How to design a better pulse oximeter: Specifications

Article By : Robert Finnerty

It is more important than ever to design medical devices that are more convenient to use and consume less power.

It is more important than ever to design medical devices that are more convenient to use and consume less power. This article covers the fundamentals of SpO2 measurement and explores how to reduce design complexity, simplify mechanical design, and decrease overall power consumption.

Traditionally, peripheral blood oxygen saturation (SpO2) is a measurement taken at the peripherals of the body on the finger or ear, most commonly with a clip device to determine the ratio of oxygen saturated hemoglobin to total hemoglobin. This measurement is used to tell how well red blood cells are transporting oxygen from the lungs to other parts of the body. Normal SpO2 levels vary from 95% to 100% in a healthy adult. Levels below this range indicate a condition known as hypoxemia. This means that the body is not transporting enough oxygen to maintain healthy organs and cognitive function.

A person suffering from hypoxemia may experience dizziness, confusion, shortness of breath, and headaches. Several medical conditions can cause poor blood oxygenation and may require continuous or intermittent monitoring at home or in a clinical setting.

SpO2 is one of the most common vital signs recorded within a clinical setting. Some conditions that require continuous SpO2 monitoring include asthma, heart disease, COPD, lung disease, pneumonia, and COVID-19 induced hypoxia.

One of the ways to determine whether symptomatic COVID-19 patients need hospitalization is by monitoring their SpO2 levels. If those levels fall below the baseline number (usually under 92%), they need to be checked into an emergency room.

The Recent Link Between COVID-19 and Hypoxia

Very recently, COVID-19 patients have been diagnosed with a particularly insidious condition known as silent hypoxia. Silent hypoxia can do severe damage to the body before any of the typical COVID-19 respiratory symptoms, like shortness of breath, occur. An article on the National Center for Biotechnology Information website1 states “the ability to detect this silent form of hypoxia in COVID-19 patients before they begin to experience shortness of breath is critical for preventing the pneumonia from progressing to a dangerous level.”

SpO2 monitoring is also a key indicator in diagnosing sleep apnea. Obstructive sleep apnea causes the airways to become partially or fully blocked during sleep. This can be observed as long pauses in breathing or periods of shallow breathing causing temporary hypoxia. If untreated over time, sleep apnea can increase the likelihood of heart attack, stroke, and obesity. It is estimated that sleep apnea affects between 1% to 6% of the total adult population.

The Urgent Need for a Better Pulse Oximeter Now and in the Future

As patient care trends toward ambulatory and in-home monitoring, there is a need to develop vital sign monitoring devices that will not impede users from completing daily tasks. In the case of SpO2, monitoring areas other than the finger and ear will present a host of design challenges. The recent emergence of silent hypoxia makes the case for development of more portable clinical-grade pulse oximeter units even more compelling.

This article will explain some of the fundamental principles of SpO2 measurement and how optical AFEs reduce design complexity for medical grade SpO2 devices. For example, integrated high performance automatic ambient light rejection reduces the burden on mechanical and electronic design. Similarly, high dynamic range at lower power consumption reduces the number of photodiodes or LED current in a design to determine slight variations in patient SpO2 level efficiently. Finally, digital integrator options allow users to enter an extremely efficient power consumption mode to enable longer run times in portable PPG solutions by disabling analog blocks in the optical signal path.

Oxygen Saturation

Oxygen saturation is the percentage of oxygen saturated hemoglobin within the blood with respect to the total available hemoglobin. The gold standard for measuring oxygen saturation is the atrial blood oxygenation measurement, SaO2. However, this method requires laboratory-based blood gas analysis of a blood sample. The calibration section covers this in greater depth.

SpO2 is an estimate of the oxygen saturation levels measured at the peripherals of the body, using a pulse oximeter. Until recently, the most common way to measure oxygen saturation has been to use a pulse oximeter positioned on the finger.

A pulse oximeter works on the principal that absorption of light in oxygenated hemoglobin (HbO2) and deoxygenated hemoglobin (RHb) differ significantly at specific light wavelengths. Figure 1 shows the extinction coefficient of HbO2, Hb, and methemoglobin (MetHb) across the visible and infrared light spectrum. The extinction coefficient is a measurement of how strongly a chemical substance absorbs light at a given wavelength. From Figure 1, it can be seen that HbO2 absorbs more red light (600 nm) and allows more infrared light (940 nm) to pass through. RHb absorbs more light at infrared wavelengths, which allows more red light to pass through than in HbO2.

The most basic pulse oximeter consists of two LED (one red 660 nm LED and one infrared (IR) 940 nm LED) and a single photodiode (PD) in a reflective or transmissive configuration (see Figure 4). The pulse oximeter will pulse the red LED and measure the resulting signal on the PD. Repeat this for the IR LED and finally with both LEDs off to get a baseline for any ambient external light sources. This generates a photoplethysmography (PPG) signal for both wavelengths.

Figure 1. Extinction factor of light through hemoglobin.

Figure 2. Basic pulse oximeter circuit.

The signal contains DC and AC components. The DC component is due to constant reflective matter such as skin, muscle and bone, and venous blood. When a body is at rest and motion is less of a factor, the AC component comprises mainly of reflected light from the pulsation of artery blood. The AC component depends on heart rate and artery thickness, with more reflected or transmitted light in systolic (pump) than the dystopic (relaxation). During the systolic phase, blood is pumped from the heart, which increases atrial blood pressure. The increase in blood pressure expands the arteries and leads to an increase in atrial blood volume. This increase in blood causes an increase in light absorption. Blood pressure drops during the diastolic phase and therefore so does the absorption of light. Figure 3 shows the diastolic trough and systolic peaks caused by the beating heart.

The Beer-Lambert law explains that light decays exponentially when travelling though absorptive material. This can be used to determine the level of oxygenated hemoglobin to total hemoglobin.

The intensity of light absorbed at the diastole and systole are related by:

Where α measures the absorption rate of light in atrial blood and d2 is the AC amplitude of the PPG signal (see Figure 3). Idiastole is equal to the DC component labelled d1.

Figure 3. Light attenuation through tissue.

By computing AC and DC from a PPG signal, we are able to determine the change in absorption of light in atrial blood –α.d2 caused by blood pumping from the heart, with no contribution from other tissue.

The ratio of the AC component to the DC component is known as the perfusion index, which is the ratio of the pulsating blood flow to the nonpulsatile static blood flow. The goal of a PPG-based heart rate or SpO2 measurement system is to increase the AC to DC signal ratio:


The perfusion index for infrared and red wavelengths can be used to calculate the ratio of ratios (RoR), which is the ratio of PIred to PLir. As the absorption of the light at a given wavelength is proportional to the

In theory, the RoR can be substituted into the following formula to compute SpO2:


EHbO2,red = extinction coefficient of HbO2 at 600 nm,

EHbO2,ired = extinction coefficient of HbO2 at 940 nm

ERHb,ired = extinction coefficient of RHb at 940 nm,

ERHb,red = extinction coefficient of RHb at 600 nm

However, the Beer-Lambert law cannot be used directly as there are a number of variable factors in every optical design that cause variations to the RoR to SpO2 relationship. These include mechanical baffle design, LED to PD spacing, electronic and mechanical ambient light rejection, PD gain errors, and many more.

To obtain clinical grade accuracy from a PPG-based SpO2 pulse oximeter, a lookup table or algorithm must be developed for the correlation between RoR and SpO2.


Calibration of the measurement system is required to develop a high accuracy SpO2 algorithm. To calibrate an SpO2 system, a study must be completed where a participant’s blood oxygen levels are medically reduced, monitored, and overseen by a medical professional. This is known as a hypoxia study.

The SpO2 measurement system can only be as accurate as the reference. Reference options include medical grade finger clip pulse oximeters and the gold standard co-oximeter. The co-oximeter is an invasive method of measuring the oxygen saturation of blood that yields high accuracy, but in most cases is not convenient to administer.

The calibration process is used to generate a best fit curve of RoR value calculated from the optical SpO2 device to the co-oximeter SaO2 measurement. This curve is used to generate a lookup table or equation for calculating SpO2.

Calibration will be required for all SpO2 designs as RoR is dependent on a number of variables such as LED wavelength and intensity, PD response, body placement, and ambient light rejection, which will differ with each design.

An increased perfusion index and, in turn, a high AC dynamic range on the red and IR wavelengths will increase the sensitivity of the RoR calculation and, in turn, return a more accurate SpO2 measurement.

During a hypoxia study, 200 measurements equally spaced between 100% and 70% blood oxygen saturation need to be recorded. Subjects are chosen with a variety of colored skin tone, and an equal spread of age and gender. This variation in skin tone, age, and gender accounts for differing perfusion index results from a spread of individuals.

The overall error for transmissive pulse oximeters must be ≤3.0% and ≤3.5% for reflective configuration.

In the next article, we’ll cover design considerations such as transmissive vs reflective, sensor positioning, perfusion index, motion artifacts, and specifics of designing with an optical AFE. 


  1. Toshiyo Tamura. “Current Progress of Photoplethysmography and SpO2 for Health Monitoring.” Biomedical Engineering Letters, February 2019.
  2. Jihyoung Lee, Kenta Matsumura, Ken-Ichi Yamakoshi, Peter Rolfe, Shinobu Tanaka, and Takehiro Yamakoshi. “Comparison Between Red Green and Blue Light Reflection Photoplethysmography for Heart Rate Monitoring During Motion.” 2013 35th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), July 2013.


This article was originally published on Embedded.

Robert Finnerty is a systems applications engineer at Analog Devices where he works in the Digital Healthcare Group based in Limerick, Ireland. He works closely with the Vital Signs Monitoring Group, focusing on optical and impedance measurement solutions. Rob joined the precision converters group within ADI in 2012 and has focused on low bandwidth precision measurement. He holds a bachelor’s degree in electronic and electrical engineering (B.E.E.E) from National University of Ireland Galway (NUIG). He can be reached at


Leave a comment