rules is needed and some kind of organization needs to police the compliance to these rules.
Now let's look at the home automation market segment we have been investigating. If you remember from the first blog in this series, I suggested that ZigBee is an ideal protocol for home automation because of the mesh networking capabilities and mature profiles. Within ZigBee, there are many ecosystems: iControl, wink, SmartThings, just to name a few.
These ecosystems provide device makers the benefit of their brand recognition, a list of devices that are all interoperable, and their set of rules. Maybe you’re not a fan of the rules, but it’s precisely the rules that are keeping the devices all working with each other.
There are 3 main types of ecosystems in the ZigBee home automation space
First, on one side there is the proprietary, or closed ecosystem. These ecosystems exist because they have non-standard implementation from special requirements. You see this type of closed ecosystems typically in lighting or commercial and industrial applications.
Second, on the complete opposite end of the spectrum, you will find completely open ecosystems. If you comply to the standard, for example ZigBee HA 1.2, then you can get on the network. However, there is a caveat. The gateway will let your device join the network, but it may or may not recognize all the features or attributes of your device.
Finally, the majority of the ecosystems out there fall somewhere in between. These ecosystems can accept other standards compliant devices. However, in order to fully utilize the features and functions on the devices, you will need to be approved by the ecosystems. Sometimes, that is also referred to as “white listing” your devices.
Once you decide on the type of ecosystem, what are the steps to joining an ecosystem and which should I choose? Even within the ZigBee ecosystems, each is different. This is because within the standard profiles like HA1.2, there are optional features that an ecosystem may choose to implement.
But the basic steps are:
There are many additional non-technical factors that go into selecting an ecosystem: brand recognition, type of devices in the ecosystem, or cost to join. These are things that only the device maker can prioritize and decide.
Ecosystems are complex and involve many players. I shared some key concepts and questions about ecosystems. There are great benefits to being a part of an ecosystem. But at the end of the day, the answer to joining an ecosystem is not always “yes”.
Check out our connected home reference designs here, and don’t forget to read the previous blogs below.
Earlier this week, technology solution providers and leaders from more than 200 cities came together at the 2016 Smart Cities Innovation Summit where they shared ideas on how to use technology to address the needs of their communities. The IoT is integral to creating a smart city, and the role of that data was the focus of the Sensor Networks panel.
Who owns the data captured by sensors?
Tarik Hammadou, Cofounder and CEO, VIMOC Technologies: The data is owned by the city. The city provides the services to the customers. The data is the asset that the city owns. The service provider charges for the service or API, not the data. Cities need a source of revenue to sustain smart cities. The only source of revenue is the data. Then they'll use that data to provide services to the customers and see if it provides the customer with value.
James Stansberry, Senior Vice President and General Manager, Internet of Things, Silicon Labs: The US government believes they own it, so the city owns it, but it's a slippery slope. There are lots of different data types out there, like Siri or Alexa. This type of data should be personal and is owned by the person. Data like the amount of pedestrian bike traffic is collected and belongs to the city. There are usually indicators that they city owns certain types of data, for instance you'll have to opt in.
How do smart cities protect their data and build a secure network?
James: Security is a big issue. Consider it like an onion. Let's start at the center at the chip level, then you get to the software level, then to the protocol stack, and so on. Your network needs to be secure throughout. No unknown or un-secure devices should even go on your network. That's how people hack into your network. There are lots of thoughts around security standards - don't ignore them.
What's the biggest unexpected benefit or challenge when building smart cities?
Scarlett King, Director, IoT Solutions - Smart Connected Communities and Cities Lead, Bosch: Support of the sensor technology used in smart cities is often an afterthought, but it's critical to be successful and can be a failing point if overlooked. The expert needs to bring this to the city's attention: how do you need to implement or deploy your program and be able to run it for years to come? Cities need to think in billions of sensors, not millions or thousands. Their job would be made easier if all the technology were on the same protocol and used the same standards.
What are you most excited about in the future for smart cities?
Scarlett: The speed of data. How quickly can you get me valuable insights, and therefore impact or add value to our customers.
Tarik: A smart city gives you data to understand things better, which means we can make helpful things.
IoT continues to grow and be a part of the future of a more connected world. We look forward to being a part of the integration of technology in everyday life to grow cities and make them healthier, smarter, and more efficient.
The worldwide focus on energy conservation in the last decade has led to tremendous growth in smart meter deployments, which help manage energy distribution and consumption more efficiently than traditional utility meters. The smart meter is a central device that bridges home energy management systems to longrange back haul communication to utility companies.
Utility providers are adding more intelligence to smart meters to differentiate their offerings and allow their customers to make energy choices that best meet their needs. Bidirectional wireless communication enables accurate, real-time utility pricing information to be sent to consumers based on their energy consumption. While “time-of-day” pricing is common today, consumers can now make informed decisions on when to use major appliances in their homes to reduce their energy bill.
The choice of wireless communications technology for a smart meter is a major decision that requires careful consideration of various design choices. The main factors in choosing the optimal communications technology include cost of deployment, security, regulatory compliance, range and power consumption. Several communications technologies are available for wireless connectivity such as Wi-Fi, Bluetooth, ZigBee and sub-GHz wireless. For long-range applications such as the backhaul communication from the meter to a data concentrator and to other meters, sub-GHz technology is a popular choice because of its superior propagation characteristics, long-range performance, low-power operation and the broad access to unlicensed sub-GHz spectrum throughout the globe.
Over the next few years, significant numbers of smart meters are expected to be deployed worldwide. The United Kingdom is one of the most-watched markets for smart meter rollouts with the goal of installing smart gas and electric meters in every home and small business by 2020.
These deployments in the UK will use ZigBee as well as sub-GHz wireless communications to connect 30 million properties to the smart grid. Italy has the second largest number of gas meters in Europe at 21 million and plans to replace 80 percent of these with smart meters over the next five years. GrDF, a natural gas distributor in France, plans to install 11 million smart gas meters over six years starting in 2015. The 169 MHz and 868 MHz frequency bands are widely expected to be used in all EU smart meter implementations.
According to a report by Pike Research, the installed base of smart meters in China will grow to 377 million units by 2020. The deployment will be split between various wired and wireless technologies, and a large portion of these meters is expected to use sub-GHz wireless devices in the 470-510 MHz band. TEPCO in Japan has announced plans to install 27 million smart meters over the next decade.
Sub-GHz Wireless in Smart Meters
Sub-GHz wireless technology is present in almost all smart meters today. It is also easy to retrofit traditional meters with sub-GHz wireless communications modules and upgrade services or software over the air. The most common use of the sub-GHz link is for communication between meters and from the meter to a data collector or concentrator.
The sub-GHz network is typically a proprietary network in an unlicensed ISM band such as 902-928 MHz in the US. An emerging trend has been the use of sub-GHz to communicate with in-home appliances. Industry alliances such as ZigBee and Wi-SUN are in the process of standardizing the sub-GHz communication protocol from the physical to the application layer for home energy monitoring systems. These alliances aim to realize interoperability between home appliances and smart meters from any manufacturer with the goal of accelerating the pace of adoption.
The graphic below shows a typical wireless smart meter system connecting the consumer to the utility:
Let’s take a closer look at some of the key considerations in designing a sub-GHz wireless solution for smart metering.
One of the primary advantages of using sub-GHz wireless in any application is the long-range capability of this frequency band. Long-range systems reduce the cost of deployments as fewer concentrators and/or repeaters are required to serve the same number of smart meters. RF waves at lower frequencies can travel longer distances for a given output power and receiver sensitivity. This phenomenon can be seen by using the Friis formula for path loss and is governed by the laws of physics.
Pr〖=PtGtGr (λ/4πR)^2 〗where Pr is the received power, Pt is the transmitted power, Gt and Gr are the antenna gains at the transmitter, and receiver, R is the distance between antennas and λ is the wavelength. As a general rule of thumb, a 6 dB increase in link budget will double the range in an outdoor, line-of-sight environment. Thus, the achievable range in the 169 MHz band is better than the 868/915 MHz bands assuming all else is equal.
As wireless system manufacturers try to squeeze every last dB of performance to get the best link budget, it is important to take into consideration other parameters to make an informed decision on trade-offs. In the commonly used GFSK modulation, lower data rates offer better sensitivity and hence longer range. However, the time to transmit a packet at lower rates means that the transmitter and receiver will have to stay in active modes for a longer period of time, which can increase the overall power consumption.
Increasing the transmit power of the radio is an easy way to increase range (“whoever shouts the loudest wins”), but this approach comes at the cost of higher power consumption. While several wireless ICs provide solutions with an integrated power amplifier (PA), the efficiency of the PA is a key differentiator. For example, Silicon Labs’ Si446x EZRadioPRO transceivers require only 18 mA to output +10 dBm or 85 mA to output +20 dBm in the 915 MHz frequency band.
As range tests are highly sensitive to the environment and device parameters, it is often tricky to achieve an accurate, apples-to-apples comparison between RF transceiver solutions from different vendors. Care should be taken to ensure that the radio parameters such as frequency, transmit power, bandwidth, packet structure, antenna, and the method of calculating Bit Error Rate (BER) or Packet Error Rate (PER) are all comparable. The table below shows ideal link budgets for different data rates based on currently available transceiver solutions.
Ultra-Low Power Consumption
Another key design consideration for a wireless smart meter is power consumption. Low-power operation is a critical concern for battery-powered water and gas meters and slightly less so for electric meters. Battery-powered meters typically use Lithium-Thionyl-Chloride (LiSoCl2) batteries, which need to last for 15-20 years with low duty cycle operation.
These LiSoCl2 batteries are significantly more expensive than the cost of the other components in the smart meter (roughly 7X-10X the cost of the transceiver IC). Material and Silicon Labs Rev 1.0 4 labor costs involved in replacing spent batteries frequently are also much higher than the cost of adding expensive batteries upfront. In gas and water meter systems, embedded components tend to spend a majority of time in low-power sleep or standby states so the current consumed in this state needs to be extremely low in the range of tens of nA. The active transmit and receive currents also need to be low especially at low data rates as they lead to longer transmission and reception times.
For example, it will take 1.25 seconds (s) to transmit 1500 bytes at 9.6 kbps and only 0.024s at 500 kbps. PA efficiency is a key parameter that affects link and power budgets. A higher transmit power will increase the communication range at the expense of battery life. Some other parameters that should be considered are fast signal detection within a few bits of preamble, fast state transition times to wake up and go back to sleep, and the ability of the radio to autonomously duty-cycle the device without interrupting the host microcontroller (MCU) for every wakeup and sleep event.
Another factor affecting system power consumption is the ability of the radio transceiver to offload the host MCU from typical packet handling functions such as preamble and sync word detection, Manchester coding and CRC calculations. Performing a majority of these functions in the radio allows the host MCU to spend less time processing the packet and frees up memory and MIPS to perform other functions or remain in a low-power state.
As communication from the meter is periodic and typically on a very low duty-cycle, the low standby and receive current is a key benefit. Silicon Labs’ EZRadioPRO transceivers, for example (as shown in the diagram below), consume only 50 nA in standby mode. In addition, the device supports a fully configurable autonomous low duty cycle mode, enabling extremely low system power consumption.
One of the more challenging aspects of using sub-GHz wireless technology is regulatory and standards compliance. For a designer trying to create a worldwide smart meter product, the 2.4 GHz band is available globally, and only the transmission power needs to be adjusted based on the region’s regulatory requirements. However, sub-GHz frequencies vary depending on the region, making it more challenging for hardware and software designers. The ISM band, which is typically where sub-GHz radios operate, is license free. Each country allocates spectrum independently, and some of the common smart meter frequencies are shown in the table and graphic below.
Licensed sub-GHz bands are available in several countries for utilities that may be concerned with interference from other wireless devices. Typically these licensed bands have more stringent regulatory compliance requirements.
For example, in the United States, FCC part 90 applies to a licensed frequency band around 460 MHz. Compliance with certain spectrum mask requirements such as mask D requires extremely good phase noise and narrowband performance at low data rates. More recent integrated sub-GHz transceivers, such as Silicon Labs’ Si446x EZRadioPRO devices, meet the regulatory requirements in this frequency band as well.
Silicon Labs Rev 1.0 6 Most smart meter designs have dedicated hardware for each region or frequency band rather than a common design to meet all worldwide requirements.
It is possible to limit the changes to just component values by selecting a common front end matching topology such that the same layout can be used for various regions. Typically the low-level PHY parameters must be optimized for each region to meet regulatory standards requirements. Significant care must be taken to design a fully compliant solution with the lowest possible BOM cost. Harmonics and spurious emissions have caused many a sleepless night for designers.
To ensure success at the compliance testing lab, it is critical to have accurate power control and filtering on the board to minimize emissions. Packet lengths, data rates and protocol choices such as frequency hopping and number of channels are largely constrained by regulatory requirements. Standards
Compliance In addition to regulatory compliance, designers must be aware of several wireless standards and industry alliances such as IEEE 802.15.4g/e, Wireless M-Bus, ZigBee and Wi-SUN. All indications are that the industry is moving towards a standard solution based on IEEE 802.15.4 (g/e), which will eventually lead to more choices for consumers as it will enable interoperability among end products. This interoperability ultimately will give consumers the freedom to choose their preferred smart energy products, regardless of the utility that provides their energy. In Europe, Wireless M-Bus is the popular choice for smart meters but there is no official certification process. In the rest of the world, proprietary implementations dominate the deployments today.
A worldwide standard acceptable to major players in the smart grid space will help increase the pace of deployments. IEEE 802.15.4g specifies the physical layer, which is typically supported by the wireless transceiver itself. A majority of MAC layer functions are implemented in a software stack that runs on the host processor. For some applications that do not require interoperability, a standards-based software stack may not be optimal especially in terms of memory requirements and architectural choices such as a star or mesh network. Network latency and power consumption are key factors in determining the final implementation.
In these cases, it is common to see proprietary software stacks or a hybrid solution that uses parts of 802.15.4g with a proprietary implementation of the upper layers. Silicon providers today offer standards-based and proprietary stacks that are optimized to run on their MCUs and wireless transceivers and allow for some customization as well.
The key is to provide a clean and simple user interface that hides the complexity of the PHY and MAC within the stack.
Range, power consumption, and standards compliance are some of the factors that define a sub-GHz wireless design. Fast signal detection, ultra-low power standby currents in the tens of nanoamps and faster Silicon Labs Rev 1.0 7 state transition times combined with a robust software solution are a few building blocks that enable new ways to improve smart meter efficiency at the system level.
While Europe and the US are leading the way in deploying smart metering systems, the high-volume growth in this market is yet to come from emerging economies such as the BRIC nations. China and India, the world’s most populous countries with a huge need for secure, energy-efficient metering solutions, are experiencing a growing trend toward the adoption of “smart” sub-GHz wireless communications for smart meters.
In the second post of “Greater China Region IoT Hero Series,” Silicon Labs interviewed Yin Shaoxiang, Manager of Shenzhen Fitcare Electronics Co., Ltd., a leading high-tech company in China, specializing in R&D, manufacturing, and sales of wearable sports and fitness electronics and systems.
With Silicon Labs’ ultra-low power EFM32™ micro-controllers (Gecko Series), heart rate monitor (HRM), and ultraviolet light (UV) sensor solutions, Fitcare’s fitness wristbands, cardio watches, earlobe/finger clip heart rate sensors, and other wearable devices achieve high precision and are comparable to professional fitness devices. Fitcare has been a leader in the Chinese sports and fitness market.
Now let’s learn more about Fitcare and how it has changed the Chinese IoT market.
Please tell us about Fitcare’s company culture as well as its unique strategy for the IoT market
Fitcare’s slogan is “Keep Moving, Keep Fit.” This is unique among hi-tech companies. Our main product line is professional wearable sports and fitness electronics. Our goal is to provide our customers with truly useful reference information and to help them improve their training and exercise results. This is vastly different from manufacturers whose focus is on wearable electronics for general leisure purposes.
Secondly, “precision comes first” is our product design guideline, and “sports and fitness” is our focus. Therefore, our products are far more specialized than various wearable wristbands and smart watches in the market, and we’re able to lead a new area in the IoT market. In fact, since its establishment, Fitcare has been providing high-quality OEM and ODM services. All our founding members have a background in engineering and have been working in the field for many years. We have established great relationships with brand-name clients worldwide and have also accumulated rich experience in the development and production of wearable devices.
Nowadays, like people in America and Europe, Chinese people are paying more attention to fitness. This means a growing and promising business opportunity in a market of a billion people. Fitcare takes this challenge and aims to become the No. 1 “sports and fitness wearable device brand” in China.
Please tell us about Fitcare heart rate watches’ features and benefits.
HW652, Fitcare’s latest generation of heart rate watches, supports low-power Bluetooth transmission and dynamic heart rate monitoring. It adopts Silicon Labs’ advanced Si1144 HRM optical sensor modules and energy-efficient EFM32 Wonder Gecko MCUs, and is equipped with the heart rate monitoring algorithm.
Also, we put a lot of efforts on optimizing optical structures, improving the precision of sensor modules through green LEDs, and adjusting the algorithm carefully for various situations with Silicon Labs. Therefore, the pulse under the skin can be detected accurately. Its error percent is less than 3%, which is comparable to chest straps used by professional athletes.
In addition, the heart rate watch can also monitor users’ daily activities, such as the number of steps taken, calorie consumption, distance, and sleep quality, transmit real-time heart rate data, and store fitness data offline. It is compatible with Bluetooth 4.0 iOS, Android devices, and most sports apps, allowing users to record and track exercise data without wearing traditional chest straps.
Compared with chest straps, heart rate watches are easier to use. When it is as accurate as chest straps, it will gradually change consumers’ preference. Fitcare’s high-precision heart rate watches aims to follow this trend and bring its users “comfort as well as accuracy.”
What made you choose Silicon Labs as your semiconductor partner?
As said before, “precision comes first” is our product design guideline. Silicon Labs helps us fully prepare for designing professional sports and fitness wearable devices. They not only provide us with Si1144 HRM optical sensor modules, which have the best precision in the industry, but also equip energy-efficient 32-bit EFM32 Gecko series MCUs with its own algorithm. This makes the precision of heart rate watches reach the professional level.
Secondly, since wearable devices require endurance, components’ power consumption is also a key factor when we evaluate our suppliers. Silicon Labs is known for its low power technology, and it can provide the most energy-efficient MCUs in the market.
In addition, for wearable devices that support HRM, Silicon Labs has developed a complete reference design platform, including required hardware, Simplicity Studio software development tools, as well as schematics, documents library, API files, and hardware/optical design guidebooks.
Silicon Labs’ American expert team and also its local Chinese team gave us fast, professional technical support and helped us reduce products’ power consumption and accelerate the product development process. Therefore, we are able to launch the first high-precision heart rate watch in the Chinese market.
Please tell us about your opinion on IoT development and also the future of the Chinese IoT market
Looking forward, the Chinese IoT market will have three huge opportunities: smart home, Internet of Cars, and smart wearable products.
Fitcare knows that as a leader in smart home technology, Silicon Labs provides low-power EFR32 (wireless SoC chip) running dual protocols of Bluetooth BLE and ZigBee, which will bring innovative applications combining wearable products and smart home.
Take wearable products as an example. Its next step is to no longer only serve as smart phone accessories for general leisure purposes, but is to gradually enter the professional athlete market, which will make the products more valuable and popular. We will need to develop products such as smart fitness wristbands that can be connected to the cloud management platform, smart medical wristbands that support remote healthcare, and heart rate watches that can help professional athletes trace and manage their training data, etc.
In order to meet the requirement of these IoT and design needs, we will need more precise HRM sensors and easier-to-use low-power Bluetooth wireless connection solutions. These are what Silicon Labs is best at. Therefore, we will continue our partnership with Silicon Labs and explore the new territory of smart wearable devices together.
Visit Fitcare to learn more about their awesome products.
We had a blast last weekend at ATX Hack for Change. This hackathon was created as a part of the National Day of Civic Hacking initiative to bring together great minds to work on social good projects in their local communities. Members of the community and the City of Austin pitched project proposals to solve unique issues. Developers then spent the weekend hacking away to find solutions and present their results.
Projects ranged from construction permit wizards to smart sprinkler systems and addressed problems that were prevalent in many different areas. Some projects finished the weekend with a new scope on how to solve their problem, while others left with a completed production-grade application. Most of these projects will continue to be built up, incorporated into other platforms, and develop real usability. Everyone at the hackathon came together to utilize their skills and create something that made a big change and a lasting impact.
Silicon Labs was proud to be a sponsor of the event, and our employees were excited to get involved as well. Our corporate band provided entertainment and showed that even executives can rock out. Michele Grieshaber, our CMO, delivered an insightful keynote speech about the world of innovation. We even had one of our own, Deirdre Walsh, be a project champion that led a team in developing an app to make local giving smarter with big data.
“Austin has an extremely smart community and great companies. It is amazing what you can do when you take that talent and apply it to do civil and social good,” Deirdre commented on the weekend as a whole, “These are smart people who are passionate about going above and beyond their jobs to help others.”
As a company that values unique innovation for a more connected world, it was great to be a part of an event that united people from all walks of life to support each other, work together, and apply their skillsets to directly improve the community.
In our previous post, Top Five Considerations for Designing a Beacon Product, Quickly, Securely, and Effectively, we talked about the fact that many OEMs who have never used wireless technology before are now adopting Bluetooth and adding beacons to their products. This can be simple, but it’s more likely a little bit of a challenge.
The beacon designer needs to consider:
Now, About Security
Recent press coverage has highlighted consumer fears about beacons “tracking your every move.” In reality, typical beacons do not collect data since they are one-way devices—they only broadcast.
It’s the smartphone that invades people’s privacy by knowing where the beacon is located and what services it offers through the smartphone applications.
By the way, smartphones can do this without beacons by using GPS, Wi-Fi, and cellular. It’s just that beacons are lower power and lower cost to deploy, thus enabling a new usage model and kicking up some valid concerns that have been around a long time.
Keep in mind that users can also disable all these “location-based services” in their device settings.
IT professionals and business people are also concerned about beacon security risks. But again, this implies functionality that beacons don’t normally have: they are typically one way devices, sending but not receiving information.
That said, beacon security in some applications is important, not only for on-going operation but also for beacon deployment.
For beacon broadcasting, unless it is part of a proprietary network, it is unencrypted by definition. However, IT professionals need to protect incoming beacon network traffic (if any) to the same levels as they do the rest of the devices on their network, or at least to the level the beacon hardware can accommodate.
In some beaconing applications where proximity to the beacon may have tangible value, such as reward points or entering a store, the provider may need to implement safeguards against spoofing as well. This could include timestamps, ephemeral IDs, or randomized security keys, generated with each proximity event and validated by the back-end system.
Security for Device Management Functions
Beacons go through a configuration process when they are deployed or upgraded in the field. This is likely when the beacon is most susceptible to security breach.
In these cases, the beacon can use Bluetooth’s security features (i.e., pairing, authentication, encryption, etc.), and other security measures such as strong password protection.
For deployment, access to configuration services can be limited to a short time window. After the window expires the device becomes a normal broadcast beacon and no longer advertises its internal services.
This process is illustrated below.
Whitepaper on Developing with Bluetooth BLE Beacons
Our experts have put some very relevant information in a whitepaper on developing with Bluetooth beacons. The goal is to help you get to market quickly with the right, stable solution.
It covers a lot of territory:
Senior Manager, Field Marketing, Silicon Labs
Joe Tillison joined Silicon Labs in February 2015 with the company’s acquisition of Bluegiga Technologies, a leading provider of Bluetooth and Wi-Fi modules and software. Previously he worked at Bluegiga as a director of business development for the Americas West region. Before joining Bluegiga in 2012, Joe managed the wireless portfolio strategy as technology director at Avnet Electronics Marketing. Prior to Avnet, he was a technical marketing manager and FAE at Memec. Joe spent his early career as a hardware designer on a number of electronics platforms for NASA and military spacecraft at Lockheed Martin Space Systems. Joe has a BSEE from the University of Oklahoma and an MSE from the University of Colorado, and he has published numerous articles and conference presentations on wireless technologies.