ADHESIVES: The Importance of CTE for Assembly Reliability
Hydrogel Material Improves Success of Transplanting Islet Cells
Skin-worn Wearable Devices: Tomorrow’s Vision, Today’s Realities
New 3D Printing Method Promises New Era for Medical Implants
New Method Manufactures Heart Valves in Minutes
Lasers: The Perfect Tool for Industry 4.0
Sonic Cyber Attack Shows Security Holes in Ubiquitous Sensors
Open Core Engineering: New Freedom for Machine Automation Programming
Industry 4.0: Virtual Twin Controls Production
Features

To measure pulsed oximetry (SpO2), a red LED and an infrared LED are used. The measurement is taken via a finger because of its strong capillary concentration. Pulse oximetry is a noninvasive method used by physicians to assess and quickly control the respiratory function of a patient. The ratio of red and infrared light through the photodiode indicates the percentage of oxygenated hemoglobin versus deoxygenated hemoglobin in blood. The oxygen saturation in the blood is also called SpO2. Oximetry is, therefore, based on the measurement of the light absorption of hemoglobin in blood capillaries and specifically on the rate of oxyhemoglobin (oxygenated hemoglobin) and deoxyhemoglobin (deoxygenated hemoglobin) of each red blood cell. A 98 percent SpO2 means that each red blood cell is loaded with 98 percent oxyhemoglobin and 2 percent deoxyhemoglobin.

Module solutions (analog front end, photodiode, and LED in the same package) are ideal for surface-constrained applications that do not require optimization of the optical measurement. Therefore, the ADPD142, which includes a red LED and an IR LED, allows SpO2 measurement on the finger. Its successor, the ADPD144, offers an improved mechanical design, which reduces internal light pollution (the direct light between the LED and the photodiode). It provides an average measurement error of 2.6 percent over 24,425 sample measurements, making it FDA compliant. The ADPD144’s package measures 5 × 2.8 mm, with a height of 1.35 mm.

To maximize the modulation index and thus the quality of the measured signal, there must be a minimum spacing between LED and photodiode, which may not be optimal in a module, where space is constrained. Thus, for applications such as sport watches, with additional constraints due to the movement, perspiration, and the displacement of the skin-watch contact, only solutions with an external LED and photodiode from the photometric front end should be used.

Software includes drivers of the photometric sensors and accelerometers as well as a motion compensation algorithm. This algorithm now runs on a Cortex M3 core, the ADuCM3027, which takes only 1.5 MIPS for 13 KB ROM and 7.8 KB of RAM. This is a major breakthrough because until recently, this type of algorithm required floating point calculations, and thus a Cortex M4 processor type, which is more power hungry and more expensive.

It is important to note that skin color and tattoos affect the quality of the measured reflected signal. A location other than on a tattoo should be used to obtain an accurate measurement. For darker skin, the modulation index should be slightly reduced to optimize the optical design.

Ultralow Power Platform

The following discussion reviews how determine the power consumption of the watch discussed earlier. The following assumes that a motion compensation algorithm is executed on the Cortex M3 ADuCM3027, and some characteristics are determined to establish the LEDs’ power consumption.

The ADPD103 sends an LED pulse train on one or two time slots. This allows it to send a different number of pulses from one LED to another. Its power consumption is the sum of the AFE and LED. Here is an example of these conditions:

  • Fs = 100 Hz; 2 slots; pulse period A = 20 μs; pulse period B = 40 μs;
  • Number of pulses A = 4; number of pulses B = 8
  • Maximum current in LED A = 25 mA; maximum current in LED B = 100 mA
  • Pulse duration A = 3 μs; pulse duration B = 3 μs
    • Thus, the effective current in the LED_A = (3 * 4 / 10000) * 25 mA = 30 μA
    • Thus, the effective current in the LED_B = (3 * 8 / 10000) * 100 mA = 240 μA
  • Current in the A channel of the AFE = Fs((20+Pulse Count * Pulse Period) * Vddpeak+0.13) = 100((20 + 4 * 20) * 0.0093 + 0.13) = 106 μA
  • Current in the B channel of the AFE = Fs((20+Pulse Count * Pulse Period) * Vddpeak+0.20) = 100((20 + 8 * 20) * 00093 + 0.20) = 187 μA
  • The total current of ADPD103 (including consumption of both LEDs) is 563 μA.

As indicated above, the motion compensation algorithm, which has a running frequency of approximately 1.5 MHz, needs only 1.5 MIPS to operate. The ADuCM3027 consumes 38 μA/MHz, and the microcontroller consumes 57 μA. The ADXL362 uses 2 μA with a sampling frequency of 100 Hz. Therefore, the AFE and LED, Cortex M3, and accelerometer system consumes 622 μA in this example. This low consumption maximizes usage time without recharging the built-in LiPo battery in this watch. In standby, the ADPD103 consumes 3.5 μA. Its successor will decrease this value to 1 μA.

It should be noted that this example shows a power calculation that does not correspond to a precise application. A design may produce better or worse results, depending on the targeted application, the current through the LEDs, and the sampling frequency having a direct link with the consumption of the system.

Conclusion

The integration of wearable healthcare technology and embedded vision technology is key to achieving a true home health monitoring system. The two systems together monitor the activities of an individual and enable the individual to monitor his or her vital signs as well. The advent of sensor networks and wearable sensor technologies will be critical to enabling independent living for the elderly.

This article was written by Vidushi Kshatri, strategic marketing manager at Analog Devices, Norwood, MA. For more information, click here.

« Start Prev 1 2 Next End»