(This post was inspired by JD Albert’s “10 Questions to Your Best Manufacturing Solution.” This is the electrical engineering point of view.)
As the Internet of Things ecosystem grows, the continuous rollout of new smart product categories (smart shoes, connected beds, smart cups?) shows no sign of slowing down.
Making your electronic device as cutting-edge as possible is tempting in this environment, where even new toasters are considered outdated unless they connect to your bathroom mirror to determine how dark you’ll want today’s bagel. But as with any other temptation, using leading-edge parts comes with risks.
You, too, might have a great, even disruptive, product to debut, but realistically it needs to get to market before someone else beats you to it. This is one consideration when deciding which passive components, processors, and other electronic parts you’ll put in it.
Think beyond the prototype for a good experience in production.
But here’s the most important thing to remember when designing your IoT product: Don’t let the part selection be driven by the newest device or easiest prototyping solution! So, what should drive it? Ask yourself the eight questions below if you want to choose the right parts for your device.
Have you defined your product’s end-to-end experience? If not, you may be including features (and parts) that add unnecessarily to the user’s attention burden. Make sure you’ve done your due diligence with research and interaction design colleagues before beginning to think about the technology that will support your product’s features. This is a crucial first step.
Too much memory? Faster processor than you need to do the job? You may end up paying a significant penalty when you get down the road into production. Use up all your connectivity and all of your sensor inputs in your first generation product design? You could be heading down the road to a full redesign (and all the costs that incurs) if you need one more feature to get to market successfully. It’s a balancing act, and the more information you have upfront to make decisions, the better.
(Read more about aligning with research and interaction design colleagues in the article I co-wrote with Bill Horan: “8 Mistakes That Can Derail Your IoT Project.”)
This is a key question. Picking the latest and greatest part can provide lower unit costs, higher performance, and all the cachet that comes with the leading edge technology. But it’s also a risk. How leading edge you want to be often hinges on how much risk you — and your company — can afford.
If an electronic component is new, it may not have software drivers fully written and tested or hardware fully debugged. Chances are the manufacturing issues that happen while making the part at high volumes haven’t been worked out, either. And if demand for that component turns out to be low, the manufacturer may discontinue it sooner rather than later. All of this can cause unexpected delays and costs that wreck your timeline and budget and damage your potential market share.
To shorten your delivery timeline on a new product release, work with a hardware and software architecture that’s familiar to you and has a known track record. The software and feature debugging process tends to wreak the most havoc on product release timelines.
The newest, fastest, most efficient processor has no track record. A slightly older, but very popular, processor will still be available 18 months from now if your product proves to be such a hot seller that you need to make more of it. Finding out that a big manufacturer like an automaker is using the processor you’ve been eyeing as its go-to component is a nice guarantee of continuing availability.
Another consideration is memory: The memory market, especially RAM, is extremely volatile. Since their volume is driven by the consumer electronics market, SDRAM, DDR2, DDR3, and RAM prices can change significantly quarterly or even monthly. Parts can become obsolete or completely unavailable within 12 to 18 months. Do your homework.
Many consumer electronics today are unfortunately considered disposable, and the passive components — capacitors, resistors and inductors — are usually as cheap as possible. But for more permanent devices such as home appliances, you’re better off investing in higher reliability parts.
Most of the cost of a security camera, for example, is in its installation. If the camera fails in ‘only’ five or six years and the customer needs to pay for another truck and technician to come out, they’ll never buy from you again. For a product like this one that consumers expect to be durable, spending a small percentage more on a quality capacitor could save the reputation of your company.
For a product that’s expected to be durable, spending a small percentage more on a quality part could save the reputation of your company.
It’s always a good idea to do analysis on component failure rates for temperatures and environments that the product will be used in. Making assumptions that the least expensive commercial-grade component will be sufficient can cost you much more in the long run.
Parts for consumer devices are generally rated for zero to 70 degrees Celsius. If you’re making a WiFi-controlled sprinkler system that needs to survive outdoors, you’ll need something tougher — probably automotive-rated parts (-40 to +150 degrees Celsius.)
Environmental specifications can significantly restrict your component choices. In some areas of a design, there may even be no choices at all! You need to know early on if your environment is going to limit your choice of components; redesigning after you’ve failed compliance testing, or worse — when the product is in the customer’s hands — is much more costly than a design review upfront.
It’s an unfortunate reality: Some suppliers will tell you parts will be available when they are not. Some will inflate the actual demand for the part you want, or up the pricing after you’ve already designed it into your product. Others make great parts at reasonable cost, but have lead times of 26 or even 52 weeks. This uncertainty can kill a startup trying to enter the Internet of Things market, which has been turning on development cycles of six to 12 months.
Find a contract manufacturer who is reliable, honest, and fast. Have detailed discussions with your suppliers, distributors, and contract manufacturers early in your design process to understand your parts selections before you start thinking about going into production. This is something to start working out before the very first prototypes.
Involve your parts manufacturer, parts distributors, and third-party contract manufacturer as soon as you have a handle on what you want to do. Sharing design information early with your supply chain partners helps them manage inventory, helps you manage your expectations, and helps you get a realistic handle on what will actually be available for your product release date.
Most contract manufacturers will want — or require — to manage the parts purchasing and inventory needed for production, in exchange for an added cost percentage. Work with your manufacturing partner to arrange pricing, order dates, and volume early — this is vital to having your device ship on time. (See #6!)
Depending on the volume of the product, this can involve an outlay of capital to hold parts or secure production. Knowing that you have to plan for these expenses upfront keeps you from running out of funds before you ship.
How will your device connect to your customer’s desired platform / system? For example, if it’s a home automation system, the customer will probably be interested in integrating with Apple’s HomeKit, or Google / Nest’s Weave and Brillo systems. There are a dizzying array of standards, and some have specific requirements for the hardware (Apple’s MFI program, for example), and certification requirements to pass.
Adding support for an ecosystem after a product is designed or just before its release can be difficult, if not impossible.
Adding support for an ecosystem after a product is designed or just before its release can be difficult, if not impossible. Knowing the ecosystems your customer is likely to use, and having an early plan to get through the hurdles of hardware, software, and certification processes for the standards you want to support will help keep your project on track. A compatible product is more desirable in an increasingly connected world.
Leading Edge Isn’t Always Best
The temptation to go with the most leading-edge parts is so widespread, it merits its own section in this post! Yes, that cutting-edge processor can give you the fastest, most capable device. But you’ll also be on the leading edge of driver debugging, and your timeline will bear the brunt of any early kinks in the new part’s manufacturing process.
Good design and manufacture can elevate a product above the sum of its parts.
Some products are truly revolutionary must-haves that can shrug off manufacturing delays. But be honest with yourself and your company about whether yours is one of them. Many excellent products use tried and true technology that is almost as small, fast, and efficient as the newest of the new. Good design and manufacture can elevate a product above the sum of its parts.
And if you really must have the latest and greatest electronic processors, be realistic about timing. If a manufacturer promises you it will be in production by November, well … plan for January. Stay away from 1.0 parts if the device needs to be on the market within the next six months. And above all, and this is worth repeating, don’t let the electronic parts of your device be an afterthought!
Learn more about our electrical engineering expertise.