Is this IMU damaged?

[Dany]

Hello.

Tried to get the compass to calibrate using an MPU9255, it works as long as i do not assign the BMP-part. But i cannot get it to calibrate.

I have an ICM20948 to try too. When i connect this IMU to my Paspberry i can see that the boat does a (physically impossible) somersault. So i go into the dialogs an click some of the imu.* parameters. The "hickup" that you see seems to occur with the same interval, a minute or so. I'm happy to time it if it helps.

Anyone seen this? Is it a bad IMU?

Thank you for your patience and time!

/Dany

   

[Jean-Marc Douroux]

Aren't you using an SD card with a new image?

If so, it takes a lot more than a 240° rotation to cancel the calibration already stored on the SD card, which can result in a tilt of the heading. I wonder if this is not what is happening to you. Remember that the image is the copy of the SD card tested on Sean's boat. With a new image, you have to remove the computer from the boat to calibrate the accelerometers and rotate the IMU until the old calibration is canceled and replaced. Concretely, it is better not to install the computer and the IMU in the boat until you have the points that make a circle on the sphere.

It is then after having fixed the calculator to the boat and carried out the alignment of the inertial sensors with the boat flat (level) that 240° is required to calibrate the compass

[seandepagnier]

I would suggest trying another imu. The plot you see does indeed seem very strange, but without more informations it could be software or hardware.

What is your setup? Maybe another program is interfering with the i2c device every 60 seconds? Is pypilot possibly resetting? Did you check the log?

[Dany]

I would suggest trying another imu.  The plot you see does indeed seem very strange, but without more informations it could be software or hardware.

What is your setup?  Maybe another program is interfering with the i2c device every 60 seconds?  Is pypilot possibly resetting?  Did you check the log?

This is what i find in the log (below). Sorry, but i cannot interpret it . This comes repeatedly.
Tried the -a flag too to copy the entire long lines. I am too inexp in terminal. Hrm 

However; it comes in even if i disconnect the IMU so i'd guess it isn't related (?).

Ordered a pimonori version. Will end up in the mail in a couple of days.
Don't put too much effort into this meanwhile.

My setup is standard OpenPlotter on RPi5, latest image + updates (not much changed), a motorcontroller on UART and windspeed though RS485 HAT, no RPS yet.

TIA!

Code:

zeroconf exception:
Sep 24 11:54:57 ironic pypilot[1031]: zeroconf addresses ['192.168.1.104', '127.0.0.1', '10.42.0.1'] 3
Sep 24 11:54:47 ironic pypilot[1031]: store_file /home/pi/.pypilot/pypilot.conf 774549.548 imu.gyrobias
Sep 24 11:53:57 ironic pypilot[1031]: e is {'192.168.1.104': ServiceInfo(type='_pypilot._tcp.local.', name='pypilot._pypilot._tcp.local.', addresses=[b'\xc0\>
Sep 24 11:53:57 ironic pypilot[1031]: zeroconf exception:
Sep 24 11:53:55 ironic pypilot[1031]: zeroconf addresses ['192.168.1.104', '127.0.0.1', '10.42.0.1'] 3
Sep 24 11:53:46 ironic pypilot[1031]: store_file /home/pi/.pypilot/pypilot.conf 774489.513 imu.gyrobias
Sep 24 11:52:55 ironic pypilot[1031]: e is {'192.168.1.104': ServiceInfo(type='_pypilot._tcp.local.', name='pypilot._pypilot._tcp.local.', addresses=[b'\xc0\>
Sep 24 11:52:55 ironic pypilot[1031]: zeroconf exception:
Sep 24 11:52:52 ironic pypilot[1031]: zeroconf addresses ['192.168.1.104', '127.0.0.1', '10.42.0.1'] 3
Sep 24 11:52:46 ironic pypilot[1031]: store_file /home/pi/.pypilot/pypilot.conf 774429.479 imu.gyrobias
Sep 24 11:51:52 ironic pypilot[1031]: e is {'192.168.1.104': ServiceInfo(type='_pypilot._tcp.local.', name='pypilot._pypilot._tcp.local.', addresses=[b'\xc0\>
Sep 24 11:51:52 ironic pypilot[1031]: zeroconf exception:

Added: When i switch to the MPU9255 i get loads of relevant error messages indication the device is not working properly. Apparently, the ICM20948 works much better. I do see something similar to the "jumps" with this one too but it's not at all the same.