Unexpected loss in altitude

atsphenomenon · Jan 4, 2020

So my mm also has finally shown the auto altitude loss on its 70th flight. Can someone please help me understand why this happened.
After losing signal and autotriggering rth, the ac decided mid flight to start dropping altitude - look at the journey thru the checkpoints K/L/M/N. I hastily started pushing the thrust up and realised quickly that i should cancel rth first. Did that and was then able to get it to go up again. Rth altitude was set higher so that ain't it. There was no notification when it started losing altitude either. It's only after i cancel rth and try to pull it up when it shows the not enough power error.
1) why did it decide to start losing altitude
2) why the esc/not enough power errors on pulling it up (wind velocity was negligible - 6mph.

atsphenomenon · Jan 4, 2020

Please see from 12:00 for the next 30-40 seconds when the ac loses altitude from 200+ ft to almost 60 ft on rth.

Fly Dawg · Jan 4, 2020

@atsphenomenon There were 2 instances of RTH in this flight. The first one, you commanded the descent. The last one you did not. I don't see a reason for that from this data. Can you supply the mobile .dat file? See this link for instructions:
How to retrieve a V3.DAT from the tablet

Below is only the second RTH. It shows the descent, but unknown as to why. Only after cancelling RTH did you regain control.

atsphenomenon · Jan 4, 2020

Fly Dawg said:
@atsphenomenon There were 2 instances of RTH in this flight. The first one, you commanded the descent. The last one you did not. I don't see a reason for that from this data. Can you supply the mobile .dat file? See this link for instructions:
How to retrieve a V3.DAT from the tablet

Below is only the second RTH. It shows the descent, but unknown as to why. Only after cancelling RTH did you regain control.

View attachment 89723

Thanks fly dawg. You're right, I'm referring to the second rth, especially the flight behaviour from 11:30 onwards.

I'm on Android and the folder DJI>DJI.go.v5>FlightRecord>MCDatFlightRecord is empty. What am I missing?

Fly Dawg · Jan 4, 2020

atsphenomenon said:
I'm on Android and the folder DJI>DJI.go.v5>FlightRecord>MCDatFlightRecordspty. What am I missing?

It should be in the folder DJI>DJI.go.v5>FlightRecord and you are looking for FLY087.DAT

atsphenomenon · Jan 4, 2020

Fly Dawg said:
It should be in the folder DJI>DJI.go.v5>FlightRecord and you are looking for FLY087.DAT

That folder is completely empty.

atsphenomenon · Jan 4, 2020

And the folder before that doesn't show any dat files. I'm sure I'm missing something, can't figure out what

Fly Dawg · Jan 4, 2020

atsphenomenon said:
That folder is completely empty.

There have been a few instances when the .dat wasn't recorded for whatever reason that is. But it is odd that there are not any .dat files at all. Let me see if @sar104 can add to this.

sar104 · Jan 4, 2020

Very interesting pitch and roll oscillations just before the uncommanded descent started:

This looks like a firmware bug to me.

sar104 · Jan 4, 2020

Fly Dawg said:
There have been a few instances when the .dat wasn't recorded for whatever reason that is. But it is odd that there are not any .dat files at all. Let me see if @sar104 can add to this.

No - that does seem to be a problem in a few cases. I wonder if it is related to the processing power of the mobile device being used.

Fly Dawg · Jan 4, 2020

@atsphenomenon In the interim, what firmware version are you running. From all the mini files I have looked at, I am stating to wonder about a "bug". On numerous occassions when RTH is initiated, a brief signal loss occurs for whatever reason that is.

atsphenomenon · Jan 4, 2020

Fly Dawg said:
@atsphenomenon In the interim, what firmware version are you running. From all the mini files I have looked at, I am stating to wonder about a "bug". On numerous occassions when RTH is initiated, a brief signal loss occurs for whatever reason that is.

Thanks for the reply @sar104 . Could you help explain the pitch and roll oscillations a bit more for me to understand. I do see em going haywire in the graph, but what does it actually mean?
Also, i would doubt the .dat file missing is a function of the mobile device's processing power. I'm on a one plus 5T with 6 gb of ram and a fairly powerful processor.

As for the firmware, the app + ac + rc + batteries are all on the latest version. In fact i had updated the app from DJI website to get the 1.0.4 version, but today uninstalled and reinstalled it from the play store once it started showing me the latest 1.0.4 version since i read somewhere that it's preferable to have the 1.0.4 from the playstore rather than the DJI website.

atsphenomenon · Jan 4, 2020

And i just googled the meaning of those pitch and roll oscillations from the graph, and I'm cold to the bone. That's a scary thing to happen to an AC! It's basically just vibrating vehemently along 2 of the 3 axis. That can't be good

Fly Dawg · Jan 4, 2020

@atsphenomenon The pitch and roll oscillations are a bit odd. It almost appears like the aircraft was fighting a cross wind. They are quite uniform, and suddenly stabalize before RTH was cancelled. Note the forward speed as well.

sar104 · Jan 4, 2020

atsphenomenon said:
Thanks for the reply @sar104 . Could you help explain the pitch and roll oscillations a bit more for me to understand. I do see em going haywire in the graph, but what does it actually mean?
Also, i would doubt the .dat file missing is a function of the mobile device's processing power. I'm on a one plus 5T with 6 gb of ram and a fairly powerful processor.

As for the firmware, the app + ac + rc + batteries are all on the latest version. In fact i had updated the app from DJI website to get the 1.0.4 version, but today uninstalled and reinstalled it from the play store once it started showing me the latest 1.0.4 version since i read somewhere that it's preferable to have the 1.0.4 from the playstore rather than the DJI website.

Unfortunately I don't have an explanation for the pitch and roll oscillations. I've checked one other Mini uncommanded descent incident and it doesn't have those oscillations. I still think that the firmware is the most likely problem.

atsphenomenon · Jan 4, 2020

Fly Dawg said:
@atsphenomenon The pitch and roll oscillations are a bit odd. It almost appears like the aircraft was fighting a cross wind. They are quite uniform, and suddenly stabalize before RTH was cancelled. Note the forward speed as well.

View attachment 89747

I'm extremely certain that wind couldn't have been a factor here, though i understand what you're saying - it's probably a bug. No reason for the speed to oscillate like that either. And it's uncanny how all of these happen within the small window of the unprecedented descent starting and the rth cancellation.

On the plus side, I'm thanking my stars that this didn't happen anytime last week when i was flying over the ocean, while taking the AC much further out!

atsphenomenon · Jan 4, 2020

sar104 said:
I still think that the firmware is the most likely problem.

What do you suggest I do, raise a ticket with DJI?

sar104 · Jan 4, 2020

atsphenomenon said:
What do you suggest I do, raise a ticket with DJI?

That's what I would do. They need to be aware of these incidents.

atsphenomenon · Jan 4, 2020

sar104 said:
That's what I would do. They need to be aware of these incidents.

Thanks for all the insights @sar104
The graphs really help, would love to know what you're using to get all these details.

I'll keep you posted on DJIs response, as and when I'm able to figure out how to raise a support ticket and then get a response.

Cheers!

sar104 · Jan 4, 2020

atsphenomenon said:
Thanks for all the insights @sar104
The graphs really help, would love to know what you're using to get all these details.

I'll keep you posted on DJIs response, as and when I'm able to figure out how to raise a support ticket and then get a response.

Cheers!

Some background to log analysis is linked in my signature.