Lost DJI Mavic Pro during Litchi mission

[sar104]

Sorry to hear about your loss. I had my own story recently and I know how this feels.

Just looked at your mission. You lost connection rather soon just before WP2 where your elevation is set to 102' (you used ground level offset). WP2 is also where you meet the railroad tracks (Portlock Road). I am looking at a power transmission line following that road across the tracks and the line is pushed rather high (probably to allow for clearance to the rail activity). I don't know if it managed to clear the lines: 100' +/- possible altimeter error which can be significant depending on weather/temperature conditions (GPS altitude information is not very accurate at all so the Mavic relies on the barometric altimeter). I would definitively take a look in that area.

Rules from someone that got burnt by WP missions as well:
1) Don't mix ground level settings with relative to takeoff settings.
2) Make sure to TURN OFF "smart RTH"
3) NEVER design a mission that pushes the maximum flight time (such as yours). Note you had some critical cell deviations - may have been low battery temperature in the beginning.
4) Check, double-check and tripple-check your mission plan before attempting the flight. I any doubt - don't fly it.

Why would you turn off smart RTH? That sounds like a good way to lose the aircraft.

 

[Akemi]

Maybe there's a lesson in that?

Sorry to hear about your loss. I had my own story recently and I know how this feels.

Just looked at your mission. You lost connection rather soon just before WP2 where your elevation is set to 102' (you used ground level offset). WP2 is also where you meet the railroad tracks (Portlock Road). I am looking at a power transmission line following that road across the tracks and the line is pushed rather high (probably to allow for clearance to the rail activity). I don't know if it managed to clear the lines: 100' +/- possible altimeter error which can be significant depending on weather/temperature conditions (GPS altitude information is not very accurate at all so the Mavic relies on the barometric altimeter). I would definitively take a look in that area.

Rules from someone that got burnt by WP missions as well:
1) Don't mix ground level settings with relative to takeoff settings.
2) Make sure to TURN OFF "smart RTH"
3) NEVER design a mission that pushes the maximum flight time (such as yours). Note you had some critical cell deviations - may have been low battery temperature in the beginning.
4) Check, double-check and tripple-check your mission plan before attempting the flight. I any doubt - don't fly it.

She absolutely cleared the lines because a train engineer spotted it traveling north towards the waypoint where it was going to make a smooth easterly turn. Since I have DJI Refresh, just recovering her dead would allow me to replace her. But without experts in physics weighing in as to where they believe it is, I'm at the mercy of luck.

 

[Meta4]

But without experts in physics weighing in as to where they believe it is, I'm at the mercy of luck.

You've had an expert physicist give you all the help he could.
There's very little here for an expert to go on.
The terrain height + trees is about all there is.

 

[Roamer105]

Why would you turn off smart RTH? That sounds like a good way to lose the aircraft.

If you plan a long mission and keep "smart RTH" the mission will never complete.

 

[sar104]

She absolutely cleared the lines because a train engineer spotted it traveling north towards the waypoint where it was going to make a smooth easterly turn. Since I have DJI Refresh, just recovering her dead would allow me to replace her. But without experts in physics weighing in as to where they believe it is, I'm at the mercy of luck.

However - if it was as low as the engineer indicated then it probably didn't make it over the Atlantic Avenue overpass.

 

[sar104]

If you plan a long mission and keep "smart RTH" the mission will never complete.

OK - but that implies that the aircraft possibly won't have enough battery reserve to complete the mission. I'd rather have it abort the mission and return than run out of gas before getting home.

 

[Roamer105]

However - if it was as low as the engineer indicated then it probably didn't make it over the Atlantic Avenue overpass.

I would have to agree with that. It looks like the pavement surface of the overpass at the point of flyover is at about 56'. In addition there are two obstacles: First, the protection fence which is at a minimum 20 feet high (above that pavement) and second there is a street lamp (on the overpass) that could be directly in the flight path. That lamp will clearly exceed the 100' flight level of your Mavic.


View attachment 59879

 

[m0j0]

OK - but that implies that the aircraft possibly won't have enough battery reserve to complete the mission. I'd rather have it abort the mission and return than run out of gas before getting home.

Going to agree. Seems like a real easy way to find your way to the classifieds

 

[Gryphon962]

I'm very late to this thread, and all of you know far more than I do about Litchi, and waypoints, and what might go wrong. However, I'm a retired military helicopter pilot with almost 4000 flight hours, much of it spent down at 100 feet or so, and I've worked in the GPS program.

My suggestion to all those using Litchi to fly autonomous routes - notwithstanding that beyond LOS flying is specifically prohibited by part 107 - is to plan the route such you'd be confident sitting blindfold in the aircraft with no access to controls. If you would feel okay doing that, its probably safe. But I doubt many of us would put our pink bodies in an aircraft that is going to fly above houses, trees, and freeway ramps at 100' agl as determined by a GPS, IMU, compass, altimeter and guidance software that isn't certified by the FAA at all.

Anything below 250' agl without sensors (or eyes) to detect obstructions you cant see on google earth is asking for trouble.

 

[sar104]

I'm very late to this thread, and all of you know far more than I do about Litchi, and waypoints, and what might go wrong. However, I'm a retired military helicopter pilot with almost 4000 flight hours, much of it spent down at 100 feet or so, and I've worked in the GPS program.

My suggestion to all those using Litchi to fly autonomous routes - notwithstanding that beyond LOS flying is specifically prohibited by part 107 - is to plan the route such you'd be confident sitting blindfold in the aircraft with no access to controls. If you would feel okay doing that, its probably safe. But I doubt many of us would put our pink bodies in an aircraft that is going to fly above houses, trees, and freeway ramps at 100' agl as determined by a GPS, IMU, compass, altimeter and guidance software that isn't certified by the FAA at all.

Anything below 250' agl without sensors (or eyes) to detect obstructions you cant see on google earth is asking for trouble.

Yes and no. Firstly, these aircraft do have sensors to avoid obstacles. Secondly, at least one of the main uses for BVLOS automated missions is search and rescue, and 250 ft AGL is too high to resolve the kind of detail needed. Thirdly, depending on location and mission priority of course, it may be reasonable to accept a higher level of risk for an inexpensive unmanned aircraft than for a helicopter.

 

[Gryphon962]

Yes and no. Firstly, these aircraft do have sensors to avoid obstacles. Secondly, at least one of the main uses for BVLOS automated missions is search and rescue, and 250 ft AGL is too high to resolve the kind of detail needed. Thirdly, depending on location and mission priority of course, it may be reasonable to accept a higher level of risk for an inexpensive unmanned aircraft than for a helicopter.

You make 3 very good points, however:

1. Does the MP do 33.6mph with sensors on? From what I read in here, its 22 mph in P-mode, 32 mph P-mode OA off, 42 mph Sport mode. So I was assuming the OP had sensors off to travel at 33 mph.
2. Fully accept the SAR mission justification for BVLOS but I assume that those missions are flown under public UAS rules. Or should be
3. Fully accept that higher risk can be tolerated, but I'd argue it's only to the extent necessary. Planning to fly anything at 100' agl over peoples homes and over freeways isn't reasonable.

 

[sar104]

You make 3 very good points, however:

Does the MP do 33.6mph with sensors on? From what I read in here, its 22 mph in P-mode, 32 mph P-mode OA off, 42 mph Sport mode. So I was assuming the OP had sensors off to travel at 33 mph.

Sensors obviously need to be on - I wasn't addressing this case in particular.

Fully accept the SAR mission justification for BVLOS but I assume that those missions are flown under public UAS rules. Or should be

No - they can be flown under Part 107 with an SGI waiver, which is what I've done.

Fully accept that higher risk can be tolerated, but I'd argue it's only to the extent necessary. Planning to fly anything at 100' agl over peoples homes and over freeways isn't reasonable.

Agreed, but that's true whether it is automated BVLOS flight or manual VLOS flight or anything in between.

 

[Bigbird48]

I guess his is lost never to be found as we haven't heard a word fron OP in a couple of months.To bad. I'm not understanding why theres not enough info in his flight logs to figure out what happened. Was it because it was a Litchi mission and the log are different or what.

 

[sar104]

I guess his is lost never to be found as we haven't heard a word fron OP in a couple of months.To bad. I'm not understanding why theres not enough info in his flight logs to figure out what happened. Was it because it was a Litchi mission and the log are different or what.

The flight log ended early when the connection was lost. What might have happened after that can only be inferred from the mission profile.

 

[AMann]

Sorry to hear about your loss. I had my own story recently and I know how this feels.

Just looked at your mission. You lost connection rather soon just before WP2 where your elevation is set to 102' (you used ground level offset). WP2 is also where you meet the railroad tracks (Portlock Road). I am looking at a power transmission line following that road across the tracks and the line is pushed rather high (probably to allow for clearance to the rail activity). I don't know if it managed to clear the lines: 100' +/- possible altimeter error which can be significant depending on weather/temperature conditions (GPS altitude information is not very accurate at all so the Mavic relies on the barometric altimeter). I would definitively take a look in that area.

Rules from someone that got burnt by WP missions as well:
1) Don't mix ground level settings with relative to takeoff settings.
2) Make sure to TURN OFF "smart RTH"
3) NEVER design a mission that pushes the maximum flight time (such as yours). Note you had some critical cell deviations - may have been low battery temperature in the beginning.
4) Check, double-check and tripple-check your mission plan before attempting the flight. I any doubt - don't fly it.

Why turn off of RTH? If you set it at the maximum height, it could be your fail safe -Assuming that height setting would put it higher than anything it was flying over in a straight line from any point along the Litchi route.

 

[AMann]

I guess his is lost never to be found as we haven't heard a word fron OP in a couple of months.To bad. I'm not understanding why theres not enough info in his flight logs to figure out what happened. Was it because it was a Litchi mission and the log are different or what.

Maybe he can’t reply. It sounded like after they realized it may have been his drone at the train incident, the police were being awfully polite about it and helping him out, a little too polite...

 

[Olav van Gerven]

Now, Im late and the only reason I read this kind of threads is, that I want to learn from mistakes others have made. They have spent money for that, I don't have to do it too.

What I am wondering about if I look at the flight map: The flight was relativly low above traintracks. Traintracks in Germany mostly have powerlines to feed the train. They have a voltage up to 3 KV and will cause a magnetic interference with radiosignals. The interference will be even stronger when a train passes.
Another point with i.m.o. possible interference to any signal could have been the telephonetower that is mentioned.
Could it be, that at one point the drone got disturbed by the interferences or that the coninues flight through the magnitcfield induced by the trainpowerlines caused the system to crash in which consequence, the drone has crashed?

Like I said, not to solve this mystery but for me to understand.

 

[sar104]

Now, Im late and the only reason I read this kind of threads is, that I want to learn from mistakes others have made. They have spent money for that, I don't have to do it too.

What I am wondering about if I look at the flight map: The flight was relativly low above traintracks. Traintracks in Germany mostly have powerlines to feed the train. They have a voltage up to 3 KV and will cause a magnetic interference with radiosignals. The interference will be even stronger when a train passes.
Another point with i.m.o. possible interference to any signal could have been the telephonetower that is mentioned.
Could it be, that at one point the drone got disturbed by the interferences or that the coninues flight through the magnitcfield induced by the trainpowerlines caused the system to crash in which consequence, the drone has crashed?

Like I said, not to solve this mystery but for me to understand.

Magnetic fields are created by current rather than voltage - in fact the high voltages are used to keep the current lower. It's unlikely to have caused any problems. Cell towers with microwave dishes can cause problems directly in the beam, but it's usually transient.

 

[Olav van Gerven]

Thanks for the replies, like I said, just trying to learn and avoid mistakes made by others (still left enough to make myself, I guess)