Coordinates inaccurate from takeoff

[oshamtn]

I'm hoping to find my MA2, but given the high mountainous and overgrown terrain, I need accurate coordinates to start my search. The coordinates for takeoff and the remote landing site are approximately 19 miles off, far northwest of the location where I was flying. I know there's a built-in buffer for civilian GPS, but surely it has not expanded this far. I'll be brief with the numbers: Takeoff is recorded at 36.30N and -105.37W. It was actually 36.18, -105.23. The location where I am looking is described in the system as 36.29, -105.41, when it's actually closer to 36.17, -105.26. The drone was connected to 18 satellites at its last report, so that one should have been accurate. It was probably on or close to the ground at that point. Even if there's a reason why these coordinates might not have been accurate, that does not explain the inaccuracy of the takeoff coordinates. Any idea where the accurate coordinates might be found? As far off as these are, they do seem to be relatively aligned, but I'm not going to extrapolate from 19 miles out. Would DJI have access to those? The link on AirData is . Many thanks for any information and insight.

 

[GFields]

I'm hoping to find my MA2, but given the high mountainous and overgrown terrain, I need accurate coordinates to start my search. The coordinates for takeoff and the remote landing site are approximately 19 miles off, far northwest of the location where I was flying. I know there's a built-in buffer for civilian GPS, but surely it has not expanded this far. I'll be brief with the numbers: Takeoff is recorded at 36.30N and -105.37W. It was actually 36.18, -105.23. The location where I am looking is described in the system as 36.29, -105.41, when it's actually closer to 36.17, -105.26. The drone was connected to 18 satellites at its last report, so that one should have been accurate. It was probably on or close to the ground at that point. Even if there's a reason why these coordinates might not have been accurate, that does not explain the inaccuracy of the takeoff coordinates. Any idea where the accurate coordinates might be found? As far off as these are, they do seem to be relatively aligned, but I'm not going to extrapolate from 19 miles out. Would DJI have access to those? The link on AirData is . Many thanks for any information and insight.

• Post your flight data to get help
• Go to DJI Flight Log Viewer | Phantom Help
• Follow the instructions there to upload your flight record from your phone or tablet.
• That will give you a detailed report on the flight data.
• Come back and post a link to the report it provides.

You are also welcome to just post the txt file here.

Upload and view DJI Phantom, Mavic, and Spark flight logs online

@Meta4 is great at offering explanations from the data.

 

[Elthree]

In general I would expect the GPS coordinates later on to be more accurate.
Before you took off did you wait long enough for the GPS to stabilise and then record a good HP?

 

[Meta4]

• Post your flight data to get help

I'm on the case now.
Everything I need is available from the OP's Airdata report.
I should have something for him soon.

 

[Meta4]

I'm hoping to find my MA2, but given the high mountainous and overgrown terrain, I need accurate coordinates to start my search.

You lost signal with the drone 900 feet higher than the launch point and about 90 feet higher than the terrain below it.
Your loss of signal action was set to RTH, so the drone would have attempted to RTH when signal was lost.
However, you reset the drone's homepoint to the current location of the drone at 8:58.4 when the drone was 12296 feet away from home.

The location to which the drone would have headed (the new home point) was 36.29184 -105.41733.

It looks like the wind was quite low so the drone should have autolanded straight down from where it was and be somewhere close to 36.29184 -105.41733.

 

[oshamtn]

In general I would expect the GPS coordinates later on to be more accurate.
Before you took off did you wait long enough for the GPS to stabilise and then record a good HP?

Yep.

 

[Meta4]

The coordinates for takeoff and the remote landing site are approximately 19 miles off, far northwest of the location where I was flying.

You had good GPS to start the flight.
I'm curious about what makes you think there was a 19 mile inaccuracy in your launch point location.

I know there's a built-in buffer for civilian GPS, but surely it has not expanded this far.

There's no "buffer" in consumer grade GPS, but there may be an inaccuracy of 6-10 feet.

The location where I am looking is described in the system as 36.29, -105.41, when it's actually closer to 36.17, -105.26.

Why do you think that?
I can't see any reason to distrust the numbers in your recorded flight data.

 

[oshamtn]

You lost signal with the drone 900 feet higher than the launch point and about 90 feet higher than the terrain below it.
Your loss of signal action was set to RTH, so the drone would have attempted to RTH when signal was lost.
However, you reset the drone's homepoint to the current location of the drone at 8:58.4 when the drone was 12296 feet away from home.

The location to which the drone would have headed (the new home point) was 36.29184 -105.41733.

It looks like the wind was quite low so the drone should have autolanded straight down from where it was and be somewhere close to 36.29184 -105.41733.

Thanks, but the problem is those coordinates are approximately 19 miles from the point shown on the map at the end of the flight. The coordinates shown from my TO location are also approximately 19 miles from that location. Neither of those locations are actually in the mountains. Both are on adjacent plains far outside the range of the MA2.

 

[Meta4]

Thanks, but the problem is those coordinates are approximately 19 miles from the point shown on the map at the end of the flight. The coordinates shown from my TO location are also approximately 19 miles from that location. Neither of those locations are actually in the mountains. Both are on adjacent plains far outside the range of the MA2.

If you are correct, yours is the first casein hundreds I've examined where GPS has been more than a few feet out.
I can't understand how anything could cause the inaccuracy you describe and would tend to trust the actual numbers.

 

[oshamtn]

You had good GPS to start the flight.
I'm curious about what makes you think there was a 19 mile inaccuracy in your launch point location.


There's no "buffer" in consumer grade GPS, but there may be an inaccuracy of 6-10 feet.

Why do you think that?
I can't see any reason to distrust the numbers in your recorded flight data.

I checked both locations on my Garmin GPS unit on the first trip I made to look for the drone. My iPhone confirmed the Garmin numbers. I then plotted both locations on USFS maps for Carson National Forest, Camino Real Ranger District, as well as the appropriate Geological Survey Maps (Osha Mountain and Shady Brook). I then plotted the DJI system's coordinates for takeoff and landing on the USFS and GS maps. Both show a discrepancy of approximately 19 miles. These coordinates would have had me taking off and landing out on the adjacent plains, which are just over 7,000 ft above sea level -- not in the mountains at more than 9,000 ft.

 

[oshamtn]

If you are correct, yours is the first casein hundreds I've examined where GPS has been more than a few feet out.
I can't understand how anything could cause the inaccuracy you describe and would tend to trust the actual numbers.

I agree. That's why I'm asking. If the system was able to accurately plot the course of the MA2 (and I had two spotters, so I know that course is accurate), then surely the accurate coordinates are buried in there somewhere.

 

[oshamtn]

If you are correct, yours is the first casein hundreds I've examined where GPS has been more than a few feet out.
I can't understand how anything could cause the inaccuracy you describe and would tend to trust the actual numbers.

I don't understand it, either. That's why I'm asking. The mystery most useful to solve is this: If the DJI system can plot the course accurately on the map, which it did (and I had two spotters upslope, and I watched the video display on a fullsize iPad as the MA2 flew upslope until it cut off), then the accurate coordinates could well be buried in the system somewhere. But the numbers that display are erroneous, and not just by a little bit.

 

[mobilehomer]

Just for fun, I downloaded the KML and imported it into Earth. Drone is 2.33 miles from takeoff point. Does this look correct?

 

[Meta4]

I checked both locations on my Garmin GPS unit on the first trip I made to look for the drone. My iPhone confirmed the Garmin numbers.

The GPS receiver in the drone uses exactly the same satellites as your phone and Garmin.
It's a Ublox module very reliable and used in millions of devices worldwide.
I've used GPS professionally since the early days of GPS and can't think of any way that GPS data for a whole flight would be transposed to a location 19 miles away.
Are you saying that your flight didn't occur where shown on Google Earth?

 

[oshamtn]

Just for fun, I downloaded the KML and imported it into Earth. Drone is 2.33 miles from takeoff point. Does this look correct?

Yes, that looks right. Wonderful -- I think you nailed it, mobilehomer!
Any way you can display this location on a more detailed map overlay?
Also, just curious: Did you get the coordinates for the takeoff location?
At this point I'm more focused on getting back up there and searching with these realistic coordinates before the snow flies -- but I'm still perplexed as to why the KML coordinates are so different from the coordinates that display in the CSV file, and why the latter coordinates are so wrong.

 

[mobilehomer]

Yes, that looks right. Wonderful -- I think you nailed it, mobilehomer!
Any way you can display this location on a more detailed map overlay?
Also, just curious: Did you get the coordinates for the takeoff location?
At this point I'm more focused on getting back up there and searching with these realistic coordinates before the snow flies -- but I'm still perplexed as to why the KML coordinates are so different from the coordinates that display in the CSV file, and why the latter coordinates are so wrong.

The coordinates are from the detail tab on your Airdata page. My guess is that the map loaded incorrectly when you looked at it.

 

[oshamtn]

Then it did so on Apple Maps and Google Maps multiple times. Same thing on Phantom. The CSV file displays the wrong coordinates, as well.

 

[Meta4]

Yes, that looks right. Wonderful -- I think you nailed it, mobilehomer!
Any way you can display this location on a more detailed map overlay?
Also, just curious: Did you get the coordinates for the takeoff location?

Your launch point was at 36.30694 -105.37993

At this point I'm more focused on getting back up there and searching with these realistic coordinates before the snow flies

These "realistic co-ordinates" are exactly the same ones that I gave you in post #5 and the same ones shown in your recorded flight data.

-- but I'm still perplexed as to why the KML coordinates are so different from the coordinates that display in the CSV file, and why the latter coordinates are so wrong.

The co-ordinates are not wrong.
The KML co-ordinates are exactly the same as those in the CSV.
The data for the KML file comes from the same recorded flight data.

the alleged 19 mile error is not real (it would be impossible for it to be) and comes from something in your working.
There is no 19 mile error.
Forget about it and use the location that I gave you originally.

 

[oshamtn]

You are correct that the points are displayed on the map properly, both home point and landing point and the course of the flight. But the coordinates describing those points are inaccurate. They are both out west of Taos Pueblo, not on the west slopes of the Sangre de Cristos. Plot those coordinates on the USFS or geological service maps and see where you are. Mobilehomer found the correct coordinates in the KML file.

 

[Meta4]

But the coordinates describing those points are inaccurate. They are both out west of Taos Pueblo, not on the west slopes of the Sangre de Cristos.

Check the way you calculated that ... it's just not right.
And read my (updated) post above.