Low GPS signal / bad vision system data leading to fly-away and crash

[Zbip57]

In the end, my main point and all I am saying is that what happened to me is completely counter intuitive and that unless you've read about it or experienced it first hand, there is no way you can be prepared for it. Hence my recommendation for a better warning in the documentation. Heck, even someone like Ian in London got it wrong initially!

That is so true! Four seconds really isn't enough time to react to something so completely unexpected and counter intuitive. And yet, kudos to you, you came sooooo close to actually saving it. My first thought was that, because it was facing you at the time, you mistakenly gave it full right stick in an attempt to make it move to YOUR right. That's a classic panic move. But you actually did properly input full left stick. So close...

It's also difficult because every newer model has different features and behaves differently, or DJI makes firmware changes which alter the way the drone reacts. The user manual often fails to explain exactly what's going on, even after reading it several times, or the manual hasn't been updated to describe what changed with the firmware update.

I had a near miss with my first DJI Phantom, the original Phantom-1, when it unexpectedly reacted in completely the opposite direction that I had intended. It all made sense much later. But at the time, my brain froze up and it took me waaaaay too long to eventually figure out how to bring it back. That incident prompted me to safely conduct experiments (15 videos) in a wide open football field until I fully understood how all the different flight modes actually functioned. And even then there were surprises. And, of course, newer models now react differently.

Here's that "runaway" video from 11 years ago. Back then, Home-Lock mode only worked when the controller is set to GPS-mode. When switched to ATTI mode, if Home-Lock mode is selected it reacts as though operating in Course-Lock mode. That's what was going on here and all made sense (much later)...

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.

Accept third party cookies

 

[Zbip57]

Here's another bizarre question concerning @djules75 's chinese canyon misadventure, one I'd like to hear @BudWalker's opinion on. It concerns the Mini 3 Pro's published top speed in N-mode of 10m/s (~22mph), as specified by DJI in the User Manual.

I had always assumed the top speed was a function of the pitch tilt-angle limit in that mode. It can only tilt so far, hence can only generate so much forward airspeed. To me, that means if it was flying at max 10m/s airspeed while fighting against a 5m/s headwind, it's only making 5m/s forward groundspeed. But if it turned around and flew downwind at 10m/s max airspeed with a 5m/s tailwind, it would be flying at 15m/s groundspeed.

But I've been told that's not how it works, because the specified maximum speed actually applies to groundspeed! Depending on how much excess tilt-angle is available, it might still be able to fly 10m/s upwind. And, even with a 5m/s tailwind, at full stick input, the flight controller would still limit the groundspeed to only 10m/s.

The flight controller measures groundspeed by the rate of change in GPS location. Obviously if the windspeed is too great, the drone will get blown away. But within the drone's physical airspeed capability, its groundspeed will always be limited to no more than 10m/s in N-mode, regardless of whether it's flying upwind or downwind.

There have been many previous discussions about that. See:
mavicpilots.com/threads/question.138227/page-5#post-1562751
mavicpilots.com/threads/has-anyone-flown-in-30-40-mph-winds-with-m4p.142340/#post-1603468

So here's my question, or maybe @BudWalker already explained this and I just didn't understand it at the time.

In @djules75's case, the drone switched to using it's downward optical sensor for position hold while in sticks-centred hover, and it fixated on the moving water current of the fast flowing river. Now it's no longer using GPS to monitor it's position, or to measure it's groundspeed. As far as the drone knows, it's standing perfectly still while it's actually being carried away downstream, moving backwards at nearly 10m/s.

When @djules75 gave it full forward stick input (in N-mode), the drone accelerated forward (upstream) but eventually hit its 10m/s groundspeed limit. As that groundspeed is now being referenced relative to the 10m/s river current flowing downstream, that means the drone is not making any forward progress at all relative to the rock walls on either side of the river. It looks like that on the video.

So what if the river current had been flowing even faster at 15m/s downstream? Does that mean the drone, with full forward stick input being applied, would have been carried downstream at 5m/s because of the drone's limited "groundspeed" of 10m/s? Even though the drone is effectively flying backwards at 5m/s reverse airspeed?

This is making my head hurt...

 

[BudWalker]

Here's another bizarre question concerning @djules75 's chinese canyon misadventure, one I'd like to hear @BudWalker's opinion on. It concerns the Mini 3 Pro's published top speed in N-mode of 10m/s (~22mph), as specified by DJI in the User Manual.

I had always assumed the top speed was a function of the pitch tilt-angle limit in that mode. It can only tilt so far, hence can only generate so much forward airspeed. To me, that means if it was flying at max 10m/s airspeed while fighting against a 5m/s headwind, it's only making 5m/s forward groundspeed. But if it turned around and flew downwind at 10m/s max airspeed with a 5m/s tailwind, it would be flying at 15m/s groundspeed.

But I've been told that's not how it works, because the specified maximum speed actually applies to groundspeed! Depending on how much excess tilt-angle is available, it might still be able to fly 10m/s upwind. And, even with a 5m/s tailwind, at full stick input, the flight controller would still limit the groundspeed to only 10m/s.

The flight controller measures groundspeed by the rate of change in GPS location. Obviously if the windspeed is too great, the drone will get blown away. But within the drone's physical airspeed capability, its groundspeed will always be limited to no more than 10m/s in N-mode, regardless of whether it's flying upwind or downwind.

There have been many previous discussions about that. See:
mavicpilots.com/threads/question.138227/page-5#post-1562751
mavicpilots.com/threads/has-anyone-flown-in-30-40-mph-winds-with-m4p.142340/#post-1603468

So here's my question, or maybe @BudWalker already explained this and I just didn't understand it at the time.

In @djules75's case, the drone switched to using it's downward optical sensor for position hold while in sticks-centred hover, and it fixated on the moving water current of the fast flowing river. Now it's no longer using GPS to monitor it's position, or to measure it's groundspeed. As far as the drone knows, it's standing perfectly still while it's actually being carried away downstream, moving backwards at nearly 10m/s.

When @djules75 gave it full forward stick input (in N-mode), the drone accelerated forward (upstream) but eventually hit its 10m/s groundspeed limit. As that groundspeed is now being referenced relative to the 10m/s river current flowing downstream, that means the drone is not making any forward progress at all relative to the rock walls on either side of the river. It looks like that on the video.

So what if the river current had been flowing even faster at 15m/s downstream? Does that mean the drone, with full forward stick input being applied, would have been carried downstream at 5m/s because of the drone's limited "groundspeed" of 10m/s? Even though the drone is effectively flying backwards at 5m/s reverse airspeed?

This is making my head hurt...

That's my expectation.

 

[Zbip57]

That's my expectation.

Oof. Mind blown.

I guess that means I was wrong when I said this issue can only ever affect the drone while it's in hands-free hover and you can always take control with manual stick inputs to drive the drone where you want it go...

Now we also have to worry about the speed limiter! Sheesh.

 

[Zbip57]

So what could @djules75 have done differently to escape this trap?

One solution is, simply never fly into areas where it will be impossible to recover your drone should it crash. But that ain't gonna happen, since the whole beauty of these things is that drones give ordinary humans the ability to fly into and capture photos and videos of such otherwise inaccessible areas.

He was prevented from flying forward by the limited "groundspeed" measured against the fast flowing downstream current. Switching to Sport-mode would have given a bit more speed. But what if the downstream river current was flowing faster than even the Sport-mode groundspeed limit?

Should he have tried to climb to a height out of range of the downward facing VPS sensors? If the drone then regained GPS confidence, you're golden. But even if it didn't recover sufficient GPS reception, and once escaped out of VPS range, it would have been left in ATTI mode. As an experienced FPV pilot, @djules75 would have been fine with that.

But, drats, foiled again! Isn't there yet another DJI safety "feature" that actually prevents the drone from climbing beyond VPS range unless the drone has good GPS reception?

Wouldn't it be nice if DJI added a switch to the controller, like with the old Phantom models, allowing you to manually switch to ATTI-mode whenever desired? That way, like in the olden days, you could go out into a nice wide empty field to practise using ATTI mode to realize it's not as terrifying or "uncontrollable" as kids these days seem to think.