Sebastian,
Let’s try to figure out lost fix first. Can you try to reproduce it? Is it time based (you leave it for some time and it loses fix) or is it boot based (sometimes it works after reboot, sometimes not)?
After you are able to reproduce it, it would be great to turn everything off except for Navio+ and see if it depends on the presence of other equipment.
I was not able to open two of your logs, only the one with GPS working. Can you try sending them again, probably in an archive?
From what I observed today on field (and on my desk), GPS does not get any sats at all from the start, not really loosing fix.
My tracker has an APM2.5 and a Navilock 652ettl GPS (ublox6 based) and it gets a fix with 5 to 6 sats 3 meters away from the window in my working room. The copter is right behind the window and does not show 1 sat.
On the field today, I plugged out the wifi adapter and power to the FPV TX and waited and waited, but no sats and no fix after several minutes.
I will try to access the GPS with u-center. Perhaps I see something, that gives me a clue.
U-center is up and running. The copters GPS sees plenty of sats (GPS and GLONASS), but was not able to get a fix so far. While the US sats show a constant “wave” of received signal strength, the russian sats are “flickering”. Every second receiption bar is left out. I have my rover powered up, too. It is right besides my copter on my work table, but it does not get a fix, either. Of course, my old and worn Navilock GPS has a lock already…
Ah now, the rover shows three, no four green sats. I also see the sats positions on the world map and the compass rose for the rover. Nothing for the copter…
Rover: (The fifth sat got green in the second I took the screenshot)
Copter:
U-Center is running for 40min now, still no fix. I put a metal jar lid under the GPS antenna as a groundplane, lets see what happens…
Two of the US sats are shown one bar darkblue and one bar lightblue in the Satellite Signal History window.
Ha, and a japanese sat wants to play, too.
Something is very wrong with the GPS. SNR levels are at 20-30 and should be close to 50. Looks like antenna malfunction. Could you please inspect the connector and the cable to antenna?
Do you by any chance have another MCX GPS antenna to try with?
If you have a voltmeter, could you measure the voltage between the center pin on MCX connector and connector body?
The voltage on the connector of the Navio or resistance on the antenna connector?
Voltage on the connector. (antenna disconnected)
3.3V with RPi shutdown, but power connected.
Changing between ca. 2V and 3.3V with Arducopter running.
Can you as well measure 3.3V voltage on the pads under the board?
My only version at the moment is that it is an antenna or u-blox malfunction.
I measured 3.3V voltage on the GPIO header, is this ok, too? If you need the measurement from the pads, I will have to do it tomorrow.
It was changing between 3.1 and 3.28 Volts.
On the GPIO header is RPi 3.3V supply, we do not use it on Navio+. Probably fluctuation is caused by your measurement tool? Is it a trusted one? It is very unlikely that fluctuation like this occurs on 3.3V rail.
Can you make a test with a different antenna? We can send you a new one right now, but delivery will take some time.
It did not fluctuate that much with other measurements. I will leave it be for today and start again tomorrow.
I will look for a new antenna from a local supplier, do you have any brand recommendations?
Not really, look for MCX connector and GPS/Glonass support.
I suspect the antenna cable broke close to the connector. I altered the cable routing and suddenly I got a fix after only a few seconds. It is not a final conclusion, but I will order a new antenna and see, if it gets better then.
Only problem I have, all the antennas I could find are meant for car navigation and have 5m connection cables…
The thing with RF cables is that they have a minimal turn radius, after which the losses increase and probably you have bent it too sharply. Could you try flattening it and observing the SNR levels?
I straightend the cable, but receiption is very poor. If I move the cable, I sometimes get spikes of higher SNR, but not for long. Right now, I have three sats visible with a SNR of 15 and lower.
Ok, it really looks like a damaged antenna. We will send you a replacement.
I am just back from a outside test (no flight, unfortunatly).
First boot, with everything running (WLAN, RC, FPV-TX 25mW), no GPS lock after 10min. The same setup, no changes in layout, no groundplane under the antenna, ran fine with APM2.5 and 3DR GPS/compass.
Second boot, just leaving the FPV-TX unpowered, killing Arducopter and mavproxy, starting u-blox tcp server, connecting with u-center, already a lock with 5 sats, improving to 8 or 9 before the third reboot.
Third boot, connecting with Missionplanner, already a lock (10-11 sats, hdop around 1.5m), but the GPS position was moving all over the place, sometimes a few 100m away. My antenna tracker nearly got a seizure, trying to follow the reported positions.
I remembered the problems I had with Ardurover and GPS with NavEKF enabled. I searched in the full parameter list, but it is possible that there is no parameter in AC3.3-rc3 to disable NavEKF anymore?
I cancelled all plans for a testflight, because my Skywalker just had a hard “landing” (aka crash, no major damage 
) a few minutes earlier and I did not fancy a second crash in one day.
@schuermannsebastian Now this is getting really strange. From your first test it sounds like FPV transmitter seems to interfere with GPS. NEO-M8N has a wider front-end to receive Glonass signals, so there is a bigger chance of in-band jamming.
The parameter should be there, but we were flying it completely fine with EKF enabled.
Could you try to achieve good GPS SNRs in u-center first and then try turning equipment on one by one and observing satellite levels?
EKF was just one thing I wanted to try. It was called AHRS_EKF_USE and this parameter is gone for sure. Perhaps it has some obscure other name now.
Today I got some mixed results. I added a round groundplane 70mm in diameter. 1mm plywood with aluminium tape.
All test with FPV-TX unpowered.
First boot, no lock with APM/Missionplanner after about 7min, killed mavproxy and APM, started u-blox server and u-center. It had a lock as soon as the connection was established, with 6 sats.
Reboot, APM has a lock, too, but only 8 sats max, sometimes going down to 4 or even loosing lock, with the copter stationary on the ground.
Got frustrated, so I tossed my DLG a few times.
Next boot, again no fix with APM, but as before, a fix with u-center (killing APM/mavproxy before). This time u-center had a lot of sats visible and finally before reboot, it had a fix with 15 sats. Max SNR was around 40.
Reboot and APM had a fix with 15 sats. The number was going up and down all the time, not 5 seconds the same number of sats. 12 sats was the lowest I observed. The position reported to Missionplanner was moving in a 10m radius.
I am just testing with u-center on my desk again. Connected over ethernet cable and pluging/unpluging the wifi adapter.
It really makes no difference.
Is it possible that APM sends some kind of wrong configuration messages to the receiver?
@schuermannsebastian Have you compiled the code yourself from master? Could you try same thing with our .deb?
