Litchi locked me out, crashed!

[BudWalker]

Good observation BudWalker. So nothing really to do with Litchi. Just happened to be the app in use at the time.. DJIGo would have do the same with forward sensor on.

.....

Just a theory so far. When the weather clears here I'll try to reproduce the situation and see.

 

[Jswest86]

Just a theory so far. When the weather clears here I'll try to reproduce the situation and see.

Reviewing the flight log in csvview, specifically the relativealt sig, I could see step changes in altitude and believe there is a good chance your theory is correct.


Sent from my iPhone using MavicPilots

 

[BudWalker]

Just an update to confirm that Obstacle Avoidance contributed to this incident. The value frontSensorDistance has been added to CsvView/DatCon. During the RTH frontSensorDistance shows that the Mavic thought there was an obstacle in front of it and kept climbing.


Looking closer at an interval where the frontSensor wasn't detecting an obstacle it can be seen that the Mavic stopped rising in that interval

 

[m0j0]

Ok so once again OA contributed to the problem. So if the mode switch was toggled to sport mode would this have fixed the problem? It almost seems like you should only enable OA if you are flying close quarters, otherwise it is more of a liability. Definitely on autonomous flights it is critical to have it turned off. Additionally, if you observe any odd behavior during flight the first corrective measure would be to flip the mode switch to sport.


Sent from my iPad using Tapatalk

 

[RallySoob]

I just dont see need for litchi, i can still set way point missions in go app i just got fly to do it ofcourse. Ide love to try the 20$ app i paid for obviously but im scared if sumn happend dji wouldnt honor my refresh cuz of it. Or possibility of . Sux dji couldnt just make go app better so peeps didnt feel the need

 

[m0j0]

I just dont see need for litchi, i can still set way point missions in go app i just got fly to do it ofcourse. Ide love to try the 20$ app i paid for obviously but im scared if sumn happend dji wouldnt honor my refresh cuz of it. Or possibility of . Sux dji couldnt just make go app better so peeps didnt feel the need

Litchi is incredible. I haven't used it on the Mavic yet but i think if you actually understand how to use it, there should not be a problem. This problem was an OA Issue. The same issue would have occurred It with DJI GO.

Monday morning quarterbacking...The solution may have been to flip the mode switch to cancel OA.






Sent from my iPad using MavicPilots

 

[Aeroworks]

Whenever you have the app crash. especially in Litchi. Immediately flip the sport mode to on then off. This will cancel any previous commands or waypoints and especially park the mavic in a hover wherever it is. Then hit the return to home button or launch litchi again and try to recover video. Remember, even tho the app crashes your remote still has positive control over the craft. Only exception is if litchi is having it fly a waypoint it will keep going unless you use sport mode to override it.

Also, if you suspect a sensor issue or flying high and into the sun. turn the front sensor off. and always turn off "fly backwards" option unless you need it. otherwise if it detects something in front of it. it will just keep flying backwards.

 

[Logger]

Whenever you have the app crash. especially in Litchi. Immediately flip the sport mode to on then off. This will cancel any previous commands or waypoints and especially park the mavic in a hover wherever it is. Then hit the return to home button or launch litchi again and try to recover video.

Not "Whenever" and especially not whenever with Litchi. But I agree a useful technique to use when you are not happy with what the aircraft is doing. I do try to remind myself of its availability just before I start each Litchi flight.

I frequently get app crashes with Android Litchi Beta.These app crashes occur a short time after Loss Of Signal (LOS) on distant WPT routes on my farm. At this point, the Mavic is very happily flying the route. Pretty much the last thing I would want to do, when this occurs is to flip to sport mode and back. So not a whenever for me.
This particular fault is in the Litchi app on my Android phone. Nothing at all wrong with the Mavics firmware, merrily flying the WPT mission. If the aircraft regains sufficient signal to receive a Mission Terminate message (sports mode toggle) from me on the back side of a route, I will have just caused myself a world of pain. Limited telemetry, no video feed, and a hovering Mavic. If I just sit on my hands, the Mavic continues the flight plan, eventually regains signal, I restart Litchi. The biggest inconvenience is that Litchi does not allow me to reload the current Mission & I have lost video feed for while.

Remember, even tho the app crashes your remote still has positive control over the craft. Only exception is if litchi is having it fly a waypoint it will keep going unless you use sport mode to override it..

...unless the aircraft has lost signal, when the remote will not have positive control & sport mode will do nothing.

IMO - first thing to do after an app crash is nothing, take a deep breath, review the telemetry on the remote, assess if you are happy with what the aircraft is doing. If so sit on our hands. If not, consider taking over by toggling sports mode or whatever other techniques you have available.

 

[f3honda4me]

Just an update to confirm that Obstacle Avoidance contributed to this incident. The value frontSensorDistance has been added to CsvView/DatCon. During the RTH frontSensorDistance shows that the Mavic thought there was an obstacle in front of it and kept climbing.
View attachment 4653

Looking closer at an interval where the frontSensor wasn't detecting an obstacle it can be seen that the Mavic stopped rising in that interval
View attachment 4654

Do we know if the RTH was directly into the sun? If not, this could be a similiar bug to what the P4P had.

 

[BudWalker]

Do we know if the RTH was directly into the sun? If not, this could be a similiar bug to what the P4P had.

Don't know how direct it has to be but it was 15:00 local time and the AC was facing west. A better answer would have to include Latitude, Longitude, heading, time of day, date, pitch and maybe a few more things

EDIT: I'm not familiar with the P4P bug. But, in this incident the Mavic's front sensor thought there was something in front of it. For a brief interval when it didn't think something was in front it stopped rising. Seems compelling that the incident was caused by the Mavic trying to avoid the (false) obstacle in front of it.

 

[stratos]

Flying today with my VR goggles and the Litchi app. Flew for awhile back and fourth around my house. Went NE of my house over a hay field for one last pass. Wide open in the country, few houses, nothing else. Was headed back toward my house, probably about 200-300 yards out and about 200 feet high, I think. I got a low battery warning coming toward my house, battery was at 25%. Mavic just stopped in the air, wouldn't do anything. So I punched Return to Home and it started rising. My RTH height was set at 300 feet. But it just kept going.....up, up, up!! I know the height on Litchi app was set at Max, which was 1639 feet I believe, though I had no intention of climbing that high. But the drone did. As it got higher and higher I couldn't get any control, so I yanked the cord from the phone with the Litchi app and ran in the house and grabbed the phone with the DJI Go 4 app. Headed up the road in my truck as it was connecting, trying to get it do anything, but to no avail. Once I connected, I could see that it had risen all the way to 1639 feet and the battery was down to 9%!! Then it tried to emergency land, but 9% wasn't near enough, I believe it got to 400 feet or so and dropped. Into some soft grass, so the damage wasn't as bad as expected.

Broke one prop (No biggie), and it looks like one landing leg is a bit broke where it attaches to the body (No biggie), but the camera is trashed. Looks like it landed right on the camera (of course) and jammed it up into the plastic cage looking thing about.....the foil looking area under those plastic pieces is dented in and the camera is hanging loose and has a broken wire.

I had read somewhere that using the Litchi app doesn't void the warranty, guess we'll see what DJI has to say, can't understand why I suddenly had NO CONTROL at all when I tried RTH, still had plenty of battery for that. Why the heck would it lock me out, and climb to maximum height? One lesson is to make sure and set that to 400 feet, didn't bother to do it, wasn't planning on going very high. If that max height was set lower maybe when it locked me out it wouldn't have risen so high, then when it finally decided to land on it's own, it would have had plenty of juice to do so....then probably just a nasty landing and nothing else.

No DJI Refresh, State Farm insurance, but hope I don't have to make a claim less than a month into it lol.

Any thoughts on what the heck happened?

Well, there are many wrong things about this one. Flying with goggles without spotter? Really, second, you should have just turned on the sport mode to cancel the return to home and get rid of the googles and land manually. RTH 1639 feet? Really? . You got not DJI Go App in this phone, really? anyway, this is a prime example of what not to do when flying.

 

[ElGuapo]

I just dont see need for litchi, i can still set way point missions in go app i just got fly to do it ofcourse. Ide love to try the 20$ app i paid for obviously but im scared if sumn happend dji wouldnt honor my refresh cuz of it. Or possibility of . Sux dji couldnt just make go app better so peeps didnt feel the need

I have been using Autopilot extensively in my old Phantom 3 PRO / Phantom 4 without any issues and recently with the Mavic without complaints. I ran couple of complicated missions on the Phantoms and in both cases the missions completed with success. Not long ago I purchased Litchi because it was on sale, I do not like how the waypoint missions work, but that's a separate issue.

 

[ElGuapo]

Thank you for looking at that for me.....makes me feel a bit better knowing that the Mavic appeared to be ignoring my RTH commands, hopefully DJI will consider that a warranty issue. Shipping it today, fingers crossed.

Let us know how it goes with DJI.

 

[SteveT]

Still waiting on my mavic. Question is, in litchi, what would pushing the pause button do?

 

[f3honda4me]

Don't know how direct it has to be but it was 15:00 local time and the AC was facing west. A better answer would have to include Latitude, Longitude, heading, time of day, date, pitch and maybe a few more things

EDIT: I'm not familiar with the P4P bug. But, in this incident the Mavic's front sensor thought there was something in front of it. For a brief interval when it didn't think something was in front it stopped rising. Seems compelling that the incident was caused by the Mavic trying to avoid the (false) obstacle in front of it.

There was a bug where OA tried to go over an object that wasn't there, nor did the on screen indicate there was any OA alert. Only happened during RTH under specific circumstances. Was fixed in the last update.

 

[royalhghnss]

RTH 1639 feet? Really?

No, not really. Try reading the post again.

 

[BudWalker]

There was a bug where OA tried to go over an object that wasn't there, nor did the on screen indicate there was any OA alert. Only happened during RTH under specific circumstances. Was fixed in the last update.

I get it. If this incident occurred while facing east in the afternoon then it would be like the P4P bug. But, since it could've been the sun we don't know if it's like the P4P bug.

 

[Jr. Stafford]

Let us know how it goes with DJI.

It went like this: Patience with DJI Repair pays off!!

 

[Jr. Stafford]

Just an update to confirm that Obstacle Avoidance contributed to this incident. The value frontSensorDistance has been added to CsvView/DatCon. During the RTH frontSensorDistance shows that the Mavic thought there was an obstacle in front of it and kept climbing.
View attachment 4653

Looking closer at an interval where the frontSensor wasn't detecting an obstacle it can be seen that the Mavic stopped rising in that interval
View attachment 4654

If so, it was definitely the sun causing it. I was in the WIDE OPEN with a hayfield in front of the Mavic.....absolutely NOTHING else there....except for the sun which it should have been facing directly into.

 

[m0j0]

I would turn all of the sensors off when flying litchi. I wouldn't want a programmed sequence to be effected by a false obstacle.




Sent from my iPad using Tapatalk