Hi guys, do you think it would be a good idea to shield the accellerometer using aluminum foil? There is a lot of noise on it when I turn the motors on. I tried filtering it, but the signal gets to slow with the filter on, so I cant detect any small changes on the orientation. Where in the board is the accellerometer is located ? I wanna use the smallest piece of aluminum as possible to try to reduce the risk of any short circuit.
Thanks!
Hi asura.
Could you tell me what Navio version (Navio+ or Navio 2) you use? And how do you detect the noise?
Log file with detected noise and photo of your drone will be really helpful. Could you please provide it?
I have the Navio2 board. I detected the noise simply by printing the read value of the accel. Is very noticable the difference when the motors are on and off. When its off it works great, very small deflections that can be eliminated by a simple low pass filter, and everything is fine. When the motors are on, the readings are really random with a lot of spikes in both directions, making my drone go crazy.
We need more info to solve your problem:
- Log files
- Screenshot of normal values
- Screenshot of noise values
- Photo of your drone
- Version of ArduPilot you are running
After you provide this information, try to unplug battery and power Raspberry via micro-usb port. Will you have noise in this case?
Hi, sorry about the delay.
Im not using ardupilot, Im using python codes to read the sensors.
Therefore I dont have log files. What I can say is that usually the (sensor readings - offset) stay around 0.0XXX (like 0.03321, 0.041231 and so forth) when the motors are off. As soon as the motors are turned on and given some power, they fluctuate to even X.XX (usually around 1.23123, 1.03213, 0.942131) , so clearly rending the date unusable. This happens if I put a light low pass filter with a cut of frequency of 1 Hz. If I reduce the frequency to 0.1 Hertz the noise is greatly reduced, but then the input gets way to slow variation, so not decting sudden moves on the drone.
This is how it is set up right now:
Its inside the body of a vtol drone
Hi, Asura
Are you sure that it is not vibration noise?
yep 
I tried giving it a good ol’ shake with the motors of and the input was all as expected. Simulated a flight cenario with the drone hanging from a wire with a fan on to emulate some wind and still readings were fine. The gyro itself can work fine with the motors on, it also has noise but its not nearly as big as the accellerometer , so I can filter it all out. I did needed the accellerometer inputs for vertical landing and taking off, thats why I am seeking some help from you guys
When motors are actually rotating you are getting different vibration frequencies compared to what you get when you shake it with your hands. Considering the very soft mount whole thing could be resonating and ruining accelerometer measurements.