Self - test way too off

[borstan]

Hello there, I have a problem with self - test values that show more than 40% difference from the factory values. In other examples, I saw that output should be as close to 0 as possible. Could someone help me solve this one? Thank you for any advice! I am using Sparkfun ESP32 Thing, MPU9250, Interrupt connected to the GPIO35.

CODE: MPU9250_MS5637_ESP32 Basic Example Code (I commented all lines with MS5637 since I am not using it)

---

00:01:52.300 -> Scanning... 00:01:52.300 -> I2C device found at address 0x0C ! 00:01:52.300 -> I2C device found at address 0x68 ! 00:01:52.300 -> done 00:01:52.300 -> 00:01:52.300 -> MPU9250 9-axis motion sensor... 00:01:52.300 -> MPU9250 I AM 71 I should be 71 00:01:53.286 -> MPU9250 is online... 00:01:53.557 -> x-axis self test: acceleration trim within : -48.5% of factory value 00:01:53.591 -> y-axis self test: acceleration trim within : -51.7% of factory value 00:01:53.591 -> z-axis self test: acceleration trim within : -49.1% of factory value 00:01:53.591 -> x-axis self test: gyration trim within : -51.5% of factory value 00:01:53.591 -> y-axis self test: gyration trim within : -26.7% of factory value 00:01:53.591 -> z-axis self test: gyration trim within : -36.8% of factory value 00:01:54.607 -> Calibrate gyro and accel 00:01:54.978 -> accel biases (mg) 00:01:54.978 -> 27.34 00:01:54.978 -> -53.83 00:01:54.978 -> 126.89 00:01:54.978 -> gyro biases (dps) 00:01:54.978 -> 8.69 00:01:54.978 -> -0.16 00:01:54.978 -> -18.47 00:01:56.371 -> MPU9250 initialized for active data mode.... 00:01:56.371 -> AK8963 I AM 48 I should be 48 00:01:57.424 -> AK8963 initialized for active data mode.... 00:01:57.424 -> Mag Calibration: Wave device in a figure eight until done! 00:02:19.389 -> Mag Calibration done! 00:02:19.389 -> AK8963 mag biases (mG) 00:02:19.423 -> 223.61 00:02:19.423 -> 270.64 00:02:19.423 -> 180.19 00:02:19.423 -> AK8963 mag scale (mG) 00:02:19.423 -> 1.09 00:02:19.423 -> 1.01 00:02:19.423 -> 0.92 00:02:21.422 -> X-Axis sensitivity adjustment value 1.18 00:02:21.422 -> Y-Axis sensitivity adjustment value 1.19 00:02:21.422 -> Z-Axis sensitivity adjustment value 1.14 00:02:22.405 -> ax = 0.00 ay = 0.00 az = 0.00 mg 00:02:22.439 -> gx = 0.00 gy = 0.00 gz = 0.00 deg/s 00:02:22.439 -> mx = 0 my = 0 mz = 0 mG 00:02:22.439 -> q0 = 1.00 qx = 0.00 qy = 0.00 qz = 0.00 00:02:22.439 -> Gyro temperature is 31.4 degrees C 00:02:22.439 -> Yaw, Pitch, Roll: 3.80, 0.00, 0.00 00:02:22.439 -> Grav_x, Grav_y, Grav_z: 0.00, 0.00, 1000.00 mg 00:02:22.439 -> Lin_ax, Lin_ay, Lin_az: 0.00, 0.00, -1000.00 mg 00:02:22.439 -> sumCount = 0 00:02:22.439 -> sum = 0.00 00:02:22.439 -> rate = nan Hz 00:02:22.914 -> ax = -17.88 ay = 13.43 az = 1303.41 mg 00:02:22.914 -> gx = 0.56 gy = 1.30 gz = 0.88 deg/s 00:02:22.982 -> mx = 1 my = -19 mz = 22 mG 00:02:22.982 -> q0 = 0.39 qx = 0.05 qy = 0.02 qz = -0.92 00:02:22.982 -> Gyro temperature is 31.5 degrees C 00:02:22.982 -> Yaw, Pitch, Roll: 229.92, 6.76, -0.15 00:02:22.982 -> Grav_x, Grav_y, Grav_z: 2.64, 117.78, 993.04 mg 00:02:22.982 -> Lin_ax, Lin_ay, Lin_az: -20.52, -104.35, 310.37 mg 00:02:22.982 -> sumCount = 1020 00:02:22.982 -> sum = 35.70 00:02:22.982 -> rate = 28.57 Hz 00:02:23.457 -> ax = -42.36 ay = 58.11 az = 1344.97 mg 00:02:23.457 -> gx = -2.86 gy = 0.56 gz = 1.82 deg/s 00:02:23.491 -> mx = 23 my = -10 mz = -4 mG 00:02:23.491 -> q0 = 0.50 qx = 0.03 qy = 0.02 qz = -0.86 00:02:23.491 -> Gyro temperature is 31.6 degrees C 00:02:23.491 -> Yaw, Pitch, Roll: 244.07, 4.41, -0.15 00:02:23.491 -> Grav_x, Grav_y, Grav_z: 2.65, 76.95, 997.03 mg 00:02:23.491 -> Lin_ax, Lin_ay, Lin_az: -45.00, -18.85, 347.94 mg 00:02:23.491 -> sumCount = 1010 00:02:23.491 -> sum = 0.52 00:02:23.491 -> rate = 1944.59 Hz

[kriswiner]

Check your code with this:

https://github.com/kriswiner/MPU9250/blob/master/Documents/AN-MPU-9250A-03%20MPU-9250%20Accel%20Gyro%20and%20Compass%20Self-Test%20Implementation%20v1%200_062813.pdf

The self test uses only the 8 LSBits.

On Sat, May 11, 2019 at 3:19 PM borstan notifications@github.com wrote:

Hello there, I have a problem with self - test values that show more than 40% difference from the factory values. In other examples, I saw that output should be as close to 0 as possible. Could someone help me solve this one? Thank you for any advice! I am using Sparkfun ESP32 Thing, MPU9250, Interrupt connected to the GPIO35.

CODE: MPU9250_MS5637_ESP32 Basic Example Code (I commented all lines with MS5637 since I am not using it)

00:01:52.300 -> Scanning... 00:01:52.300 -> I2C device found at address 0x0C ! 00:01:52.300 -> I2C device found at address 0x68 ! 00:01:52.300 -> done 00:01:52.300 -> 00:01:52.300 -> MPU9250 9-axis motion sensor... 00:01:52.300 -> MPU9250 I AM 71 I should be 71 00:01:53.286 -> MPU9250 is online... 00:01:53.557 -> x-axis self test: acceleration trim within : -48.5% of factory value 00:01:53.591 -> y-axis self test: acceleration trim within : -51.7% of factory value 00:01:53.591 -> z-axis self test: acceleration trim within : -49.1% of factory value 00:01:53.591 -> x-axis self test: gyration trim within : -51.5% of factory value 00:01:53.591 -> y-axis self test: gyration trim within : -26.7% of factory value 00:01:53.591 -> z-axis self test: gyration trim within : -36.8% of factory value 00:01:54.607 -> Calibrate gyro and accel 00:01:54.978 -> accel biases (mg) 00:01:54.978 -> 27.34 00:01:54.978 -> -53.83 00:01:54.978 -> 126.89 00:01:54.978 -> gyro biases (dps) 00:01:54.978 -> 8.69 00:01:54.978 -> -0.16 00:01:54.978 -> -18.47 00:01:56.371 -> MPU9250 initialized for active data mode.... 00:01:56.371 -> AK8963 I AM 48 I should be 48 00:01:57.424 -> AK8963 initialized for active data mode.... 00:01:57.424 -> Mag Calibration: Wave device in a figure eight until done! 00:02:19.389 -> Mag Calibration done! 00:02:19.389 -> AK8963 mag biases (mG) 00:02:19.423 -> 223.61 00:02:19.423 -> 270.64 00:02:19.423 -> 180.19 00:02:19.423 -> AK8963 mag scale (mG) 00:02:19.423 -> 1.09 00:02:19.423 -> 1.01 00:02:19.423 -> 0.92 00:02:21.422 -> X-Axis sensitivity adjustment value 1.18 00:02:21.422 -> Y-Axis sensitivity adjustment value 1.19 00:02:21.422 -> Z-Axis sensitivity adjustment value 1.14 00:02:22.405 -> ax = 0.00 ay = 0.00 az = 0.00 mg 00:02:22.439 -> gx = 0.00 gy = 0.00 gz = 0.00 deg/s 00:02:22.439 -> mx = 0 my = 0 mz = 0 mG 00:02:22.439 -> q0 = 1.00 qx = 0.00 qy = 0.00 qz = 0.00 00:02:22.439 -> Gyro temperature is 31.4 degrees C 00:02:22.439 -> Yaw, Pitch, Roll: 3.80, 0.00, 0.00 00:02:22.439 -> Grav_x, Grav_y, Grav_z: 0.00, 0.00, 1000.00 mg 00:02:22.439 -> Lin_ax, Lin_ay, Lin_az: 0.00, 0.00, -1000.00 mg 00:02:22.439 -> sumCount = 0 00:02:22.439 -> sum = 0.00 00:02:22.439 -> rate = nan Hz 00:02:22.914 -> ax = -17.88 ay = 13.43 az = 1303.41 mg 00:02:22.914 -> gx = 0.56 gy = 1.30 gz = 0.88 deg/s 00:02:22.982 -> mx = 1 my = -19 mz = 22 mG 00:02:22.982 -> q0 = 0.39 qx = 0.05 qy = 0.02 qz = -0.92 00:02:22.982 -> Gyro temperature is 31.5 degrees C 00:02:22.982 -> Yaw, Pitch, Roll: 229.92, 6.76, -0.15 00:02:22.982 -> Grav_x, Grav_y, Grav_z: 2.64, 117.78, 993.04 mg 00:02:22.982 -> Lin_ax, Lin_ay, Lin_az: -20.52, -104.35, 310.37 mg 00:02:22.982 -> sumCount = 1020 00:02:22.982 -> sum = 35.70 00:02:22.982 -> rate = 28.57 Hz 00:02:23.457 -> ax = -42.36 ay = 58.11 az = 1344.97 mg 00:02:23.457 -> gx = -2.86 gy = 0.56 gz = 1.82 deg/s 00:02:23.491 -> mx = 23 my = -10 mz = -4 mG 00:02:23.491 -> q0 = 0.50 qx = 0.03 qy = 0.02 qz = -0.86 00:02:23.491 -> Gyro temperature is 31.6 degrees C 00:02:23.491 -> Yaw, Pitch, Roll: 244.07, 4.41, -0.15 00:02:23.491 -> Grav_x, Grav_y, Grav_z: 2.65, 76.95, 997.03 mg 00:02:23.491 -> Lin_ax, Lin_ay, Lin_az: -45.00, -18.85, 347.94 mg 00:02:23.491 -> sumCount = 1010 00:02:23.491 -> sum = 0.52 00:02:23.491 -> rate = 1944.59 Hz

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/kriswiner/MPU9250/issues/371, or mute the thread https://github.com/notifications/unsubscribe-auth/ABTDLKVBIJVRRPTK4ZDS24TPU5A5LANCNFSM4HMJC4CA .

[borstan]

Hello Kris, I read true the sheet you sent me but I am not very clear which are the trim values I should change in your code to have it working properly.

In this test I printed out only Yaw, Pitch and Roll (data is in this order) and when I did 8 pattern I am getting quite stabile output (you said that when the MPU is stationary on the table the Pitch and Roll should be larger than 1). Problem is when I don't do the 8 pattern data pitch and roll gets >6. How can I achieve calibrated data without performing 8 pattern before every start of communication with MPU?

16:27:26.165 -> Scanning... 16:27:26.199 -> I2C device found at address 0x0C ! 16:27:26.199 -> I2C device found at address 0x68 ! 16:27:26.199 -> done 16:27:26.199 -> 16:27:26.199 -> MPU9250 9-axis motion sensor... 16:27:26.199 -> MPU9250 I AM 71 I should be 71 16:27:27.183 -> MPU9250 is online... 16:27:27.521 -> x-axis self test: acceleration trim within : -46.0% of factory value 16:27:27.521 -> y-axis self test: acceleration trim within : -43.1% of factory value 16:27:27.521 -> z-axis self test: acceleration trim within : -42.9% of factory value 16:27:27.521 -> x-axis self test: gyration trim within : -45.3% of factory value 16:27:27.521 -> y-axis self test: gyration trim within : -24.0% of factory value 16:27:27.521 -> z-axis self test: gyration trim within : -34.8% of factory value 16:27:28.502 -> Calibrate gyro and accel 16:27:28.875 -> accel biases (mg) 16:27:28.875 -> -10.44 16:27:28.875 -> -38.64 16:27:28.875 -> -87.34 16:27:28.875 -> gyro biases (dps) 16:27:28.875 -> 6.25 16:27:28.875 -> -0.16 16:27:28.875 -> -10.29 16:27:30.267 -> MPU9250 initialized for active data mode.... 16:27:30.267 -> AK8963 I AM 48 I should be 48 16:27:31.320 -> AK8963 initialized for active data mode.... 16:27:31.320 -> Mag Calibration: Wave device in a figure eight until done! 16:27:53.317 -> Mag Calibration done! 16:27:53.317 -> AK8963 mag biases (mG) 16:27:53.317 -> 101.16 16:27:53.317 -> 292.01 16:27:53.317 -> 61.78 16:27:53.317 -> AK8963 mag scale (mG) 16:27:53.317 -> 1.06 16:27:53.317 -> 0.64 16:27:53.317 -> 2.02 16:27:55.319 -> X-Axis sensitivity adjustment value 1.18 16:27:55.319 -> Y-Axis sensitivity adjustment value 1.19 16:27:55.319 -> Z-Axis sensitivity adjustment value 1.14 16:27:56.305 -> 3.80, 0.00, 0.00 16:27:56.815 -> 248.49, 0.37, -0.07 16:27:57.323 -> 248.75, 0.40, -0.64 16:27:57.799 -> 248.34, -0.32, -0.51 16:27:58.309 -> 248.69, 0.35, 0.05 16:27:58.817 -> 248.41, 0.02, -0.87 16:27:59.327 -> 250.03, 0.33, 0.64 16:27:59.836 -> 249.06, 0.46, 0.22 16:28:00.313 -> 248.04, 0.12, -0.48 16:28:00.822 -> 248.41, -0.03, 0.05 16:28:01.331 -> 249.18, 0.81, 0.39 16:28:01.840 -> 248.74, 0.35, -0.35 16:28:02.315 -> 247.66, 0.73, -0.74 16:28:02.823 -> 248.96, 0.26, 0.12 16:28:03.331 -> 248.59, 0.04, 0.30 16:28:03.807 -> 248.63, -0.49, -0.77 16:28:04.317 -> 248.75, 0.08, 0.38 16:28:04.826 -> 247.59, 0.43, -0.36 16:28:05.334 -> 249.26, -0.02, -0.30

[kriswiner]

Store the calibration corrections in the main program so you don't have to redo them every time.

On Sun, May 12, 2019 at 7:36 AM borstan notifications@github.com wrote:

Hello Kris, I read true the sheet you sent me but I am not very clear which are the trim values I should change in your code to have it working properly.

In this test I printed out only Yaw, Pitch and Roll (data is in this order) and when I did 8 pattern I am getting quite stabile output (you said that when the MPU is stationary on the table the Pitch and Roll should be larger than 1). Problem is when I don't do the 8 pattern data pitch and roll gets >6. How can I achieve calibrated data without performing 8 pattern before every start of communication with MPU?

16:27:26.165 -> Scanning... 16:27:26.199 -> I2C device found at address 0x0C ! 16:27:26.199 -> I2C device found at address 0x68 ! 16:27:26.199 -> done 16:27:26.199 -> 16:27:26.199 -> MPU9250 9-axis motion sensor... 16:27:26.199 -> MPU9250 I AM 71 I should be 71 16:27:27.183 -> MPU9250 is online... 16:27:27.521 -> x-axis self test: acceleration trim within : -46.0% of factory value 16:27:27.521 -> y-axis self test: acceleration trim within : -43.1% of factory value 16:27:27.521 -> z-axis self test: acceleration trim within : -42.9% of factory value 16:27:27.521 -> x-axis self test: gyration trim within : -45.3% of factory value 16:27:27.521 -> y-axis self test: gyration trim within : -24.0% of factory value 16:27:27.521 -> z-axis self test: gyration trim within : -34.8% of factory value 16:27:28.502 -> Calibrate gyro and accel 16:27:28.875 -> accel biases (mg) 16:27:28.875 -> -10.44 16:27:28.875 -> -38.64 16:27:28.875 -> -87.34 16:27:28.875 -> gyro biases (dps) 16:27:28.875 -> 6.25 16:27:28.875 -> -0.16 16:27:28.875 -> -10.29 16:27:30.267 -> MPU9250 initialized for active data mode.... 16:27:30.267 -> AK8963 I AM 48 I should be 48 16:27:31.320 -> AK8963 initialized for active data mode.... 16:27:31.320 -> Mag Calibration: Wave device in a figure eight until done! 16:27:53.317 -> Mag Calibration done! 16:27:53.317 -> AK8963 mag biases (mG) 16:27:53.317 -> 101.16 16:27:53.317 -> 292.01 16:27:53.317 -> 61.78 16:27:53.317 -> AK8963 mag scale (mG) 16:27:53.317 -> 1.06 16:27:53.317 -> 0.64 16:27:53.317 -> 2.02 16:27:55.319 -> X-Axis sensitivity adjustment value 1.18 16:27:55.319 -> Y-Axis sensitivity adjustment value 1.19 16:27:55.319 -> Z-Axis sensitivity adjustment value 1.14 16:27:56.305 -> 3.80, 0.00, 0.00 16:27:56.815 -> 248.49, 0.37, -0.07 16:27:57.323 -> 248.75, 0.40, -0.64 16:27:57.799 -> 248.34, -0.32, -0.51 16:27:58.309 -> 248.69, 0.35, 0.05 16:27:58.817 -> 248.41, 0.02, -0.87 16:27:59.327 -> 250.03, 0.33, 0.64 16:27:59.836 -> 249.06, 0.46, 0.22 16:28:00.313 -> 248.04, 0.12, -0.48 16:28:00.822 -> 248.41, -0.03, 0.05 16:28:01.331 -> 249.18, 0.81, 0.39 16:28:01.840 -> 248.74, 0.35, -0.35 16:28:02.315 -> 247.66, 0.73, -0.74 16:28:02.823 -> 248.96, 0.26, 0.12 16:28:03.331 -> 248.59, 0.04, 0.30 16:28:03.807 -> 248.63, -0.49, -0.77 16:28:04.317 -> 248.75, 0.08, 0.38 16:28:04.826 -> 247.59, 0.43, -0.36 16:28:05.334 -> 249.26, -0.02, -0.30

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/kriswiner/MPU9250/issues/371#issuecomment-491601173, or mute the thread https://github.com/notifications/unsubscribe-auth/ABTDLKWVYNYAIMHTBPQ2LJTPVATNTANCNFSM4HMJC4CA .

[borstan]

Okay I'll do that, and what do you think about pitch, roll and Yaw values? I thought after calibration pitch will get stabilized at 360 if I keep the mpu stationary.

[kriswiner]

Pitch and roll should be 0 is stationary and flat if the sensors are properly calibrated.

On Sun, May 12, 2019 at 10:17 AM borstan notifications@github.com wrote:

Okay I'll do that, and what do you think about pitch, roll and Yaw values? I thought after calibration pitch will get stabilized at 360 if I keep the mpu stationary.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/kriswiner/MPU9250/issues/371#issuecomment-491613018, or mute the thread https://github.com/notifications/unsubscribe-auth/ABTDLKVHD7YWS57AFGRS52LPVBGKTANCNFSM4HMJC4CA .

[borstan]

Kris, thank you for your advice. I put my calibrated data into the main program and calibration process works fine now, while not moving, I am getting pitch and roll values under 1 (even tho I have constant changes of sign of Pitch while not moving - don't know if it is an issue or not).

But I did an observation and find out that: 1) When I move MPU in the Yaw direction while resting on the table everything works fine. 2) When I move in the Pitch direction pitch starts increasing with the Yaw value. 3) While I do roll I can achieve 30° and then Yaw starts increasing,

Do you know what might be a problem here?

Input to my Quaternion update is like this:

MahonyQuaternionUpdate(ax, ay, az, gx PI / 180.0f, gy PI / 180.0f, gz * PI / 180.0f, my, mx, mz);

PS: video (https://youtu.be/rD4n9S4dzKM)

[kriswiner]

" MahonyQuaternionUpdate(ax, ay, az, gx PI / 180.0f, gy PI / 180.0f, gz

• PI / 180.0f, my, mx, mz);"

This cannot be correct. The data needs to be fed in NED or ENU order, a RH coordinate system.

So first decide which accel axis will be designated as North. The determine which of the remaining axes is pointing toward North, East and Down. The input the data into the Update function accordingly.

I know the above cannot be correct because the gyro and mag z-axes are opposite each other, for example.

On Tue, May 14, 2019 at 4:51 AM borstan notifications@github.com wrote:

Kris, thank you for your advice. I put my calibrated data into the main program and calibration process works fine now, while not moving, I am getting pitch and roll values under 1 (even tho I have constant changes of sign of Pitch while not moving - don't know if it is an issue or not).

But I did an observation and find out that:

1. When I move MPU in the Yaw direction while resting on the table everything works fine.
2. When I move in the Pitch direction pitch starts increasing with the Yaw value.
3. While I do roll I can achieve 30° and then Yaw starts increasing,

Do you know what might be a problem here?

Input to my Quaternion update is like this:

MahonyQuaternionUpdate(ax, ay, az, gx PI / 180.0f, gy PI / 180.0f, gz

• PI / 180.0f, my, mx, mz);

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/kriswiner/MPU9250/issues/371?email_source=notifications&email_token=ABTDLKWBER3AYECROOZVRWTPVKRVPA5CNFSM4HMJC4CKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODVLHJWI#issuecomment-492205273, or mute the thread https://github.com/notifications/unsubscribe-auth/ABTDLKSNPPARVGMWVY2ZYN3PVKRVPANCNFSM4HMJC4CA .

[robertoostenveld]

Hi @borstan, you may want to check #370.

[borstan]

Thank you, guys. I used the help from the #370 and now it is much better I am feeding data in this order (MahonyQuaternionUpdate(ax, ay, az, gy PI / 180.0f, gx PI / 180.0f, gz * PI / 180.0f, my, mx, -mz)).

But while having a pitch close to +-90° Yaw is getting a bit crazy. I tried to print ax, ay, az; gx, gy, gz; mx, my, mz and compare it to some values that Kris founds as reasonable.

Observation: I think that accelerometer is well calibrated as the ax and ay don't go above 10 miliGs and az oscillates around 1000. But I am concerned about the Mag values, are they in norm? I already did the calibration and put the MagBiases and scales into the code.

Do you know what might be a problem?

Thank you for any help.

TAB: Calibrated raw data output (MPU was not moving)

ax	ay	az	gx	gy	gz	mx	my	mz
1.65	0.37	998.05	0.00	0.02	0.04	95	276	228
2.62	1.10	1000.49	0.00	0.04	0.12	92	308	228
1.65	-0.12	998.78	0.08	0.06	-0.02	95	304	216
1.53	-1.83	998.84	-0.02	-0.05	0.02	96	278	210
2.14	0.55	999.57	-0.03	0.02	-0.02	104	284	212
1.89	3.23	997.56	0.18	0.02	0.05	108	290	214
3.17	-1.59	994.81	0.00	0.01	0.05	95	300	209
2.38	-0.43	1000.12	0.08	-0.01	0.02	102	298	222
3.11	0.24	999.27	0.01	-0.02	0.02	99	294	218
1.89	-0.18	998.05	0.09	-0.03	-0.07	113	308	224
-0.12	-0.31	998.54	0.01	0.05	0.04	92	292	220
2.08	-0.12	998.78	0.00	-0.06	0.01	104	292	216
1.40	-1.40	998.29	-0.02	-0.06	0.08	98	296	216
1.04	0.73	999.02	-0.07	-0.03	0.05	107	300	224
1.59	0.85	996.28	-0.01	-0.04	0.03	98	292	220
2.20	-0.24	997.99	0.05	0.05	0.00	95	288	216
1.34	-0.79	1001.04	-0.02	-0.02	0.14	87	282	214
0.55	1.95	999.33	-0.05	-0.01	0.06	98	284	220
0.61	-0.49	997.38	0.05	-0.02	0.05	90	318	222
1.28	0.31	999.27	-0.04	0.02	0.05	101	284	220
1.22	3.36	999.76	-0.01	0.02	0.02	90	294	210
-0.06	0.67	1000.67	-0.05	-0.04	0.08	101	304	216
2.32	-1.40	997.62	0.00	-0.02	-0.05	93	302	214
0.79	0.43	997.68	-0.02	-0.04	0.07	98	280	209
0.24	-0.43	1000.37	0.03	-0.03	-0.03	96	294	233
0.73	0.61	996.40	0.05	-0.07	0.02	102	286	222
3.60	1.10	998.60	0.01	0.02	0.05	101	288	209
0.31	-0.31	997.80	-0.01	0.00	0.01	98	300	209
1.89	0.00	1000.92	0.03	-0.02	0.06	93	290	207
3.36	-0.12	999.27	-0.02	-0.07	0.11	90	310	222
1.95	-0.37	998.05	0.04	-0.02	0.03	105	298	229
1.53	1.16	998.84	0.02	0.01	0.02	113	296	197
2.44	0.24	999.57	0.04	-0.01	0.06	96	282	222
2.93	-0.31	997.38	0.04	-0.08	-0.05	96	286	218
2.62	-1.34	996.40	-0.01	-0.02	0.00	104	288	231
2.75	-0.31	1001.16	-0.02	-0.05	0.03	90	282	222
2.44	-0.85	1000.37	0.08	-0.03	0.13	99	298	214
2.75	-1.10	1001.40	0.05	0.02	0.13	99	302	222
-0.37	-0.67	1000.24	0.02	-0.05	0.11	99	278	222
2.99	2.26	999.45	-0.02	-0.02	-0.05	119	296	231
2.38	0.79	1000.18	-0.15	-0.07	0.00	98	308	220
2.26	-1.46	1000.37	-0.08	-0.11	-0.03	96	278	184
1.77	-0.55	1000.61	0.08	-0.03	0.05	101	300	212
1.16	0.67	997.31	0.05	0.07	-0.03	95	296	224
1.40	-0.43	998.72	-0.02	-0.07	0.05	110	300	220
1.89	0.31	1001.59	0.04	-0.01	0.05	95	284	209
1.40	-0.61	998.84	-0.06	-0.09	0.08	101	296	205
1.71	-0.92	999.15	-0.03	-0.08	0.07	99	294	218
3.36	0.79	996.58	0.01	-0.06	0.01	98	296	212
3.23	-0.61	1000.31	-0.01	0.00	0.01	98	280	209
0.00	0.55	994.87	0.05	-0.04	0.02	92	280	197
1.34	-1.22	997.25	-0.01	0.02	0.05	105	302	210
2.87	-0.43	997.31	-0.02	0.02	0.11	110	296	209
2.56	0.98	996.22	-0.07	-0.07	0.01	104	304	224
-0.24	0.06	997.25	-0.02	0.01	0.04	95	308	228
-0.06	0.67	999.33	0.09	0.02	0.00	107	284	212
0.98	1.10	999.39	0.05	0.00	-0.05	95	296	212
2.32	0.43	997.62	-0.02	-0.04	0.05	101	296	209
2.99	0.24	999.76	0.06	0.00	-0.02	102	302	222
3.91	0.67	999.57	0.05	-0.05	-0.02	98	288	216
2.99	-0.12	998.78	0.00	-0.08	-0.02	104	284	220
0.98	0.55	994.26	0.00	0.00	0.04	89	300	224
0.92	-0.31	998.78	0.02	-0.02	-0.02	90	286	218
1.04	0.12	998.90	-0.03	-0.05	-0.01	86	296	220
1.10	-0.37	999.88	-0.06	0.01	0.02	96	294	207
2.14	0.85	999.39	-0.05	0.02	0.09	101	296	209
3.30	1.65	1001.10	-0.03	-0.02	0.00	99	290	226
2.87	-0.67	996.46	-0.06	-0.05	0.00	87	306	210
2.26	0.49	997.62	-0.01	-0.04	-0.04	93	286	218
2.50	-1.40	998.17	0.07	-0.03	0.06	102	310	222
3.23	-0.55	1000.61	0.02	0.01	-0.02	105	282	229
3.11	-0.43	1002.38	0.08	-0.11	0.00	99	298	218

[kriswiner]

" (MahonyQuaternionUpdate(ax, ay, az, gy PI / 180.0f, gx PI / 180.0f, gz PI / 180.0f, my, mx, -mz)*)"

sure the gyro is right

On Tue, May 14, 2019 at 3:43 PM borstan notifications@github.com wrote:

Thank you, guys. I used the help from the #370 https://github.com/kriswiner/MPU9250/pull/370 and now it is much better I am feeding data in this order (MahonyQuaternionUpdate(ax, ay, az, gy PI / 180.0f, gx PI / 180.0f, gz PI / 180.0f, my, mx, -mz)*).

But while having a pitch close to +-90° Yaw is getting a bit crazy. I tried to print ax, ay, az; gx, gy, gz; mx, my, mz and compare it to some values that Kris founds as reasonable.

Observation: I think that accelerometer is well calibrated as the ax and ay don't go above 10 miliGs and az oscillates around 1000. But I am concerned about the Mag values, are they in norm? I already did the calibration and put the MagBiases and scales into the code.

Do you know what might be a problem?

Thank you for any help.

TAB: Calibrated raw data output ax ay az gx gy gz mx my mz 1.65 0.37 998.05 0.00 0.02 0.04 95 276 228 2.62 1.10 1000.49 0.00 0.04 0.12 92 308 228 1.65 -0.12 998.78 0.08 0.06 -0.02 95 304 216 1.53 -1.83 998.84 -0.02 -0.05 0.02 96 278 210 2.14 0.55 999.57 -0.03 0.02 -0.02 104 284 212 1.89 3.23 997.56 0.18 0.02 0.05 108 290 214 3.17 -1.59 994.81 0.00 0.01 0.05 95 300 209 2.38 -0.43 1000.12 0.08 -0.01 0.02 102 298 222 3.11 0.24 999.27 0.01 -0.02 0.02 99 294 218 1.89 -0.18 998.05 0.09 -0.03 -0.07 113 308 224 -0.12 -0.31 998.54 0.01 0.05 0.04 92 292 220 2.08 -0.12 998.78 0.00 -0.06 0.01 104 292 216 1.40 -1.40 998.29 -0.02 -0.06 0.08 98 296 216 1.04 0.73 999.02 -0.07 -0.03 0.05 107 300 224 1.59 0.85 996.28 -0.01 -0.04 0.03 98 292 220 2.20 -0.24 997.99 0.05 0.05 0.00 95 288 216 1.34 -0.79 1001.04 -0.02 -0.02 0.14 87 282 214 0.55 1.95 999.33 -0.05 -0.01 0.06 98 284 220 0.61 -0.49 997.38 0.05 -0.02 0.05 90 318 222 1.28 0.31 999.27 -0.04 0.02 0.05 101 284 220 1.22 3.36 999.76 -0.01 0.02 0.02 90 294 210 -0.06 0.67 1000.67 -0.05 -0.04 0.08 101 304 216 2.32 -1.40 997.62 0.00 -0.02 -0.05 93 302 214 0.79 0.43 997.68 -0.02 -0.04 0.07 98 280 209 0.24 -0.43 1000.37 0.03 -0.03 -0.03 96 294 233 0.73 0.61 996.40 0.05 -0.07 0.02 102 286 222 3.60 1.10 998.60 0.01 0.02 0.05 101 288 209 0.31 -0.31 997.80 -0.01 0.00 0.01 98 300 209 1.89 0.00 1000.92 0.03 -0.02 0.06 93 290 207 3.36 -0.12 999.27 -0.02 -0.07 0.11 90 310 222 1.95 -0.37 998.05 0.04 -0.02 0.03 105 298 229 1.53 1.16 998.84 0.02 0.01 0.02 113 296 197 2.44 0.24 999.57 0.04 -0.01 0.06 96 282 222 2.93 -0.31 997.38 0.04 -0.08 -0.05 96 286 218 2.62 -1.34 996.40 -0.01 -0.02 0.00 104 288 231 2.75 -0.31 1001.16 -0.02 -0.05 0.03 90 282 222 2.44 -0.85 1000.37 0.08 -0.03 0.13 99 298 214 2.75 -1.10 1001.40 0.05 0.02 0.13 99 302 222 -0.37 -0.67 1000.24 0.02 -0.05 0.11 99 278 222 2.99 2.26 999.45 -0.02 -0.02 -0.05 119 296 231 2.38 0.79 1000.18 -0.15 -0.07 0.00 98 308 220 2.26 -1.46 1000.37 -0.08 -0.11 -0.03 96 278 184 1.77 -0.55 1000.61 0.08 -0.03 0.05 101 300 212 1.16 0.67 997.31 0.05 0.07 -0.03 95 296 224 1.40 -0.43 998.72 -0.02 -0.07 0.05 110 300 220 1.89 0.31 1001.59 0.04 -0.01 0.05 95 284 209 1.40 -0.61 998.84 -0.06 -0.09 0.08 101 296 205 1.71 -0.92 999.15 -0.03 -0.08 0.07 99 294 218 3.36 0.79 996.58 0.01 -0.06 0.01 98 296 212 3.23 -0.61 1000.31 -0.01 0.00 0.01 98 280 209 0.00 0.55 994.87 0.05 -0.04 0.02 92 280 197 1.34 -1.22 997.25 -0.01 0.02 0.05 105 302 210 2.87 -0.43 997.31 -0.02 0.02 0.11 110 296 209 2.56 0.98 996.22 -0.07 -0.07 0.01 104 304 224 -0.24 0.06 997.25 -0.02 0.01 0.04 95 308 228 -0.06 0.67 999.33 0.09 0.02 0.00 107 284 212 0.98 1.10 999.39 0.05 0.00 -0.05 95 296 212 2.32 0.43 997.62 -0.02 -0.04 0.05 101 296 209 2.99 0.24 999.76 0.06 0.00 -0.02 102 302 222 3.91 0.67 999.57 0.05 -0.05 -0.02 98 288 216 2.99 -0.12 998.78 0.00 -0.08 -0.02 104 284 220 0.98 0.55 994.26 0.00 0.00 0.04 89 300 224 0.92 -0.31 998.78 0.02 -0.02 -0.02 90 286 218 1.04 0.12 998.90 -0.03 -0.05 -0.01 86 296 220 1.10 -0.37 999.88 -0.06 0.01 0.02 96 294 207 2.14 0.85 999.39 -0.05 0.02 0.09 101 296 209 3.30 1.65 1001.10 -0.03 -0.02 0.00 99 290 226 2.87 -0.67 996.46 -0.06 -0.05 0.00 87 306 210 2.26 0.49 997.62 -0.01 -0.04 -0.04 93 286 218 2.50 -1.40 998.17 0.07 -0.03 0.06 102 310 222 3.23 -0.55 1000.61 0.02 0.01 -0.02 105 282 229 3.11 -0.43 1002.38 0.08 -0.11 0.00 99 298 218

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/kriswiner/MPU9250/issues/371?email_source=notifications&email_token=ABTDLKUOMWCK43ZSP2C7FH3PVM57XA5CNFSM4HMJC4CKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODVNACVY#issuecomment-492437847, or mute the thread https://github.com/notifications/unsubscribe-auth/ABTDLKUGODSKREZWKUZNW3DPVM57XANCNFSM4HMJC4CA .

[kriswiner]

Not for the MPU9250, since the accel and gyro are aligned...so if Ax is North, so is Gx.

On Tue, May 14, 2019 at 5:34 PM Tlera Corporation tleracorp@gmail.com wrote:

" (MahonyQuaternionUpdate(ax, ay, az, gy PI / 180.0f, gx PI / 180.0f, gz PI / 180.0f, my, mx, -mz)*)"

sure the gyro is right

On Tue, May 14, 2019 at 3:43 PM borstan notifications@github.com wrote:

Thank you, guys. I used the help from the #370 https://github.com/kriswiner/MPU9250/pull/370 and now it is much better I am feeding data in this order (MahonyQuaternionUpdate(ax, ay, az, gy PI / 180.0f, gx PI / 180.0f, gz PI / 180.0f, my, mx, -mz)*).

But while having a pitch close to +-90° Yaw is getting a bit crazy. I tried to print ax, ay, az; gx, gy, gz; mx, my, mz and compare it to some values that Kris founds as reasonable.

Observation: I think that accelerometer is well calibrated as the ax and ay don't go above 10 miliGs and az oscillates around 1000. But I am concerned about the Mag values, are they in norm? I already did the calibration and put the MagBiases and scales into the code.

Do you know what might be a problem?

Thank you for any help.

TAB: Calibrated raw data output ax ay az gx gy gz mx my mz 1.65 0.37 998.05 0.00 0.02 0.04 95 276 228 2.62 1.10 1000.49 0.00 0.04 0.12 92 308 228 1.65 -0.12 998.78 0.08 0.06 -0.02 95 304 216 1.53 -1.83 998.84 -0.02 -0.05 0.02 96 278 210 2.14 0.55 999.57 -0.03 0.02 -0.02 104 284 212 1.89 3.23 997.56 0.18 0.02 0.05 108 290 214 3.17 -1.59 994.81 0.00 0.01 0.05 95 300 209 2.38 -0.43 1000.12 0.08 -0.01 0.02 102 298 222 3.11 0.24 999.27 0.01 -0.02 0.02 99 294 218 1.89 -0.18 998.05 0.09 -0.03 -0.07 113 308 224 -0.12 -0.31 998.54 0.01 0.05 0.04 92 292 220 2.08 -0.12 998.78 0.00 -0.06 0.01 104 292 216 1.40 -1.40 998.29 -0.02 -0.06 0.08 98 296 216 1.04 0.73 999.02 -0.07 -0.03 0.05 107 300 224 1.59 0.85 996.28 -0.01 -0.04 0.03 98 292 220 2.20 -0.24 997.99 0.05 0.05 0.00 95 288 216 1.34 -0.79 1001.04 -0.02 -0.02 0.14 87 282 214 0.55 1.95 999.33 -0.05 -0.01 0.06 98 284 220 0.61 -0.49 997.38 0.05 -0.02 0.05 90 318 222 1.28 0.31 999.27 -0.04 0.02 0.05 101 284 220 1.22 3.36 999.76 -0.01 0.02 0.02 90 294 210 -0.06 0.67 1000.67 -0.05 -0.04 0.08 101 304 216 2.32 -1.40 997.62 0.00 -0.02 -0.05 93 302 214 0.79 0.43 997.68 -0.02 -0.04 0.07 98 280 209 0.24 -0.43 1000.37 0.03 -0.03 -0.03 96 294 233 0.73 0.61 996.40 0.05 -0.07 0.02 102 286 222 3.60 1.10 998.60 0.01 0.02 0.05 101 288 209 0.31 -0.31 997.80 -0.01 0.00 0.01 98 300 209 1.89 0.00 1000.92 0.03 -0.02 0.06 93 290 207 3.36 -0.12 999.27 -0.02 -0.07 0.11 90 310 222 1.95 -0.37 998.05 0.04 -0.02 0.03 105 298 229 1.53 1.16 998.84 0.02 0.01 0.02 113 296 197 2.44 0.24 999.57 0.04 -0.01 0.06 96 282 222 2.93 -0.31 997.38 0.04 -0.08 -0.05 96 286 218 2.62 -1.34 996.40 -0.01 -0.02 0.00 104 288 231 2.75 -0.31 1001.16 -0.02 -0.05 0.03 90 282 222 2.44 -0.85 1000.37 0.08 -0.03 0.13 99 298 214 2.75 -1.10 1001.40 0.05 0.02 0.13 99 302 222 -0.37 -0.67 1000.24 0.02 -0.05 0.11 99 278 222 2.99 2.26 999.45 -0.02 -0.02 -0.05 119 296 231 2.38 0.79 1000.18 -0.15 -0.07 0.00 98 308 220 2.26 -1.46 1000.37 -0.08 -0.11 -0.03 96 278 184 1.77 -0.55 1000.61 0.08 -0.03 0.05 101 300 212 1.16 0.67 997.31 0.05 0.07 -0.03 95 296 224 1.40 -0.43 998.72 -0.02 -0.07 0.05 110 300 220 1.89 0.31 1001.59 0.04 -0.01 0.05 95 284 209 1.40 -0.61 998.84 -0.06 -0.09 0.08 101 296 205 1.71 -0.92 999.15 -0.03 -0.08 0.07 99 294 218 3.36 0.79 996.58 0.01 -0.06 0.01 98 296 212 3.23 -0.61 1000.31 -0.01 0.00 0.01 98 280 209 0.00 0.55 994.87 0.05 -0.04 0.02 92 280 197 1.34 -1.22 997.25 -0.01 0.02 0.05 105 302 210 2.87 -0.43 997.31 -0.02 0.02 0.11 110 296 209 2.56 0.98 996.22 -0.07 -0.07 0.01 104 304 224 -0.24 0.06 997.25 -0.02 0.01 0.04 95 308 228 -0.06 0.67 999.33 0.09 0.02 0.00 107 284 212 0.98 1.10 999.39 0.05 0.00 -0.05 95 296 212 2.32 0.43 997.62 -0.02 -0.04 0.05 101 296 209 2.99 0.24 999.76 0.06 0.00 -0.02 102 302 222 3.91 0.67 999.57 0.05 -0.05 -0.02 98 288 216 2.99 -0.12 998.78 0.00 -0.08 -0.02 104 284 220 0.98 0.55 994.26 0.00 0.00 0.04 89 300 224 0.92 -0.31 998.78 0.02 -0.02 -0.02 90 286 218 1.04 0.12 998.90 -0.03 -0.05 -0.01 86 296 220 1.10 -0.37 999.88 -0.06 0.01 0.02 96 294 207 2.14 0.85 999.39 -0.05 0.02 0.09 101 296 209 3.30 1.65 1001.10 -0.03 -0.02 0.00 99 290 226 2.87 -0.67 996.46 -0.06 -0.05 0.00 87 306 210 2.26 0.49 997.62 -0.01 -0.04 -0.04 93 286 218 2.50 -1.40 998.17 0.07 -0.03 0.06 102 310 222 3.23 -0.55 1000.61 0.02 0.01 -0.02 105 282 229 3.11 -0.43 1002.38 0.08 -0.11 0.00 99 298 218

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/kriswiner/MPU9250/issues/371?email_source=notifications&email_token=ABTDLKUOMWCK43ZSP2C7FH3PVM57XA5CNFSM4HMJC4CKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODVNACVY#issuecomment-492437847, or mute the thread https://github.com/notifications/unsubscribe-auth/ABTDLKUGODSKREZWKUZNW3DPVM57XANCNFSM4HMJC4CA .

[robertoostenveld]

My experience with a calibration scheme like this one (which takes max/min values along the cardinal axes), is that waving in a figure of eight does not work as well as rotating it as a cube/dice, making sure that each axis (side of the cube/dice) at a certain moment points towards the magnetic north, about 45 degrees downward.

Note that the magnetic north field direction is not aligned along the surface of the earth, but pointing quite steep down (for me in Europe). See

which I took from here.

[kriswiner]

Yes, we have developed a calibration cube to do this all better. The idea is to have a fixed bracket such that the precision cube faces are always oriented with a reference surface. Then by taking data at the 24 unique cube positions and doing some math both the accel and mag are calibrated and the cross axis error is also compensated for. The result is < 1 degree heading accuracy or better. This device will be for sale on Tindie in a few weeks time.

On Wed, May 15, 2019 at 12:02 PM Robert Oostenveld notifications@github.com wrote:

My experience with a calibration scheme like this one https://github.com/sparkfun/SparkFun_MPU-9250_Breakout_Arduino_Library/blob/ea379ce8c930ef90f4ebc37799544d68108e3126/src/MPU9250.cpp#L622 (which takes max/min values along the cardinal axes), is that waving in a figure of eight does not work as well as rotating it as a cube/dice, making sure that each axis (side of the cube/dice) at a certain moment points towards the magnetic north, about 45 degrees downward.

Note that the magnetic north field direction is not aligned along the surface of the earth, but pointing quite steep down (for me in Europe). See

[image: The-Earths-magnetic-field-the-geomagnetic-field-Notice-that-the-southern-and-northern] https://user-images.githubusercontent.com/899043/57801015-0e743b00-7753-11e9-842f-cea40a571709.png

which I took from here https://www.researchgate.net/publication/279446083_Magnetoreception_in_Birds_and_Its_Use_for_Long-Distance_Migration/figures?lo=1 .

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/kriswiner/MPU9250/issues/371?email_source=notifications&email_token=ABTDLKRLWC4FRKMYFPUZ5E3PVRM6HA5CNFSM4HMJC4CKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODVPUK5Y#issuecomment-492782967, or mute the thread https://github.com/notifications/unsubscribe-auth/ABTDLKXLEB47QBN4QRMA3RTPVRM6HANCNFSM4HMJC4CA .