ModalAI Forum

@Eric-Katzfey From what I read in the forums, the M10 seem to be a very nice improvement to the M8N. It would be very nice if you can send them our way. Cudos to your team: They have already reached out to us with an RMA.

@Eric-Katzfey When saying that M8N units aren't working, I mean that we do receive data, GPS location is showing up on the QGroundControl map roughly were we are. But the incoming data is not accurate enough for good position data and PX4 will not go into position mode warnings like: "GPS Horizontal Pos Drift too high" and others. I confirmed that it is the GPS units by plugin the GPS devices into another drone using a Pixhawk 6C. The M9N unit that was in their before should perfectly well position data with no drifting. When plugin in the M8N from our RB5 the same warning started to appear and on the QGroundControl map the drone was drifting around again.

Today we also tested the Holybro M9N device from another drone on our RB5. It was not a perfect GPS condition, even the GPS of my cell phone was moving around (Rainy day / had to stand under cover). But using the M9N the PX4 on the RB5 reported stable position and no warnings appeared being in position mode.

Also revering to the following forum post from my colleague, using the M9N we now got magnetometer to work.
https://forum.modalai.com/topic/3410/outdoor-navigation-problem/9?_=1716987565941

@Moderator We thought of the GPS devices moving across half of the earth and had them being turned on standing outside for multiple hours now I would say. We also did a lot of VIO outdoor flying with the GPS plugged in. So the should have received all relevant GPS data by now.

Dear Team,

I would like to know if the Holybro M10 (Secondary Version using the 6-Pin JST GH) can be regarded as a plug&play alternative for the M500 as opposed to the Holybro M8N that we received our drones with?
Here is the product: https://holybro.com/products/m10-gps?variant=42991226224829

Some of the M8N that were preinstalled on the M500 don't seem to be working.

We see that you are selling the uBlox M10 for the M500 as well on your products page for the M500 still with "more details coming soon".

Regards
Tim

Drone: RB5
System-image: 1.7.1-M0052-14.1a-perf-nightly-20231025
Kernel: #1 SMP PREEMPT Thu Oct 26 05:24:02 UTC 2023 4.19.125
hw version: M0052
voxl-suite: 1.1.2
PX4 Version: 1.14.0dev

@droen8 Your video is set to private. I can not watch it.

@Moderator This is a very interesting pose as we were about to post something very similar. We had two drones hovering. One dropped out just sitting/hovering there.

Detailed description under the video: https://www.youtube.com/watch?v=H2JpUSIQqXA

Will take a look at vibration. But VIO in this outside location has been very unstable for us. On nearly every flight it VIO stopped working.

Regards.

@Eric-Katzfey said in Take-Off and Land command each fly to xy=[0,0] location instead of on the spot:

@LaborzaubererTHI Sounds like it may be similar to: https://forum.modalai.com/topic/2533/failsafe-landing-bug-in-px4-1-14/19?_=1713800679314

Indeed it does.

@wilkinsaf
@LaborzaubererTHI

How would one go about using the different coordinate systems? All of our code is currently running on an external computer, not on the drone. [Question might need separate thread].

@wilkinsaf

We are connected to the drone via MavSDK. We booted the drone at the [0,0] location. Carried it forwards to [1,0] and send the takeoff and land commands via MavSDK.
Due to the fact that I think we are connected to PX4 I suspect we are using its coordinate system.

Regards.

Dear developers,

while connected to PX4 instance over MavSDK, using the python environment, we initiate the take-off and land command.
But instead of the commands being executed on the spot, meaning where the drone is, the drones fly towards its xy=[0,0] location while executing the respectively command. So instead of flying vertical each time, they actually fly diagonal.

More detailed description:
Drone position at [1, 0, 0].
Command takeoff.
Offboard to setpoint [-2,0].
Command land.

April tag #0 is set to “unknown” and no other tag specified.

Result:
Took off and flew diagonal to location: NED = [0, 0, -1.5].
Flew backwards to [-2,0,-1.5] as commanded in offboard mode.
Land command led to a diagonal land flight at [0,0] location, where it initialy took off from.

Expected:
Take-off and land command would lead to vertical flight at the xy-location where it was when initiated.

The flight can be seen here: https://www.youtube.com/watch?v=Trz97csbsz0

Our example is very close to your example offboard_position_ned2.py found here:
https://gitlab.com/voxl-public/voxl-docker-images/voxl-docker-mavsdk-python/-/blob/master/offboard_position_ned2.py?ref_type=heads

From testing PX4 within an gazebo environment, we did not expect such a behavior.

Regards
LabMag

Drone: RB5
System-image: 1.7.1-M0052-14.1a-perf-nightly-20231025
Kernel: #1 SMP PREEMPT Thu Oct 26 05:24:02 UTC 2023 4.19.125
hw version: M0052
voxl-suite: 1.1.2