Lost control over Mavic2 - RTH was fatal decision even at 55% batt

[tcmjhb]

Sent M2 up on my usual Litchi app to get footage of a vacant plot in the sparsely populated seaside village of Cape St Francis.
After getting the shots there was a minor transmission-loss-spike [for around a second] and the mavic would not allow for complete control again.
It behaved erratically and would allow only for reduced control and then when left uncontrolled [ie: fingers not toggling at control sticks]
it would insistantly begin a rapid climb and drift away from the home point.

A point here is that the flight log shows a near constant left stick up for altitude throughout the flight, however in reality the only time i gained
any altitude intentionally was briefly after take off. Thereafter at any opportunity it was left stick down to try get the craft near the ground again.
Things were happening fast especially the altitude problem and i switched to S-mode to get some strength and speed to fly it it home safely.

Ultimately this never happened.
The flight log will reveal all to those who can analyse such details but in short the drone was so unresponsive that even after pressing
the dreaded RTH emergency function it began a too slow RTH at approx 6kmh.
Being 2km + away from the home point i calcu-guessed it wasnt going to make that flight and headed for the car. My hope was to get as close
to the drone as possible in an attempt to gain better control through closer proximity.

However even once closer it was a panicky situation for me as every-time i let the controls go the drone just shot up and further away from me.



Eventually the battery drained and after battling continually to lower the craft, 0% batt arrived - it lost transmission completely at 63m - and either dropped or emergency landed.

DJI Flight Log Viewer - PhantomHelp.com

I have been searching the surrounding bushland since 22 Dec and am losing hope of a recovery. Even with the final lat/long co-ords in hand im uncertain as to how much drift the M2 may have undergone in those final seconds.

Any ideas or advice would be much appreciated in guiding my next step.

 

[sar104]

The aircraft climbed because you repeatedly applied up throttle - the climb and descent correlates just fine with throttle input:



The main problem here was the wind speed - at the height you flew at (500 m) it was 40 - 50 mph out of the ENE.


It would be useful if you could post the DJI txt log that Litchi also creates. You will find it in the DJI directory rather than the Litchi directory: DJI » com.aryuthere.visionplus » FlightRecord

In the meantime I'll see if I can narrow down the landing point.

 

[sar104]

The aircraft was reporting zero percent battery when it disconnected, at which point it was around 30 meters AGL and descending in autoland. There is no way of knowing whether it made it to the ground or shut down in mid-air, but either way it should have ended up somewhere on the 30 meter red line segment shown below:

 

[Camino Ken]

The aircraft was reporting zero percent battery when it disconnected, at which point it was around 30 meters AGL and descending in autoland. There is no way of knowing whether it made it to the ground or shut down in mid-air, but either way it should have ended up somewhere on the 30 meter red line segment shown below:

View attachment 88966

SAR - Even though I completely understand the analysis and calculations I am still amazed every time you predict the landing/crash point of a failed flight and history has proven it to be quite accurate.

 

[sar104]

SAR - Even though I completely understand the analysis and calculations I am still amazed every time you predict the landing/crash point of a failed flight and history has proven it to be quite accurate.

It can be fairly accurate, depending on the reliability of the data and the unknown variables. This was a relatively simple case, since the aircraft was in autoland. The uncertainty is just whether it made it to the ground before the battery expired.

 

[Prismatic]

The aircraft climbed because you repeatedly applied up throttle - the climb and descent correlates just fine with throttle input:

View attachment 88964

The main problem here was the wind speed - at the height you flew at (500 m) it was 40 - 50 mph out of the ENE.


It would be useful if you could post the DJI txt log that Litchi also creates. You will find it in the DJI directory rather than the Litchi directory: DJI » com.aryuthere.visionplus » FlightRecord

In the meantime I'll see if I can narrow down the landing point.

The OP claimed that he only applied down throttle, despite the log. Is it possible for the RC to fail in such a way that it is constantly issuing a command (like up throttle) with the sticks neutral?
I realize that wind was the chief problem, but the OP seems clear about his struggle with vertical control, which I believe is not affected (much) by wind.

 

[sar104]

The OP claimed that he only applied down throttle, despite the log. Is it possible for the RC to fail in such a way that it is constantly issuing a command (like up throttle) with the sticks neutral?
I realize that wind was the chief problem, but the OP seems clear about his struggle with vertical control, which I believe is not affected (much) by wind.

I guess it's possible. I've seen a couple of logs where the RC was sending spurious inputs, but they looked quite different. These look like real stick inputs to me.

 

[Deleted member 94047]

Could it be that the throttle stick was stuck in the up position?

 

[sar104]

Could it be that the throttle stick was stuck in the up position?

It doesn't look like it could have been stuck - the throttle input was varying over the full range during the flight.

 

[tcmjhb]

Thank you all for your generous donation of time to have read my post. In particular to sar104 for your inputs and efforts to narrow down the dropzone to around 30m for me, i will investigate the area thoroughly this afternoon. If the aircrafts battery did expire before it completed auto landing, it may still be in good shape as the bush is quite thick for a drop/fall landing.
In ref. to the vertical climbing i must reiterate that besides the altitude nessessary for takeoff and 1st min of flight and conterary to the flight log viewer , for the rest of this flight it was a firm left control stick down.
Worth noting, around 2 months ago, the controller began displaying the error "stick err" upon each booting up. This error has been resolved each time by momentarily facing the 2 control sticks inward and downward then releasing, rotating the small multi-function joystick located to the right controller sticks left, clockwise once and then powering down the RC. The procedure needs to be repeated sometimes thrice before the desired "connecting" stand-by screen is displayed [minus the very piercing error alarm].
Along this same timeline i also noted the distinct lack of finesse that usually accompanies the P-mode setting. It is often the case that my returning craft is in S-mode to hurridly race against draining battery and to land safely. Once directly above "Home" i go into a rapid descend and then switch to P-mode for an accurate and more know-what-im-doing looking landing. But P-mode has displayed the same difficulties associated with S-mode when attempting up close and personal manuvoures. Jerky and not-trustworthy looking flight control. To overcome this it has been as simple as sliding the button to T-mode to complete a flight.
My apologies because now in hindsight these may have been critical points i neglected to mention.
I think i have applied these work-arounds with the RC because i religiously [obsessively] fly everyday after work and before its too dark, and the idea of sending the craft away for analysis for possibly weeks has been frightening.
But the fact now is , what craft

 

[tcmjhb]

sar104 as per request

 

[new2mavic]

Sorry about your loss. After going through the .csv file of your flight, I came to the conclusion that the wind was simply too strong so there was nothing you could have done. When you applied full forward elevator stick in sport mode, the drone did respond by pitching forward ( nose down ) to about 30 degree which is the maximum pitch angle achieved by my M2P when flying against strong head wind in sport mode and full forward stick. This is the best the drone can do.

Checking the wind speed after taking the drone to the flying altitude can avoid flying away of this kind. I don't remember seeing it in Litchi but in the DJI GO 4 screen, there is an indication of the motor power as well as the pitch angle. In the example below, the pitch angle is at it's maximum limit and the power is quite close to 100% so you know that the drone is fighting against the wind.



In the early stage of your flight, the drone was actually flying along with the wind so it didn't need to work very hard. If you were using DJI GO, you should have seen a rather low pitch angle as indicated by the attitude ball in the phantomhelp player :



The power % figure should also be lower than normal. These two signs should be treated as a warning of strong wind

 

[Rockycat]

The aircraft was reporting zero percent battery when it disconnected, at which point it was around 30 meters AGL and descending in autoland. There is no way of knowing whether it made it to the ground or shut down in mid-air, but either way it should have ended up somewhere on the 30 meter red line segment shown below:

View attachment 88966

Your skills in analysis amaze me sar104.
outstanding!

 

[tcmjhb]

Attention sar104 - another kill for your belt. Well done Boss- your accuracy was within 20 metres. After 6 days the M2 has been found! Let it be a lesson toward a happy ending, persaverance and the incredible advice from such dedicated enthusiasts as those who have commented on this thread. So many thanks, the feeling is quite unreal at present

 

[Prismatic]

@tcmjhb , can you tell whether the drone landed, or fell from some altitude? Also, I'd have that RC serviced, or even replaced.

Finally, note the posts above about detecting high winds in flight, and managing the flight accordingly. With or without actual RC faults, that flight was very risky.

 

[sarniac]

Attention sar104 - another kill for your belt. Well done Boss- your accuracy was within 20 metres. After 6 days the M2 has been found! Let it be a lesson toward a happy ending, persaverance and the incredible advice from such dedicated enthusiasts as those who have commented on this thread. So many thanks, the feeling is quite unreal at present

Wonderful news - happy for you!

I've had to hunt for my Mavic Pro twice using Find My Drone and I found it both times.

Please update on the condition of your Mavic2 once you've had a chance to inspect.

 

[Deleted member 94047]

Yet another rescue by sar!

 

[sar104]

Attention sar104 - another kill for your belt. Well done Boss- your accuracy was within 20 metres. After 6 days the M2 has been found! Let it be a lesson toward a happy ending, persaverance and the incredible advice from such dedicated enthusiasts as those who have commented on this thread. So many thanks, the feeling is quite unreal at present

That's great news. What kind of shape is it in?

If the recorded throttle inputs (black trace) in this graph are definitely not a correct reflection of the stick position then you have a bad RC and will need to replace it.

 

[Gindra]

Great News! sar104 rocks! I'm constantly in awe!

 

[offtheback]

Great News! sar104 rocks! I'm constantly in awe!

sar104 remains true to his handle letters once again!?