Is this the GPS Firmware Bug? Yikes!

[rrmccabe]

It happened so fast, I can't honestly say I know what happened, but I assumed it was loss of satellites. I was doing not-terribly-precise, figure-eights in my backyard, a few feet above the ground, when the Solo pitched violently from side to side, and then darted away. It was unexpected, and lasted for maybe two seconds. My brain wasn't quick enough to think about Pause or anything else. I just instinctively pulled back on the right stick. But the Solo continued moving forward.

The Solo's behavior wasn't attributable to wind. That I do know.

Yea I know how that goes Sarah. My brain cant process that fast in those situations.

All I can tell you is this. If a UAV is working properly and it loses satellites it should start to slowly drift with whatever wind you have. I have lost satellites with no wind and did not even know until I say it on screen.

Program your B button to manual fly mode and activate that while hovering. That is exactly what you should expect when you lose satellite.

 

[Sarah B]

Thanks for the advice, Rick.

Joking! Thanks, Rich. I WILL program the B button to manual fly mode.

I don't know as much about these helicopters as you and the others, but I own two Phantoms and have lost satellites a gazillion times. I know what that's like, and it isn't what happened with the Solo. Unless whatever happened was attributable to operator error, and I'm pretty certain it wasn't, there was a glitch in the Solo's behavior. I saw the videos of the 3DR spokesperson and the person whose helicopter flew into his house, and I thought I had experienced the same sort of problem.

 

[rrmccabe]

LMAO

Well it should act just like your Phantom when it switches from GPS to manual. If not, 3DR needs to work on it as there is no room for an intermediate fly around the yard while this is happening !

 

[Raybro]

So far, no glitches. None that were my own fault. The only thing I can report is my bonehead dog chewed up the foam block that goes between the battery and chassis. Had to make a new one.

 

[rrmccabe]

I can report is my bonehead dog chewed up the foam block that goes between the battery and chassis.

Did you submit a trouble ticket?

 

[nacjr]

So far, no glitches. None that were my own fault. The only thing I can report is my bonehead dog chewed up the foam block that goes between the battery and chassis. Had to make a new one.

This piece of foam came in my Solo box:

Is it supposed to be used somewhere?

 

[Raybro]

Place between the battery contact side and chassis to store the Solo and battery in the supplied box. It keeps the battery from making contact. You don't want to store the Solo with the battery connected. It's probably not a good practice to do anyway.

 

[nacjr]

Place between the battery contact side and chassis to store the Solo and battery in the supplied box. It keeps the battery from making contact. You don't want to store the Solo with the battery connected. It's probably not a good practice to do anyway.

Thanks, mystery solved.

nc

 

[Raybro]

LMAO

Well it should act just like your Phantom when it switches from GPS to manual. If not, 3DR needs to work on it as there is no room for an intermediate fly around the yard while this is happening !

Took it out, got the Selfie down pat, no glitches this time...
Flew a little in manual and had some winds aloft even though it was clam at ground level. Taking it back out of manual, well, I never had this many buttons on my P2.....gotta get very familiar with them. I the fly button and it went into orbit mode. I eventually got back to standard fly mode. I could have just hit the "A" button. Living and learning......

 

[Rothgarr]

Regardless if it's a Solo or Phantom, etc... THIS is why you don't fly around/over people.

 

[Lon Snyder]

There's supposed to be a firmware update out any day now that will address Solo's erratic behavior when it loses GPS. There does appear to be a real problem that is going to be fixed. Yes it's true people have been flying Solo too close to people and property. Notwithstanding that, when it loses GPS it should not just dart off in an unexpected direction.

 

[DodgeP]

I tried to fly today but it is a little to windy for my taste.

 

[thefrog]

I hope this "GPS bug" is software issue and not the hardware problem caused by the interference from GoPro or SOLO onboard computers. If this is indeed hardware issue, I would rather consider to return SOLO back to BB now and wait until new batch of SOLOs with fixed hardware will be out. Good GPS receiver does not loose 3D lock in such dramatic way as some SOLOs did with few trees around. Will test it on SOLO and Galaxy S6, I would expect S6 with much smaller, lower gain GPS antenna will loose GPS 3D lock much faster than SOLO in bad reception areas.

 

[DodgeP]

I just picked up a Galaxy Tab S yesterday so I could download Tower but I haven't experienced any of the GPS issues that others have.. Fingers Crossed!!!

 

[Sarah B]

I may have figured it out. In my backyard, with seven satellites, the horizontal dilution of precision--something I'd never heard of before joining this forum--is nearly four. Earlier today, flying in the parking lot of a college soccer field, I had thirteen satellites, and the HDOP was 1.8.

I don't know what kind of algorithm 3dr uses to combine the two numbers and deem whether it's safe to take off, but I'm thinking they have it wrong. I think that yesterday, when my Solo developed a whacked mind of its own, the Solo and I thought it had a good satellite lock, but it didn't. I think GPS orientation got screwed up, and Solo momentarily thought it should be someplace else.

This would also explain the trade show flyaway, and the Solo that zoomed backward into a house. There's also the video here of the Solo acting crazy in the canopy of trees. And I've read where people have twelve satellites, but don't get a clear-to-fly message. Maybe that's because the HDOP is too high, despite all the satellites.

It looks like, when there's momentary satellite confusion, rather than going into manual-flight mode, the Solo thinks it's somewhere it isn't, and zooms to where it thinks it should be.

And if I'M completely off kilter, I'm CERTAIN someone here will let me know.

 

[1013D]

....And I've read where people have twelve satellites, but don't get a clear-to-fly message. Maybe that's because the HDOP is too high, despite all the satellites.
.....

that happened to me, and happened again today, but does not seem to be directly related to HDOP only.
today I had HDOP of 2.4 and 10 sats, and waited and waited and waited and never got the ability to start the motors. so I power cycled the solo only, and then after 20 seconds I had 8 sats and HDOP of 2.8 and was able to start the motors and fly......so that issue (of not getting the all clear to start the motors) is likelly something else.

 

[sixshooterstang]

Answers to all of your problems lay in github. Looks like there was an issue with the GPS timer resetting If you don't know how GPS works the satellites have extremely precise atomic clocks. When a receiver is started up it synchronizes with satellites and the internal timer begins to measure how long it takes each timestamp from each satellite to reach the receiver. Using this information a distance can be measured between each satellite in view and a position can be triangulated. If the timer resets and becomes inaccurate then an incorrect position will be reported and the aircraft will fly off trying to aggressively correct its position.

Since there are only headings and not much for details on github i cannot say for sure if this is the case but it is a possible explanation.

Seems they also had some issues with the EKF (glitch filter that uses the IMU to detect gps glitches) and with EKF failsafe.

https://github.com/3drobotics/ardupilot-solo/pulls?q=is:pr+is:closed

If you go down the page there are several new and fixed things in the works on there that will likely be an update here soon as Ian Hudson mentioned on Facebook.

 

[thefrog]

Answers to all of your problems lay in github. Looks like there was an issue with the GPS timer resetting If you don't know how GPS works the satellites have extremely precise atomic clocks. When a receiver is started up it synchronizes with satellites and the internal timer begins to measure how long it takes each timestamp from each satellite to reach the receiver. Using this information a distance can be measured between each satellite in view and a position can be triangulated. If the timer resets and becomes inaccurate then an incorrect position will be reported and the aircraft will fly off trying to aggressively correct its position.

Since there are only headings and not much for details on github i cannot say for sure if this is the case but it is a possible explanation.

Seems they also had some issues with the EKF (glitch filter that uses the IMU to detect gps glitches) and with EKF failsafe.

https://github.com/3drobotics/ardupilot-solo/pulls?q=isr+is:closed

If you go down the page there are several new and fixed things in the works on there that will likely be an update here soon as Ian Hudson mentioned on Facebook.

gps_glitch filter was implemented long time ago, and none of my Pixhawks never have any issues with it. Looking at the code and changes at the github it more appears the SOLO have some issues with the GPS reception for some reasons and they are trying to fix it with the more aggressive filtering, which in worst case will completely disengage GPS. Again, I have no issues with my SOLO. I only fly in open space with no trees, hills or buildings.

 

[thefrog]

Quick no-fly test under heavy canopies and hills around - nothing spectacular. Solo locked to 8-9 satellites after while with 1.2 HDOP and Galaxy S6 to 17-19 satellites including Glonass. Solo position however drifted slowly by about 30ft and back within minute, while the Galaxy S6 drifted on google maps by about 15ft in the same time. I would say this is normal and expected considering S6 uses twice as much satellites.

 

[sixshooterstang]

If that is truly the problem mine will eventually be getting an antenna upgrade. Did it once to a naza as an experiment and got solid improvements indoors and out.

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.

Accept third party cookies