Reducing motor noise
Author: Bao Nguyen
Date: September 9, 2019
When looking for a new pump for the eDNA project, we found this tiny pump on Amazon that delivers the desired flow-rate as well as pressure. However, the noise created by the motor is enough to corrupt all of our I2C sensors, thus hang the whole system.
Measure to reduce motor noise in power line has been implemented by Mitch Nelke in the previous version of the OpenSampler. eDNA electronic inherent this design completely from Mitch. However, the motor now has the ability to corrupt data even on a separate power supply. The motor was powered with a bench power supply, while the Feather M0 and sensors are powered via laptop USB. Looking around the internet suggested that the spark at DC motor brushes are causing the interference. I found more suggestion on StackExchange , GM3 Noise Suppressor, Polulu.
The first step that I took is adding one 0.1uF capacitor between the motor terminals. The motor is then run for 3 hours as an effort to shape the commuters to reduce sparks.
| 0.1uF capacitor added |
|---|
 |
So we add more capacitors between the terminal and the casing. After that, a ferrite bead was also added on the power cable to reduce noise. The result was promising since the sensor would work around 5 inches away from the motor without hanging the whole system. This solution lasts us for around 2 months before the noise starts to corrupt the pressure sensor.
| More capacitors and ferrite bead |
|---|
 |
Another suggestion by Kawin and Dr.John S Selker is to use a Faraday cage to block the interference. We tried with a Fanta aluminum soda can and it seems to work for a couple days before interference happens again. A solution that I should have thought of at the start, is just to move the sensor further away from the motor. The sensor is then moved to the other side of the machine, furthest away from the motor.
| Before | After |
|---|---|
 |
 |
By changing just the location of the sensor, it still did not work until we purchase TitanRF on Amazon. 3 layers of this material around the motor is good enough for us to have a stable connection. We have been running the machine for 1 weeks now without any slowdown or fail in I2C communication.
Keep in mind when installing this material, the whole fabric is conductive. Tape the motor with Kapton tape to prevent short across the motor terminal.
| Added Faraday Fabric |
|---|
 |
*update: System still works fine as of December 20th, 2019