oxullo / Arduino-MAX30100

Arduino library for MAX30100, integrated oximeter and heart rate sensor
GNU General Public License v3.0
192 stars 134 forks source link

Arduino-MAX30100

Build Status

Arduino library for the Maxim Integrated MAX30100 oximetry / heart rate sensor.

MAX30100

Disclaimer

The library is offered only for educational purposes and it is not meant for medical uses. Use it at your sole risk.

Notes

Maxim integrated stopped the production of the MAX30100 in favor of MAX30101 and MAX30102. Therefore this library won't be seeing any further improvement, besides fixes.

IMPORTANT: when submitting issues, make sure to fill ALL the fields indicated in the template text of the issue. The issue will be marked as invalid and closed immediately otherwise.

Hardware

This library has been tested with the MikroElektronika Heart rate click daughterboard:

http://www.mikroe.com/click/heart-rate/

along with an Arduino UNO r3. Any Arduino supporting the Wire library should work.

The only required connection to the sensor is the I2C bus (SDA, SCL lines, pulled up).

An example which shows a possible way to wire up the sensor is shown in extras/arduino-wiring.pdf

Note: The schematics above shows also how to wire up the interrupt line, which is currently not used by the library.

Pull-ups

Since the I2C interface is clocked at 400kHz, make sure that the SDA/SCL lines are pulled up by 4,7kOhm or less resistors.

Architecture

The library offers a low-level driver class, MAX30100. This component allows for low level communication with the device.

A rather simple but working implementation of the heart rate and SpO2 calculation can be found in the PulseOximeter class.

This high level class sets up the sensor and data processing pipelines in order to offer a very simple interface to the data:

The PulseOximeter class is not optimised for battery-based projects.

Examples

The included examples show how to use the PulseOximeter class:

Troubleshooting

Run the MAX30100_Tester example to inspect the state of your rig. When run with a properly connected sensor, it should print:

Initializing MAX30100..Success
Enabling HR/SPO2 mode..done.
Configuring LEDs biases to 50mA..done.
Lowering the current to 7.6mA..done.
Shutting down..done.
Resuming normal operation..done.
Sampling die temperature..done, temp=24.94C
All test pass. Press any key to go into sampling loop mode

Pressing any key, a data stream with the raw values from the photodiode sampling red and infrared is presented. With no finger on the sensor, both values should be close to zero and jump up when a finger is positioned on top of the sensor.

Typical issues when attempting to run the examples:

I2C error or garbage data

In particular when the tester fails with:

Initializing MAX30100..FAILED: I2C error

This is likely to be caused by an improper pullup setup for the I2C lines. Make sure to use 4,7kOhm resistors, checking if the breakout board in use is equipped with pullups.

Logic level compatibility

If you're using a 5V-based microcontroller but the sensor breakout board pulls SDA and SCL up to 3.3V, you should ensure that its inputs are compatible with the 3.3V logic levels. An original Atmel ATMega328p considers anything above 3V as HIGH, so it might work well without level shifting hardware.

Since the MAX30100 I2C pins maximum ratings aren't bound to Vdd, a cheap option to avoid level shifting is to simply pull SDA and SCL up to 5V instead of 3.3V.

Sketchy beat frequency readouts

The beat detector uses the IR LED to track the heartbeat. The IR LED is biased by default at 50mA on all examples, excluding the Tester (which sets it to 7.6mA). This value is somehow critical and it must be experimented with.

The current can be adjusted using PulseOximeter::setIRLedCurrent(). Check the _MAX30100Minimal example.

Advanced debugging

Two tools are available for further inspection and error reporting:

Both tools have additional information on the README.md in their respective directories.

Tested devices

This combination works without level shifting devices at 400kHz I2C clock rate.

As above, working at 400kHz

Even if this combination works (MAX30100 communication), the slower clock speed fails to deliver the required performance deadlines for a 100Hz sampling.

Troubled breakouts

This breakout board: http://www.sunrom.com/m/5337

Has pullups on the Vdd (1.8V) line. To make it work, the three 4k7 pullups must be desoldered and external 4.7k pullups to Vcc of the MCU must be added.