Sudden MM descent into pond...

[FleetPhil]

Side question: I installed CSVView on my Mac but I can't see how to translate the TXT file to CSV - the 'TXTlogToCSVtool' program seems to be Windows only? Are there any other solutions?

 

[RogerDH]

You were rather quick there to get airborne, didn't have enough strong sats ... got the HP eventually, but on the other side of the house along side with the driveway.
The Mini were then later battered in harder winds on it's highest altitude above the trees, this produced alot of wind warnings & messages about max power load reached.

Messages:

View attachment 93613

It was an uncommanded descent in the end though ... but this time without the characteristic up pitch. I sincerely get a picture of a tired & exhausted Mini no longer wanting to fly
After the negative throttle at 559sec it continues to descent even though both positive throttle & elevator are applied. Would have been nice to see the DAT log also ...

View attachment 93614

Does the data rule out vortex ring state?

 

[slup]

Does the data rule out vortex ring state?

No ...

 

[John Lord]

Short flight in my garden with my 3 week old MM - no damage or issues on the previous 20 or so flights. On this flight I started a descent but when I tried to level out it failed to respond to left stick input until it hit my pond...

Details are in the attached TXT, but in summary:

9m 17s: I had 55% batteries holding at 31ft altitude and commanded a descent
9m 20s: Altitude 21ft and left stick back to neutral
9m 21s: Altitude 14ft and descending rapidly with stick still neutral
9m 22s: Left stick full up as I realise what is happening but MM still descending at ~8ft/sec
9m 23s: Altitude negative as the MM hits the pond

In hindsight I should be careful about descending over water, but I don't understand why there was no response to my input at 9m 22s... is there any other analysis I can do to try and understand why?

Thanks, Phil

Unfortunately DJI will most likely say it was Pilot error as that is what they say most of the time, even when it can be proven from the data and eyewitnesses that it was not. as you might have gathered I have a very low opinion of DJI Care Refresh and Customer Support and not without good reason. DJI produce great products. If only their Customer Support was up to the standard of their products.

 

[RogerDH]

Short flight in my garden with my 3 week old MM - no damage or issues on the previous 20 or so flights. On this flight I started a descent but when I tried to level out it failed to respond to left stick input until it hit my pond...

Details are in the attached TXT, but in summary:

9m 17s: I had 55% batteries holding at 31ft altitude and commanded a descent
9m 20s: Altitude 21ft and left stick back to neutral
9m 21s: Altitude 14ft and descending rapidly with stick still neutral
9m 22s: Left stick full up as I realise what is happening but MM still descending at ~8ft/sec
9m 23s: Altitude negative as the MM hits the pond

In hindsight I should be careful about descending over water, but I don't understand why there was no response to my input at 9m 22s... is there any other analysis I can do to try and understand why?

Thanks, Phil

When it was descending, was it wobbling? (Vortex ring state is also known as the Wobble of Death.)

 

[FleetPhil]

No - it looked very stable - just descending fairly quickly ignoring the control inputs

 

[BudWalker]

Side question: I installed CSVView on my Mac but I can't see how to translate the TXT file to CSV - the 'TXTlogToCSVtool' program seems to be Windows only? Are there any other solutions?

Use the converter at phantomhelp.com and then download the Verbose .csv


Next, submit that .csv to CsvView

 

[FleetPhil]

Got it! Thanks

 

[sar104]

I ask this in a spirit of pure ignorance and enquiry. If the aircraft manages to achieve initial takeoff and climb, then the props are presumably undeformed. So is the deformation something which occurs during flight, exacerbated by sustained high rotational speeds making the prop blades deform to a finer pitch due to air pressure against the lower surface from leading to trailing edge?

I'm thinking along the lines of minimising full throttle operations - could that mitigate or delay the onset of deformation?

Does the plastic the props are made of have any "memory"? I know nothing about plastics, do they fatigue like some metals? Would some sort of periodic replacement schedule?

I guess the initial deformation would be at the tips where the pressure is highest and the material the thinnest? Do the motors have enough headroom (sorry, don't know the technical term) in terms of RPM to allow, say, snipping a few mills off the prop ends?

As a personal anecdote, years ago I used to race Control Line model planes. We used to make our own CF props by using a commercial, inferior-materials, plastic prop to make a mold. Our shed-made props were turning well over 30,000 rpm in flight. It can't be too hard.

Mike

They are likely deformed for the entire flight. In the logs that have been posted of this behavior, one or more of the motors is typically close to maximum, but the props are more effective when they are not descending into disturbed air - i.e. when they are climbing or moving laterally. When they do start to descend they seem to lose lift.

 

[EyesWideShut]

They are likely deformed for the entire flight. In the logs that have been posted of this behavior, one or more of the motors is typically close to maximum, but the props are more effective when they are not descending into disturbed air - i.e. when they are climbing or moving laterally. When they do start to descend they seem to lose lift.

I like that point. I remember years ago when flying a big Mikrokopter Okto XL that with it's eight 12 inch props that moved a lot of air, that the descent was much more stable if I gave it some forward movement into cleaner air. I can imagine the mini being so much lighter, loosing some lift capabilities while descending thru it's own turbulence.

 

[David2111]

This is pretty much a carbon copy of my incident. I sent the drone back to DJI and posted my flight details on this forum. DJI said the drone was OK. Nice guys on this forum said I experienced an uncontrolled decent. DJI are now sending me a new drone under warranty. Since I can't trust the drone, I will be selling it and going back to my Mavic Air.

A golden rule for me is if I can't control something i have put in the air then I won't put it in the air.

DJI have a control problem with the MM.

 

[Wild Drone Pilot]

Short flight in my garden with my 3 week old MM - no damage or issues on the previous 20 or so flights. On this flight I started a descent but when I tried to level out it failed to respond to left stick input until it hit my pond...

Details are in the attached TXT, but in summary:

9m 17s: I had 55% batteries holding at 31ft altitude and commanded a descent
9m 20s: Altitude 21ft and left stick back to neutral
9m 21s: Altitude 14ft and descending rapidly with stick still neutral
9m 22s: Left stick full up as I realise what is happening but MM still descending at ~8ft/sec
9m 23s: Altitude negative as the MM hits the pond

In hindsight I should be careful about descending over water, but I don't understand why there was no response to my input at 9m 22s... is there any other analysis I can do to try and understand why?

Thanks, Phil

Wow! When will DJI fix this bug? I have a new one from yesterday almost crashed because of it!
The altitude on the app for the Mavic MINI was way off. at around 30 feet it said 162 Feet
And when landing at 3 feet it said 60 foot.

 

[RodPad]

Yes, I'm really wondering how this is going to turn out. If ends up being props, there are going say no problem found. And will they make a firmware change for a prop problem.

@Don K,
Yours doesn't sound like a prop problem, seem like a barometer problem?

Rod

 

[Zeke]

Because you have Java 32 bit and it needs 64 bit! Go to the java download page and manually and install Windows Offline (64-bit) , then retry.

Thanks!!

 

[Zeke]

Because you have Java 32 bit and it needs 64 bit! Go to the java download page and manually and install Windows Offline (64-bit) , then retry.

I can't win... it only reads .dat files? No CSV or txt?

 

[sar104]

I can't win... it only reads .dat files? No CSV or txt?

On a Mac it won't convert txt logs because that requires the Windows executable TXTlogToCSVtool. @BudWalker explained that, and the workaround, in post #27.

 

[Zeke]

On a Mac it won't convert txt logs because that requires the Windows executable TXTlogToCSVtool. @BudWalker explained that, and the workaround, in post #27.

It's not a Mac, it's windows, but I figured it out. Can I delete the DA files off the phone or do I need to leave them?

 

[sar104]

It's not a Mac, it's windows, but I figured it out. Can I delete the DA files off the phone or do I need to leave them?

That's entirely your choice.

 

[cybervet]

Is it possible the flexibility of the arms to play a role in the uncommanded descents? I mean if the arms are flexible that would make the propellers flex too and change the AOA.

 

[sar104]

Is it possible the flexibility of the arms to play a role in the uncommanded descents? I mean if the arms are flexible that would make the propellers flex too and change the AOA.

It's been demonstrated fairly conclusively that it's the props - the back ones, in particular, end up requiring far too high motor speeds to produce the necessary lift, and swapping them out significantly reduces the motor speeds.