My P3P crashed, DJI not cooperating

[MikeyOnline]

Ive looked at the file creation times of the video and the photos taken just prior to me landing the Phantom. the video started at 12:36PM, so six minutes in would be around 12:41PM. The last picture I tool before bringing it in was at 12:51. It appears I did a video first then took a series of pictures from 12:43 to 12:51 then brought her in to land.

I see that you started the video early on (55s into the flight) which is good because I wanted to see how far off your altitude was at that point. Going 6 minutes into the video, I see that your altitude is reported as 23.4m or 77 feet while the actual was closer to 45 or 50 feet max. A discrepancy of 20 feet is about what I've seen in my last two flights, so your drift definitely became worse as the flight moved on since the starting error looks closer to 20 feet and the error when you crashed was closer to 50 feet.

I've not flown out a full battery lately so I'm going to have to keep an eye on my altitude drift to see if it becomes worse at the end of longer flights.

Mike

 

[GadgetGuy]

Here is a link to the video shot earlier before the crash

Dropbox - DJI_0222.MOV

Did your video record settings in the app also record an .SRT sidecar file, along with the video, that you can make available? That will synch up some telemetry with the video, as a subtitle text overlay, if you did. It's another DJI Go setting that can be quite useful. I have mine turned on all the time, so the .SRT file is recorded whenever video is recorded.

 

[Luap]

 

[GadgetGuy]

DJI has the worst customer service of any commercial entity I have ever dealt with. And it's not just addressing issues such as yours, just getting answers to basic questions like upgrading the firmware is like pulling teeth. In part, I think the people who handle these matters haven't mastered English, but it goes beyond language skills.

I also had a loss of control incident resulting in my P3P falling into Biscayne Bay near my house in Miami Beach, where the current probably took it to the Atlantic ocean. Here is brief version of what happened. About half a mile from the Home Point in a huge and empty parking lot, I lost the signal. A few seconds later, the bird started an automatic return to home and I lost sight of it because of background clouds as it rose to the failsafe altitude of 400 feet. Eventually I received a warning on my device that the aircraft was descending. I expected to find it above my head, but instead it was about to land about 100 yards away in the middle of a busy intersection. I canceled RTH and took control manually, but the bird would not come towards me. I could make it fly in an arc, as if it were in Home Lock mode, but not toward the HP. It was as if there were an invisible barrier that it wouldn't cross. I struggled and tried everything, but it simply refused to approach until the battery ran out and it dropped in the water.

I requested a replacement from DJI (my bird was less than five months old) and provided the flight logs as well as data from Healthy Drones, which showed no user error whatsoever. DJI refused, insisting it was user error because I had continued to fly the drone until the battery ran out. Imagine that logic! Of course I continued to fly the drone until the battery ran out; I was trying to bring it home. DJI completely ignored the fact that, inexplicably, the bird just refused to approach the Home Point, and could not provide any evidence of pilot error. DJI also offered me a 15% discount on a new one. I refused; then it offered 20% as a final offer. In the meantime, I filed a protest with Paypal, which said it couldn't make a decision and we had to resolve it on our own. As I am a lawyer and was furious with DJI, I considered filing a class action lawsuit, since incidents such as these are common. But I decided it wasn't worth it and reluctantly took the offer. I looked at other options, such as Yuneec, but wasn't convinced they would be any better. Good luck in your dealings with DJI.

For what it's worth, there is now a brand new, multiple page agreement, which must first be accepted to use the latest DJI GO iOS app, after installation, which requires agreeing to mandatory arbitration and explicitly prohibits class action lawsuits, among many other terms. DJI is becoming more savvy, with American drone lawyers at the helm.

 

[GadgetGuy]

I agree ... there are ALWAYS freeky things that can happen, but you can ALWAYS cancel the RTH function and manually override and land the aircraft. I test my RTH every now and then to make sure it is working, and cancel as soon as I can see the aircraft as the GPS is not dead on and it will land in my pool if you let it land on it's own. Also .. if you were following the boat, why not use the "Follow Me" function that AUTOMATICALLY updates the home point? When I am in the boat, I always use the "Follow Me" function to update the home point. But even then, I don't really trust that I am in the exact same spot.

I have never used Follow Me, but does it really also change the Home Point dynamically, rather than simply use the controller GPS position to follow, while not changing the Home Point, unless you also choose to do so? I might want to use Follow Me on a boat, but still want to set and leave my Home Point on dry land on the shore. Are you saying that cannot be done?

 

[GadgetGuy]

I see that you started the video early on (55s into the flight) which is good because I wanted to see how far off your altitude was at that point. Going 6 minutes into the video, I see that your altitude is reported as 23.4m or 77 feet while the actual was closer to 45 or 50 feet max. A discrepancy of 20 feet is about what I've seen in my last two flights, so your drift definitely became worse as the flight moved on since the starting error looks closer to 20 feet and the error when you crashed was closer to 50 feet.

I've not flown out a full battery lately so I'm going to have to keep an eye on my altitude drift to see if it becomes worse at the end of longer flights.

Mike

There is no doubt that altitude drift increases over time when using any barometer, which is why they need to be periodically recalibrated during use over prolonged time periods by real pilots and others whose lives depend upon the accuracy of their data. The real question is how long is the flight (ours are very short compared to real pilots), and is that drift significant enough to worry about on a Phantom. I suspect not, as long as the original barometer reading was fully calibrated before launch, by whatever time is necessary for that to occur, whether 1 minute or 3 minutes or some number in between. Weather changes also change the pressure, so if the weather changes significantly during your flight, that will also impact the barometer readings, introducing error.

 

[henick]

I get compass calibration warnings... in a measurable scenario and if my IMU falls out of 1400 range or if I get any odd behavior I run a check and possibly recalibrate that. Never have I gotten a barometer warning or for that matter even ever checked for discrepancies against VPS prior to this thread.

I do now. It's a little annoying but can't help but be slightly curious...

 

[GadgetGuy]

I get compass calibration warnings... in a measurable scenario and if my IMU falls out of 1400 range or if I get any odd behavior I run a check and possibly recalibrate that. Never have I gotten a barometer warning or for that matter even ever checked for discrepancies against VPS prior to this thread.

I do now. It's a little annoying but can't help but be slightly curious...

Just knowing that it isn't 100% accurate is very helpful, to avoiding potential problems that might occur, if you otherwise assume it is completely accurate at all times.

 

[hawkeye62]

There is no doubt that altitude drift increases over time when using any barometer, which is why they need to be periodically recalibrated during use over prolonged time periods by real pilots and others whose lives depend upon the accuracy of their data. The real question is how long is the flight (ours are very short compared to real pilots), and is that drift significant enough to worry about on a Phantom. I suspect not, as long as the original barometer reading was fully calibrated before launch, by whatever time is necessary for that to occur, whether 1 minute or 3 minutes or some number in between. Weather changes also change the pressure, so if the weather changes significantly during your flight, that will also impact the barometer readings, introducing error.

OK, but, who uses altitude reading to determine a safe altitude? Not me, I use common sense to tell if I am too low and need to increase altitude. Come on guys, enough already with barometer accuracy. I am a private pilot and I **** well don't look at indicated altitude when I am landing, I look at the ground.

Regards, Jim

 

[Meta4]

if you were following the boat, why not use the "Follow Me" function that AUTOMATICALLY updates the home point? When I am in the boat, I always use the "Follow Me" function to update the home point. But even then, I don't really trust that I am in the exact same spot.

Are you sure that follow-me updates home point?
That sounds very unlikely but easy to check if you look at any flight record where you have done it.

 

[GadgetGuy]

Are you sure that follow-me updates home point?
That sounds very unlikely but easy to check if you look at any flight record where you have done it.

I am also almost certain that it doesn't.

 

[GadgetGuy]

OK, but, who uses altitude reading to determine a safe altitude? Not me, I use common sense to tell if I am too low and need to increase altitude. Come on guys, enough already with barometer accuracy. I am a private pilot and I **** well don't look at indicated altitude when I am landing, I look at the ground.

Regards, Jim

When you are flying low over water, 3 miles away, above the surf, and don't want to land, it becomes very important.

 

[hawkeye62]

When you are flying low over water, 3 miles away, above the surf, and don't want to land, it becomes very important.

Right, just increase altitude, you know, that throttle thingy. And, he wasn't three miles away from the rock.

Regards, Jim

 

[GadgetGuy]

Right, just increase altitude, you know, that throttle thingy. And, he wasn't three miles away from the rock.

Regards, Jim

Indeed. Ironically, the stick movement cancelling the elevation ascension during RTH is only relevant with the dumb Smart Return Home, as with the true No Signal RTH, the stick movement has no effect, until after you have regained signal control, and want to cancel the now unnecessary RTH!

 

[MikeyOnline]

Interesting. Looks like VPS might have picked up the boat as it went under the Phantom and it is indicating that the Phantom is 3m above the boat. That would give a rough estimate of the height above the water as about 28 feet. Maybe that's a little low, but certainly closer than the baro altitude. Right now, I'm only interested in finding out how far off the altitude reading can really get, worst case. For example, if it's not uncommon to see a 20m discrepancy, I'd like to set my RTH about 20m higher than I feel is already giving me a safe margin. Right now, I set my RTH at 60m knowing that's about 20m above any obstacles in my neighborhood where I fly most. If we can really expect to see readings that can be off by 20m toward the end of the flight, I might bump that up to 80 or even 100m.

Mike

 

[hawkeye62]

Indeed. Ironically, the stick movement cancelling the elevation ascension during RTH is only relevant with the dumb Smart Return Home, as with the true No Signal RTH, the stick movement has no effect, until after you have regained signal control, and want to cancel the now unnecessary RTH!

WTF? He had strong signal for the entire episode.

Regards, Jim

 

[henick]

While I agree you use the land and wind in actual assessments, I dislike having incorrect instruments reading at me. Our jump plane has duct tape holding the back window on and a piece of 2x4 propping the pilots seat in place. Don't mind that a bit. In fact, a cessna withOUT some length of duct tape is more suspect. When the instruments don't read right though, critical judgement can change under false pretense. Not really a fan of that. IMO of course.

That stupid wrong number wouldn't affect seeing the ground much .. besides flying BLOS when you don't have that luxury and are relying strictly on questionable telemetry ... like I said earlier it'd be clipping a treetop if you didn't leave a sufficient buffer would be my concern... anyway talked to death but my point remains.

Losing your bird because you may be misled by a faulty or inaccurate altitude reading would suck big floppy donkey nose.

 

[Meta4]

If we can really expect to see readings that can be off by 20m toward the end of the flight, I might bump that up to 80 or even 100m.

Anything more than a few feet is abnormal.
Just go back through your last 5 or 6 flight records and see what the difference is between launch and landing heights.
Of course allowing for any landings at other places or the extra few feet for a hand catch.

 

[MikeyOnline]

I looked at about a half dozen recent flights and the ones that are in the 6 to 9 minute range read about 15 feet high on landing. There were two longer ones in the 14 to 16 minute range and those were off by about 25 feet on landing (too high). Not sure what that means, but I guess it's data. It's a small sampling but from those few flights, it looks like the longer you fly, the higher the altitude reads (above actual) on landing.

Mike

 

[Meta4]

I looked at about a half dozen recent flights and the ones that are in the 6 to 9 minute range read about 15 feet high on landing. There were two longer ones in the 14 to 16 minute range and those were off by about 25 feet on landing (too high). Not sure what that means, but I guess it's data. It's a small sampling but from those few flights, it looks like the longer you fly, the higher the altitude reads (above actual) on landing.

Sounds like recalibrating the IMU is called for.
Do that on a good level surface and recheck.
That usually improves the altitude reading accuracy.