Tom Nordman, Pasi Rahikkala, Tommi Hakkila
This paper explores the challenges that come with designing connected devices into increasingly smaller products, specifically antenna integration and how system-in package (SiP) modules can help.
The Size Challenge
As we wirelessly connect more and more devices to the Internet, electronics engineers face several challenges, including how to package a radio transmitter into their existing device real estate and how to make increasingly smaller devices. They’re also striving to meet consumer demand for Internet of Things (IoT) products that are ergonomically easy to use and unobtrusive.
Size expectation is one of the most frequently asked questions when considering IoT devices, along with radio performance and price.
Ideally, engineers would like to use IoT components that are as small as possible, have great RF performance, and are affordable. These characteristics do not typically converge in IoT component offerings, and that presents a challenge for solution providers. Fortunately, the size of a silicon die has been getting smaller and smaller over the years as the industry adopts new silicon manufacturing processes. The industry has been solving the space issue for IoT implementations by combining the MCU and RF frontend into system-on-chip (SoC) configurations (i.e., making wireless MCUs available.) However, the trend toward SoCs has not solved the physics of the RF transmitter—the antenna. Antenna design is often left for customers to sort out, or they may be guided to choose ready-to-use wireless modules with an integrated antenna. The space required for an antenna is a challenge that comes with designing small IoT devices. It needs to be efficient while also enabling reliable wireless connections. For this reason, the focus of this paper is highlighting the specific concerns around antenna integration.
Why an SoC?
When the IoT boom started to blossom during the 2000s, the industry was called machine-to-machine (M2M), and the components offered for IoT connectivity were mainly general packet radio service (GPRS) modems, Bluetooth serial cable replacement, or sub-GHz proprietary radios. These designs had two main components for connectivity: the MCU and the radio modem. And the required space for basic IoT functionality was typically at its smallest—50 mm on each dimension—meaning the devices were about the size of a mobile phone. When the semiconductor industry moved to processes where the required MCU and RF functionality could be packaged into the same die, new possibilities for developers began to emerge. Now developers could implement the functionality of an IoT device in the same IC/SoC. The IoT component architectures shifted to wireless MCUs due the obvious benefits—engineers could design IoT devices with a single component and save significant space, but they could also save money because of the lower component costs. When selecting the architecture for modern IoT devices, it’s obvious the SoC-based systems will lead the way thanks to their size advantage. However, the trend toward SoCs has not solved the physics of the RF transmission—the antenna.
What About the Antenna? How Much Space?
The new era of highly integrated SoCs leaves developers with some questions: What about the antenna? How much space should I reserve for the antenna? What kind of antenna should I choose, and should I use a module with the antenna already integrated? The antenna question is complex at many levels as we need to consider not only size and efficiency but detuning questions as well, especially across designs that may have different housings but with the same antenna architecture.
It has been common to use printed circuit board (PCB) trace antennas, such as inverted-F, for IoT designs due their low bill of material (BOM) costs. But these PCB antennas have significant size requirements, normally in the range of 25 mm x 15 mm, ultimately making the resulting IoT devices enormous. These antennas have another downside if used in a module as well: They are sensitive to the detuning caused by the housing materials and require specific consideration in the product assembly to work optimally. In SoC designs, antenna tuning is part of the normal design flow and requires a certain amount of expertise. In these designs, a printed antenna does not differ from other antenna types. These printed PCB antennas have significant size requirements, normally in the range of 25 mm x 15 mm, ultimately making the resulting IoT devices enormous.
Antenna manufacturers have been offering “chip antennas” for quite some time to simplify design efforts, but there are size benefits as well. These chip antennas are offered mainly in two different forms:
- Antennas that are not coupled to the GND plane and will require a relatively large clearance area (or that is free from the ground, traces, and components). Examples of such antennas are monopole and inverted –F type antennas.
- Antennas that are coupled to the GND plane and require either a relatively small clearance area beneath the antenna or don’t require clearance area at all.
Both antenna types have space requirements on either clearance area and/or ground plane and PCB size. The required space for the RF part of an IoT design should also include the necessary clearance area because no components or traces can be placed here.
This means that when designers are doing size estimations for IoT devices, they need to pay attention to the necessary dimensions for the antenna and the required clearance areas, but they also need separation distances from the antenna with the edge of the housing.
When designing tiny IoT devices the size of a coin cell battery, there is always a compromise with antenna efficiency. The smaller the size we try to achieve, the less efficiency we can have for the RF performance. Devices less than 10 mm on each dimension begin to achieve performance in the 2.4 GHz band, giving users Bluetooth connectivity of approximately 10 meters with a mobile phone, which is acceptable for most personal IoT devices. However, when the dimensions are closer to 20 mm in each direction, the RF efficiency increases significantly, giving a practical range of 20–40 meters with a mobile phone depending on the conditions.
When dimensions reach 40 mm, the optimum efficiency of several antennas that tune with the ground plane size reaches maximum performance. This means with the Bluetooth 4.2 protocol, a practical range between two identical devices is around 60–400 meters. When using 802.15.4 protocols such as Zigbee, the range can be up to +500 meters in line of sight. Depending on the application and size targets, a designer needs to look at the antenna performance and efficiency in relation to PCB size, as most of the chip antennas use the PCB ground plane as part of the antenna configuration. In addition, the position of the antenna/module in the design is important, and designers need to consider the clearance areas, grounding for the optimum location of the module in the design.
Antenna clearance area, ground plane stitching, and position of the BGM121 SiP module from the PCB edges impacts RF Performance.
What About the External Antenna?
According to Silicon Labs market data, when considering several different antenna packaging options, nearly 50 percent of IoT customers evaluate the performance and feasibility of external antenna (antennas integrated into the housing example via U.Fl connector). However, approximately 10 percent of these evaluated designs deploy the external antenna, and 90 percent of the customers choose modules with a built-in chip antenna. What is the reason behind this? Why would engineers not widely deploy external antennas on their designs? The answer has two main dimensions. First, the mechanics of an external antenna is not design friendly; they look ugly and break easily if the IoT device is dropped. These external antennas also significantly increase BOM and assembly costs. Also, when comparing the efficiency of a well-built RF design with a chip antenna versus an external antenna through a U.Fl antenna connector, there is nobenefit in using an external antenna. The benefit of the external antenna is obvious if the housing of the device is metallic, forming a faraday cage that makes it impossible for the RF signals to penetrate the device. An external antenna is also justified if the absolute best performance is required and assembly costs and mechanical designs allow for its usage.
How Packed Should the Housing Be?
When engineering the IoT device with an antenna, the mechanics and housing play a key role in avoiding or causing antenna detuning. The RF radiation, when bursting out of the antenna, is impacted by the proximity of the materials. The antenna will detune if it touches the metal or even plastic. For this reason, the antenna must be separated from physical contact with housing plastic or metals. There are big differences in the types of antennas and their sensitivity to detuning. Monopole-type antennas are more sensitive than GND coupled antennas. Some of the latest packaging innovations for Silicon Labs’ SiP module solve the detuning issue because the antenna is already within the substrate and detuned to the proximity of the plastic housing. This sets designers free to place the SiP module freely on their designs, reducing the size of the devices significantly.
Possibilities of System in Package (SiP)
Silicon Labs has combined its IoT SoC experience with Bluegiga antenna design experience by creating a SiP module that offers the benefits of an SoC module combined with an ultra-small footprint. The total design footprint size, including the antenna clearance area, is slightly over 50 mm2. This means that it leaves space for other components in the design, finally making it feasible to design truly compact IoT devices.
For example, Silicon Labs’ BGM12x SiP modules are designed for the smallest design footprint for Bluetooth low energy technology. Its 6.5 mm x 6.5 mm size offers a complete implementation and includes an ARM Cortex®-M4Fcore-based MCU, plenty of flash and RAM, an integrated antenna, and an ultra-small clearance area of 5.0 mm x 3.0 mm to enable high-performance applications. The SiP module also integrates all required passive components in practice, leaving the designer free from all RF-related design worries if layout guidelines are followed. The SiP modules are ideal for wearable sports and fitness devices and home automation systems -- any end device that must be slim and small.
SoC vs. Module?
When using an SoC with an antenna matched for the actual end-application design, an experienced designer can meet the required performance and cost goals. The optimization of antenna performance and the SoC RF layout requires understanding of RF design rules and quite expensive equipment. As the number of wireless applications and devices increases, the required regulatory certification may become too expensive and time-consuming. This is especially true when the product portfolio increases to tens or even hundreds of devices, each requiring separate certification-related product management. An increasingly popular solution is to use pre-certified, ready-to-use wireless modules, especially for IoT-related designs.
In RF modules, the arduous task of antenna matching has been done with no application-specific tuning necessary in most cases. However, sometimes the product design cannot avoid the placement of an attachment screw, LCD, or battery close to the antenna, causing significant changes in the antenna resonant frequency and overall performance. The modules themselves cannot be easily modified to compensate for the detuning.
The following figure shows the impact a plastic material placed on top of a monopole-type chip antenna. The plastic material on top of the antenna clearly shifts the resonant frequency and moves it out of the target frequency range. The problem regarding compensation is that with a typical module the matching components are inside the module making adjustment impractical.
Impact of plastic on the return loss of a typical monopole chip antenna
Often an external antenna can be used with a module. This is still easier than the total design effort of using an SoC and includes the certifications when the selected antenna is chosen from the list of accepted antennas included in the certification reports. An external antenna, either with a connector or a ceramic chip antenna, provides more flexibility regarding the placement of the antenna and its performance. However, matching the impedance of an external antenna is quite difficult and requires expensive equipment. An improved level of flexibility is provided by SiP modules where the small size allows a part of the antenna structure to be placed outside of the module. This typically consists of a short copper trace, the dimensions of which can, in most cases, be copied from the manufacturer’s datasheet recommendations. The resonant frequency of the antenna depends to some extent on the actual length of this external track. Thus, the detuning effect caused by objects near the antenna can be compensated simply by adjusting the length of the external copper trace. This provides a very effective method for detuning compensation without adding a single item to the BOM list.
Antenna Robustness and SiP Modules Covered with Substrate
The proximity of any material will always have some effect on the resonant frequency of the antenna. The strength of this effect depends on the actual antenna model used. For example, Silicon Labs has an antenna solution for SiP modules that minimizes the required total size of the antenna while still making the antenna exceptionally insensitive against the effects of nearby objects. Any plastic object can be placed as close to the modules as necessary without having to re-tune the antenna, including full overmold with protective conformal coating or epoxy mass. High performance is possible even with a metallic object touching the module, which is impossible with almost any other RF module with an integrated antenna, although such a setup may need some antenna re-tuning for optimal performance.
Well-designed SiP modules are practically immune to any dielectric close to or in touch with the module. Thus, the thickness of the application board PCB, conformal coating, plastic casing or any kind of molding on top of the module do not have a significant negative impact on the module’s RF performance.
Example: BGM121 SiP Module antenna return loss
When placed near a metallic object, even the SiP module will experience some detuning. In these cases, the modules can be easily tuned externally by adjusting the width of the antenna clearance area (see the figure below). This allows placing the module into any kind of mechanical design, even with a battery or display under the antenna area while still achieving the highest possible radiated performance. There are no restrictions whatsoever where these modules can be used if they are not surrounded by metal, which would prevent RF radiation.
The detuning for SiP modules is done simply by adding a suitably dimensioned clearance area (a copper-free area on the PCB) as indicated in the figure below, which shows the clearance area dimensions.
Silicon Labs SiP module antenna allows compensation of detuning by adjusting the size of the clearance area width.
The shape of the radiation pattern (see figure below) does not depend on the shape of the application PCB. The SiP module can provide an effective omnidirectional radiation pattern in all designs, thus ensuring reliable operation in Bluetooth applications regardless of the PCB design surrounding the module.
3D radiation pattern of BGM121 SiP Module