Phantom 4 motors randomly turned off - bug or feature?

[Dingoz]

Phantom 4 Advanced. Today I had the drone on a table inside with props on and motors spinning (just idling). Controller and GO4 app all connected as well at same time.

The avoidance sensors were beeping at me as about 1m in front of the drone was an object. The sound was annoying me. So, in the Go4 app I turned off the sensors. All good the beeping stopped.

After a while I decided to turn the sensors back on. It wouldn't. I set the sensor slider to green (on) and it looked like it was on but once I closed the screen and went back in to check it was deselected again. So I repeated this a few times trying to turn on the sensors again. Same thing happened over and over (looked on but back out and back in - it was still off).

It is at this point and randomly I got a warning message (now in atti mode) and then all of a sudden the motors turned off and props stopped spinning. The drone had shut the motors off without input from controller sticks or Go4 app. Incidentally, I could now turn on the sensors again properly (the setting stuck this time while motors were stopped).

At the time I just thought that's really odd and like some bug perhaps. Then a scary thought made me consider this some more... is there any 'feature' in the phantom 4's that could possibly automatically and without input from sticks or Go4 app (auto land for eg), turn off the motors while spinning/ideling (like a timeout feature if motors started but not taken off after a certain time its auto stops the motors). Alternatively what caused the motors to randomly shut down without being instructed to. BTW battery was at 60% and cells were perfectly aligned at the time (new battery).

The scary part if this can not be explained leads me to think software bug in Go4 app caused motors shut down while spinning (what if it was in the air at the time).

I am going to try and repeat the steps that led to this just out of curiosity but hoping someone can explain a simple cause.

 

[Dingoz]

Tried to repeat the steps that led to this without success. This time I could turn the main sensor switch on and off while props were spinning repeatedly and the setting stuck. Not sure what else was different other than perhaps number of sats locked at any given time. Todays tests had full gps lock whereas yesterdays random motor shutdown I think did not have full lock.

In any event, at this stage i must assume that it might be possible for something to glitch/bug out (Go4 app most likely from the initial incident) that could cause the motors to instantly shut down when they shouldn't. If in the air, well you know what happens next All I can learn from this is if in flight and a button or function appears to not take or work properly in the Go4 app then do not keep trying. Land and restart everything!!!

It does not surprise me that tech of this nature app+controller+drone combo could bug out. What surprises me is such a bug out would somehow instruct the drone to shut off the motors. I would have thought very little if anything in code would shut motors off (other than stick command or app land command or faulty esc/board perhaps - but this is not the case - just a random app bug at the time and boom - motors off.

 

[Deleted member 62848]

Phantom 4 Advanced. Today I had the drone on a table inside with props on and motors spinning (just idling). Controller and GO4 app all connected as well at same time.

The avoidance sensors were beeping at me as about 1m in front of the drone was an object. The sound was annoying me. So, in the Go4 app I turned off the sensors. All good the beeping stopped.

After a while I decided to turn the sensors back on. It wouldn't. I set the sensor slider to green (on) and it looked like it was on but once I closed the screen and went back in to check it was deselected again. So I repeated this a few times trying to turn on the sensors again. Same thing happened over and over (looked on but back out and back in - it was still off).

It is at this point and randomly I got a warning message (now in atti mode) and then all of a sudden the motors turned off and props stopped spinning. The drone had shut the motors off without input from controller sticks or Go4 app. Incidentally, I could now turn on the sensors again properly (the setting stuck this time while motors were stopped).

At the time I just thought that's really odd and like some bug perhaps. Then a scary thought made me consider this some more... is there any 'feature' in the phantom 4's that could possibly automatically and without input from sticks or Go4 app (auto land for eg), turn off the motors while spinning/ideling (like a timeout feature if motors started but not taken off after a certain time its auto stops the motors). Alternatively what caused the motors to randomly shut down without being instructed to. BTW battery was at 60% and cells were perfectly aligned at the time (new battery).

The scary part if this can not be explained leads me to think software bug in Go4 app caused motors shut down while spinning (what if it was in the air at the time).

I am going to try and repeat the steps that led to this just out of curiosity but hoping someone can explain a simple cause.

Yes good read,i have had it happen 3 times before a early morning flight,started motors went to take off and it shut down,had to restart 3 times but keept goin after that(thankfully),gave no warnings nothing motors just stop,was pretty cold so put it down the temp,not done it again since but worried for a while,not something someone would want mid flight...

 

[Dingoz]

Yes good read,i have had it happen 3 times before a early morning flight,started motors went to take off and it shut down,had to restart 3 times but keept goin after that(thankfully),gave no warnings nothing motors just stop,was pretty cold so put it down the temp,not done it again since but worried for a while,not something someone would want mid flight...

Hopefully in your case it was cold temp (at least something to go by, maybe). In my case it was middle of the day and not hot nor cold... just right .

 

[Dingoz]

After my weird event I purged the Go4 app flight records for all non actual real flights (test on desk etc) so lost in app .txt flight log. I did copy all I could find .dat wise from bird/controller/phone. I found a some .dat files in a folder named flight logs and opened the one that related to the incident in notepad.Below is the contents:

"2017-09-22 13:21:52" "Flight Mode Set to Atti" "Set flight mode to GPS to ensure safety."
"2017-09-22 13:23:47" "Obstacle Avoidance Disabled. Fly with caution."
"2017-09-22 13:23:51" "Obstacle Avoidance Disabled. Fly with caution."
*** It would be in this section that I was trying to turn back on flight sensors - and the setting would not take/stick on ***
*** Then all of a sudden the motor just turn off ***
"2017-09-22 13:30:26" "Tips" "Start Auto Landing"

Is this last entry "Start Auto Landing" what happened. The bird entered auto landing mode (might explain why motors shut off) as it was on a table not moving at all. What is not explained still if this is the case is what triggered auto landing to start (it was absolutely not me within the phone app as I know 100% I was going in and out of sensor settings at the time and never pressed or slid to start the auto landing feature.

In the event it was in flight, at least if this is the case, it was trying to land not just dropping out of the sky. In atti mode it would be trying to land wherever it was at the time. Still I did not trigger it manually and there was no onscreen "landing" pop ups or info at all. Odd but a little less worry some as it was trying to land So now I am only left with the worry that a bug in app could cause the bird to randomly land in the ocean.

 

[BudWalker]

@Dingoz , @ianzone it's possible that the .DAT log file on the P4 will have some data that explains these incidents. To see how to retrieve the .DAT look here

How to retrieve a V3 .DAT File

It will likely be too large to post here so you'll need to use Dropbox or GoogleDrive and post a link.

 

[Dingoz]

@Dingoz , @ianzone it's possible that the .DAT log file on the P4 will have some data that explains these incidents. To see how to retrieve the .DAT look here

How to retrieve a V3 .DAT File

It will likely be too large to post here so you'll need to use Dropbox or GoogleDrive and post a link.

Thanks. I have retrieved all of the dat files and converted them to csv. Can upload them if someone else is curious.

It certainly looks like an auto landing was triggered/performed (see below section - no stick input and motors were shut down). And unless I am mistaken this is a concern with the DJI logs as finding out what triggered auto-landing in the app is impossible? Does DJI log keystrokes/button presses within the app. Does any log state "auto landing command received from GO4". Does any log state "app error". Or is there any scenario where auto landing will just kick in on its own (at 60% battery left).

RC:Aileron RC:Elevator RC:Rudder RC:Throttle Motor:Speed:RFront Motor:Speed:LFront Motor:Speed:LBack Motor:Speed:RBack
0 0 0 0 1512 1404 1454 1445
0 0 0 0 1369 1297 1304 1333
0 0 0 0 1369 1297 1304 1333
0 0 0 0 1369 1297 1304 1333
0 0 0 0 1094 1024 1036 1018
0 0 0 0 1094 1024 1036 1018
0 0 0 0 1094 1024 1036 1018
0 0 0 0 924 779 803 791
0 0 0 0 924 779 803 791
0 0 0 0 924 779 803 791
0 0 0 0 156 86 82 0
0 0 0 0 156 86 82 0
0 0 0 0 156 86 82 0
0 0 0 0 49 0 0 37
0 0 0 0 49 0 0 37
0 0 0 0 49 0 0 37
0 0 0 0 0 0 0 0

 

[sar104]

Thanks. I have retrieved all of the dat files and converted them to csv. Can upload them if someone else is curious.

It certainly looks like an auto landing was triggered/performed (see below section - no stick input and motors were shut down). And unless I am mistaken this is a concern with the DJI logs as finding out what triggered auto-landing in the app is impossible? Does DJI log keystrokes/button presses within the app. Does any log state "auto landing command received from GO4". Does any log state "app error". Or is there any scenario where auto landing will just kick in on its own (at 60% battery left).

RC:Aileron RC:Elevator RC:Rudder RC:Throttle Motor:Speed:RFront Motor:Speed:LFront Motor:Speed:LBack Motor:Speed:RBack
0 0 0 0 1512 1404 1454 1445
0 0 0 0 1369 1297 1304 1333
0 0 0 0 1369 1297 1304 1333
0 0 0 0 1369 1297 1304 1333
0 0 0 0 1094 1024 1036 1018
0 0 0 0 1094 1024 1036 1018
0 0 0 0 1094 1024 1036 1018
0 0 0 0 924 779 803 791
0 0 0 0 924 779 803 791
0 0 0 0 924 779 803 791
0 0 0 0 156 86 82 0
0 0 0 0 156 86 82 0
0 0 0 0 156 86 82 0
0 0 0 0 49 0 0 37
0 0 0 0 49 0 0 37
0 0 0 0 49 0 0 37
0 0 0 0 0 0 0 0

Would need to see the full log, not just that small excerpt.

 

[Dingoz]

Would need to see the full log, not just that small excerpt.

[DELETED]

 

[BudWalker]

Dropbox - 2017-09-22_13-20-10_FLY035.csv

This is from the tablet. Can you provide the .DAT from the AC itself? The instructions are here

How to retrieve a V3 .DAT File

It would be best if you could provide the .DAT instead of the .csv. I'll also want to look at the eventLog stream that can be obtained with by directing DatCon to produce it.

 

[sar104]

This is from the tablet. Can you provide the .DAT from the AC itself? The instructions are here

How to retrieve a V3 .DAT File

It would be best if you could provide the .DAT instead of the .csv. I'll also want to look at the eventLog stream that can be obtained with by directing DatCon to produce it.

Looks like that is actually the DAT file, but converted to csv with DatCon.

 

[BudWalker]

Looks like that is actually the DAT file, but converted to csv with DatCon.

Yes, that's a .DAT file. But, it's the .DAT file from the tablet, not the P4 itself.

 

[sar104]

Yes, that's a .DAT file. But, it's the .DAT file from the tablet, not the P4 itself.

Right - I keep forgetting that the mobile device also has DAT files and that's obviously why it's missing a number of data fields that I expected to see. There's nothing in it that explains the event anyway - we do need the full DAT file and the event log.

 

[Dingoz]

Yes, that's a .DAT file. But, it's the .DAT file from the tablet, not the P4 itself.

Ok just trying to work out which dat file from the RC assistant is the one that matches the time I was testing and this event occurred. Its a bit odd. In RC assistant I see the logs have date/time and assume the date/time closest matching the other dat file. The strange thing is I do not have a dat file for the real flight I did on the same day before testing. Confused. Eg:

2017-09-22 09:28 - 20 minute real flight that morning (no dat file in RC assistant matching this date/time or near about)
2017-09-22 09:34 - 20 minute real flight that morning (no dat file in RC assistant matching this date/time or near about)
2017-09-22 13-27 - I think this is the one (test at desk after real flight that morning but dat files for 22/09/17 start with time at 13-27)
2017-09-22 13-50 - Another test with a different battery shortly after

 

[Dingoz]

Before I go troubling people to look at the logs, I had a look as best I can (and what I think is the right log) and does this make sense:

- Idling on desk with motors spinning (no gps lock so in atti mode - no home point set)
- RC disconnect* (not sure what caused this but I suspect bug not actual loss 1m away from drone)
- Auto Landing (triggered by drone - no GPS and in atti mode - is that normal behaviour. Is it supposed to land or hover in this scenario.. If hover would it then have triggered auto land / motors off as drone was not actually in the air and no movement due to being on desk.
- Motors Off.

The RC Disconnect seems like the culprit to the rest of the events. But what triggered it I do not know. Perhaps it did lose signal 1m away from drone on desk. At the time (and it appears before RC disconnect log) I was going in and out of the sensor setting trying to turn obstacle sensors back on and the setting was not taking/sticking. Then rc disconnect - auto land/motors off.

Dropbox - Fly35.xlsx
Dropbox - Fly35.csv
Dropbox - FLY035 Event Log.txt

 

[RémyHeli]

Dat file might not have current date/hour if you haven't powered on RC before aircraft

 

[BudWalker]

Before I go troubling people to look at the logs, I had a look as best I can (and what I think is the right log) and does this make sense:

- Idling on desk with motors spinning (no gps lock so in atti mode - no home point set)
- RC disconnect* (not sure what caused this but I suspect bug not actual loss 1m away from drone)
- Auto Landing (triggered by drone - no GPS and in atti mode - is that normal behaviour. Is it supposed to land or hover in this scenario.. If hover would it then have triggered auto land / motors off as drone was not actually in the air and no movement due to being on desk.
- Motors Off.

The RC Disconnect seems like the culprit to the rest of the events. But what triggered it I do not know. Perhaps it did lose signal 1m away from drone on desk. At the time (and it appears before RC disconnect log) I was going in and out of the sensor setting trying to turn obstacle sensors back on and the setting was not taking/sticking. Then rc disconnect - auto land/motors off.

Dropbox - Fly35.xlsx
Dropbox - Fly35.csv
Dropbox - FLY035 Event Log.txt

Checking the RC:sigStrength shows the signal was good when the disconnect occurred. Likewise, RC:appLost shows AC still talking to the app.

 

[RémyHeli]

I wonder if there is some sort of time limit in iddle mode after power on and no fly, wathever blades mounted or not.

In test scenario I would include watching elasped time from iddle start point.

 

[sar104]

Before I go troubling people to look at the logs, I had a look as best I can (and what I think is the right log) and does this make sense:

- Idling on desk with motors spinning (no gps lock so in atti mode - no home point set)
- RC disconnect* (not sure what caused this but I suspect bug not actual loss 1m away from drone)
- Auto Landing (triggered by drone - no GPS and in atti mode - is that normal behaviour. Is it supposed to land or hover in this scenario.. If hover would it then have triggered auto land / motors off as drone was not actually in the air and no movement due to being on desk.
- Motors Off.

The RC Disconnect seems like the culprit to the rest of the events. But what triggered it I do not know. Perhaps it did lose signal 1m away from drone on desk. At the time (and it appears before RC disconnect log) I was going in and out of the sensor setting trying to turn obstacle sensors back on and the setting was not taking/sticking. Then rc disconnect - auto land/motors off.

Dropbox - Fly35.xlsx
Dropbox - Fly35.csv
Dropbox - FLY035 Event Log.txt

That seems to be correct. The reported RC disconnect caused the shutdown. Disconnect was at 619.33 s, even though the RC signal was reporting full strength. After that:

630.413 : 2849703971 : 31407 [L-FLYMODE][Ctrl<6>] REQ_RC_LOST AUTO_LANDING ctrl_auto_landing
630.453 : 2849881893 : 31409 [L-FMU/MOTOR] Stop. reason:ACT.landing

At around 630.55 the motor command values went to zero and 0.7 s later the motors were stopped.

 

[Dingoz]

I wonder if there is some sort of time limit in iddle mode after power on and no fly, wathever blades mounted or not.

In test scenario I would include watching elasped time from iddle start point.

Not sure but I do not think this is correct. This disconnect / shut down happened not long into starting up the motors. After this I changed batteries and did the same thing and it ran for much longer than the first episode and never shut down..