Pulse-oximeter

Owned by Matt Ball
Last updated: Jan 17, 2009 by bdg4Version comment
2 min read

Background

The pulse oximeter is one of the most common medical instruments, and is the most frequently requested piece of medical equipment. Prices have dropped for this product to the point where nearly every patient in a US hospital will be connected to such a device. However, in the developing world, pulse oximeters are still difficult to find. The most important function is to display heart rate. Displaying oxygen concentrations is a secondary desire. The device must use non-disposable sensors. It should operate on a finger of an adult. It should provide audible (beeps) for each heart beat and ideally also provide a digital display of heart rate.

Past work:

A team has designed a low-cost circuit that beeps on every heart beat. Now someone needs to add the display of heart rate. Probably needs a PIC programmer.

Additional specs

  • Device available for kit design or local production only
  • Cost: <$8 in quantities of 1

Current Work

  • Device will use a PIC and a 9V battery
  • Probe Design: Looking for way to create robust, reusable probe, somewhat resistant to ambient light
  • LED considerations: needs to be in IR spectrum, photodiode detector needs to match IR LED, Low Power
  • Approximate price list(if mass produced): IR LED Mass Production: $.12
  • Photodiode - OP950,  935 nm peak receptance, 4 pF input capacitance(relevant for feedback capacitor), i forward about 1 mA)
  • LED = LNA2802L(950 nm, 20 degree)
    • Voltage Forward; typical 1.3 V, typical 50 mA. So we'll use 5 V power supply, resistor needed = (5V - 1.5V)/.05 = 70 ohms
  • Transimpedance circuit --- didn't use capacitor in initial test, webster's book claims its very important to use a capacitor with a very specific value to minimize gain peaking and improve stability. Will have to calculate next time.
  • Opamp for transimpedance circuit - OPA347 for now assume using 15 k feedback resistor
  • note: if we switch the feedback resistor to 100k we'd need a 2.2 pF capacitor - any thoughts on the feedback resistor, all I've found thus far is bigger is better... but there seems like there should be an upper limit
  • feedback capacitor = 1/(4*pi*Rf*fc)*(1+sqrt(1+8*pi*Rf*Ci*fc) where fc is unit gain frequency of op amp, C1 i total input capacitance = photodiode junction capacitance + opamp input capcitance, Rf is the feedback resistance
  • Rf = 15k
  • opamimp input capacitance = 6 pf (3 pf differential mode)
  • fc = 166805 hz
  • photodiode junction capacitance = 4 pf
  • need feedback capacitor  38 pf
  • Cost Analysis:
  • Probably leaning towards atmel microcontrollers, low power(selectable voltage range), sleep mode, CHEAP, easier to program...http://www.mouser.com/Search/Refine.aspx?Keyword=556-ATTINY88-AU - only 1.54 if you buy at least 100, that's an excellent price... would require buying a programmer but this is not a very expensive addition
  •  1/(2*pi*(330*10^3*470*10^3*.1*10^(-6)*.1*10^(-6))^(.5)) = 4.04 hz(240 bpm) cut of frequency, 2 pole LP filter seen out http://ourworld.compuserve.com/homepages/Bill_Bowden/opamp.htm
    • two .1 uf caps, one 330kohm resister, one 470 k

Recommended Resources

  • Webster, Design of Pulse Oximeters (Available at Perkins)
  • Moyle, Pulse Oximetry (Available at Duke as Electronic Resource - Ebook)

Contact

if interested, contact Ben Grant (benjamin.grant@duke.edu)