I've been looking through the forums and have come across instructions to use both versions 1.6.5 and 1.8 of the PX4 firmware ( suggests 1.6.5 while says to upgrade to 1.8). Could someone please clarify which firmware to use as I have experienced some issues calibrating the compass while using the 1.8 version firmware and am considering reverting to the older version. Thanks
Thank you for your interest in the Intel Aero Ready to Fly drone.
On /thread/127056 Intel aero waypoint - QGC failing to send mission waypoints the issue appeared when uploading the flight plan in QGroungControl, after setting the waypoints. This issue was solved by flashing the latest PX4 firmware version (1.8).
On the other hand, on /thread/127761 Intel Aero RTF - CRITICAL NAVIGATION FAILURE the error is not received when uploading the flight plan, but after taking off. We recommended the 1.6.5 PX4 version to see if the same behavior is to be seen.
Could you please give me more information about the calibration issue that you have encountered?
I reverted to the 1.6.5 firmware but I am still experiencing significant compass error and drift after calibrating the drone numerous times. The compass appears to be 90 or 180 degrees off at times but when the vehicle is rotated, the offset compass will drift to a different offset. On the 1.8 version of the firmware, I was additionally experiencing issues with the EKF filter, which would throw warnings through QGroundControl about its internal state if the drone moved significantly.
Could you please confirm that you have made an Airframe reset after downgrading the firmware?
You notice this behavior only while doing the compass calibration? Could you please provide us a print screen of the view in QGroundControl?
Thank you in advance,
Yes I did the airframe reset - currently I am using the Intel Aero Quadrotor X setting. After investigating the EKF filter parameters I decided to try the separate Attitude Estimator by setting the SYS_MC_EST_GROUP to "local_position_estimator,attitude_estimator." That opened up several parameters in the Attitude Q Estimator section, one of which is the ATT_W_MAG that controls the filter magnetometer weight. Setting this parameter to 0 allows the filter to run without the magnetometer. When I try this, the drift is almost entirely eliminated and the heading seems to respond well although it is off by some angle depending on the orientation of the drone when it is first powered on. This seems to indicate a problem with the compass sensor, its calibration, or the relevant parameters in the heading filter.
I can take some screenshots in QGroundControl but it will be difficult to capture the compass errors or drift. If you would like to see a screenshot of the parameters set on the vehicle I can provide that.
Which software version are you using? Have you updated to the latest (1.6.2)? If not, please update all the components. You can download it from https://downloadcenter.intel.com/download/27833/Intel-Aero-Platform-for-UAVs-Installation-Files here and by following the instructions from https://github.com/intel-aero/meta-intel-aero/wiki/02-Initial-Setup here
Please also send us the output of get_aero_version.py
Can you send us a picture of the drone (top view)?
Is the behavior also seen with Ardupilot?
Please do not modify any settings after the frame reset and calibration process.
Yes I have updated to 1.6.2 and updated all the other components. Here is the output of aero-get-version.py:
BIOS_VERSION = Aero-01.00.16
OS_VERSION = Poky Aero (Intel Aero Linux Distro) 1.6.2 (pyro)"
AIRMAP_VERSION = 1.8
FPGA_VERSION = 0xc2
AeroFC firmware version = 1.6.5
I have not tried the ardupilot firmware but I have tried the version 1.8 of PX4 with similar results. Today I took the drone to a remote flying field with no large structures or metal nearby and performed an airframe reset and calibration. Without adjusting any parameters afterwards, I experienced the same compass drift as before. When attempting to fly the vehicle, I found that position hold mode failed to keep the drone in a stationary position and even when controlling the vehicle via qgroundcontrol, the vehicle would begin to drift soon after taking off.
Is it possible the magnetometer on the flight controller board is defective? Is there any way to make use of the magnetometer onboard the intel comput board with the flight controller? Maybe a small c program to read the imu and generate mavlink messages that forward the measurements to the flight controller? I can also try installing the ardupilot firmware on the flight controller and see if that performs better. Thanks for your help.
It is not possible to use the onboard magnometer without making code changes.
From the picture that you have provided, we can see that the orientation of the GPS module is wrong, please turn the GPS module 90 degrees to the left. The "GPS" letters need to be read parallel to the bottom of the drone.
Is the magentometer the vehicle uses located in the gps module as well? After correcting the orientation and recalibrating everything I still observe the compass drift I described earlier. If the magnetometer is housed in the gps module, perhaps I can try replacing this component?
When trying to debug this some more, I tried rotating the gps module relative to the drone while monitoring the reported orientation of the vehicle in QGroundControl and didn't see any change. If the magentometer in the gps module is the one being used by the system I think this indicates there is some fault with my compass hardware. Please let me know if this reasoning makes sense. Thanks.
The Intel Aero Flight Controller includes a mag, baro, and accel that are used by the flight stack. I suggest we do more troubleshooting before you replace any components.
When you calibrate all of the components via QGroundControl, do you see any errors when performing the compass calibration?
Are your compass errors seen in the outdoors with a good GPS signal? Could you share a video of the drone's current behavior?Looking forward to your reply!
After troubleshooting some more, I decided to replace the gps/compass module since the compass from the module seems to be routed into the flight controller. After replacing the module, the compass issues went away and I was able to calibrate with both 1.6.5 and 1.8 px4 firmware. Sending waypoints did not work very well with 1.6.5 but seems to work with the latest firmware (I.e. 1.8). Since the module did not work out of the box, is there any way I can rma my compass module? Thanks for your help.
Thank you for the information!
If you are looking for RMA, please reach out the place of purchase and they will guide through the process.
Thank you very much,