Strange things happened with booth versions- 4.3.5 and 4.5.0 from custom build server.
We have in some points GPS loss.
But not total- from 20 sats I have 11-13 and HDOP up to 1.2.
And I have gps or compass glitch, EKF variance, Althold.
But GPS pos is ok, Nsats ok, hdop ok.
If I leave copter in Althold for 10-20 sec- I have “Glitch cleared” and I can switch to Loiter for long time.
When I try to fly again- I have in few sec Glitch- Althld again.
Ok. No problem. I see in log abnormal speed “pulses” so EKF generate Glitch.
But why in same area and same time I have no problem with RTL???
Copter just flying home without any errors.
I tested lot of times. Forward fly in Loiter stable ALthld and even DeadReconig started. Back flight in RTL -no problem.
I tryed relax EKF.
In fact I make copter stronger to variances and innovations.
But it doesn’t fix this strange.
Also who can help me switch off Compass realighment in fly?
I switch off compass calibrating in flight, in EK3 Compass YAW reset… it looks like I switched off everything about.
But first I see copter YAW rotation, compass changed course, after that “GPS or Compass error”
I know that it’s GPS error!
How can I say it to AP?
How I can say "if you see GPS or Compass error- be sure. It’s GPS. Leave compass alone!
My GPS is unthrust usually.
So compass and baro are 1-st priority in flight.
And it’s very bad that I have “Emergency YAW realigned” - I flying 83 deg course, after realign I have course 186… WHY???
I can’t have a thrue compass?
I make Comp-Mot calibration and adjust it manually. So at any copter movement in Althold (bigger curent and interference)I have up to 1-2 deg short term deviations only.
And wnen I fly in good GPS area- I can do anything with copter- no variances, no errors…
What you describe is not an issue related to the firmware versions, or at least highly unlikely.
You should probably start your own discussion instead of continuing this in the 4.4 Beta discussion.
A .bin log would make it easier to identify the cause.
GPS glitches are not necessarily an issue with the GPS itself, but more a disagreement about position and heading. High vibrations, poor HDOP, poor compass calibration or interference can all cause that message.
HDOP should be below 1.0 to be good, typically 0.8 to 0.6
Ensure your current sensor is accurate (and voltage) then do a flight with circles and figure 8’s and some up/down throttle. Then we can use Magfit to refine the compass calibrations.
I am able to do Radio Calibration just fine, ie, I can fly the drone but when I go to detect New Sensors, I get only the 13 “default” sensors and no GPS sensors.
Is this part of the bug * Serial passthrough fixed or am I missing a setting somewhere.
Serial passthrough is described here on the wiki but the short-form explanation is that it allows easier configuration of some sensors and devices (e.g. GPS, gimbal, etc) by allowing an external non-ardupilot configuration program (for example UBlox’s ucenter) to pass whatever serial data it wants back-and-forth with the sensor.
I guess the “detect New Sensors” must be a page in the TBS Crossfire telemetry system? I wonder if maybe @andyp1per knows about this.
The “detect sensors” is actually a feature of the Radiomaster TX16S transmitter.
Dave, excuse my ignorance but I’m not 100% sure which board it is. I think I saw QGC showing it as PX4 but I am not convinced. It is a RadioLink Pixhawk. Is there a way I can check within QCG or MissionPlanner to see the exact board and version?
This can be seen in the banner which is displayed whenever the ground stations requests the full parameter list from the vehicle. This is an example of the banner displayed for one of my copters… so it’s a CubeOrange running ArduPilot-4.5.0-dev.
Radiolink are bad actors. If the message shows fmuv2 then you will not get CRSF telemetry. Best option is to replace the Flight Controller with a supported one. Or, you could try flashing the fmuv3 version of firmware.
Go here Fmuv3 download the .apj file and use Mission Planners “load custom firmware” button to flash it.
Thanks a million times over for your support.
I noticed in the message banner as @rmackay9 suggested to check for the board version, that the board has 2mb flash and to load fmuv3, which I’ve done and FINALLY was able to detect the GPS sensors.
