“Optical pulse sensor is a sensor that uses light sensing technology, one of semiconductor technologies, to measure pulse. This kind of light sensing technology is to make the light source (LED) illuminate the living body, and use the photosensitive part-photodiode (hereinafter referred to as “photoelectric Di”) or phototransistor to measure the light transmitted or reflected in the living body. Arterial blood contains hemoglobin with light absorption properties, so by sensing the amount of light in time series, the change in hemoglobin amount, that is, the pulse signal, can be obtained.
Optical pulse sensor is a sensor that uses light sensing technology, one of semiconductor technologies, to measure pulse. This kind of light sensing technology is to make the light source (LED) illuminate the living body, and use the photosensitive part-photodiode (hereinafter referred to as “photoelectric Di”) or phototransistor to measure the light transmitted or reflected in the living body. Arterial blood contains hemoglobin with light absorption properties, so by sensing the amount of light in time series, the change in hemoglobin amount, that is, the pulse signal, can be obtained.
In recent years, in smart bracelets and smart watches equipped with commercially available optical pulse sensors, considering the wearability and load between the skin and the skin, most mainstream products use reflective light sensors that use green light. Since the penetration depth of green light to the organism is shallow, it is not easily affected by tissues other than blood. In addition, the absorption coefficient of hemoglobin is large, so pulse signals with large pulsating components can be measured.
This article will introduce the characteristics and application of ROHM optical pulse sensor “BH1790GLC”, which is very suitable for wearable devices.
1 Specifications required for pulse sensors for wearable devices
1.1 Low power consumption
Since wearable devices need to be worn on the body, there are restrictions on the size and weight of the device itself, and it is difficult to increase the battery capacity. Therefore, it is very important to make it work with low power consumption. Figure 1 shows the current consumption of the pulse sensor. In the pulse sensor using the conventional technology, the current consumption of the LED drive unit and the analog front end (hereinafter referred to as “AFE”) unit is relatively large. In order to reduce the current of the LED drive unit, the BH1790GLC improves the sensitivity of the photosensitive unit, so that the pulse signal can be obtained even under the condition of low LED brightness, and integrates the AFE unit into one chip, thereby effectively reducing the current consumption.
The following is a specific introduction to the method of increasing the sensitivity of the photosensitive part. The previous technology uses a transimpedance amplifier circuit (hereinafter referred to as “TIA circuit”) to convert the current generated by the photoelectric Di into a voltage. The TIA circuit is a circuit that uses an amplifier and a resistor to convert the current generated by the photoelectric Di into a voltage. However, the current generated when the light hits the photoelectric Di is very small, and increasing the sensitivity requires increasing the resistance value. Therefore, the noise of the amplifier and the thermal noise of the resistor have always been problems to be solved urgently.
BH1790GLC uses an integrating charge amplifier to achieve higher sensitivity (Figure 2). The integral charge amplifier filters the noise during the charging period by charging the current of the photoelectric Di to the capacitor for a certain period of time to convert the current into a voltage, thereby reducing the noise. Therefore, low-noise photometry can be achieved, and the sensitivity of the photosensitive part can be improved. If the sensitivity of the photosensitive part is higher, a smaller photosensitive element can be used to fully measure light, so the photoelectric Di and AFE parts are easier to integrate on one chip. Moreover, the pulse can be measured under low-brightness conditions, which can reduce the current consumption of the LED drive unit. BH1790GLC adopts an integrating charge amplifier, which reduces current consumption by up to 85% compared with the previous technology.
Figure 1 Current consumption of the pulse sensor
Figure 2 Current-voltage conversion amplifier circuit example
1.2 Infrared filter characteristics
Since wearable devices are also used outdoors, they need to use light sensors that filter out interference light such as infrared rays that easily penetrate the human body. The photoelectric Di using ordinary Si PCB board has higher sensitivity at infrared wavelength (850nm), so it is easily affected by interference light.
The photoelectric Di equipped with BH1790GLC reaches its peak sensitivity near 530nm in the green band. This Di is realized by utilizing the property that the shallower the distance from the Si surface to the PN junction, the more the peak of sensitivity shifts to the short wavelength end, and the photoelectric Di arranged on the shallower part of the Si surface.
Not only that, BH1790GLC is also equipped with two types of optical filters, a color film and a multilayer film filter, on the Si PCB to filter out red light and infrared components. The color film filter has the characteristic of filtering out red light, and the multilayer film filter has the characteristic of filtering infrared light, which makes the photosensitive part only allow the light of the green wavelength band to pass. (image 3).
Figure 4 shows the actual measurement result of pulse signal using BH1790GLC and ordinary photoelectric Di. When measuring the pulse signal in an environment where interference light is generated, the noise of ordinary photoelectric Di is increased due to the superposition of the interference light component and the pulse signal, while the BH1790GLC is very little affected by the interference light, and the pulse can be obtained stably. This makes it possible to obtain high-quality pulse signals in sunny beaches and parks, and is a pulse sensor that is very suitable for wearable devices.
Figure 3 Spectroscopic characteristics of the photosensitive part of BH1790GLC
Figure 4 Comparison of pulse signals under interference light environment
2 Pulse sensor system
This time, a bracelet-type pulsimeter using BH1790GLC to measure pulse rate was produced. The pulse sensor part consists of a pulse sensor (ROHM BH1790GLC), an LED (ROHM SMLE13EC8T), an acceleration sensor (Kionix KX-022), and a microcontroller (Lapistone Semiconductor ML630Q791) (Figure 5). The communication with the outside is carried out through the Bluetooth LE module (Lapistone Semiconductor Corporation MK71050-03) mounted on another PCB board.
The pulse signal measured using the manufactured pulsimeter is shown in Figure 6.
Figure 5 Pulse meter using BH1790GLC
Figure 6 Pulse signal measured with BH1790GLC
3 Pulse algorithm
It is clear from the results in Fig. 6 that due to the different capillary density, the pulse signal levels at different measurement locations vary greatly. Larger pulse signals can be obtained from fingertips and earlobes, but smaller pulse signals can be obtained from wrists wearing smart bracelets. In addition, the wrist is also a part of frequent activities in daily life, and the impact of dynamic noise is greater. Therefore, it is difficult to accurately calculate the pulse rate based on the pulse signal of the wrist.
In response to this issue, ROHM uses an acceleration sensor to develop a pulse rate algorithm with a built-in body motion noise cancellation function. Body motion noise is caused by sensor position deviation and blood flow changes caused by physical activity, so the noise component is related to the signal of the acceleration sensor. Using this phenomenon, an algorithm was created that extracts body motion noise components from the acceleration sensor and eliminates the noise interference contained in the pulse signal. Fig. 7 is a comparison between the pulse rate curve of the actual treadmill exercise and the measurement result of the electrode type heart rate monitor. Demonstration machines equipped with ROHM algorithm show better tracking performance than heart rate monitors, the influence of receiver dynamic noise is less, and the pulse rate can be calculated with high precision.
Figure 7 Pulse rate measurement results during treadmill exercise
4 Future Outlook
The pulse meter has been popularized as an application that uses pulse signals, and the development of applications such as obtaining pressure information through pulse fluctuation analysis and obtaining blood pressure information through waveform analysis is also accelerating. When these functions are mounted on wearable devices and can be measured stably, it will be just around the corner to realize early detection of disease signs based on daily changes in body state. Today, ROHM is also working to support the development of pulse sensors that measure biological information.
To perform pressure measurement or obtain blood pressure information through pulse, the time resolution of the pulse signal needs to be improved. So ROHM trial-produced a pulse sensor that increased the sampling frequency to 1024 Hz. As shown in Figure 8, it was confirmed that the pulse sensor can detect pulse signals with high resolution and high accuracy. In the future, ROHM will also work to develop algorithms for calculating pressure and blood pressure information using this pulse sensor.
Figure 8 Comparison of pulse signals between BH1790GLC and 1024Hz frequency products