Pursuing a refund - Flyaway not found...

[ianwood]

The compass heading is only relevant if there was a commanded direction of travel. If the OP put in forward elevator and the GPS course did not reflect the heading, that could indicate a compass issue. But there was no commanded direction of travel that corresponded to the actual movement. At least, according to the logs. The Phantom moved on its own.

Looking at the video, you can see from the orientation of the Phantom, that it flew to the left and not forward. We don't know why it did that. And it doesn't mean the Phantom actually did that.

 

[III% Streve]

Ok... makes sense.. but that commanded movement could have also come from attempting to maintain its position based on GPS couldn't it? Like... lift off with a slight drift doe to wind of gound contact? Mine frequently drifts a little on takeoff, and quickly stabilizes once out of ground effect.

 

[BudWalker]

That's what I was thinking. I.e. that GPS based stabilization results in non-OP commands to move the P3 back to position. But, I'm not totally confident about this because it would seem the deviations are smaller than GPS accuracy.

Regardless of whether this fly-away was due to compass error I'm still interested to know if it's possible to determine P3 heading by using data other than the compass heading.

But, is it possible to determine the P3 heading from the flight data that is independent from the measured compass heading? I'm attempting to do this by writing a program that computes speed and heading from OSD.longitude and OSD.latitude and then comparing that to OSD.xSpeed, OSD.ySpeed, and OSD.headingCompass. The problem is that OSD.xSpeed and OSD.ySpeed may have already been derived from (i.e. is dependent on) OSD.longitude, OSD.latitude. Does anybody know where OSD.xSpeed and OSD.ySpeed come from? In the program that I wrote it does seem like OSD.xSpeed and OSD.ySpeed are independent of OSD.latitude and OSD.longitude. It'd be nice if they were somehow derived from IMU measurements, although I don't see how this is possible.

What is OSD.yaw measuring? I'm assuming that OSD.pitch and OSD.roll are relative to what the P3 thinks is level. Is OSD.yaw relative to the front of the P3?

In essence if 1) the direction of flight is known from the GPS longitude, latitude data and 2) the P3's orientation to that direction of flight can be derived without using the compass heading then it should be possible to determine the P3 heading without using the compass heading.

 

[TAZ]

So I have an estimate of where you will find your bird:

How did I calculate this? Well, I looked at your FlightRecord.csv. I made a few assumptions. We know your GPS coordinates over time. Towards the end of the flight based on last known location and direction of travel, I utilized excel trend functions (known time, known Lat, known long) and the battery level decrease over the known flight time. With this data, I projected where the bird would end up when your battery hit 10% (critical land). Then I, used the changes of Lat/Long, as no one knows for sure if the Compass Heading was correct. In addition, the lat/long changes over time will appropriatly reflect both direction, and prevailing wind (assuming that wasn't variable)

Unfortunately given the distance you traveled in 1:13minutes...this projection resulted in your bird flying 14km away before it ran out of battery. Pretty unlikely to find it...and certainly wind or heading variations would make this a pretty large searchable area.

However, I then overlayed a polygon in Google Earth Pro...and realized that your projected flight path actually travels over constantly increasing Ground Level. Assuming the bird did not continue to climb (which it seems to be holding about 57m)...it likely crashed where this flight path polygon intersected the ground. I lowered the flight path height estimate by 20ft (guessing that surrounding obstructions could reach that high)...and where the flight path estimate polygon (the yellow shade triangle) intersects the ground....is likely where the bird hit an obstruction.

This area is relatively small.

If you'd like to see the models I used, you can upload this Google Earth Pro .kmz file and check out my polygons in more detail: https://onedrive.live.com/redir?resid=9D9108EB917328CA!1575&authkey=!AD0F4815oUMGUKU&ithint=file,kmz

I used a couple models...using larger sample sizes of your known GPS track...and surprisingly it didn't vary the ultimate location by much. Perhaps this will help.

For reference sake, here is a zoomed out view of the estimate:

EDIT...I use Google Earth Pro. Not sure if the polygons show in earlier versions. However, you can download Google Earth Pro for free now. Just use GEPFREE as the license key when you install it.

 

[III% Streve]

Thats some pretty detailed work! The only variable that i see as important, but missing, is the the constant, yet slow drift clockwise.

 

[TAZ]

Thats some pretty detailed work! The only variable that i see as important, but missing, is the the constant, yet slow drift clockwise.

I completely agree. That is the biggest variable (unless we are to consider variable wind). However, I did look at the data we do have and even accounting for a clockwise drift...the search area does seem within error.

An additional variable is the height of obstructions. I only used 20'...which is probably too low. Given trees in the area 40' might be more appropriate. When I get home I'll make an update that refines the area a bit. However, on the path given...I'd say most likely locations are in the SE side of the red box...and within the lower portions of the overlap of the red and yellow areas.

 

[flyNfrank]

This thread went wrong in a hurry. Other then when I'm beta testing app's, I do spend a lot of time on this P3 forum. I do not go to any other forums on this website unless someone moved my thread without giving notice.

I hope the rudeness is over with as there has been way too much of it. I hope the moderator's wake up and look at the IP addresses (which btw need to be viewable by all) and get rid of those posting with 2nd accounts.

Now as for the Image below it is of the OP's 1.13 minute flight. I use this as a tool when evaluating my own flights. The only thing I wanted the OP to see was at the very end you can see both the Altitude and Speed are nosing over together. This gives me some indication the flight may have ended at this point as the flight log did.

 

[djivinz]

i had the same problem (immediately flying left after take off), but i didn't use auto-takeoff. my bird did not respond to my request to return to home (other than turn to face home as it continued on its crazy flight away from me). i lost connection to it about a mile away. i jumped in my car and drove to the last known spot, and i was able to reconnect. it had drifted a hundred feet or so from that spot and was slowly drifting further away. in my panic, i didn't think of switching to ATTI mode (though i have read that it allows the pilot to regain control), but will if it happens in the future. i learned a big lesson in that it pays to not turn off your controller and to take it with you when searching for a lost Phantom. i also purchased a retractable dog leash to attach to the bottom of my Phantom to prevent this type of fly-away in the future. once i get it up and confirm that it is stable, i will detach the leash and continue flying.

From a newb perspective, I am taking the lessons of the others unfortunate experience. Here it would be to not use auto take off and switch in ATTI mode.

To switch to this mode, does multi flight mode need to be enabled? It is just the three step switch on the left hand to move in the middle, right?

 

[Leon001]

From a newb perspective, I am taking the lessons of the others unfortunate experience. Here it would be to not use auto take off and switch in ATTI mode.

To switch to this mode, does multi flight mode need to be enabled? It is just the three step switch on the left hand to move in the middle, right?

That's what I think. Better to have a damaged drone than no drone at all.

 

[atki]

From a newb perspective, I am taking the lessons of the others unfortunate experience. Here it would be to not use auto take off and switch in ATTI mode.

To switch to this mode, does multi flight mode need to be enabled? It is just the three step switch on the left hand to move in the middle, right?

yeh multi flight mode enabled in the app and then you can use the tx toggle switch to select which mode you want. You'll also need to toggle IOC in the app if you want to use that.

 

[atki]

That's what I think. Better to have a damaged drone than no drone at all.

Leon there doesnt need to be a damaged drone if you use Atti mode, good mode to practice in a nice open area and just have it in the back of your mind to switch into if things go pear shaped.

 

[Leon001]

Leon there doesnt need to be a damaged drone if you use Atti mode, good mode to practice in a nice open area and just have it in the back of your mind to switch into if things go pear shaped.

Sorry. I should qualify my statement. By the sounds of things I could have activated atti mode to stop my drone in its tracks at any point before I lost connection with it. In that situation, I'm not sure how well I could have guided the it back (it being more difficult) and it may have been damaged.

All I mean is the risk is worth it. Beyond question.

 

[Leon001]

This thread went wrong in a hurry. Other then when I'm beta testing app's, I do spend a lot of time on this P3 forum. I do not go to any other forums on this website unless someone moved my thread without giving notice.

I hope the rudeness is over with as there has been way too much of it. I hope the moderator's wake up and look at the IP addresses (which btw need to be viewable by all) and get rid of those posting with 2nd accounts.

Now as for the Image below it is of the OP's 1.13 minute flight. I use this as a tool when evaluating my own flights. The only thing I wanted the OP to see was at the very end you can see both the Altitude and Speed are nosing over together. This gives me some indication the flight may have ended at this point as the flight log did.

Thanks for the input.

Clarify for me, why would a stable speed coupled with a stable altitude indicate that the done went down at that point?

Please correct me if I'm not understanding your point.

Regards

 

[SENC]

This thread went wrong in a hurry. Other then when I'm beta testing app's, I do spend a lot of time on this P3 forum. I do not go to any other forums on this website unless someone moved my thread without giving notice.

I hope the rudeness is over with as there has been way too much of it. I hope the moderator's wake up and look at the IP addresses (which btw need to be viewable by all) and get rid of those posting with 2nd accounts.

Now as for the Image below it is of the OP's 1.13 minute flight. I use this as a tool when evaluating my own flights. The only thing I wanted the OP to see was at the very end you can see both the Altitude and Speed are nosing over together. This gives me some indication the flight may have ended at this point as the flight log did.

Interesting, Frank. Are you pointing to the very last datapoint or two where it looks like the red and blue lines are beginning to drop? What tool is this?

 

[Leon001]

@III% Streve & @TAZ

What is the source of the discrepancy between the distances of the drones flight?

 

[JamesBandes]

Please post the flight log from your device.

Why not cancel RTH and try to fly it back in Atti mode? I feared this so I purchased a good GPS Tracker and holder.

I would like to start this post and this thread by saying that I absolutely loved flying the Phantom 3. It was like a dream to get my hands on it and I felt like a 10 year old on Christmas day when the package arrived a week ago.

Since receiving the P3 I have flown it every day, and have used it in the most prosaic manner I could have done. In fact, until today I hadn't flown it further than 300 or so metres, not higher than 70.

Today I made two flights, the first lasting no longer than a few minutes, but the second was the last time I've seen my P3.

The flights were within several minutes of one another.

On the first flight, once I'd reached the furthest point of the flight, I hit the home button and returned safely. All was well and good.

On the second flight, I took off and immediately found the P3 being dragged to my left and then started to accelerate and ascend upwards - all the while my fingers were off the controls. I tried to steer the P3 toward me, ensuring that the drone's front end was pointing toward me, but the drone wasn't reacting to the RC. The P3 must have been no further than 50 metres from me when I pressed the return to home button. Needless to say, and to my surprise, the drone carried on along it's unknown trajectory. The only lead I have is the direction the drone was heading in prior to my live camera view and flight log on the pilot app losing connection to the drone. I have been unable to locate the drone for the past 16 hours.

Now here's the issue, I did all my preflight checks for my first flight and everything was fine. I even tested the return to home feature. Additionally, I have flight logs on the app of both flights. Both show the home-points and the RTH feature being activated. Yet for some reason the drone malfunctioned and ignored this - EVEN THOUGH THE FLIGHT LOG SHOWS IT WAS ACTIVATED.

I feel that I've taken every precaution and have proof that this fly away was not pilot error. The logs show I activated RTH and had my home point set, and the first successful flight, ending only minutes before, shows that everything was working properly up until that point. I typically get 17 satelites while flying from that spot and today was no different.

After reading other posts, DJI seem to have a reputation for ascribing flyaways to pilot error (which may be true in some cases) but I really don't see how anyone could question my evidence. Any thoughts?

Additionally, can someone advise me how to record my screen or download the flight logs so everyone is able to see them on the forum.

I really want to get this sorted in some way because this is an amazing product. But I can't bare the thought of having to shoulder this loss myself as I've acted with the careful handling equipment like this deserves.

Regards,

I would like to start this post and this thread by saying that I absolutely loved flying the Phantom 3. It was like a dream to get my hands on it and I felt like a 10 year old on Christmas day when the package arrived a week ago.

Since receiving the P3 I have flown it every day, and have used it in the most prosaic manner I could have done. In fact, until today I hadn't flown it further than 300 or so metres, not higher than 70.

Today I made two flights, the first lasting no longer than a few minutes, but the second was the last time I've seen my P3.

The flights were within several minutes of one another.

On the first flight, once I'd reached the furthest point of the flight, I hit the home button and returned safely. All was well and good.

On the second flight, I took off and immediately found the P3 being dragged to my left and then started to accelerate and ascend upwards - all the while my fingers were off the controls. I tried to steer the P3 toward me, ensuring that the drone's front end was pointing toward me, but the drone wasn't reacting to the RC. The P3 must have been no further than 50 metres from me when I pressed the return to home button. Needless to say, and to my surprise, the drone carried on along it's unknown trajectory. The only lead I have is the direction the drone was heading in prior to my live camera view and flight log on the pilot app losing connection to the drone. I have been unable to locate the drone for the past 16 hours.

Now here's the issue, I did all my preflight checks for my first flight and everything was fine. I even tested the return to home feature. Additionally, I have flight logs on the app of both flights. Both show the home-points and the RTH feature being activated. Yet for some reason the drone malfunctioned and ignored this - EVEN THOUGH THE FLIGHT LOG SHOWS IT WAS ACTIVATED.

I feel that I've taken every precaution and have proof that this fly away was not pilot error. The logs show I activated RTH and had my home point set, and the first successful flight, ending only minutes before, shows that everything was working properly up until that point. I typically get 17 satelites while flying from that spot and today was no different.

After reading other posts, DJI seem to have a reputation for ascribing flyaways to pilot error (which may be true in some cases) but I really don't see how anyone could question my evidence. Any thoughts?

Additionally, can someone advise me how to record my screen or download the flight logs so everyone is able to see them on the forum.

I really want to get this sorted in some way because this is an amazing product. But I can't bare the thought of having to shoulder this loss myself as I've acted with the careful handling equipment like this deserves.

Regards,

awful. Did you try canceling RTH and flying it back in Atti mode? I bought a GOOD GPS tracker with a holder just in case.

 

[Leon001]

Why not cancel RTH and try to fly it back in Atti mode? I feared this so I purchased a good GPS Tracker and holder.


awful. Did you try canceling RTH and flying it back in Atti mode? I bought a GOOD GPS tracker with a holder just in case.

Honestly, i had never used or heard about atti mode before joining this forum.

 

[TAZ]

@III% Streve & @TAZ

What is the source of the discrepancy between the distances of the drones flight?

I think Steve (but he'll have to clarify) was assuming the bird went down shortly after telemetry ended...where my model shows where the ground (or obstructions) would have risen into the path of the bird, had it continued to hold its altitude.

 

[ianwood]

Let's not jump to conclusions:

• Nobody knows if switching to ATTI would have helped in this case.
• Nobody knows if this event was caused by atuo take off.

Until we know what caused it, we cannot say what would prevent or stop it.

 

[TAZ]

Now as for the Image below it is of the OP's 1.13 minute flight. I use this as a tool when evaluating my own flights. The only thing I wanted the OP to see was at the very end you can see both the Altitude and Speed are nosing over together. This gives me some indication the flight may have ended at this point as the flight log did.

@flyNfrank...I don't see a great change in those values. I am thinking your chart might be exaggerating the slope due to the fact that the last few pieces of data are 1.2 seconds apart (where most previous data points are .1 seconds apart). Is your X axis a function of time? If not, the last few data points might be inappropriately compressed...exaggerating the down slope. Speed and altitude don't seem to be varying much, when you take time into account.

Here's a chart of speed...charted as a function of time.