search

kriswiner / EM7180_SENtral_sensor_hub

(Affordable) Ultimate Sensor Fusion Solution
https://www.tindie.com/products/onehorse/ultimate-sensor-fusion-solution/
96 stars 37 forks source link

magnetic anomaly detected #45

Open Rababermaber opened 5 years ago

Rababermaber commented 5 years ago

Hello,

whatever I am doing I always get the report that there is a magnetic anomaly detected. I am using a Teensy 3.6 and added the "i2_t3" library and optimized the USFS library to work with it.

I also noticed that the INT pin switchtes states ~5700 times per second. :?

kriswiner commented 5 years ago

Ok, and this is a problem, how?

On Thu, Nov 22, 2018 at 3:46 PM Aaron notifications@github.com wrote:

Hello,

whatever I am doing I always get the report that there is a magnetic anomaly detected. I am using a Teensy 3.6 and added the "i2_t3" library and optimized the USFS library to work with it.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/kriswiner/EM7180_SENtral_sensor_hub/issues/45, or mute the thread https://github.com/notifications/unsubscribe-auth/AGY1qkSDoc6b_KWSivGB9XdZDCUGteG3ks5uxzdAgaJpZM4Yv-U8 .

Rababermaber commented 5 years ago

Issue #29 I thought so because I read in the Issue linked above that if "magnetic anomaly detected" the magnometer is ignored.

kriswiner commented 5 years ago

Shouldn't be an issue if you calibrate the sensors properly and use warm start... Are you not getting accurate orientation estimation?

On Thu, Nov 22, 2018 at 3:55 PM Aaron notifications@github.com wrote:

Issue #29 https://github.com/kriswiner/EM7180_SENtral_sensor_hub/issues/29#issuecomment-411199798 I thought so because I read in the Issue linked above that if "magnetic anomaly detected" the magnometer is ignored.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/kriswiner/EM7180_SENtral_sensor_hub/issues/45#issuecomment-441135168, or mute the thread https://github.com/notifications/unsubscribe-auth/AGY1qltS4yBXrDBCo6c-MxOUeqWr-XRMks5uxzlvgaJpZM4Yv-U8 .

Rababermaber commented 5 years ago

currently without warm start kind of good - I just got the sensor a few hours ago so I am going to fully test it tomorrow

when shaking the sensor hard the calculated YAW value shitfs quiet a bit and then stays a its "new" location

Rababermaber commented 5 years ago

The shaking and the magnetic anomaly detected problems are still present, also when testing today the YAW value slowly increased over time and by correcting to the "correct" heading my robot just performed circles...

my current problems are similar to the problems described in issue #15 linked below Issue #15

kriswiner commented 5 years ago

Not sure what the shaking problem is exactly. How did you calibrate the sensors? Some more details could be useful to determine what kind of help you might need.

On Fri, Nov 23, 2018 at 6:56 AM Aaron notifications@github.com wrote:

The shaking and the magnetic anomaly detected problems are still present, also when testing today the YAW value slowly increased over time and by correcting to the "correct" heading my robot just performed circles...

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/kriswiner/EM7180_SENtral_sensor_hub/issues/45#issuecomment-441260865, or mute the thread https://github.com/notifications/unsubscribe-auth/AGY1qnbsJSwtV4i41j7ooVclFkTtw8M1ks5uyAyCgaJpZM4Yv-U8 .

Rababermaber commented 5 years ago

For the magnometer calibration I'm waving the device in an figure 8, but the AlgorithmStatusRegister is never showing me that the magnometer calibraton is completed, instead all the time it shows me that there is a "magnetic anomaly", like I described it in the beginning.

My shaking problem is, that if I shake the sensor, like turning it 90° rapidly (and then bringing it back to its original position), the YAW value jumped to another position (usually about +-20°) and then stays there.

kriswiner commented 5 years ago

This all sounds like you are not calibrating the megnetometer (and possible the accelerometer) sufficiently. Can you simply check the calibration yourself? Maybe this https://github.com/kriswiner/MPU6050/wiki/Simple-and-Effective-Magnetometer-Calibration would help: it describes a simply way to check the calibration quality without relying on a calibration bit.

On Sun, Nov 25, 2018 at 7:36 AM Aaron notifications@github.com wrote:

For the magnometer calibration I'm waving the device in an figure 8, but the AlgorithmStatusRegister is never showing me that the magnometer calibraton is completed, instead all the time it shows me that there is a "magnetic anomaly", like I described it in the beginning.

My shaking problem is, that if I shake the sensor, like turning it 90° rapidly (and then bringing it back to its original position), the YAW value jumped to another position (usually about +-20°) and then stays there.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/kriswiner/EM7180_SENtral_sensor_hub/issues/45#issuecomment-441449236, or mute the thread https://github.com/notifications/unsubscribe-auth/AGY1qs2cM6eA_02p-BGMIrVLoy_3PNymks5uyrjwgaJpZM4Yv-U8 .

Rababermaber commented 5 years ago

I am trying to calibrate the sensor without any success! I am using your code (WarmStartandAccelCal) and followed the steps described in the wiki section F, but no matter how long I tried to calibrate the magnetometer it never showed me any good results. Also the software and hardware yaw values nearly always differed for more than 10°/20°/... . When restarting the board after saving some "calibration data" the yaw value never stayed constant when heading the sensor in the same direction that it faced before the restart. I have absolutely no idea what i could have done wrong D:

Thank you for your help

kriswiner commented 5 years ago

It would help to know which MCU you are using. And if you would send a copy of your sketch (zip file, don;t cut and paste here. if you need to you can send an e-mail to tleracorp@gmail.com.

On Mon, Nov 26, 2018 at 12:36 PM Aaron notifications@github.com wrote:

I am trying to calibrate the sensor without any success! I am using your code (WarmStartandAccelCal) and followed the steps described in the wiki section F, but no matter how long I tried to calibrate the magnetometer it never showed me any good results. Also the software and hardware yaw values nearly always differed for more than 10°/20°/... . When restarting the board after saving some "calibration data" the yaw value never stayed constant when heading the sensor in the same direction that it faced before the restart. I have absolutely no idea what i could have done wrong D:

Thank you for your help

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/kriswiner/EM7180_SENtral_sensor_hub/issues/45#issuecomment-441788630, or mute the thread https://github.com/notifications/unsubscribe-auth/AGY1qgUAUL7vez2VsdSvaTWjh5BxrO5Nks5uzFCxgaJpZM4Yv-U8 .

Rababermaber commented 5 years ago

EM71280_MPU9250_BMP280_M24512DFC_WS_Acc_Cal.zip

There is not much code I could send you, I am mostly using the code provided in this Git. The code I linked above is exactly the one I used when trying to calibrate the sensor (I have not changed much other than the INT pin and some I²C stuff). I am using a Teensy 3.6 and I am using this sensor I bought from you #link; .

kriswiner commented 5 years ago

OK, but this sensor does not use the MPU9250, so why are you using a sketch that supports the MPU9250?

On Tue, Nov 27, 2018 at 3:26 AM Aaron notifications@github.com wrote:

EM71280_MPU9250_BMP280_M24512DFC_WS_Acc_Cal.zip http://Sketch

There is not much code I could send you, I am mostly using the code provided in this Git. The code I linked above is exactly the one I used when trying to calibrate the sensor (I have not changed much other than the INT pin and some I²C stuff). I am using a Teensy 3.6 and I am using this sensor I bought from you #link https://www.tindie.com/products/onehorse/ultimate-sensor-fusion-solution-lsm6dsm-lis2md/#specs ; .

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/kriswiner/EM7180_SENtral_sensor_hub/issues/45#issuecomment-442024714, or mute the thread https://github.com/notifications/unsubscribe-auth/AGY1qpL6QCIT9w4wWFoA4R2q_33_d-1Xks5uzSFvgaJpZM4Yv-U8 .

Rababermaber commented 5 years ago

Ok I see, but that explains why the WarmStart did not work. But for all my other testing i was using the code provided for the sensor baord. Is there an WarmStart sketch I could use for my sensor?

kriswiner commented 5 years ago

I would start with getting the sensors properly calibrated and getting sensible heading behavior out of the EM7180 before I would worry about warm start.

Can you calibrate the sensors? Can you confirm the sensors are calibrated? When the sensors are properly calibrated, do you see reasonable orientation estimation? If not, specifically what is the problem (a plot of yaw vs actual heading or time would help).

It's pretty hard to know from the little you have told me what the problem(s) might be.

On Tue, Nov 27, 2018 at 1:57 PM Aaron notifications@github.com wrote:

Ok I see, but that explains why the WarmStart did not work. But for all my other testing i was using the code provided for the sensor baord. Is there an WarmStart sketch I could use for my sensor?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/kriswiner/EM7180_SENtral_sensor_hub/issues/45#issuecomment-442232881, or mute the thread https://github.com/notifications/unsubscribe-auth/AGY1qm1EEM5Tt3832rr66-8aWXOvOA-Rks5uzbVfgaJpZM4Yv-U8 .