Identifying IMUs on PixHawk Orange Cube from Mission Planner tlogs

h_jackson · May 5, 2022, 1:31pm

Hi everyone,

I have a PixHawk Orange Cube.
Recorded some tlogs using Mission Planner version 1.3.70.

Now trying to analyse the tlogs, have used Mission Planner version 1.3.76 to convert the tlogs to csv files, and the identity of the specific IMUs is not seeming straightforward.

The 3 IMU readings we see as:
mavlink_raw_imu_t
mavlink_scaled_imu2_t
mavlink_scaled_imu3_t

Initially we assumed they were listed in the order IMU1, IMU2 & IMU3, but mavlink_raw_imu_t is the only line which lists 9DOF data (has mag data) which means it is what the Orange Cube calls IMU3 (as listed here: cubepilot-docs/the-cube-module-overview.md at master · CubePilot/cubepilot-docs · GitHub)

Does Mission Planner rename the IMUs, and how do I tell which of the chips is associated with each identifier?

Thanks in advance

Michael_Oborne · May 10, 2022, 11:14am

i havnt looked hard into this, but the raw_imu message will contain the data from the current primary sensor/ the primate instance

github.com

ArduPilot/ardupilot/blob/c33f6f3a8d4d94f4603eeace31e3cc38eb991fc1/libraries/GCS_MAVLink/GCS_Common.cpp#L1769

      
        
            
            
const Vector3f &accel = ins.get_accel(0);
            const Vector3f &gyro = ins.get_gyro(0);
            Vector3f mag;
            if (compass.get_count() >= 1) {
                mag = compass.get_field(0);
            } else {
                mag.zero();
            }
            
            
mavlink_msg_raw_imu_send(
                chan,
                AP_HAL::micros64(),
                accel.x * 1000.0f / GRAVITY_MSS,
                accel.y * 1000.0f / GRAVITY_MSS,
                accel.z * 1000.0f / GRAVITY_MSS,
                gyro.x * 1000.0f,
                gyro.y * 1000.0f,
                gyro.z * 1000.0f,
                mag.x,
                mag.y,

so this means it will report mag instance 0, acc instance 0 and gyr instance 0.
these could be technicly from any of the sensors that are inited at instance 0

i would check the HWID’s to confirm the order