Managing sensor integration woes, optical design in medical wearables

Covid-19 has provided a renewed push for the commercialization of health-care wearable devices. This is evident from the development of wireless temperature-sensing tags and patch monitors to detect the early symptoms of coronavirus.

One recent example is a partnership between SkyWater Technology, Linear ASICs, SensiML, and Upward Health. Foundry service provider SkyWater Technology recently joined hands with analog chipmaker Linear ASICs to mass-produce a wireless-connectivity–enabled IC with temperature-sensing capabilities.

The other partners in this collaboration are SensiML Corp., a subsidiary of QuickLogic, which provides software for ultra-low–power IoT devices, and Upward Health, an in-home and virtual health-care service provider. The sensor platform will eventually be integrated into contact-tracing applications to detect early-stage symptoms of Covid-19.

The patch — a small, circular-shaped adhesive device — is placed on the skin and wirelessly paired to a mobile device. (Source: SkyWater Technology) Click for larger image.

EnSilica Ltd. has also produced a single-chip medical sensor with wireless connectivity for a wide range of sensors, including electrocardiogram (ECG), temperature, electromechanical, and bioimpedance sensors. The wireless medical sensor ASIC platform can work with multiple processor and DSP configurations and, as an option, can incorporate machine-learning accelerators such as Arm Ethos-U55 to support AI applications.

The sensor chip’s specialized analog front end (AFE) works with single- and multiple-lead ECG with clinical-grade accuracy. It can accurately measure ECG, heart rate, respiratory rate, temperature, and pulse oximetry.

While the above sensor chip launches point toward increasing traction of wearable designs for medical and health-care use cases, what are the major design challenges that developers are confronting to make these ultra-thin devices a design reality? This article expands on the two major issues in medical wearable design: integration woes and optical design.

Integration woes
While sensor modules are widely available to address real-estate and integration challenges in ultra-thin medical wearables, chipmakers are also trying to ease the integration of sensor ICs into the system by offering ultra-compact form factors and supporting a range of operating voltages.

Take the example of the digital temperature sensor unveiled by sensor chipmaker ams, the AS6221, which requires no calibration or linearization and provides measurement outputs via eight I²C interfaces. The temperature sensor IC — available in a 1.5 × 1-mm WLCSP package — draws 6 µA at an output data rate of 4 Hz and is targeted at fitness-monitoring wristbands and smartwatches.

According to ams, the AS6221 is the industry’s most accurate digital temperature sensor, with a measurement accuracy of ±0.09°C over a temperature range from 20°C to 42°C. Other temperature sensors, according to the company, can achieve accuracy no better than ±0.10°C.

Another sensor chipmaker, Integrated Device Technology (IDT), now a wholly owned subsidiary of Renesas, has released a biosensor module to ease integration woes in wearable designs serving medical use cases. The OB1203 module comes in a 4.2 × 2 × 1.2-mm package that includes two LEDs, drivers, the sensors, and a signal-conditioning chip that outputs all sensor data on an I²C bus.

Besides the sensor measuring heart rate and blood oxygen levels, the module’s additional sensors include an ambient light sensor, RGB color sensor, proximity sensor, and pulse oximeter biosensor. Moreover, the design platform features clinical-grade heart rate and blood oxygen saturation (SpO₂) algorithms along with an easily customized Android app, and it requires three data lines to operate.

Enhanced optical design
One of the value propositions that IDT claims for its new sensor module is that it’s optically enhanced. That’s also a key focus of the Health Sensor Platform 3.0 (HSP 3.0) launched by Maxim Integrated. The enhanced optical architecture significantly improves the signal-acquisition quality of sensors in both consumer and regulatory-approved clinical devices.

The HSP 3.0 platform enables wearable devices to monitor cardiac heart and respiratory issues for the management of ailments like chronic obstructive pulmonary disease, infectious diseases such as Covid-19, sleep apnea, and atrial fibrillation. Besides wrist-based wearables, the reference design can also be adopted for other dry-electrode form factors like chest patches and smart rings.

The MAXREFDES104# — the part number for the HSP 3.0 reference design — claims to save at least six months of development time. It facilitates a ready-to-wear wearable device that monitors ECG, heart rate, SpO₂, and body temperature.

The sensor board in the HSP 3.0 reference design incorporates an AFE that integrates PPG and ECG measurements in a single chip. (Source: Maxim Integrated) Click for larger image

Andrew Baker, managing director of the Industrial and Healthcare Business Unit at Maxim Integrated, said that a key challenge in health-care wearable devices is optomechanical design. That includes the arrangement of LEDs and photodetectors.

Therefore, the reference board has implemented the learning acquired over the past couple of years in terms of component placement relative to one another. It places three photodetectors at the bottom. Two photodetectors are focused on capturing the heart rate using green LEDs, while the third one, which is farther away from the LEDs, is used for blood saturation measurement.

“Optical design is all about signal-to-noise ratio [SNR], and the MAXREFDES104# offering an SNR of 110 dB is critical in making two measurements simultaneously,” Baker added. The AFE in the sensor board combines ECG and optical functions, previously implemented on two separate chips.

That means wearable designs can perform synchronous PPG and ECG measurements, even with independent sample rates; it also enables more power optimization. The AFE chip, the MAX86176, offers an SNR of 110 dB to add SpO₂ saturation capability and over 110-dB common-mode rejection ratio for dry-electrode ECG applications.

After unboxing, engineers can load it on the PC using a Bluetooth Low Energy dongle within minutes and start streaming data from the sensor board for data collection and analysis. It comes with source code design files, so designers can take the platform and adapt it according to their specific use case.

As mentioned earlier, these features can save wearable design engineers up to six months in development time. However, if engineers don’t have expertise in optical design as well as in electrode design and ECG, it can take more than six months.