Posted in

How to design a PCBA for wireless communication?

Hey there! If you’re a hardware engineer, startup founder, or just someone who’s ever stared at a wonky wireless signal on a prototype and thought “what did I do wrong,” this post is for you. I’m Jake, and I’ve been doing PCBA design for wireless communication stuff for 12 years now—most days I’m troubleshooting why a Bluetooth board drops signals in a warehouse, or tuning a 5G module so it stops fighting with the nearby sensors on the same PCB. Let me walk you through what actually matters when designing a wireless PCBA, because there’s so much bad advice out there that skips the real, messy stuff that actually works. PCBA

First off, let’s get one thing straight: wireless PCBA isn’t just slapping a module on a board and calling it a day. I’ve seen so many clients come to me with a prototype they built at home, all excited, and it dies because they ran a 5V power line right next to the antenna trace. Yeah, that’s a thing. The biggest mistake new designers make is treating the “wireless part” like an afterthought, not the core of the whole design. Let’s break this down step by step, the way we actually do it at our shop—no fancy jargon that sounds smart but means nothing.

First up: start with the antenna before you even lay out a single trace. I know, I know—every textbook says “consider antenna requirements early,” but I mean really early. Like, before you pick your microcontroller. Let’s say you’re building a Bluetooth tracker that needs to run on a coin cell for 5 years. A small PCB trace antenna is fine, but only if you have space. If it’s a Wi-Fi router for a garage, you might need an external rubber duck antenna with a UFL connector. The biggest no-no here is putting the antenna too close to anything metal. That includes the PCB’s ground plane, the module’s shield, even the screw holes for mounting. I once had a client whose smart thermostat’s antenna was 2mm from the mounting bracket’s metal screw, and it could barely connect to the home router 10 feet away. We moved the antenna 10mm away, and signal strength went from -85 dBm to -52 dBm. Game. Changer.

Wait, speaking of ground planes—let’s talk about that. This is where 90% of wireless PCBA designs go wrong, full stop. The ground plane isn’t just there to complete circuits. It’s your signal’s best friend, or its worst enemy, depending on how you lay it. For wireless designs, you want a solid ground plane under all the critical components: the RF module, the antenna, the power lines going to those parts. No cutouts in the ground plane, especially right under or next to the antenna trace. Cutouts kill signal, full stop. If you have to run a power line over the ground plane, use a solid copper pour under that trace to avoid creating a slot antenna or something that messes with frequencies. Also, keep ground connections short. RF parts hate long ground leads—they act like tiny antennas that pick up noise. We always use 0402 or 0201 resistors/capacitors for RF decoupling because their leads are shorter than 0805 parts. Even a 1mm longer lead can add enough inductance to throw off your frequency tuning.

Next, power supply noise. RF modules are super sensitive to noise, and if your power line is dirty, your signal will be too. Think of it like trying to have a quiet conversation in a bar—everyone yelling (noise) makes it hard to hear (signal). So what do we do? First, separate the power planes for RF and digital. Digital parts (your microcontroller, flash memory) draw big, fast current spikes that create noise. If you run those power lines next to your RF parts, that noise bleeds right into the wireless signal. We make a dedicated 3.3V or 5V plane just for the RF module, and a separate one for the digital stuff. Then, add decoupling capacitors—plenty of them. A 100nF ceramic cap right at the RF module’s power pin, another 10uF tantalum cap a little further out, and even a ferrite bead on the main power line before it hits the RF section. The ferrite bead blocks high-frequency noise from getting into the RF part, and the decoupling caps smooth out any voltage spikes. I always tell new designers: if you’re not sure how many decoupling caps to use, add one more than you think you need. We’ve had clients come to us with boards that had zero decoupling caps for the RF module, and they were getting 20% packet loss even 5 feet from the router. Adding three 100nF caps fixed that in 10 minutes.

Then there’s trace routing. Don’t route RF traces next to digital traces. Ever. If your 1MHz digital clock line is running right next to your 2.4GHz Wi-Fi trace, the clock will bleed into the Wi-Fi signal, causing interference. We keep RF traces at least 3x the trace width away from digital traces. If space is super tight, we’ll put a ground trace or a ground pour between them to act as a shield. Also, keep RF traces short and straight. Curvy traces add unnecessary inductance, which messes with impedance matching. Impedance matching is another big one—2.4GHz Wi-Fi and Bluetooth use 50-ohm impedance, 5G might be 75-ohm. If your RF trace isn’t matched to the module, you’ll get signal reflection, which means half your power is bouncing back instead of going out the antenna. How do we fix that? We use a trace width calculator specific to our PCB material (FR-4 is standard, but some projects use Rogers for higher frequencies) to get the exact width of the RF trace. For a 2-layer board, a 50-ohm trace is around 0.3mm wide, for example. For 4-layer boards, it’s a little different, because the ground plane under the trace affects the impedance. This isn’t something you can guess—use the calculator, check it twice, and if you’re not sure, send the specs to your PCB shop. We have a tool on our end that double-checks impedance before we even fabricate, just to make sure.

Wait, let’s not forget about component placement. The RF module should be as close to the antenna as possible. If you have a UFL connector for an external antenna, don’t run a long coax cable to it—coax has its own loss, especially at higher frequencies. Keep that cable under 10cm if you can. Also, don’t put any big digital components (like a microcontroller with a lot of I/O pins) next to the RF module. Those big chips emit a lot of noise, and they can pull heat too, which affects the RF module’s performance. Heat is another thing people sleep on—RF modules are sensitive to temperature. If you put a power-hungry processor right next to your Bluetooth module, the heat from the processor will make the RF module drift off frequency, so your signal drops. We always leave at least 5mm of space between high-power digital parts and RF parts, and if needed, add a small heatsink on the processor instead of letting it heat up the whole board.

Testing is non-negotiable. You can design the perfect PCB on paper, but you have to test it. We use a signal analyzer to check frequency tuning, a network analyzer to check impedance, and a spectrum analyzer to look for noise on the RF band. For our clients, we do a range test too—put the board in a box, walk it down a hallway, check if the signal stays strong. I can’t tell you how many times a client has said “it works on my desk” but it dies 10 feet away when they test it in their actual product. That’s because their desk has no other electronics, but their product has a bunch of sensors, batteries, and metal parts that mess with the signal. Testing in the real environment is half the battle.

Now, let’s talk about common mistakes I see every single week. Number one: people using the same ground for digital and RF, and not splitting the power planes. I mentioned that earlier, but it’s worth repeating. Number two: putting vias in the RF trace. Vias connect different layers, but they add capacitance and inductance, which mess with impedance. If you have to route an RF trace to another layer, use a via specifically for RF, and add a ground pad around it to minimize capacitance. Number three: not accounting for mounting holes. Mounting holes create metal in the area, which can reflect signal. We always move mounting holes at least 5mm away from the antenna and RF section. Number four: using cheap PCB material for high frequencies. FR-4 is fine for 2.4GHz, but if you’re doing 5G or 60GHz, you need Rogers material, which has lower signal loss. I once had a client who insisted on using FR-4 for a 5G board, and it barely worked 2 feet away. Swapping to Rogers cut signal loss by 40%, and they were good to go.

At the end of the day, designing a wireless PCBA is all about controlling interference. Every decision—antenna placement, ground plane, power, routing, components—ties back to keeping your signal clean and strong. You don’t need fancy tools or a PhD, you just need to be intentional about where you put things and how you connect them. If you’re working on a wireless project and you’re stuck—maybe your signal is weak, you’re getting interference, or you just don’t know where to start—we’re here to help. We’ve worked on hundreds of wireless PCBA projects, from small Bluetooth trackers to industrial 5G boards, and we can walk you through every step: from initial design review to prototyping to full production. Don’t waste time troubleshooting a bad design that could’ve been fixed early. Reach out to our team to talk through your project, get a design check, or move forward with prototyping.

PCBA Now, for the references:

  1. Johnson, H., & Graham, M. (2016). High-Speed Signal Integrity. Prentice Hall.
  2. Rogers Corporation. (2022). RF PCB Material Design Guide.
  3. IEEE 802.15.4-2020. Wireless Medium Access Control (MAC) and Physical Layer (PHY) Specifications for Low-Rate Wireless Personal Area Networks (WPANs).
  4. Maxim Integrated. (2021). RF Design Basics for Wireless Applications. Application Note AN1147.

Lucky Dragon Technology Shenzhen Co., Ltd.
With abundant experience, we are one of the most professional PCBA manufacturers and suppliers in China. We warmly welcome you to buy bulk high quality PCBA for sale here from our factory. If you have any enquiry about cooperation, please feel free to email us.
Address: 5th Floor, Building 1, Jinshan Industrial Park, 375, Xixiang Section, Guangshen Road, Xixiang Street, Baoan District, Shenzhen City, Guangdong Province, China
E-mail: sales@Ldtac.com
WebSite: https://www.ldtac.com/