My P3P crashed, DJI not cooperating

[flyNfrank]

I'm curious where in the log you saw a CSC. I looked for that but didn't see one. As I understand it, you'd have to see 10,000 in all 4 control columns to initiate a CSC. I don't remember the sign (neg or pos) but I believe it's three -10000's and one 10000. But in any case, I didn't see 10,000 (positive or negative) in all 4 control columns at any point during the flight (except on takeoff where you'd expect it).

Mike

If the left stick would have been to the right slightly more then what the data shows(green arrows) it would have been lights out. I wont post anymore data cause in this situation it doesn't matter. I see too many holes in this story. The biggest one being there is nothing to back up the claim on hand catching the a/c. The aircraft data in 2 data type does not show the aircraft altitude lowering at any point of then when it tumbling at the very end.

 

[GadgetGuy]

Bladestrike, a respected DJI employee who used to help out and post here, has previously stated that accurate barometer readings, which are always relative to your launch point, require a 3 minute warmup period on the ground, to properly calibrate the barometer. He recommended using a spare battery just for that purpose, and then replacing that battery with a fully charged one just before takeoff. The barometer calibration, the IMU warmup, and the satellite lock will survive the reboot during the battery swapout, enabling flying almost immediately. As soon as the Home Point is recorded, you can take off, with a properly calibrated barometer. If you don't allow sufficient time for the barometer to calibrate before takeoff, the readings won't be accurate.

 

[msinger]

The biggest one being there is nothing to back up the claim on hand catching the a/c. The aircraft data in 2 data type does not show the aircraft altitude lowering at any point of then when it tumbling at the very end.

Prior to the RTH, the AC did appear to be coming down similar to how the OP described though.

 

[Chris_P]

As I've stated before, I will be going to the locating this weekend and I will post the exact GPS location of the impact. That will put all these arguments to rest .... It seems that the flight logs are viewed as indisputable evidence with no consideration given to the fact that logs simply record information provided by the instruments and sensors on the aircraft. If these sensors malfunction, which we have seen verified proof that they do, the logs are WRONG !

I may even do one better.... I will bring the Phantom and fly around the area of the impact so the same GPS coordinates are logged by the same device. I will video the phantom hovering over the impact location and show the distance to any trees in the vicinity. That should be proof enough I think !

 

[MikeyOnline]

Prior to the RTH, the AC did appear to be coming down similar to how the OP described though.

View attachment 57146

What I observed as well. There are two clear periods of descent right before the AC started moving toward that rock face. That's based on the throttle stick being held in the down position and the baro reporting about a 12m descent. While the baro reading might be in question, I don't think there is any question that there was a descent before RTH.

Mike

 

[msinger]

While the baro reading might be in question, I don't think there is any question that there was a descent before RTH.

Is there any question about the large descent at the very end of the flight? How could the log show such a large drop in altitude after the AC crashed directly into the rocks?

 

[Chris_P]

The impact with the rock possibly shook the baro into reality ... I will try to find the mark on the rock where it hit. Judging by the damage to my camera, there must be a mark left behind ! Pictures will be posted on the weekend.

Have you seen the gouges in the plastic on my Phantom I posted earlier? Those are clearly from fast moving forward contact with something hard. Not falling from a tree.

 

[msinger]

Chris_P, where were you on the boat when your Phantom took off?

 

[Chris_P]

Yes, The flight started from the boat and I was bringing it back to the boat when the RTH engaged. Both times I was standing on the swim platform.

 

[MikeyOnline]

Is there any question about the large descent at the very end of the flight? How could the log show such a large drop in altitude after the AC crashed directly into the rocks?

That's the very part I can't understand. I'm thinking maybe the AC tries to recalibrate the baro when, due to some IMU input, it determines it is on the ground? The very end of the flight does show -16m which makes me wonder if maybe the AC was in the water during that "descent" and the baro was actually picking up water pressure, causing it to show a descent. In other words, AC hits the rock, takes a second to tumble into the water, but once it enters the water, the baro reading shows a descent due to water entering the case (water would increase pressure on the baro and I think that final -16m entry might support that).

I just located the row showing the initial impact (X-axis accelerometer shows about -12G and other parameters go wonky). It's at 981.613s in the file. In that row, I copied the lat/long and pasted it into both Google and Bing maps which show the point of impact being as indicated in the attached. Bing shows it a little farther eastward of Google, but the nearest tree even from the Bing map is 30 feet away.

Mike

 

[msinger]

the nearest tree even from the Bing map is 30 feet away

You put a lot of credence in that Bing map

 

[Chris_P]

Guys, I will fly the phantom again this weekend and hover over that spot then post the logs ..... that will solve the mystery.

I can fly over near the closest tree and see what that looks like .... This will be solved !

 

[MikeyOnline]

You put a lot of credence in that Bing map

Haha. Not really. The Google map just doesn't show a scale at that level but the Bing map does. My point was that both maps corroborate the pilots account of flying into the rock. And I do think rolling down into the water is a plausible explanation of what looks like a descent at the end. Obviously water is more dense than air so when it enters the water, the pressure in the case will rise (showing a descent) until the electronics are no longer able to transmit. I'm tempted to throw my P3P in my pool and record the altitude.

Mike

 

[MikeyOnline]

Another random tidbit. During the final phase where it looks like the AC is falling (after impact), the motor load on the motors goes up by a factor of 5x to 8x yet the motors still report between 6000 and 7000 RPM. That would seem to refute my hypothesis about it being underwater: I doubt the motors could spin at 6000 RPM underwater. Thought I had a reasonable explanation for it reporting a "fall" at the end (it entered the water and water pressure affected the baro). But every hypothesis I think I find seems to have evidence either way. So I give up (for now).

Mike

 

[Chris_P]

It didn't fall into the water right away. It hit the rock, did a few cartwheels, where my props were beaten bad (see pics I posted earlier), then fell in the water. Maybe thats what you are seeing when the motors rev up that high. All 4 props were on the AC when it was fished from the water.

Someone else posted earlier that they could see the cartwheel action from the logs ... Not sure how he did that.

"thanks as I said before, it confirms that barometer's readings were off
and it also confirms, that moment of crash is at offsetTime 981.613
I counted that bird made two or three 360 rollovers in first second after the impact
after another second all sensors went crazy
and then the bird is just rolling down into water
it must have been scary to see something like that
fortunately nobody was hurt"

 

[impilot51]

Bladestrike, a respected DJI employee who used to help out and post here, has previously stated that accurate barometer readings, which are always relative to your launch point, require a 3 minute warmup period on the ground, to properly calibrate the barometer. He recommended using a spare battery just for that purpose, and then replacing that battery with a fully charged one just before takeoff. The barometer calibration, the IMU warmup, and the satellite lock will survive the reboot during the battery swapout, enabling flying almost immediately. As soon as the Home Point is recorded, you can take off, with a properly calibrated barometer. If you don't allow sufficient time for the barometer to calibrate before takeoff, the readings won't be accurate.

I redid my very unscientific test, but this time I let it set with motors running at idle for 3 minutes before taking off. I then hovered at eye level again (6 feet), confirmed I had both the barometer and the VPS reading 6 feet. Then ascended to 400 feet, came back down, went back up, came back down. Now I only have about 1 foot of error in the barometer (read 5 feet when hovering at eye level) while VPS said 6. So there is obviously something to be said for allowing the barometer to calibrate before taking off. I only lost 1% of indicated battery life while letting it idle for 3 minutes. Might be a worthy trade off for less altimeter drift. YMMV.

 

[Meta4]

That's the very part I can't understand. I'm thinking maybe the AC tries to recalibrate the baro when, due to some IMU input, it determines it is on the ground? The very end of the flight does show -16m which makes me wonder if maybe the AC was in the water during that "descent" and the baro was actually picking up water pressure, causing it to show a descent.

Obviously water is more dense than air so when it enters the water, the pressure in the case will rise (showing a descent) until the electronics are no longer able to transmit. I'm tempted to throw my P3P in my pool and record the altitude.

I wouldn't worry too much about any negative numbers following the impact.
After a crash, the sensors show all kinds of numbers.
It's common to see on-land crashes show seriously negative altitudes following a crash even though the aircraft is at or about home point.
There would have been no transmission of any data from the time the Phantom entered the water.

 

[Meta4]

There was nothing to suggest that RTH had engaged.

The log shows that the Phantom was in RTH mode.
This would have appeared on your screen as can be confirmed by a playback of the flight record.
Your app would also have been beeping.

 

[Chris_P]

The log shows that the Phantom was in RTH mode.
This would have appeared on your screen as can be confirmed by a playback of the flight record.
Your app would also have been beeping.

Keep in mind I was landing the Phantom. My eyes were on the AC not on my screen. I was in an outdoor boating environment. Whatever beeps came from my iPad were most certainly lost in my surroundings.

 

[Noodle]

I have learnt quite a bit from reading this thread and looking at Chris's log myself, thanks to those expert members here.

Anyway, a few mistakes have contributed to this crash - (1) Didn't turn off Smart return home. (2) Didn't update the home point during the flight. (3) Missed cancelling the RTH button on the screen. Means the flight is doomed. The inaccuracy in altitude reading is known to myself as I have noted it can vary by 10ft when I landed my P3A from my experience. The fact is even if your drone did clear those short trees (as shown in the photo), it may not have enough battery to fly back to its original home point and crashed anyway. And if it did have enough battery to return to its original home point, which is in the middle of the river from the look of the log map, your drone is still doomed. Chris, I think you have to take this as a very expensive lesson. If you want to pull off another one of these flights in the future when you plan to fly the battery down to such low level, you better check and double-check all your settings before hand. I doubt you are going to receive anymore discount or compensation than the 20% they are offering you now.

Best of luck.