Well, not the best day, but at least I now know what went wrong.  Maybe a good reminder for us; for both the importance of GPS and Obstacle Avoidance (and nto to rely on it........)

Tough break, but thanks for posting the video. Good luck..

Thanks for the tip!
Hope you get it sorted out!

The recorded flight data has a lot more information that is shown in the Airdata viewer you looked at.

Your Mini 3 receives sats from 3 separate GPS systems and when flying out in the open, it should be receiving around 30 sats, so if you are only getting 13 sats, that's a clue that  your drone isn't getting good GPS reception.

But the number of satellites is only part of the picture and thinking that 13 or 15 or some other number means you have good GPS is not correct.
Your drone needs to have a good number of satellites AND a good spread of sats to provide proper triangulation and accurate positioning.
If you are flying where a significant part of the sky is blocked by trees, tall buildings, canyon walls etc, it's hard to have a good spread of satellites.

The drone's flight controller is looking at much more than just the number of sats.
It's also assessing at the quality of the positioning data and the detailed flight data shows this as a GPS reliability score out of 5.
A score of 4/5 or 5/5 means the GPS data is good for proper positioning and stable flight.
Scores less than 4/5 mean the drone will not have proper horizontal position holding ability.
You get an indication of the GPS reliability from the colour of the satellite icon in your upper screen.
When you turn the drone on and it's still acquiring GPS, the icon is red.
When it turns yellow, the drone has better GPS and approximate position data.
When it turns white, the position data is good and the drone will have proper position holding (as long as the icon stays white).

Deeper in the flight data, the app records the GPS reliability and it can be seen in much more detail.
But the important point is that the number of sats is not a measure of GPS quality and you need to consider more than thinking 12 or 13 or 15 sats means you are good to go.

Hi there, Ian. Thank you for reaching out and we sincerely apologize for the incident that happened to your unit. Hoping that you will receive the unit from our facility soon for you to go up in the sky again. We also appreciate you for sharing this video that you posted. Should you have any concerns, please don't hesitate to reach us. Thank you for always choosing DJI!

I am surprised it doesnt say ATTI mode in the log, the drone was being pushed by the wind and not holding position so, thats gotta be ATTI mode, no?

Labroides Posted at 11-19 16:10
The recorded flight data has a lot more information that is shown in the Airdata viewer you looked at.

Your Mini 3 receives sats from 3 separate GPS systems and when flying out in the open, it should be receiving around 30 sats, so if you are only getting 13 sats, that's a clue that  your drone isn't getting good GPS reception.

Good points well made.  You're spot on, as indeed much of the sky was obscured by the ravine, hills and trees.  
The main point I wanted to make here was don't just take the audible 'Home Point Updated' message as a green light to fly. It's a shame I wasn't screen recording as I can't recall if the sat count indicator was orange or white.  

Bashy Posted at 11-19 23:03
I am surprised it doesnt say ATTI mode in the log, the drone was being pushed by the wind and not holding position so, thats gotta be ATTI mode, no?

I think that would indicate if there was no sat lock; here there were at least 13 so it may have thought it had okay GPS....  

Ian in London Posted at 11-19 23:28
I think that would indicate if there was no sat lock; here there were at least 13 so it may have thought it had okay GPS....

True but it still acted like it was in ATTI mode though what with the drift.

BTW, hope its not too costly to fix, i suspect it wont be

Ian in London Posted at 11-19 23:27
Good points well made.  You're spot on, as indeed much of the sky was obscured by the ravine, hills and trees.  
The main point I wanted to make here was don't just take the audible 'Home Point Updated' message as a green light to fly. It's a shame I wasn't screen recording as I can't recall if the sat count indicator was orange or white.

This is something that i have always taken for granted, as soon as i get the "home point updated", i go, i should change that habit, thanks

Ian in London Posted at 11-19 23:27
Good points well made.  You're spot on, as indeed much of the sky was obscured by the ravine, hills and trees.  
The main point I wanted to make here was don't just take the audible 'Home Point Updated' message as a green light to fly. It's a shame I wasn't screen recording as I can't recall if the sat count indicator was orange or white.

It's a shame I wasn't screen recording as I can't recall if the sat count indicator was orange or white.  
Your recorded flight data will show the GPS reliability for all times during the flight.
That's better than the crude red/yellow/white indication.

Labroides Posted at 11-20 00:34
It's a shame I wasn't screen recording as I can't recall if the sat count indicator was orange or white.  
Your recorded flight data will show the GPS reliability for all times during the flight.
That's better than the crude red/yellow/white indication.

Yes; it shows an orange '13' for the entire flight.....  

oops,  a good reminder.   Also worth noting that the DJI return instructions I recently had were wrong.,  they waned three invoices created,  One on the parcel and two to give over to the courier,  who said they wouldn't take them as they are not needed,   They were needed and should have been in the plastic wallet on the parcel.  

Labroides Posted at 11-20 00:34
It's a shame I wasn't screen recording as I can't recall if the sat count indicator was orange or white.  
Your recorded flight data will show the GPS reliability for all times during the flight.
That's better than the crude red/yellow/white indication.

Hi,

Crude red/ yellow / white icon colouring = the 0 - 5 value in the log, they show the same value in colours and numbers. So it is not better or worse - it is the same

Red = 0, 1 and 2
Yellow = 3
White = 4 or 5

btw  3 or more  = GPS lock.
cheers
JJB

Hi Ian,

Would be nice to see your flightlog   why?  because for the message HomePoint is updated you need white icon wich is the 4 or 5 for GPS 'health'.
So when flying into the area the GPS reception got worse.Don`t know how low you flew over streaming water, bc in that case hover stability is sometimes worse due to bottom sensor funny readings....
So mayby there is more than just a GPS hover instability, in that case hover positioning performance is worse, but not always will it drift into one direction.

If you like  upload your log to > https://www.phantomhelp.com/LogViewer/Upload/
Post the uploaded link on here
More data in this conversion than in AIrData.


cheers
JJB

JJB* Posted at 11-20 03:40
Hi Ian,

would be nice to see your flightlog   why?  because for the message HomePoint is updated you need white icon wich is the 4 or 5 for GPS 'health'.

I've just uploaded, but I must say, this seems to give less info than Airdata.  It's missing some of the error messages and also doesn't show the message when the Home Point got updated, whilst thi sinfo does show on Airdata.

Anyway, here's the link; very happy to get any pointers.  Wondering if the 13 sats was a combination of GPS, Bediou and Galileo systems, which could mean that no one system gave a proper lock....

https://www.phantomhelp.com/LogViewer/8SSJSXPBRFAGRM92CCYU

Ian in London Posted at 11-20 03:45
I've just uploaded, but I must say, this seems to give less info than Airdata.  It's missing some of the error messages and also doesn't show the message when the Home Point got updated, whilst thi sinfo does show on Airdata.

Anyway, here's the link; very happy to get any pointers.  Wondering if the 13 sats was a combination of GPS, Bediou and Galileo systems, which could mean that no one system gave a proper lock....

Hi Ian

Thanks,

You started your flight with 13 satellites and 2 for GPS health or signal quality reception
GPS icon nust be coloured RED....

At 21.8s in flight GPS icon white, bc "health"now 4. With 4 or 5 HomePoint is set.
(check it in my chart of your data)

Rest of the flight always 13 sats and 4.
And check also the max height setting, wich is set to 30 meters and later in flight to your own max height setting (only possible with "good" GPS)

VPS height about 5.9 meter, Baro height about 8 meter.

So IMO not the GPS was the problem but flying above streaming water.
As said, hover instability due to bad GPS will not drift a drone from position is one direction.It will drift ofcourse on the wind in ATTI mode, but your drone was not in ATTI mode.

If i look at GoogleEarth, seems tha the drone move direction is the streaming water direction....

ADD : from 29.2s the VPS height sensor was not measuring....this until 43.6. So the height values 5.9 meters is not the actual height. ( it is copy paste from the last know height)

cheers
JJB

Ian in London Posted at 11-20 03:45
I've just uploaded, but I must say, this seems to give less info than Airdata.  It's missing some of the error messages and also doesn't show the message when the Home Point got updated, whilst thi sinfo does show on Airdata.

Anyway, here's the link; very happy to get any pointers.  Wondering if the 13 sats was a combination of GPS, Bediou and Galileo systems, which could mean that no one system gave a proper lock....

Hi,

strange   PH conversion shows way more data than AIirdata...and i never miss HP updating messages in PH.
But indeed   in your log the HomePoint updated is not in the messagel list.
Guess AirData do the same as i do....

cheers
JJB

JJB* Posted at 11-20 04:01
Hi Ian

Thanks,

According to the logs, the signal quality dropped before the start of an uncontrolled flight, or this is a consequence.

Drone.Hunter Posted at 11-20 04:47
According to the logs, the signal quality dropped before the start of an uncontrolled flight, or this is a consequence.

the moving away from position started at 31.8s in flight, GPS #13, quality #4.
13 and 4 remains the same after 31.8s.

Do post a screen shot what you are saying.

cheers
JJB

JJB* Posted at 11-20 03:37
Hi,

Crude red/ yellow / white icon colouring = the 0 - 5 value in the log, they show the same value in colours and numbers. So it is not better or worse - it is the same

Crude red/ yellow / white icon colouring = the 0 - 5 value in the log, they show the same value in colours and numbers. So it is not better or worse - it is the same
Crude = less precise.
You can see more information, like the difference between 0, 1 & 2.

JJB* Posted at 11-20 04:51
the moving away from position started at 31.8s in flight, GPS #13, quality #4.
13 and 4 remains the same after 31.8s.

I'm looking at your screenshots. I'm trying to figure it out too. At 29s, according to your screenshots, it seems to have risen high enough not to use the IR sensors. In cloudy weather and with 20 satellites, I had a yellow indicator, and in good 15 it was already normal.

Drone.Hunter Posted at 11-20 05:28
I'm looking at your screenshots. I'm trying to figure it out too. At 29s, according to your screenshots, it seems to have risen high enough not to use the IR sensors. In cloudy weather and with 20 satellites, I had a yellow indicator, and in good 15 it was already normal.

Hi,

GPS 'quality' , the value in the CVS 0 - 5, has nothing to do with the use of the bottom height sensor.
2 different systems wich does not interact.

cheers
JJB

ADD : from 29.2s the VPS height sensor was not measuring....this until 43.6. So the height values 5.9 meters is not the actual height. ( it is copy paste from the last know height)

So he didn't use the bottom IR sensors? So he was not supposed to react to the movement of water?

Sorry to see that your Mini 3 Pro took a dive.  Must have been a chore to retrieve it at that location.  Thanks for the tip.  

It is great to see one of my favorite Youtubers is human.  : )  Thanks for sharing a good reminder!  

Drone.Hunter Posted at 11-20 05:52
ADD : from 29.2s the VPS height sensor was not measuring....this until 43.6. So the height values 5.9 meters is not the actual height. ( it is copy paste from the last know height)

So he didn't use the bottom IR sensors? So he was not supposed to react to the movement of water?

Not sure   in the log the it shows that the height sensor was not active....problem is that we do not know exactly what that means in DJI logic....

or measuring but signal not accurate or trustfully enough to show data in the log
or simply not measuring

But why would it be not 'active' as the drone was well within VPS range?

I have seen many flight logs where drone does not hold position flying low over streaming water.
As in this log.

GPS bad hover stabiliy means that drone will keep it position but moves around the position.
In this log the LatLon data shows accurate data, moving line is straight line.

cheers
JJB

Но почему он не был «активным», потому что не может быть в пределах досягаемости VPS?
The screen stops showing the height after 6m. In the logs, the last value is 5.9m.

JJB* Posted at 11-20 04:01
Hi Ian

Thanks,

Wow; thanks for this.  Very interesting and does seem to point towards the drone's downward visions sensors actively following the streaming water currents rather than the GPS.   
It's interesting; should it take GPS or should it take the vision of the 'ground' below.   No right answer and I guess in most cases, the ground isn't moving.....  
I'm hoping DJI come back to me today and may shed a bit more light.  Or not.  

Ian in London Posted at 11-21 00:03
Wow; thanks for this.  Very interesting and does seem to point towards the drone's downward visions sensors actively following the streaming water currents rather than the GPS.   
It's interesting; should it take GPS or should it take the vision of the 'ground' below.   No right answer and I guess in most cases, the ground isn't moving.....  
I'm hoping DJI come back to me today and may shed a bit more light.  Or not.

Hi Ian,

My pleasure to help.

This is indeed a issue DJI should look after. Flying low over streaming water with a GPS lock and VPS height is measuring or not showing in the log that is is measuring is asking for trouble.
Drone will not hover at position (with no RC input) but will drift away....

Should test myself to simulate this, mayby hovering over a carpet wich i will slide away....

I often fly low over water (not streaming fast), and not always is the bottom sensor measuring height, well in the log no height (or copy paste last measured height to the next records) plus the TrueFalse column " OSD.isSwaveWork" is showing FALSE.
Note that the column headers are not from DJI, naming is done by the guys who decrypt the log.

Seeing many logs, i have the idea that the column should be named as "OSD.isSwaveMeasuring".

cheers
JJB

Drone.Hunter Posted at 11-20 17:06
Но почему он не был «активным», потому что не может быть в пределах досягаемости VPS?
The screen stops showing the height after 6m. In the logs, the last value is 5.9m.

Hi

Don`t understand your question.
Last height values in the log : BaroHeight -1.5 and VPS height 0.0

cheers
JJB

Ian, there are a number of tales of drones tracking moving objects that are under the drone i.e. via the VPS system.
I have seen this indoors with a Mavic Mini, I hovered the drone over something distinctive and then slowly moved what ever the thing was and the drone followed it. I might have also seen it with an M2P/Z in an indoor flight. With regards to the Mavic Mini the movement needed to be slow, the room is quite cluttered and there would be plenty of other, stationary, objects in the field of view of the VPS cameras for the VPS to lock on to, but I have read others where the drone (other models) tracked something in a river etc.. I can't remember the precise details of the M2P/Z event, if it did happen, but the speed of movement of the object would have been slow.
I also saw an M2P/Z do something that may be related in an outdoor flight, I had launched from a 'lawn' of  long grass but used a blue topped box as a launch pad. I took off so as to enable precision landing but then decided that the box was a bit risky as a  landing pad so moved it to somewhere else (quite visually different from the launch point but not far from the launch point). When I RTH'ed the drone it passed over the "somewhere else" whilst the drone was slowing, the drone distinctly hesitated over the box before carrying on to the correct landing site.

An update on why this may have crashed.....  Thanks to @JJB for your input!

Tough break, but thanks for posting the video. Good luck..

Ian in London Posted at 11-25 08:38
An update on why this may have crashed.....  Thanks to @JJB for your input!

https://youtu.be/Ow8WzVSgxYA

Hi Ian,

Thanks for my name in your post.

But, flying low over streaming AND having good GPS reception! the vision system still get confused and often results is 'fly-away' from position (without RC input).
It simply cannot hold GPS hover position.

Manual state indeed that Downward Vision System is applicable when GNSS signals are unavailable or weak. (unavailable = red gps icon), weak = amber gps icon)

So imo the manual is incorrect. Mayby something to do with the bottom height sensor wich cannot be switch off ; "Landing protection is always ON"

My advise : Do not fly too low over streaming water!  or if you do ; keep good eye on your drone and be ready for a pull-up!

see this log > https://forum.dji.com/forum.php?mod=viewthread&tid=281359    > 30 GPS and good GPS recption.

cheers
JJB

JJB* Posted at 12-17 02:48
Hi Ian,

Thanks for my name in your post.

Yes; Poor Darren @DMProductions just suffered the same fate with his Mini 3 Pro low over water..... It's a real issue.

Ian in London Posted at 12-17 03:37
Yes; Poor Darren @DMProductions just suffered the same fate with his Mini 3 Pro low over water..... It's a real issue.

Or just a limitation to use DJI drones low flying over (steaming) water, see all the important notes in the manual.

Description about the sensors and when they work or not (+what they really do) has to be changed in the manuals.

cheers
JJB