How to stop a "run away" when Mavic Pro is in your hand?

[GFields]

thanks for sharing! Thankfully no one was seriously injured. You could have had a lawsuit to worry about.

 

[tjcooper]

Thanks again for quick interest in this event. I need to emphasize some points for Meta4 so that wrong conclusions are not drawn. Here in silicon valley of california I typically need 8 satellites to get GPS to say it has locked in. Only been able to get GPS once with 6 satellites found. I took off from a empty table in ALLI mode and did not ask for precision takeoff so my RTH would have been my starting position on that table. When the "fly away" event occurred, the RTH position was BEHIND me, and the AC took off LEVEL and DIRECTLY to the LEFT of my current position. So there is no way it was doing a RTH move. Besides, I have my RTH setup to pop up to 50 feet before going home. I was under tree canopy at about 20 feet and I knew that if I had asked for RTH I would have crashed into the canopy before moving laterally.

Since I took roughly 7 minutes of VIDEO with everything working, I most likely had GPS mode lock in during that time period. Sun was too bright to see the cell phone screen so I don't know if it found GPS or not. But high probability that it did get GPS lock. I was in standard mode with no special conditions turned on. During my 7 minute flight someone asked me if the drone would fly into someone. I backed off the AC and flew it directly into my face at 5 feet elevation. The collision avoidance kicked in and stopped it at 3 feet and made its usual noise. So I have a strong belief that the system was working properly.

During the whole flight, I only hit the two control rods and the video on button. I did not get close to the RTH button or other functions. Again, if I had hit the RTH button, it would have popped up and crashed into the tree canopy. When the crash happened and I grabbed the AC, I immediately put throttle to zero, I know this does not stop the rotors, but it cuts their rotation to a couple hundred rpm. There was absolutely no decrease in rotor speed the whole time it was in my hand until I finally removed the battery. I am sure that during the panic mode trying to shut down the AC that I might have hit every button on the controller. But there was no change in the AC rotors while this happened. What really bothers me is that cycling the AC battery via the battery button had no change in rotor speed.

So I am starting to download the log files and want to check out what they say. Which log file is the most important to review?

 

[Gagey52]

Mate at the end of the day it was a learning experience for you. I hope no one was hurt and the drone is OK ?. Don’t listen to that self aware malarkey that’s why they call them drones. You have a good day.
Regards

 

[Meta4]

Here in silicon valley of california I typically need 8 satellites to get GPS to say it has locked in.

There is no magic number, sometimes it might be 6 or 8 or 10 or even more.
The numbers don't matter, having the drone show you that it is getting good location data is what matters.

I took off from a empty table in ALLI mode and did not ask for precision takeoff so my RTH would have been my starting point;

It's Atti Mode.
And launching without waiting for GPS, means that your launch point won't be the location the drone returns to on RTH, because it cannot record a home point until it gets good location data.
Your drone would have recorded a home point later in the flight, when/if it got good location data later in the flight.
And that home point would be nowhere near your launch point, unless you were hovering above it when the drone got GPS.

When the "fly away" event occurred, the RTH position was BEHIND me, and the AC took off LEVEL and DIRECTLY to the LEFT of my current position. So there is no way it was doing a RTH move.

As explained above, you have no idea where the home point was or if there was a homepoint.
Post the data and it will show what happened.
That's going to achieve a lot more than guessing without any facts.

So I am starting to download the log files and want to check out what they say. Which log file is the most important to review?

It should only take a couple of minutes to find that .txt file.
Details were given back in post #2

 

[Meta4]

I did not get close to the RTH button or other functions. Again, if I had hit the RTH button, it would have popped up and crashed into the tree canopy.

From post #14: when finished flying, I hit the RTH

I am sure that during the panic mode trying to shut down the AC that I might have hit every button on the controller.

Possibly including RTH?
The flyer's memory of events in an incident are often sketchy.
Details of everything that happened and the order in which they happened can vary from what actually happened.
That's why it's important to look at the recorded flight data to cut through the confusion and get to the facts.

 

[tjcooper]

I am having problems getting the DJI flightlog files for the "fly away" flight. The files on my cell phone are not in the same directories as the "Locating Your Flight Logs" document describes. But I do find the .txt files in similar directories on my cell phone:
This PC\H710VL\Internal storage\DJI\dji.pilot.pad\FlightRecord

The last .txt file I have is:

DJIFlightRecord_2021-05-29_[19-50-11].txt
which is for a test flight at the local park a few days earlier. Using CsvView I can see the flight path and all the parameters for that flight. I fear that when I grabbed the AC out of the air and fought to stop the rotors from turning, I did the "shutdown" action with the controller power button while I struggled to gain control I fear this might have caused the controller not to record the .txt file into the cell phone internal memory. I don't have any .txt or .dat files on the cell phone for the date of the "fly away" event (23 June 2021).
Are these same files recorded on the AC memory stick? Can I get those files from the drone itself? That would be a great help if I can find them on either the internal memory or my SD card on the AC. Should they have the same name(s) as what one would find on the cell phone memory?

 

[Meta4]

I fear that when I grabbed the AC out of the air and fought to stop the rotors from turning, I did the "shutdown" action with the controller power button while I struggled to gain control I fear this might have caused the controller not to record the .txt file into the cell phone internal memory.

Pulling the battery cannot prevented the flight record from being recorded.
If a flight is cut short, the flight record is still present and ends when the drone was lost, signal was lost or the drone crashes etc.

I don't have any .txt or .dat files on the cell phone for the date of the "fly away" event (23 June 2021).

You do but haven't found them.
When/if you do find and post the appropriate file, you'll get some help to understand what happened to your drone.

 

[Meta4]

The files on my cell phone are not in the same directories as the "Locating Your Flight Logs" document describes. But I do find the .txt files in similar directories on my cell phone:
This PC\H710VL\Internal storage\DJI\dji.pilot.pad\FlightRecord

The directory structure you should be looking for is:
DJI\dji.go.v4\FlightRecord

Check the flight records page in DJI Go 4 and see what flights it has recorded like this:

If you see the flight in question listed there, the .txt file is in there.

 

[kidroc]

..
Even turning it 90 degrees into a sideways position shuts it down.
Immediately.

After 300hrs I find it hard to imagine you've never known or tried this ?
I use it sometimes when hand catching, much easier / safer, negates the drones attempt to pull away while fiddling with controller left stick.
There are other things you can do (if it's in a stable hover), by slipping your hand in close from the front, it won't try and lift up, and you can use the landing command to land it on your palm.

After 300hrs safe flying, 7 mins can make you and drone ops look pretty bad in front of a crowd.
No mention of the party the drone hit, hopefully it just 'bounced off' his backside without injury, but bet it gave a few people there a fright and a half.

Not trying to be rude or anything but 300 hours you definitely definitely should know how this thing operates and what you did caused a problem not the drone please stay away from people I’ve been in this hobby for almost 30 years I kind of wish drones never came out. So much easier when only a handful of people could fly these things

 

[MAvic_South_Oz]

Not trying to be rude or anything but 300 hours you definitely definitely should know how this thing operates and what you did caused a problem not the drone please stay away from people I’ve been in this hobby for almost 30 years I kind of wish drones never came out. So much easier when only a handful of people could fly these things

Why are you replying to me kidroc ?
You do realise I'm not the OP ?

 

[BudWalker]

I am having problems getting the DJI flightlog files for the "fly away" flight. The files on my cell phone are not in the same directories as the "Locating Your Flight Logs" document describes. But I do find the .txt files in similar directories on my cell phone:
This PC\H710VL\Internal storage\DJI\dji.pilot.pad\FlightRecord

The last .txt file I have is:

DJIFlightRecord_2021-05-29_[19-50-11].txt
which is for a test flight at the local park a few days earlier. Using CsvView I can see the flight path and all the parameters for that flight. I fear that when I grabbed the AC out of the air and fought to stop the rotors from turning, I did the "shutdown" action with the controller power button while I struggled to gain control I fear this might have caused the controller not to record the .txt file into the cell phone internal memory. I don't have any .txt or .dat files on the cell phone for the date of the "fly away" event (23 June 2021).
Are these same files recorded on the AC memory stick? Can I get those files from the drone itself? That would be a great help if I can find them on either the internal memory or my SD card on the AC. Should they have the same name(s) as what one would find on the cell phone memory?

The dates on the .txt and .DAT log files can be off by a day. So, a .txt following DJIFlightRecord_2021-05-29_[19-50-11].txt could be the correct one. Same is true for the .DAT file.

If you want the on board .DAT take a look at these instructions.
How to retrieve a V3 .DAT File from the AC

This can be a bit tricky depending on the FW version. You might get lucky and the MP's internal drive will mount on the PC without having to use DJI Assistant 2.

From reading your description I would guess that the incident occurred when the MP switched from ATTI to GPS mode. This can happen even though the FC's location solution isn't quite good enough. As a result the FC decides the MP needs to move from it's present location. A .txt or .DAT log file will determine if this is what happened.

 

[BossBob]

Never really a good idea to take off on a rushed or unplanned flight. Preflight checks were not done so appropriate loss of signal settings not set, appropriate RTH height not set, and gps position not acquired before flight so the poor drone gets the blame for the pilots shortcomings. From the moment you start the propellers the drone will try to keep its position in 3 dimensions unless you command it to do something. That is why the motors spooled up. Leaving the sticks at neutral will not stop this from happening.
Glad nobody was hurt.

 

[mmoguls]

This falls in the category of know your drone, and what to do when things go wrong. If you don't have time to wait for satellite lock, then arrive sooner, or don't fly. Period. Safety with a group is paramount. What you saw is typical no-lock moving to GPS lock behavior. If you hand catch your Mavic 1, you can turn it over to shut off.

Glad you and no one else was hurt during this error flight. Learn from it and proceed with more care.

 

[mmoguls]

The original post refers to a mavic pro. This certainly does work on that particular drone.

 

[tjcooper]

There is getting to be a great deal of contradictory infomation in this thread. I have many times done rolls and knife edges with Mavio Pro and the AC does not shut off. If you hold it will little or no momentrum going forward it sometimes shuts off. I think there is a time constant that must be exceeded with certain accelerometer readings before shut down occurs. Anyway, this forum has many cases documented where turning Mavic Pro sideways or upside down does not shut off unit under certain circumstances.

What we all need to see are the .txt and .dat files for my fly away incident on 23 June 2021. The file structure that has been listed as correct for Mavic Pro version 4 does not match my system. I am providing screen dumps of my Huawei cell phone with NLD app on it. There is a Go.dji.v4 directory under DJI but it has no .txt or .dat files and most folders are EMPTY. What I do have is a directory under DJI called DJI_pilot_pad. This specific directory DOES HAVE .TXT and .DAT files that match previous flights used on this cell phone. Sadly I have no .txt or .dat for the 23 June 2021 date I mentioned. I have used CsvView on the .txt files in June and the video records and Google Map plots all show the park next to my house which is where the flights took place on the earlier dates.

I have used the great app called PhoneRecover to search for erased or hidden files on my cell phone. Shows the .txt and .dat I have mentioned but no erased or hiddlen files belonging to 23 June 2023. So that fact that cannot be contradicted is that my "fly away" flight did not get recorded on the cell phone. I am using the file locations mentioned by BudWalker to see if I have any records of that flight on the AC internal memory. Will keep you updated on what I find.

I will not fly my Mavic Pro until I have completely exhausted searching for the .txt and .dat files. There are threads which prove that "fly away" drones do happen and have hardware related problems that DJI confirmed when unit was returned for repair. DJI said not to attempt to fly unit after possible "fly away" because certain critical flies may get altered in the process.

 

[Kilrah]

I have many times done rolls and knife edges with Mavio Pro and the AC does not shut off.

You cannot roll or knife edge a Mavic Pro in the first place, so color me doubtful.

 

[tjcooper]

Forgot to mention that I was NOT in ALLImode (I know it is ATTimode for DJI but the "mission planner" for other drones uses ALLImode for Altitude and Level flight when GPS is not available). I was reading the user manual last night looking for when the HomePosition is set. It indicated that you are in P-mode (not sport mode) if you have visual collision avoidance operational. As I mentioned in earlier post, I was asked by members of picnic party if the drone was "safe" to fly around people. I took drone from 8 feet down to 5 feet and flew it straight into my face. The collision avoidance took over and stopped the drone. So according to the user manual I was in P-mode which is different from ATTimode. It is highly likely that because of the 7 minutes I had flown the AC since takeoff, I was most likely in GPS mode when the "fly away" finally happened. I am trying to find DJI documents that define where HomePosition is when you lift off in ATTimode and then go into P mode and most likely into GPS mode after that.

 

[tjcooper]

Kilrah,
you have never "looped" your Mavic Pro? And if you go up high enough and turn AC sideways, you will fall through the air for a few seconds in what airplane pilots call the knife edge position. Let me see if I can find some video for you. This are very quick maneuvers that only last for a second or two.

 

[Kilrah]

So according to the user manual I was in P-mode which is different from ATTimode.

But the AC can switch back to ATTI at ANY time if GPS reception becomes too poor.

I am trying to find DJI documents that define where HomePosition is when you lift off in ATTimode and then go into P mode and most likely into GPS mode after that.

It will be whenever the aircraft first gets a good enough position fix since powered up, regardless of mode.

you have never "looped" your Mavic Pro? And if you go up high enough and turn AC sideways, you will fall through the air for a few seconds in what airplane pilots call the knife edge position.

No, it does not let you do that.

I took drone from 8 feet down to 5 feet and flew it straight into my face. The collision avoidance took over and stopped the drone.

This has to be trolling at this point...

 

[Meta4]

What we all need to see are the .txt and .dat files for my fly away incident on 23 June 2021. The file structure that has been listed as correct for Mavic Pro version 4 does not match my system. I am providing screen dumps of my Huawei cell phone with NLD app on it.

It's hard to believe this is still going.
All that's needed is the .txt flight record and it usually takes less than two minutes to find.
Unfortunately your screenshots are too small to read the txt so I cannot see what's there.
It would help if you posted those full size instead of tiny thumbnail images.

** Important**
It would also help if you could answer the question I asked back in post #28.
Do you see the flight listed there in the listing of flights?
If you do, then the .txt file is in there.

There is a Go.dji.v4 directory under DJI but it has no .txt or .dat files and most folders are EMPTY ... What I do have is a directory under DJI called DJI_pilot_pad.

I think the pilot-pad directory you mentioned would be from the DJI Pilot app.
That's an app that does not work with your Mavic.
Which app do you fly with ... DJI Go4 or DJI Pilot?
I suspect you are looking in the wrong place but as I cannot read your tiny screenshots, I can't help you further.

So that fact that cannot be contradicted is that my "fly away" flight did not get recorded on the cell phone

It would be very, very unusual for that to happen.
See the question above that I marked Important.

I will not fly my Mavic Pro until I have completely exhausted searching for the .txt and .dat files.

All that's needed is the .txt file and it doesn't take days to find it.

There are threads which prove that "fly away" drones do happen and have hardware related problems that DJI confirmed when unit was returned for repair. DJI said not to attempt to fly unit after possible "fly away" because certain critical flies may get altered in the process.

This is irrelevant.
Your drone is fine but aspects of the way you flew it are the problem.
If you were to fly it, you won't alter critical files either.

From analysing data from hundreds of flight incidents, I know that drones don't "fly away".
The explanation for your incident is almost certainly very simple, involving you fighting against the drone trying to RTH or hold position.

Now ... can you answer the question and post some readable screenshots.
If you can't do that or find the file yourself, there's nothing anyone can do for you.