Drone did not RTH!

[m0j0]

Sorry, I'm not buying the... I calibrated the compass which allowed me to properly get the Mavic to RTH.

If you calibrate in the same area every time what is the point of doing that. (It's the same area)

 

[m0j0]

Drone is lost and I have sent my flight records to DJI for review. In reading posts since I lost it on 8/10, others have had similar problems but, like you, were able to return the Mavic safely. DJI is not getting real good press regarding their customer service so I am waiting to read what DJI determines caused the "fly away" and RTH failure. The Mavic was less than one week old!

DJI customer service has been perfect with me. I'm on my 6th DJI product. They have been close to flawless. I have sent my birds in twice for small issues and had great results.

However, what exactly are you expecting from this situation? I'm not saying it as a certainty but It appears that your mobile device sent bad gps data to your Mavic. The Mavic didn't "fly away" it flew where it was directed to go by your mobile device. Not the Mavic or DJI's fault. Even though it went as directed to the old location, ... you could have cancelled this easily.

Good luck

 

[Lone Droner 10]

So, when I hear "Home point updated" - Then I'm not going to experience what happened in the original post. Thanks Msinger!

My understanding is the Mavic sends data to the controller and then if you have an android or IOS device connected and are using the DJI GO 4 app and when it connects to the controller and then the device and says "the home point is updated, check it on the map" you should be fine when you activate RTH. Not sure about the post that stated if your iPad only has wifi, you do not have gps.

I am the original person who experienced the "fly away" after activating RTH and then having to select either current location or original location. As I stated in my original post, when I used the auto takeoff on the app, I then selected "precise" home point and it took off as usual. I flew it for about 1,000' hit RTH, up came the two options, selected current location and the next thing I saw was the telemetry showing me 45,000' feet away. The rest was history - off it went. I am still waiting for DJI to diagnose the flight data I sent them a week ago. This is the second Mavic I have lost due to failure of the RTH. Not real happy with DJI.

 

[msinger]

Not sure about the post that stated if your iPad only has wifi, you do not have gps.

I guess you're referring to my post? If so, I meant the issue you experienced cannot occur if someone is using a Wi-Fi only mobile device. Since it does not contain a GPS receiver, there is no way to set the home point to the current location of the mobile device.

 

[sar104]

sar104
Question before I proceed further. "Should invocation of Master Control -> Home Point to Me (little man icon), result in a new HP, thus overriding the launch HP?"

That dialog box allows you to change the coordinates of the home point, resetting it either to the current location of the aircraft or the current location of the mobile device. Somewhat obviously, the aircraft coordinates are supplied by the FC GPS, while the mobile device coordinates, if you choose that option, are supplied by the mobile device's internal GPS. If those are wrong then the home point will not be in the right place.

As you can see from the screen shots that you posted, after you reset the home point to the controller position there was no longer a home point icon (H) visible on your map. That strongly suggests that the reset was successful but that the coordinates were incorrect. The blue dot, indicating the controller, is in view in the screen shots after you reset but, without knowing what it was reporting at the moment you did the reset, it is impossible to reconstruct the events. If you posted the DAT file then we could find out, since it continuously reports home point, but you have not done that.

 

[scjerry]

That dialog box allows you to change the coordinates of the home point, resetting it either to the current location of the aircraft or the current location of the mobile device. Somewhat obviously, the aircraft coordinates are supplied by the FC GPS, while the mobile device coordinates, if you choose that option, are supplied by the mobile device's internal GPS. If those are wrong then the home point will not be in the right place.

As you can see from the screen shots that you posted, after you reset the home point to the controller position there was no longer a home point icon (H) visible on your map. That strongly suggests that the reset was successful but that the coordinates were incorrect. The blue dot, indicating the controller, is in view in the screen shots after you reset but, without knowing what it was reporting at the moment you did the reset, it is impossible to reconstruct the events. If you posted the DAT file then we could find out, since it continuously reports home point, but you have not done that.

sar104
OK, I'm dubious about the assumption (belief) in incorrect coordinates from my mobile device. I agree that precision could be an issue. I can't find the DAT files, only God awful .txt ones, and only use the CVS files from AirData, which report the GPS coordinates from the flight controller to 8 decimal places, then I can match launch and landing values. My MD's, (I have multiple) report to 6 decimal places and agree with those from the FC to within less than a couple meters (probably the last 2 decimal places), e.g. I stand over my live Mavic at landing and launch and take a measurement using a GPS app.

I've conducted about 20+ short flights to test the ability of DJI Go 4 to reset the HP to my MD location. They're all on AirData. I've done a lot of due diligence and I won't bore the list with the details. Less than a few were partially successful. Here's a failure: Airdata UAV - Flight Data Analysis for Drones . Even invoking the Follow Me mode, then attempting to reset the HP didn't work. Then even RTH to the launch HP was off by a couple yd. The partial successes were deemed so,since, they too, didn't RTH to the HP precisely.

I believe that the Go app is corrupting the HP data that I assume resides in the controller (RC) data store. I assume this location because it is the only place that is immune from an MD disconnect/failure and allows the AC to go home in case of the MD failure. To that belief (here I am with beliefs again), there is Litchi. See Airdata UAV - Flight Data Analysis for Drones . It executed perfectly and landed w/i inches of the launch position. Same hardware, no interference warnings, no spurious no-fly-zone warnings and update demands in mid-flight, just good execution. Right now, their IOS version does not have HP-reset to MD.

I've reported this issue to DJI through their DJI Go 4 (lower right screen) Me-> DJI Support -> Bug Report. I will not hold my breath, but will use Litchi without dynamic HP setting, instead. They dynamically update their Google Maps I/F, so I can use all of their mapping modes. DJI Go 4 does not.

BTW, my version of DJI Go 4, v.4.1.9 running under iOS and my iPad does not have a drop down option for selecting the current position of the drone as a new HP under the Master Control menu, but it does under Advanced Flight Modes and I missed it. I obviously would have chosen that for comparison. Back to some testing.

 

[sar104]

sar104
OK, I'm dubious about the assumption (belief) in incorrect coordinates from my mobile device. I agree that precision could be an issue. I can't find the DAT files, only God awful .txt ones, and only use the CVS files from AirData, which report the GPS coordinates from the flight controller to 8 decimal places, then I can match launch and landing values. My MD's, (I have multiple) report to 6 decimal places and agree with those from the FC to within less than a couple meters (probably the last 2 decimal places), e.g. I stand over my live Mavic at landing and launch and take a measurement using a GPS app.

I've conducted about 20+ short flights to test the ability of DJI Go 4 to reset the HP to my MD location. They're all on AirData. I've done a lot of due diligence and I won't bore the list with the details. Less than a few were partially successful. Here's a failure: Airdata UAV - Flight Data Analysis for Drones . Even invoking the Follow Me mode, then attempting to reset the HP didn't work. Then even RTH to the launch HP was off by a couple yd. The partial successes were deemed so,since, they too, didn't RTH to the HP precisely.

I believe that the Go app is corrupting the HP data that I assume resides in the controller (RC) data store. I assume this location because it is the only place that is immune from an MD disconnect/failure and allows the AC to go home in case of the MD failure. To that belief (here I am with beliefs again), there is Litchi. See Airdata UAV - Flight Data Analysis for Drones . It executed perfectly and landed w/i inches of the launch position. Same hardware, no interference warnings, no spurious no-fly-zone warnings and update demands in mid-flight, just good execution. Right now, their IOS version does not have HP-reset to MD.

I've reported this issue to DJI through their DJI Go 4 (lower right screen) Me-> DJI Support -> Bug Report. I will not hold my breath, but will use Litchi without dynamic HP setting, instead. They dynamically update their Google Maps I/F, so I can use all of their mapping modes. DJI Go 4 does not.

BTW, my version of DJI Go 4, v.4.1.9 running under iOS and my iPad does not have a drop down option for selecting the current position of the drone as a new HP under the Master Control menu, but it does under Advanced Flight Modes and I missed it. I obviously would have chosen that for comparison. Back to some testing.

The DAT files are easy to get: How to retrieve a .DAT

DJI GO 4 v 4.1.9 on the iPad certainly does have the menu (screen shot below) for selecting a new HP, as described on p. 46 of the v1.6 or v1.8 Mavic manual. You get to it by swiping left in camera view.



Interesting that your mobile devices appear to be reporting correct positions. That does cast some suspicion onto the app itself, and there have been reports, mostly on Android, of the app sending incorrect position data, although none that I'm aware under iOS. In terms of expectations, a couple of yards is well within the expected accuracy of the GPS system.

So - just as the simplest possible test, without Follow Me or any other complications:

1. Power everything up standing next to the aircraft;
2. Wait for the message that the HP has been set, check the map that it is at the correct (aircraft) location;
3. Walk away 50 yds or so with the controller;
4. Choose "Home Point: Me" and wait for confirmation;
5. Check the Home Point location on the map.

Does that work, or do you get the Home Point somewhere distant and inapporpriate?

 

[scjerry]

OK, well here I am awake at this hour waiting for Irma reports.We have some interests in Florida.

Yes, I saw the Home Point: Aircraft display in my docs, too, with no narrative or explanation or mention of the HP reset under the Main Control menu, hence my "BTW" at the end of my note.

What an odd interface!

So, I have some tolerance for s/w bugs, but none for poor documentation and sloppy human interfaces, although given the complexity of dealing with over 20 processors, and multitasking with asynchronous events, the DJI drone system is a marvel, giving credit where credit is due.

Having done some programming in my previous life, I became a DJI SDK developer (in name only) so I could gain some insight from the SDK docs. Here's one example from: Flight Controller - DJI Mobile SDK Documentation


Returning Home
The aircraft can automatically return-to-home (RTH) in a number of scenarios:

• Smart RTH: Commanded to by the pilot through the application or the remote controller
• Failsafe RTH: If the wireless link is lost between the remote controller and aircraft
• Low Battery RTH: If the battery drops below a threshold that is enough to get home, but not enough to require immediate emergency landing

When automatically going home, the aircraft will rise to a minimum altitude, fly to the home location (home point) using GPS positioning, then land.

The home point is automatically set as the location the aircraft first takes off from after power on. After that, the home point can be updated through APIs in the DJI Mobile SDK but is limited to being within 30m of initial take-off location, current aircraft's location, current mobile location, or current remote controller's location (for remote controllers with GPS capability).

Note: If the GPS signal is not sufficient during take-off to record a home location, the home point will be recorded when the GPS signal is strong enough. When taking off in poor satellite signal environments, developers should ensure the home point being set is within the user's expectations.

I thought about this yesterday and agree with attempts to reset the HP at greater distances that I was using in my tests.

Also, the launch HP isn't recorded/reset unless it takes off from a fresh power-on state. Just launching it again after a landing w/o a power down doesn't reset the HP.

OMG!

 

[sar104]

OK, well here I am awake at this hour waiting for Irma reports.We have some interests in Florida.

Yes, I saw the Home Point: Aircraft display in my docs, too, with no narrative or explanation or mention of the HP reset under the Main Control menu, hence my "BTW" at the end of my note.

What an odd interface!

So, I have some tolerance for s/w bugs, but none for poor documentation and sloppy human interfaces, although given the complexity of dealing with over 20 processors, and multitasking with asynchronous events, the DJI drone system is a marvel, giving credit where credit is due.

Having done some programming in my previous life, I became a DJI SDK developer (in name only) so I could gain some insight from the SDK docs. Here's one example from: Flight Controller - DJI Mobile SDK Documentation


Returning Home
The aircraft can automatically return-to-home (RTH) in a number of scenarios:

• Smart RTH: Commanded to by the pilot through the application or the remote controller
• Failsafe RTH: If the wireless link is lost between the remote controller and aircraft
• Low Battery RTH: If the battery drops below a threshold that is enough to get home, but not enough to require immediate emergency landing

When automatically going home, the aircraft will rise to a minimum altitude, fly to the home location (home point) using GPS positioning, then land.

The home point is automatically set as the location the aircraft first takes off from after power on. After that, the home point can be updated through APIs in the DJI Mobile SDK but is limited to being within 30m of initial take-off location, current aircraft's location, current mobile location, or current remote controller's location (for remote controllers with GPS capability).

Note: If the GPS signal is not sufficient during take-off to record a home location, the home point will be recorded when the GPS signal is strong enough. When taking off in poor satellite signal environments, developers should ensure the home point being set is within the user's expectations.

I thought about this yesterday and agree with attempts to reset the HP at greater distances that I was using in my tests.

Also, the launch HP isn't recorded/reset unless it takes off from a fresh power-on state. Just launching it again after a landing w/o a power down doesn't reset the HP.

OMG!

I don't have any particular complaints about the interface. The home point reset was moved from settings to the action screen, probably because it is not a setting, but an action. It is clearly documented in the manual. As for when the home point is set, that has always been at power on for all DJI aircraft and, again, is clearly documented.

Back on the specific problem you reported - I noticed that back in January there were multiple complaints of similar behavior on the P3S with an early build of GO 3.1.2 . It was clearly an app problem, not a hardware problem, and DJI fixed it with an updated build of the same version. Given that you are quite certain that the app was displaying the correct controller location it sounds suspiciously like a similar problem. I have three different GO app versions that I can test with both Mavic and Phantom, so I'll see if I can replicate the issue.

 

[scjerry]

I'll conduct more experiments as soon as the weather lifts.

 

[scjerry]

I don't have any particular complaints about the interface. The home point reset was moved from settings to the action screen, probably because it is not a setting, but an action. It is clearly documented in the manual. As for when the home point is set, that has always been at power on for all DJI aircraft and, again, is clearly documented.

Back on the specific problem you reported - I noticed that back in January there were multiple complaints of similar behavior on the P3S with an early build of GO 3.1.2 . It was clearly an app problem, not a hardware problem, and DJI fixed it with an updated build of the same version. Given that you are quite certain that the app was displaying the correct controller location it sounds suspiciously like a similar problem. I have three different GO app versions that I can test with both Mavic and Phantom, so I'll see if I can replicate the issue.

Have you been following these threads? :

Home Point to RC not registered
Serious bug in the Mavic's "Home Point Me" function - caused loss of drone!

I've successfully submitted a bug report and now have a number from DJI.

 

[sar104]

Have you been following these threads? :

Home Point to RC not registered
Serious bug in the Mavic's "Home Point Me" function - caused loss of drone!

I've successfully submitted a bug report and now have a number from DJI.

Not those specifically, but I have seen a number of reports that seem to indicate a recurrence of the previous home point error bug or something similar. It seems very likely that it is what you experienced too. I haven't been able to replicate it, but I haven't updated to the latest firmware. Please report back on DJI's response.

 

[ablack]

I hope someone can clear this up for me, but doesn't the Mavic pro controller have its own GPS built in allowing for flight without a mobile device and RTH?

 

[msinger]

doesn't the Mavic pro controller have its own GPS built in allowing for flight without a mobile device and RTH?

No. The Mavic itself does contain a GPS receiver though.

 

[scjerry]

Not those specifically, but I have seen a number of reports that seem to indicate a recurrence of the previous home point error bug or something similar. It seems very likely that it is what you experienced too. I haven't been able to replicate it, but I haven't updated to the latest firmware. Please report back on DJI's response.

sar104

Reporting on DJI's response(s) to my bug report. But before I do, note that the author of the thread Finlayson ( Serious bug in the Mavic's "Home Point Me" function - caused loss of drone! ) got a gift certificate for a replacement drone after reporting the same problem that I have, i.e., Home Point Update Not Recorded.

Now back to DJI's response. First, I placed all data that I could retrieve into a DropBox folder : Dropbox - Mavic Log Files. In spite of this effort, I received the following directive from DJI support: " Kindly try to do the online repair request Jerry, and send us the error that occur on the website. You have to send us back the drone first before the result of data analysis. Kindly update us if you already have a case number."

So like a lemming, I did. Now I don't have my $1K drone for an indeterminate time.

I don't believe it is a flight controller problem, which is the only reason they could possibly want my physical drone for testing. I believe, (belief systems again, right?) it is some iOS programmer's fault using DJI's SDK in developing the iOS DJI Go 4 app. Not sure how they will reproduce the exact conditions that I report, but suspect this testing could have been done without my drone.

Just to summarize, see Dropbox - Mavic Log Files.

Home Point Me is not the same as Home Point Recorded. RTH is to Home Point Recorded, which may be the home point of a previous flight, several miles away.

 

[sar104]

sar104

Reporting on DJI's response(s) to my bug report. But before I do, note that the author of the thread Finlayson ( Serious bug in the Mavic's "Home Point Me" function - caused loss of drone! ) got a gift certificate for a replacement drone after reporting the same problem that I have, i.e., Home Point Update Not Recorded.

Now back to DJI's response. First, I placed all data that I could retrieve into a DropBox folder : Dropbox - Mavic Log Files. In spite of this effort, I received the following directive from DJI support: " Kindly try to do the online repair request Jerry, and send us the error that occur on the website. You have to send us back the drone first before the result of data analysis. Kindly update us if you already have a case number."

So like a lemming, I did. Now I don't have my $1K drone for an indeterminate time.

I don't believe it is a flight controller problem, which is the only reason they could possibly want my physical drone for testing. I believe, (belief systems again, right?) it is some iOS programmer's fault using DJI's SDK in developing the iOS DJI Go 4 app. Not sure how they will reproduce the exact conditions that I report, but suspect this testing could have been done without my drone.

Just to summarize, see Dropbox - Mavic Log Files.

Home Point Me is not the same as Home Point Recorded. RTH is to Home Point Recorded, which may be the home point of a previous flight, several miles away.

Agreed - everything I've seen indicates that this is not an aircraft firmware problem, but rather a Go app problem. I find it hard to believe that they are not aware of that too, especially since it appears that they fixed that specific bug in the last update. I thought that you were also aware of that prior to sending it back to DJI.

 

[scjerry]

Agreed - everything I've seen indicates that this is not an aircraft firmware problem, but rather a Go app problem. I find it hard to believe that they are not aware of that too, especially since it appears that they fixed that specific bug in the last update. I thought that you were also aware of that prior to sending it back to DJI.

I conducted 4 test flights before returning it. Jury still out on aforementioned behavior with latest updates for both Go 4 & FW. I'm nothing if not careful.

 

[sar104]

I conducted 4 test flights before returning it. Jury still out on aforementioned behavior with latest updates for both Go 4 & FW. I'm nothing if not careful.

Well good luck - I hope it gets resolved.

 

[BudWalker]

@scjerry started a thread over on PhantomPilot about this issue. A .DAT was obtained for two of the flights exhibiting the behavior that's being discussed here. I posted my findings there also. see
[TOOL][WIN] Offline TXT FlightRecord to CSV Converter

In summary, the Mavic landed where it did and didn't travel back to to the HP because it was within 20 meters of the HP. Looking at FLY095 it can be seen from the eventLog stream that RTH was initiated at 163 secs

163.044 : 944416899 : 10226 [L-RC]rc cmd:STD_GO_HOME
163.185 : 945048350 : 10233 [L-FLYMODE][Ctrl<7>] REQ_RC_GO_HOME AUTO_LANDING ctrl_auto_landing

Since the AC was within 20 meters of the HP the response was for flightAction to enter the TOO_CLOSE_GOHOME_LANDING. Shortly after that flyCState entered the AutoLanding state

 

[scjerry]

@scjerry started a thread over on PhantomPilot about this issue. A .DAT was obtained for two of the flights exhibiting the behavior that's being discussed here. I posted my findings there also. see
[TOOL][WIN] Offline TXT FlightRecord to CSV Converter

In summary, the Mavic landed where it did and didn't travel back to to the HP because it was within 20 meters of the HP. Looking at FLY095 it can be seen from the eventLog stream that RTH was initiated at 163 secs

163.044 : 944416899 : 10226 [L-RC]rc cmd:STD_GO_HOME
163.185 : 945048350 : 10233 [L-FLYMODE][Ctrl<7>] REQ_RC_GO_HOME AUTO_LANDING ctrl_auto_landing

Since the AC was within 20 meters of the HP the response was for flightAction to enter the TOO_CLOSE_GOHOME_LANDING. Shortly after that flyCState entered the AutoLanding state
View attachment 89047

Thanks again for your response.

I found this (which I posted before) from the SDK website: Flight Controller - DJI Mobile SDK Documentation. I didn't find a reference to a minimum HP reset distance in the Mavic manual, but I probably missed it, and I should have taken it into consideration when conducting my tests:
Returning Home
The aircraft can automatically return-to-home (RTH) in a number of scenarios:

• Smart RTH: Commanded to by the pilot through the application or the remote controller
• Failsafe RTH: If the wireless link is lost between the remote controller and aircraft
• Low Battery RTH: If the battery drops below a threshold that is enough to get home, but not enough to require immediate emergency landing

When automatically going home, the aircraft will rise to a minimum altitude, fly to the home location (home point) using GPS positioning, then land.

The home point is automatically set as the location the aircraft first takes off from after power on. After that, the home point can be updated through APIs in the DJI Mobile SDK but is limited to being within 30m of initial take-off location, current aircraft's location, current mobile location, or current remote controller's location (for remote controllers with GPS capability).

Note: If the GPS signal is not sufficient during take-off to record a home location, the home point will be recorded when the GPS signal is strong enough. When taking off in poor satellite signal environments, developers should ensure the home point being set is within the user's expectations.

It's a bit murky, i.e., current mobile location? So actually the new HP to Me must be more than 30 meters (not 20) from launch HP? Does this restriction apply using the new dynamic HP in Active Track? So the dynamic HPs will be more than 30m apart?

In any event, thanks for the data analysis education. I will await the return of my MP from DJI repair. I expect they won't find anything wrong, given your results and the new updated s/w, but will retest upon its return.