SRED questionnaire

What scientific or technological uncertainties did you attempt to overcome – uncertainties that could not be removed using standard practice? (Maximum 350 words)

Is it possible to design a low-cost compact battery-powered pulse oximetry device using modern highly-integrated components that can be built and repaired exclusively with low-precision low-tech methods? What would such a device look like?

Current devices are tending towards high-volume high-precision automated assembly and zero field-repairability. This is not ideal for situations where it's impractical to send the device back to the manufacturer for service. On the other hand, modern components make it possible to bring size, power consumption and cost of devices to a much lower level than was previously feasible. Our objective was to find the point of intersection between low-tech assembly and repair methods (manual assembly via reflow, low-precision PCB manufacture) and high-tech highly integrated components.

Can we build a multi-wavelength pulse oximetry device that has the size, cost and complexity of a dual-wavelength device, but can additionally detect carboxyhemoglobin and methemoglobin?

What would an optimal open hardware platform for the development of advanced (multi-wavelength) pulse oximetry applications look like? Is this a goal compatible with the previous objective of using low-tech tooling?

A major objective of our work is to deliver designs that are not only useable in their intended application but also useable for research and further development. By having a reference design which is known to work, but is parametric in terms of things like acquisition rate, excitation wavelengths and software processing methods, it's possible to use the same base system for other kinds of measurements with minimal development effort. A flexible low-cost open source pulse oximetry platform makes it very simple to include pulse oximetry as a feature/component in more complex devices, such as vital signs monitoring systems.

What is the cost/precision tradeoff point that minimizes cost and complexity while remaining useable in emergency medicine use?

How do we optimize for multiple price/functionality points with minimal design effort?

(x) What work did you perform in the tax year to overcome the scientific or technological uncertainties described in Line 242? (Summarize the systematic investigation or search) (Maximum 700 words)

After investigating the method of operation of existing devices, we considered how the system can be structured to reduce cost and complexity (minimize number of parts) while sticking to parts that are both widely available and feasible to assemble using low-tech methods. This excluded all parts with BGA packages and many parts with QFN packages, and all packages requiring high density (pin pitch below 0.5mm) PCBs or vias. A further requirement was to keep maximum flexibility as to the number of different wavelengths we would be emitting on, as well as the form factor of the probe that would be attached to the patient.

We decided to use a microcontroller's onboard ADC to reduce cost, assembly complexity and part count. While dedicated pulse oximetry chipsets do exist, they're more expensive than the entire remaining device, limited to two emitters, and generally available only in form factors incompatible with our requirements. We designed a power supply circuit consisting of two stages, a boost converter followed by an LDO regulator so that the input voltage of the system could be either higher or lower than the system operating voltage without affecting function. We settled on a charlieplexed (multiplexed by changing direction of current) array of four LED emitters, expecting that having four wavelengths would allow us to distinguish between the four varieties of hemoglobin that we are interested in detecting. We tested the absorption of different wavelengths of light through the skin, and the different possible geometries of the emitter-receiver(LED-photodiode) configuration such that a clean signal could be obtained even when the emitters were not on the same die and were thus separated by up to several millimeters. We built a series of device prototypes, testing several microcontrollers, emitters, amplifiers and receivers. We found a way to construct the device using a single-sided, single-layer circuit board that could be manufactured with a laser printer and three widely-available chemicals. All assembly could be done either by hand or by automated robot.

We put effort into making the system adaptable to components (specifically emitters and receivers) of variable quality. The on-times of the emitters and the gain and acquisition time of the analog to digital conversion are controllable in software, so changing components or using components with inconsistent properties could be accomodated with no design changes and minor software changes.

We considered adaptability to research, hospital emergency and field emergency settings. We investigated the optimal degree of modularity of the device, and split it up into a core device, optional battery power, optional LCD screen, optional audio feedback and optional external data link.

We also looked into reducing power consumption to a minimum in the field emergency configuration. We investigated the power consumption impact of each component, and determined when each component can be powered down without affecting operation.

What scientific or technological advancements did you achieve as a result of the work described in Line 244? (Maximum 350 words).###

We found a way to construct a pulse oximetry research platform using low-tech assembly methods.

We designed and built a modular open source reference design for pulse oximetry applications and research.

We discovered a configuration of individual emitters that can be used for multi-wavelenght pulse oximetry.

We found ways to bring the costs of pulse oximetry devices down to the point where they can be used in many situations that are currently cost-prohibitive, specifically in underserved areas of the world.