I replaced the GPS module in my 520e. The remote sees satellites but the 520e does not. Today, I learn from a rotational experimentation that the compass and gps are not on the same board. I have checked continuity from the gps board to the motherboard. I have checked every connector that I can see to be plugged in. I have reset and updated software. Does anyone have any input? Does anyone have schematics or block diagrams? I'm popping smoke for any help.
GPS not working
- Thread starter wmattp
- Start date
"the compass and gps are not on the same board"
I'm not an electronics expert, and have not dealt with GPS problems on the HPlus family. But I think the compass is actually on the GPS board for Hplus, H520, H520e.
Can you describe the rotational experiment that concluded it is not?
Thanks for getting back so quickly. Having the upper shell uncrewed, I rotated the 520e and watched the compass rose in the controller display move. I then rotated only the upper shell, to which the gps is attatched. The compass did not move. This told me that the compass is NOT in the gps module. I see the pics you sent. This is the exact gps that I have. Now I'm really confused.
"Now I'm really confused."
Me too. Your experiment sounds reasonable to me. I will give it a try on mine but it will be more than an hour before I can do so. I will come back with the results as quickly as I can, but an hour may even be a little optimistic.
"I will come back with the results as quickly as I can,.."
O.k. The results are in. With surprising results.
My results inside the house, surrounded by magnetic interference, was the same as yours. So I took it outside, in the middle of a large open area. That's where I learned a couple of things.
INDOORS: (First time)
I got the same result as you. Fast response to moving the drone itself. No response from moving only the top shell/GPS module.(but see comments under "INDOORS: (second time)" below.)
OUTDOORS:
I took an original H520 out to a large open area and repeated your experiment. I got a different response.
With the top shell in place, it looked just like the indoor response. Fast changes that matched the movement of the drone.
But a big difference when moving the top shell/GPS unit only. The compass would move SLOWLY to the new heading. A 90 degree turn took about 15 or 20 seconds for the compass display to move slowly to and match the new heading.
INDOORS: (second time)
It looked like the outdoor response, but with even longer time for the reading to catch up with the heading of the top shell. I wonder if this is what was happening the first time, and I did not notice the very, very slow movement.
Conclusion:
What you see short term and rapid response on the controller compass is probably a calculation. It is likely a best estimate calculation based on the best available compass readings and movement sensed by the IMU. The full system seems to use the actual compass signal, averaged over time (the running average time based on current compass reading quality?) as the basis for the heading. It then uses IMU input to calculate a fast reading that matches the short term movements of the aircraft. This is just my guess. The smart guys will need to help us understand if this is even close to correct.
You will need to check yours outside, but I suspect there is nothing wrong with your compass and we need to look at more areas related to GPS acquisition.
The next area to look at might be the GPS almanac.
If the new GPS unit is actually "NEW", or if it has been in storage for a few years, it is likely it does not have a current GPS almanac in it's memory. You will need to turn the drone on and leave it somewhere with a clear view of the sky for as much as an hour to acquire the almanac.
The GPS satellites broadcast the almanac every 12 minutes, so theoretically you might get a lock in just a few seconds (outside). There is a good chance of getting a lock by 15 minutes. If no lock by an hour, something is wrong.
@WTFDproject i like the theory you proposed and your deductive process on this one.
I believe the H series also includes a 3D compass chip on the F/C module and may compare the readings between the two. When they move together and in close agreement the readout could be faster. When only the one on the GPS board is moved they would differ and updating may take longer. Especially since the IMU is not detecting motion.
@h-elsner is more knowledgeable on these things and may have a better answer.
I believe the H series also includes a 3D compass chip on the F/C module and may compare the readings between the two. When they move together and in close agreement the readout could be faster. When only the one on the GPS board is moved they would differ and updating may take longer. Especially since the IMU is not detecting motion.
@h-elsner is more knowledgeable on these things and may have a better answer.
h-elsner
Premium Pilot
- Joined
- Mar 23, 2016
- Messages
- 2,366
- Reaction score
- 2,106
- Location
- Bavaria / Germany
- Website
- h-elsner.mooo.com
I have no experience with H520E but the principles are always the same for PX4 based drones.
The compass and the GPS are on the same module but different systems. The compass is connected to I²C bus and the GPS is connected to a dedicated serial connection (UART).
Once the I²C bus has a problem the GPS is working, the compass not.
The rotation will also be detected by IMU. Compass (3D magnetometer), 3D accelerometer and 3D gyroscope are the components to detect the attitude of the drone in all 3 axis.
You can connect the drone via micro USB to your PC and check the MAVlink message RAW_IMU. There you can see al six values and can check which changes and wich not. I use QGroundControl but maybe the Yuneec derivate (Datapilot?) of QGC can do this too.
The compass and the GPS are on the same module but different systems. The compass is connected to I²C bus and the GPS is connected to a dedicated serial connection (UART).
Once the I²C bus has a problem the GPS is working, the compass not.
The rotation will also be detected by IMU. Compass (3D magnetometer), 3D accelerometer and 3D gyroscope are the components to detect the attitude of the drone in all 3 axis.
You can connect the drone via micro USB to your PC and check the MAVlink message RAW_IMU. There you can see al six values and can check which changes and wich not. I use QGroundControl but maybe the Yuneec derivate (Datapilot?) of QGC can do this too.
h-elsner
Premium Pilot
- Joined
- Mar 23, 2016
- Messages
- 2,366
- Reaction score
- 2,106
- Location
- Bavaria / Germany
- Website
- h-elsner.mooo.com
Another message is SYS_STATUS. Here we should see what HW sensors are present and working (and if there is a second GPS/compass).
Example is from H Plus.
Example is from H Plus.
Last edited:
Thank you. I will try that to see if it works.
I let it sit outside for over an hour. No gps data. The camera works. I see other sensor data from the vehicle. No gps
Question: During your Indoor experiment, did you see gps data from the vehicle displayed on the remote?
Yes, but it depends a lot on which room. How many floors, how much insulation, construction materials and furniture. And how many trees. For me, it eventually shows a few and today got enough to lock in. Yesterday it never locked in while indoors.
If you saw zero GPS satellites while sitting outside for an hour, it is a problem.
Did you see any drone satellite count at all while sitting outside?
Do you still see controller satellites while indoors? While outdoors?
I feel dumb because I should have asked this before. But;
Why did you replace the original GPS module?
(Was it by any chance acting EXACTLY like the "new" one?)
Why did you replace the original GPS module?
(Was it by any chance acting EXACTLY like the "new" one?)
The controller sees gps everywhere in my testing. The 520e does not sees gps, according to the controller. The 520e seems to be well connected to the controller.
Similar threads
