Performance test report - Renesas OB1203
2 Thursday, February 18, 2021 Revision 1.0
Test Method
1. Calibrate simulator settings to human skin color
2. Verify Perfusion Index reading of OB1203SD-RL-EVK
3. Test Limits of OB1203SD-RL-EVK performance and accuracy
4. Results analysis
Background
Pulse oximeter accuracy testing must be done on humans which is expensive and time consuming.
Therefore the electro-optical characteristics of the devices used must be tested to ensure that adequate
performance is available to meet real world performance requirements. Once the hardware platform
and software algorithm meet requirements calibration using human subject test data is straightforward.
Simulator Calibration
The OB1203SD-RL-EVK device was applied to a light skinned person who has participated in many
studies and therefore has a well-known light transmission level. In this case the measured ADC counts
were 36,000 infrared and 35,000 red, with the LEDs set at 13 mA (250mA max.) infrared and 9mA
(125mA max) red utilizing the provided auto gain control function. With no input signal the IR led is set
at 165mA and the red led at 126mA.
From studying hundreds of people Physio Monitor has determined that the darkest skinned subjects
recruited for human testing are about 20X darker than the reference light skinned person.
From the above the ADC counts on the darkest person can be predicted assuming that the LEDs go to
maximum at the low level;
Measured counts * maximum LED current / (measured LED current * 20)
Light Skin Darken Light Skin Current Change Predicted Counts Add
Counts div by 20 Current Light to dark For Dark Skin Crosstalk
IR 36000 1800 13 12.69 22846 24196
Red 35000 1750 9 13.89 24306 25106
The next step was to adjust the setup on the AECG100.
1. The OB1203SD-RL-EVK was mounted on the AECG100 as shown.
A weight on top of the board ensures that the OB1203SD-RL-EVK stays in contact with the fixture. A
lithium battery to power the OB1203SD-RL-EVK is used for convenience. The LCD display is was located
for convenient viewing.