Introduction to Developing a New Wireless Product

This article is by John Teel, an award-winning electronics design engineer and the founder of Predictable Designs (a company that helps entrepreneurs and makers develop new electronic hardware products). Download his free cheat sheet 15 Steps to Develop Your New Electronic Hardware Product.

There are two routes to incorporating wireless functionality into a custom PCB design: a self-contained module, or a chip design. In almost all cases, it’s best to start off with a module solution. Wireless functions can be one of the trickiest types of circuits to design so modules are very commonly used.

Later you can migrate to a chip design once you reach production volumes high enough to warrant the extra development costs. Although if your profit margins are high enough you may choose to stick with a wireless module indefinitely.

A module solution will get your product to market faster, but at the cost of lower profit margins. When launching a new product focus on minimizing your risk and cost to market, not maximizing your profit. Maximizing profit will come later.

Selecting and optimizing an antenna

An antenna is a length of wire, or some structure, to which RF energy is applied and this energy then radiates out into space. There are many different types of antennas with names like monopole or dipole. Sometimes they may be labeled as ½ wave or ¼ wave. Patch, dish and Yagi are other names you will encounter. Don’t let these terms overwhelm you. Knowing all of them isn’t usually necessary.

Bluetooth Classic, Bluetooth Low-Energy, WiFi, and ZigBee all operate at a carrier frequency of 2.4GHz. The FCC designation for this frequency band is the Industrial, Scientific and Medical (ISM) band.

Almost any antenna designed to work in the 2.4GHz ISM band will work for any of the wireless standards such as BLE, WiFi or Zigbee and you don’t have to limit yourself to antennas advertised for a particular protocol. A quick search on a distributor’s web site, such as Digikey.com or Mouser.com, for 2.4GHz antennas will show many pages of antennas, of all different types.

Most antennas are designed to have an impedance of 50 ohms. This means that the PCB layout should match the impedance of the RF transceiver (module or chip) to the antenna. By matching the impedance the maximum signal power is transferred between the antenna and the transceiver. Without proper impedance matching much of the signal will be lost along the antenna feedline.

Keep in mind this is complex impedance and not simple resistance. It has nothing to do with the resistance of the PCB trace. Instead, you need to design your PCB so that the impedance between the antenna feedline and nearby ground planes is 50 ohms.

Most designs incorporate a pi-matching network (made of inductors and capacitors) to allow adjustment of this impedance for optimum tuning. For maximum performance tuning will need to be done inside the final product enclosure.

To calculate the feedline (a.k.a. transmission line) impedance I like to use a free tool called AppCAD. You’ll need to enter in the specs for your PCB such as the metal thickness, dielectric constant and thickness, feedline width, and the distance to the underlying ground plane. Within AppCAD you’ll use either the “microstrip” calculator or the “co-planar waveguide” calculator depending on whether you have a ground plane surrounding your feedline on the same layer. In most cases, you’ll adjust the width of your feedline to obtain 50 ohms.

For short-range wireless protocols such as Bluetooth, WiFi, ZigBee, and Z-wave having a less than perfect antenna design is usually acceptable (at least initially). The operating range and data rate may be somewhat reduced but everything will still work.

On the other hand, the antenna design for something like a Global Positioning System (GPS) is extremely critical. You are after all trying to receive incredibly weak signals from multiple satellites in space! If your antenna design isn’t perfectly optimized in a GPS design then it simply won’t work at all. It’s not like you can move the satellites closer for your initial testing.

PCB trace vs. chip antenna

When it comes to small on-board antennas there are usually two ways to go: a PCB trace antenna, or a chip antenna.

Chip antennas are more compact than PCB trace antennas and they are easier to tune accurately. I generally recommend starting with a chip antenna since they are easier to design in your product.

PCB trace antennas have various names that sometimes describe their shape. For example, an inverted F antenna looks like the letter F and a microstrip patch antenna looks like a square or rectangular patch.

The primary advantage of a PCB trace antenna is they are essentially free since they are made entirely from a trace on the PCB. They will increase the board size some so they aren’t entirely free, but almost.

If you plan to use a PCB trace antenna then the PCB insulating material becomes more critical. FR-4 is the most common PCB material but the dielectric constant of the material is not well controlled and may even vary significantly from lot to lot. The result is the antenna may no longer be properly tuned on some boards.

Rogers Corporation makes a PCB material called RO4350B, which is a suitable low cost replacement for FR-4 when building RF circuits. RO4350B provides tight control of the dielectric constant and is low loss, two things that generic FR-4 does not have.

Modules with a built-in antenna

On modules with a built-in antenna the complex RF design details have already been addressed by the module manufacturer. The real concern for the designer is placing the module so that the antenna is kept away from metal objects, like other traces, ground planes or a battery. Most modules will specify a ground clearance area around the antenna.

The ESP8266 based WiFi module is an example of a product that includes the radio transceiver, antenna, TCP/IP protocol stack and a convenient serial interface to the host microcontroller. The antenna is integrated into the module as a PCB trace antenna.

The usual recommendation is to place the RF module near a board edge with the antenna kept clear of other metal structures. This is done to keep objects away from the antenna’s near field radiation pattern which could couple the antenna to an RF ground.

Modules without a built-in antenna

Some RF modules have a small coaxial RF connector on the module for connecting an external antenna. External antennas are external in the sense that they are not located on the PCB along with the trasceiver. This arrangement is used when the designer locates the antenna in a different location or uses an antenna that is not printed or mounted onto the PCB.

A few modules don’t include either a built-in antenna or a coaxial connector. In that case the antenna is connected via a PCB trace. This means you’ll need to pay attention to the same issues (impedance matching) for laying out an antenna feedline on a custom RF design. I recommend avoiding such modules, since having to do a custom antenna feedline layout negates much of the reason for using a module in the first place.

Summary

There are many ways to implement wireless functionality into your product. My suggestion for entrepreneurs is always to take the path of least cost and risk. This means using a fully self-contained wireless module with either an on-board antenna, or a co-axial connector for use with external antennas.

If you want to learn more about how to develop your own electronic product to bring to market then check out my Ultimate Guide — How to Develop a New Electronic Hardware Product.