How to Design Small Bluetooth Medical Devices and Wearables – Seven Design Concepts to Reduce Form Factor and Costs

Building Bluetooth medical devices with small and challenging form factors is not only about picking the smallest hardware components from the market. You can also reduce the size and dimensions of your product by optimizing the bill of materials (BoM). Using SILICON LABS’ integrated Bluetooth SoCs and modules that include various peripheral functions and features needed in medical applications can save you a great deal of PCB footprint, enhance design flexibility, and reduce costs.

Silicon Labs has been working with many medical device makers for decades. As a result of these successful engagements, Silicon Labs has developed several functionalities that are relevant to wireless medical applications and integrated them into Silicon Labs Bluetooth SoCs and modules. In this blog, Silicon Labs explains seven effective Bluetooth design concepts that can save your PCB footprint, design costs, and BoM on medical applications.

Silicon Labs’ Bluetooth solutions feature robust security with up to the highest PSA Level 3 Certification and DTSec, the Cybersecurity Standard for Connected Diabetes Devices.

The Seven Bluetooth Design Concepts

1. Integrated AI/ML Accelerator

Machine Learning (ML) allows large volumes of sensor data to be processed against existing models to identify irregularities and identify what data is required for transmission back to the cloud.  For devices such as EKGs, wireless devices today transmit all of the collected data back to the cloud for processing and analysis. By identifying key data on the device, only certain subsets of data need to be transmitted back, saving precious resources and extending battery lifetime.

Artificial Intelligence (AI) can also be utilized in applications such as aging in place, where monitoring vitals, looking at changes in data, observing things like gait, and generally identifying patterns outside the norm can provide crucial diagnostic data to providers and trigger life-saving alerts to caretakers and loved ones that could drastically improve, and even save the life of a patient.

Silicon Labs BG24 Bluetooth SoC provides device makers with an integrated AI/ML accelerator with faster and more power-efficient ML inferencing, eliminating the need for an external ML processor on the board.

2. Analog Peripherals

An ADC is used on many wireless medical devices and wearables to enable sensor measurement and battery supply monitoring. Several Silicon Labs Bluetooth SoCs feature an integrated ADC with a high resolution of 12, 16, and 20 bits, saving costs, footprint, design, and testing effort. Additionally, Silicon Labs Bluetooth solutions deliver a higher effective number of bits (ENOB) for ADCs compared to many competing solutions, enabling more effective sampling.

For CGM manufacturers (and other portable medical devices) with solutions based on a discrete Analog Front End (AFE), the ADC and DAC are two critical functionalities that can be utilized on-chip, saving the need for additional components on the board. Silicon Labs BG24 features an ADC and DAC that can be coupled with external Operational Amplifiers to form a complete Analog Front End with reduced cost, space, and complexity of integration on a CGM device.

3. DC-DC Converter

A DC-DC converter is a valuable function for battery-powered medical devices. A Boost Converter allows an SoC to operate with lower-voltage batteries like alkaline and silver oxide with an input range of 0.8 to 1.7V, enabling a Bluetooth SoC to operate from the lower supply voltages. The Buck Converter allows an SoC to operate with other ~3V batteries, such as lithium coin cells, and it is used to improve energy efficiency and battery life or to reduce the size of a battery. The Silicon Labs BG27 includes both of these DC-DC Converters.

4. Coulomb Counter

To predict and prevent unexpected battery depletion during the use of critical health applications, a Coulomb Counter enables accurate battery level tracking to enhance user safety and experience. Silicon Labs BG27 features an integrated Coulomb Counter, eliminating the need for another external component on the board.

5. Low-Frequency RC Oscillator

For Bluetooth LE 2.4 GHz applications, the Bluetooth LE SoC needs to meet the specified Sleep Clock accuracy of ±500 ppm. BG22 and BG27 have an internal Low-Frequency RC oscillator 32kHz (LFRCO) that can self-calibrate against the HFXO on the device to meet the Bluetooth LE requirements. As a result, device makers can eliminate an external low-frequency crystal on the board, freeing up space and BoM costs unless the application requires a higher clock accuracy.

6. Matching Network

Silicon Labs reference radio board designs typically use an extensive number of components and multiple layers for RF and VDD filtering to provide device makers with the best possible RF performance at even the highest output power levels. However, medical devices and wearables use lower power levels, which means that you can decrease the number of PCB layers and elements from the design while maintaining an acceptable RF performance.

By optimizing the RF performance of BG22, BG24, and BG27 SoCs to match the ETSI and FCC regulations for low-power devices, you can reduce the number of bypass capacitors, filters, and the antenna matching network with a total of up to 6-14 components to decrease your footprint and BoM costs yet supporting the ETSI/FCC requirements.

7. Power Consumption

Minimizing the power consumption of a Bluetooth device enables device makers to achieve smaller device form factors. Ultra-low power consumption, both during operation and in optimized sleep modes (EM2), combined with the 1.5v supply through the DC-DC Boost provided in the BG27, allows a device to be fitted with an extremely small silver oxide 1.5v battery, minimizing the overall form factor of the device.

To learn more about the benefits of low power consumption, read the case study of the world's smallest wearable device, the smart tooth implant by Lura Health, which uses the Silicon Labs BG27 ultra-low-power Bluetooth WLCSP chip.