DJI Mavic, Air and Mini Drones
Friendly, Helpful & Knowledgeable Community
Join Us Now

Very Informative Mini 3 Pro Crash Analysis

You should read the two? threads on the DJI forum concerning this, and in particular JJB's comments, I can't remember if Labroides contributed.
It's a pity Ian doesn't post here it would be interesting to see what @slup, @Meta4 and @sar104 et al could dig out.
 
Mmm... yeah, pretty sure that was he says is only part of the cause for the drift. Would be nice to see the GPSHealth along side the not equal important GPS sat count... the PhantomHelp online log viewer have that as a symbol in the top of the screen (so no need to dig through the .csv file) & if the GPSHealth were sketchy there the drift most probably were caused by the VPS sensors locking at the below stream & followed it.
 
  • Like
Reactions: Yorkshire_Pud
The drifting aspect because according to the video the UAS drifted a considerate distance. I am trying to get into the mechanics of the unit and not just fly the drone. I am trying to absorb any information related to the internal components, functions, malfunctions causes, etc. Which intern, will make me a safest pilot, the places where I fly are quite like that particular scenario. I own a Mavic 3 and an Air 2s, but I have been using my Mini 3 to get the experience flying in tight quarters. I am a nature guy and love to fly and venture in our gorgeous Colorado Habitats.
 
...I am trying to get into the mechanics of the unit and not just fly the drone. I am trying to absorb any information related to the internal components, functions, malfunctions causes, etc ...using my Mini 3 to get the experience flying in tight quarters.

Here below you have 4 general lessons that should be made from Ian's YT posting...

1. The full downloadable user manual... read it, understand it, try all out in a open location & re-read it until the info there is second nature and totally clear & logical.

2. Understand what makes these crafts fly stable, hold position, & brake when sticks are released... & when they no longer can manage it.

3. Learn to pay attention to what's shown on screen & understand what everything means, what's good, what's bad & what is in between... flying & making stills/video isn't one or the other, it's always multi tasking.

4. Until all 3 above is full done deal... keep it near, low, & slow and in open environments without obstacles.

It's actually surprising that Ian with his drone experience fail in all 4...

And here one extra...

These thing's flies nearly by them selves... most of the time, even without knowledge... but suddenly it happens due to the pilot's ignorance.
 
Here below you have 4 general lessons that should be made from Ian's YT posting...

1. The full downloadable user manual... read it, understand it, try all out in a open location & re-read it until the info there is second nature and totally clear & logical.

2. Understand what makes these crafts fly stable, hold position, & brake when sticks are released... & when they no longer can manage it.

3. Learn to pay attention to what's shown on screen & understand what everything means, what's good, what's bad & what is in between... flying & making stills/video isn't one or the other, it's always multi tasking.

4. Until all 3 above is full done deal... keep it near, low, & slow and in open environments without obstacles.

It's actually surprising that Ian with his drone experience fail in all 4...

And here one extra...

These thing's flies nearly by them selves... most of the time, even without knowledge... but suddenly it happens due to the pilot's ignorance.
Thanks for your input. This is another example that Murphy's law can happen to the best, another reason to stay humble and grateful!
 
  • Like
Reactions: Bernard
The drifting aspect because according to the video the UAS drifted a considerate distance. I am trying to get into the mechanics of the unit and not just fly the drone. I am trying to absorb any information related to the internal components, functions, malfunctions causes, etc. Which intern, will make me a safest pilot, the places where I fly are quite like that particular scenario. I own a Mavic 3 and an Air 2s, but I have been using my Mini 3 to get the experience flying in tight quarters. I am a nature guy and love to fly and venture in our gorgeous Colorado Habitats.
I am not being awkward but I think the specific lesson to be taken from Ian's accident and others of a simillar nature is the following.
Where VPS is fuctional it can, and possibly normally does, overide GPS in terms of holding position.
There have been several threads where, despite adequate GPS, a drone is thought to have 'followed', uncommanded, something in flowing water.
How far, with respect to a GPS lat & long position, any drone permitted itself to move, or will permit itself to move, I can not say, nor can I say whether or not a significant drift with respect to a GPS location will cause some form of conflict within the drone's control system.
In my own experience I can say I have seen things that suggest to me that VPS can cause a drone to "track" a distinctive object that is in the field of view of its VPS sensor but this has been primarily in indoors flights that lack GPS.

But out of doors, with full GPS, 20+ satellites, I have seen much 'tighter' control of low level ( eye level ) postion holding in daylight than I have seen in darkNESS. In daylight VPS will work, in darkness VPS does not work so the drone would have only GPS position holding....... so, be wary when low (within VPS range) over flowing water especially if the water generates foam etc. or waves or contains debris. If it is suspected that the drone is following something in the water then I'd be inclined to send the drone up, out of VPS range, if that is possible.
 
Last edited:
The sensor override only occurs when the drone is within 10 or 20 feet of a feature the drone thinks is a ground feature. Above that, the inertial guidance system along with the gps combines together to maintain position. If this integration is considered faulty the drone switches into an inertial guidance only mode (ATTI mode) and drifts according to the existing speed, and the local wind currents.

For this reason, a pilot needs to have a plan, and have some practice, flying in ATTI mode.
 
Where VPS is functional it can, and possibly normally does, override GPS in terms of holding position.
Nope, GPS is the primary means for holding position whenever the control sticks are centred.

The vision sensor of the VPS only becomes the active means for holding position when GPS coverage is insufficient.

The User Manual for the Mini says this, "When GPS is unavailable, the Downward Vision System is enabled if the surface has a discernable surface and sufficient light."

If GPS is unavailable AND there is insufficient light, an indiscernible surface, drone too high so the VPS is out of range, THEN the drone will default to ATTI mode and is unable to automatically brake to a halt or hold position when sticks are centred.

[...] be wary when low (within VPS range) over flowing water especially if the water generates foam etc. or waves or contains debris.

I've experienced a similar issue as Ian in London, but without crashing. Similar to Ian, I was flying my Mini low over the water at the bottom of a narrow river gorge. My drone had a very limited view of the sky and, unnoticed by me, had lost GPS lock. Similar to Ian, I left my Mini hovering for a moment while I looked away to make setting adjustments on my screen. When I looked back, I was alarmed to see the drone had drifted a little ways downstream all by itself!

Without GPS, it was relying on the optical sensor of the VPS and had locked onto the foam bubbles floating on the current. In my case, no harm done. In Ian's case, the drone drifted into a tree.

If it is suspected that the drone is following something in the water then I'd be inclined to send the drone up, out of VPS range, if that is possible.

You still have full control over the movement of the drone by using normal stick inputs. It will only try to hold position on objects detected by the optical sensor while the sticks are centred. If those objects are moving, like wind ripples or foam on the water's surface, the VPS will try to follow those when sticks are centred, but only if GPS is currently insufficient.

The drone will only be following something in the water if GPS is unavailable, in which case if you also send the drone up high enough to get out of VPS range, then with GPS and VPS both not functioning the drone will switch to ATTI mode. Even that is no cause for panic as the drone will continue to respond to stick inputs. It just won't brake to a halt or hold position by itself.

The moral of the story is, always make sure your drone actually stays put in one spot and holding position before looking away to make setting changes on your screen. If in any doubt, keep your drone in sight and don't look away for long periods of time.
 
Nope, GPS is the primary means for holding position whenever the control sticks are centred.

The vision sensor of the VPS only becomes the active means for holding position when GPS coverage is insufficient ....... etc
Sorry but we will have to agree to disagree on which has precedence.
I am going by my observations. For me positon holding, at a given location, was better is daylight, when VPS would have been working, than it was at night when VPS could not work.
Relying on VPS alone, indoors, my Mavic Minis and Mini 2 s are cm perfect in hovers in both the short and long term. I have actually left one of them hovering for 10 minutes whilst I was in another room and when I came back the drone was within single digit cm of where I had left it.
Out of doors, in daylight, with 18 - 23 satellites I would say that they hover within 1ft of their 'assigned' position. it is rarely windless where I have flown such flights so they do get buffeted a bit even on a 'calm' day. At night, reliant on only GPS, the oscilations were significantly larger, maybe greater than 2ft.

It is my recollection of threads discussing drift downstream there was no concern about satellite numbers.
 
  • Like
Reactions: chiangmaicharlies
Sorry but we will have to agree to disagree on which has precedence. {...]
You're claiming the VPS takes priority over GPS? That should be simple enough to test for yourself.

My Mini gets sufficient GPS reception even inside our house. I'd need to go find an underground bunker or parking garage somewhere to test out flying without GPS. In that case, for sure the optical sensor of the VPS (and not GPS) is the system used for position holding and it is very effective and accurate. If you know for certain you have no GPS, try having the drone hovering over a piece of carpet with enough texture that the VPS can get a fix on it, then drag the carpet slowly away. In the absence of GPS, the drone will follow the carpet.

Now try the same thing while flying outside in an open area with good GPS coverage. Hover the drone over the carpet again so the optical sensor of the VPS can see it and get a good fix on the carpet. What happens now if you slowly drag the carpet out from under the drone?

If you are right and VPS takes priority, you're saying the drone will follow the carpet even if it has good GPS reception. I'm saying GPS takes priority and the drone will remain hovering in one spot and won't react to the surface moving under it.

Try it and let us know what you find. 🤔
 
Last edited:
Sorry but we will have to agree to disagree on which has precedence.
I am going by my observations. For me positon holding, at a given location, was better is daylight, when VPS would have been working, than it was at night when VPS could not work.
Relying on VPS alone, indoors, my Mavic Minis and Mini 2 s are cm perfect in hovers in both the short and long term. I have actually left one of them hovering for 10 minutes whilst I was in another room and when I came back the drone was within single digit cm of where I had left it.
Out of doors, in daylight, with 18 - 23 satellites I would say that they hover within 1ft of their 'assigned' position. it is rarely windless where I have flown such flights so they do get buffeted a bit even on a 'calm' day. At night, reliant on only GPS, the oscilations were significantly larger, maybe greater than 2ft.

It is my recollection of threads discussing drift downstream there was no concern about satellite numbers.
It's actually not that one or the other takes over. Remember that GNSS itself is not a primary data source for the fusion calculation in any situation - that's always the inertial data from the IMU. The GNSS, compass, barometer and VPS data are then folded in with varying gains (weightings) depending on estimated error margins, via a Kalman-type algorithm.
 
You're claiming the VPS takes priority over GPS? That should be simple enough to test for yourself.
I am saying that is the impression I get.

With regards to indoor gps, in my man cave none of my drones have sufficient satellite lock to establish the drones' location, 0-5 satellites is normal for me indoors, I have seen 8 momentarily once or twice, but never has they been able to establish a home point nor the drones' position. The Lat & Long columns in the relevant logs are empty.

Yes, indoors they do follow distinctive objects that are moved when they are in the VPS field of view.
Out doors, that I remember I have not tried moving an object under a hovering drone but since you have aroused my curiosity I may, weather permitting, try it.

BUT, with an M2P/Z launched so as to enable precision landing, I launched the drone from an orange landing mat on grass near my car but I hadn't pegged the mat down and a gust subsequently caught the mat and blew it over and behind a wall, it landed and settled probably 12ft away from the launch point. A subsequent RTH brought the drone back, at close to the minimum RTH height, over the landing mat and the drone hesitated over the landing mat before moving on to the correct landing point and it was a distinct hesitation.
Indoors the M2P hovers slightly less accurately than the minis but I suspect that that is, in part, due to its stronger down draught flapping things or creating greater turbulence in air in the room.
 
Lycus Tech Mavic Air 3 Case

DJI Drone Deals

New Threads

Forum statistics

Threads
130,592
Messages
1,554,178
Members
159,596
Latest member
da4o98