A question I often get asked when adding wireless connectivity to an existing product is whether to use commercially available hardware or design something custom. In my experience, the answer is often simpler than people expect.
There’s a tendency to assume that adding wireless capability means designing a new circuit board, selecting a wireless chipset, and integrating everything into the existing electronics. However, that isn’t always necessary. In fact, depending on the application, I might recommend avoiding custom hardware altogether.
The first thing I look at is what the product actually needs to accomplish. I usually start by asking:
- What data are we transmitting, and how frequently?
- What wireless range do we need?
- Are we talking about Wi-Fi, Bluetooth, cellular, or another technology?
- Where will the product be sold and operated?
- Are there power or physical space constraints?
- What interfaces does the existing product support?
- How many units are we planning to manufacture?
- What’s the target cost per unit?
- Are there specific security or environmental requirements?
Other considerations include throughput, latency, reliability, and regulatory compliance. These questions help determine whether we’re dealing with a straightforward integration project or something that genuinely requires custom electronics.
Let’s Walk Through an Example
Suppose you manufacture an industrial product that collects sensor measurements. The product already works, but now customers want to access those measurements remotely.
For this example, let’s assume:
- We want to use Wi-Fi.
- We’re transmitting a small sensor payload once per minute.
- We need reasonably long-range coverage within an industrial facility.
- Annual production is approximately 500 units.
- Our target hardware cost is under $200 per unit.
- The product will operate in the United States.
- Power consumption isn’t a major concern because the product is continuously powered.
- We have plenty of room inside or near the existing enclosure.
So, do we need to design a custom wireless printed circuit board (PCB)? My first instinct would be no. I’d start by looking at commercially available industrial gateways or routers.
Many of these devices already support common industrial interfaces, including Ethernet, RS-485, and serial communication. Depending on the existing product, we may be able to connect a gateway, configure the communication interface, and start transmitting data without touching the original circuit board. For example, we could have the gateway collect sensor readings and periodically send them to a cloud API using HTTPS or MQTT.
Because we’re transmitting such a small amount of data, Wi-Fi throughput isn’t likely to be the limiting factor. I’d probably start by evaluating 2.4 GHz Wi-Fi, which can offer useful coverage advantages over 5 GHz in some environments.
In general, though, actual wireless performance needs to be tested. Industrial facilities can be challenging RF environments, especially with metal enclosures, machinery, and interference. But the point is that we don’t necessarily need to develop wireless hardware to solve this problem.
Let’s Look at the Economics
This is where production volume becomes important. Suppose we find an industrial gateway that meets our requirements and costs $150 per unit. At 500 units annually, that’s $75,000 in hardware. Now suppose we believe we can develop a custom PCB that costs $110 per unit; this saves $40 per unit, or $20,000 annually.
But there’s another side to the equation. Developing custom hardware requires engineering time, prototypes, testing, firmware integration, regulatory evaluation, and manufacturing setup. Let’s assume the custom development effort costs $60,000.
At $20,000 in annual savings, we’re looking at roughly three years just to recover the initial development cost, assuming production volume and unit costs remain constant.
And that doesn’t include ongoing maintenance, component obsolescence, or unexpected design revisions. In this particular example, I’d have a hard time recommending custom hardware based on cost savings alone. There’s also something to be said for getting a product into customers’ hands sooner rather than spending months developing electronics that may not provide a meaningful advantage.
When Would I Recommend Custom Hardware?
Now let’s change the requirements. Suppose we’re developing a battery-powered sensor that needs to operate for several years without replacing the battery. We also have a small enclosure, limited PCB space, and expect to manufacture several thousand units annually; this is a different problem. An industrial gateway is probably no longer practical. It may consume too much power, take up too much space, and cost more than the application can support. In this situation, I’d be much more inclined to consider a custom PCB.
We could select a low-power microcontroller, integrate the necessary sensor interfaces, and choose a wireless solution appropriate for the power budget. The firmware could keep most of the electronics asleep, waking periodically to collect measurements and transmit data. We’d also have more control over the physical layout, component selection, and manufacturing cost.
However, a custom circuit board doesn’t mean we need to design the RF circuitry from scratch. In many projects, I’d still recommend using a commercially available wireless module. For example, we could integrate a pre-certified Wi-Fi or Bluetooth module into our PCB and design the rest of the electronics around it. This gives us much of the flexibility of a custom design without taking on the full complexity of RF development. Usually, that is a reasonable compromise.
Don’t Forget About Regulatory Testing
Another consideration that sometimes gets overlooked is regulatory compliance. If you’re selling an electronic product in the United States, FCC requirements need to be considered. For example, the finished product may need to meet FCC Part 15 Subpart B requirements for unintentional emissions. Wireless transmitters are also subject to applicable intentional radiator requirements, such as Part 15 Subpart C or E.
Using a pre-certified wireless module can simplify the process, but it doesn’t necessarily eliminate the need to evaluate or test the finished product. The antenna, enclosure, PCB layout, and surrounding electronics can all affect compliance. And if you’re using Bluetooth technology, there are also Bluetooth SIG qualification requirements and rules governing the use of Bluetooth trademarks.
I generally recommend looking at these requirements early rather than treating them as something to address right before production. Regulatory surprises late in development can be expensive.
What If We’re Manufacturing 100,000 Units?
At higher production volumes, the economics can shift considerably. A $10 reduction in unit cost might not matter much when you’re manufacturing 500 units annually. At 100,000 units, that same reduction represents $1 million in annual savings. This can justify considerably more engineering effort.
However, I still wouldn’t automatically recommend designing a custom RF chip. There’s a big difference between developing a custom PCB around an existing wireless chipset and developing custom RF silicon. The latter involves substantial engineering investment, specialized expertise, and semiconductor development costs. Even products manufactured in the millions often use commercially available wireless chipsets. Unless the application has particularly demanding technical requirements or very unusual economics, I’d generally start with existing RF technology rather than developing custom silicon.
So, What’s the Right Approach?
In my experience, the biggest mistake isn’t necessarily choosing an off-the-shelf solution or a custom design, but rather committing to either approach before understanding the requirements.
For an industrial product that simply needs to transmit a few sensor readings every minute, an off-the-shelf gateway might solve the problem perfectly well. For a compact, battery-powered device manufactured in the thousands, a custom PCB with an integrated wireless module may make more sense. And for specialized, high-volume applications, a more customized RF design might be justified. There’s no single answer that applies to every product.
My general approach is to start with the simplest solution that meets the requirements, then introduce custom hardware only when there’s a clear technical or economic reason to do so.