Today I had my first, and hopefully last close call with my Spark, Everything was fine, I did my checks, made sure the home point was all set, 15/16 satellites where on the Drone, clear open area with no interference or metals around, so I took off, 30 seconds into my Flight, the Spark switched to ATTI Mode, got disconected from the GPS, IMU Went crazy and Compass went crazy as well, the drone was around 130 ft from me when that happened so I stopped any input and ran towards the spark to start my landing, 30 seconds later the Drone was on the ground and safe. I'll upload the flight log to see if someone can explain what happened. Thank you.

http://www.phantomhelp.com/LogViewer/ATN86NVU55GY5CFKQ24C/

Roberto, any error message in your app? Please upload the data to the dropbox and post the link here. Thanks a lot.

DJI Elektra Posted at 2017-12-29 23:29
Roberto, any error message in your app? Please upload the data to the dropbox and post the link here. Thanks a lot.

Yes, I did get 3 errors, I'll upload the pictures on my dropbox:
https://www.dropbox.com/sh/n0c89 ... 2N1Go7j_FQc9Da?dl=0

Roberto Lafarga Posted at 2017-12-29 23:39
Yes, I did get 3 errors, I'll upload the pictures on my dropbox:
https://www.dropbox.com/sh/n0c898fir68h18v/AAAdD56igX82N1Go7j_FQc9Da?dl=0

Hi, sir. I would recommend you fly in a wide place without interference. Please calibrate the IMU and compass before flight. Please ensure that you do not turn on the bluetooth devices. If it is not caused by environment, please export the flight data and the black box for further analysis.

YOU SHOULD NOT USE THE INTERNET WHILE FLYING THE spark.  Also you were in a warning Zone area ... Yaw errors are caused by many factors . From what i deduce from  the flight log, you were using the internet or a second GPS for the flight ..Also you were flying very low where GPS was weak.

S.J Posted at 2018-1-3 04:00
YOU SHOULD NOT USE THE INTERNET WHILE FLYING THE spark.  Also you were in a warning Zone area ... Yaw errors are caused by many factors . From what i deduce from  the flight log, you were using the internet or a second GPS for the flight ..Also you were flying very low where GPS was weak.

Funny thing was that, I'm flying in Mexico (no "regulations" still), hence the Class D", had airplane mode just wifi for the controller and had 16 locked Satellites with full bars when I took off, No other gps was near me, and I was on a Clear Open area, some trees, but I was Flying well above them, I'll upload all of the other data. Thanks for the Reply

S.J Posted at 2018-1-3 04:00
YOU SHOULD NOT USE THE INTERNET WHILE FLYING THE spark.  Also you were in a warning Zone area ... Yaw errors are caused by many factors . From what i deduce from  the flight log, you were using the internet or a second GPS for the flight ..Also you were flying very low where GPS was weak.

Also you were flying very low where GPS was weak....

Lol wut?!


Height has very little effect.


His 1st yaw error had the same number of sattelites than at ground level.




0m 30.7s - 16satellites -57.1ft -Yaw Error

ped078 Posted at 2018-1-3 10:36
Also you were flying very low where GPS was weak....

Lol wut?!

That doesn't explain anything .

what about       at 31.9 s     Compass Error. Exit P-GPS Mode; Yaw Error

You must not fly low where it hinders proper GPS reception even momentary

ped078 Posted at 2018-1-3 10:36
Also you were flying very low where GPS was weak....

Lol wut?!

I agree with you, also I was flying on an open Field, I would call low GPS reception if I had maybe 7 satellites, not 16. Thank's for the answer!

S.J Posted at 2018-1-3 10:42
That doesn't explain anything .

what about       at 31.9 s     Compass Error. Exit P-GPS Mode; Yaw Error

My point exactly, flying low DOES NOT hinder GPS reception, if that was the case, in car GPS systems wouldn't work. Get you're fact's straight.

I'm an electronics engineer with plenty of RF / Microwave radio experience.

Are you?

Roberto Lafarga Posted at 2018-1-3 11:02
I agree with you, also I was flying on an open Field, I would call low GPS reception if I had maybe 7 satellites, not 16. Thank's for the answer!

Roberto, I am sorry to hear about this incident and I am starting to get concerned by the large number of drones that have experience yaw errors on the latest firmware. I count 5 different instances, all with very different flight conditions and no past reports of yaw errors of this scale in earlier firmware revisions.

However, for my own analysis I wanted to get more information about what occurred. Specifically:

What version of firmware was the drone running
What version of the DJI GO App was running
Are you using the remote controller

You mentioned that when the error occurred that you started running towards the drone and attempted to land it. It does seems as if it was responsive to the landing command along with your screenshots you uploaded to dropbox showing you going full stick forward at one point as well. Did you activate auto landing or did you just pull the left stick down to bring the drone to the ground?

You had compass problems either due to bad compass or magnetic interference on the ground, this is usually seen first as magnetic yaw followed by magnetic interference IMU heading exception, gps switched off and Atti mode so your now controlling your aircraft without gps , you managed to hold off until gps returned and aircraft landed.

Why would a compass error disable GPS?
Orientation (compass data) and position (GPS data) are different but related.
When the Spark is moving and getting position information from the GPS that tells the Spark that it is travelling in a different direction from what it's getting from compass data, The Spark programming can't work with the data conflict.
The solution is to drop data from one source and since the Spark can't fly without a compass but can without GPS, it's the GPS data that gets dropped.

There have been some compass problems with Spark lately or your issue may have been related to something you did.
I'd be curious about when and where you calibrated it and where you launched from as these may be relevant factors.

ped078 Posted at 2018-1-3 11:45
My point exactly, flying low DOES NOT hinder GPS reception, if that was the case, in car GPS systems wouldn't work. Get you're fact's straight.

I'm an electronics engineer with plenty of RF / Microwave radio experience.

Gps in cars and phones receive signal in much different ways than your gps in your aircraft , gps in drones takes signal directly from satellites,so yes if you are above obstacles in the clear sky you have much better chance of receiving higher gps sat count.

Cars and phones receive signal from antenna masts signals being bounced around from strategically placed very large GPS receivers and that’s why they don’t have problems driving through cities etc.

asourceunknown Posted at 2018-1-3 11:51
Roberto, I am sorry to hear about this incident and I am starting to get concerned by the large number of drones that have experience yaw errors on the latest firmware. I count 5 different instances, all with very different flight conditions and no past reports of yaw errors of this scale in earlier firmware revisions.

However, for my own analysis I wanted to get more information about what occurred. Specifically:

Hi, Luckily I never lost any connection from the Spark to My Remote, Funny thing that that had happened with the "latest" firmware for that day, Spark was on .900 and DJI Go 4 on .20.

And nope, Manual landing, pulled my sticks down.

Thanks for the response!

hallmark007 Posted at 2018-1-3 12:35
You had compass problems either due to bad compass or magnetic interference on the ground, this is usually seen first as magnetic yaw followed by magnetic interference IMU heading exception, gps switched off and Atti mode so your now controlling your aircraft without gps , you managed to hold off until gps returned and aircraft landed.

Why would a compass error disable GPS?

Hi, Hallmark, I did the compass and IMU calibration probably a week before, because It was flying funky, did it on a flat table, with no magnetic interference. Launched the Spark on a large field too.

hallmark007 Posted at 2018-1-3 12:40
Gps in cars and phones receive signal in much different ways than your gps in your aircraft , gps in drones takes signal directly from satellites,so yes if you are above obstacles in the clear sky you have much better chance of receiving higher gps sat count.

Cars and phones receive signal from antenna masts signals being bounced around from strategically placed very large GPS receivers and that’s why they don’t have problems driving through cities etc.

Do you have a source for this information?  Having worked in space communications and satellite navigation this is the first time I’ve heard of this.

Nidge Posted at 2018-1-3 12:52
Do you have a source for this information?  Having worked in space communications and satellite navigation this is the first time I’ve heard of this.

Do I have proof about what,

Nidge Posted at 2018-1-3 12:52
Do you have a source for this information?  Having worked in space communications and satellite navigation this is the first time I’ve heard of this.

What is GPS Drone Navigation?
The global positioning system is a satellite navigation system that uses a radio receiver to collect signals from orbiting satellites to determine position, speed, and time. This navigation system is more accurate than over forms of navigation, and provides position knowledge to within a few meters. Advanced GPS systems can provide even better accuracies to within a few centimeters. The miniaturization of integrated circuits has allows GPS receivers to be highly economical, and available to everyone. GPS is a broadcast radio system that reaches almost all areas of the planet, so it is highly accessible.

Cell Phones
New-generation cell phones combine many devices into one. When traveling, one of the most convenient of these features is the built-in GPS. When visiting a place that you aren't familiar with, getting lost is a lot harder when you have turn-by-turn directions on your cell phone. Even the cell phones that don't have GPS can use cell tower position and distance to calculate your location. Cell phones function by communicating with towers connected to a base station in a configuration called a "cell." As you move through the cell, the base station monitors your cell phone's signal and transfers it to the nearest tower.

hallmark007 Posted at 2018-1-3 13:16
What is GPS Drone Navigation?
The global positioning system is a satellite navigation system that uses a radio receiver to collect signals from orbiting satellites to determine position, speed, and time. This navigation system is more accurate than over forms of navigation, and provides position knowledge to within a few meters. Advanced GPS systems can provide even better accuracies to within a few centimeters. The miniaturization of integrated circuits has allows GPS receivers to be highly economical, and available to everyone. GPS is a broadcast radio system that reaches almost all areas of the planet, so it is highly accessible.

I’m aware of positional data obtained by triangulation of cell towers but this is not accurate enough to provide, nor provide suitable refresh rates, for navigational position accuracy, it’s similar to that provided by project Skyhook and the ability to triangulate a user by known WiFi locations. For mobile navigation, land sea or air GPS is the only accurate system for use from sea level and upward. When law enforcement agencies use cellular data to determine a suspects location this only works to a realistic accuracy of 100metres.

Nidge Posted at 2018-1-3 13:32
I’m aware of positional data obtained by triangulation of cell towers but this is not accurate enough to provide, nor provide suitable refresh rates, for navigational position accuracy, it’s similar to that provided by project Skyhook and the ability to triangulate a user by known WiFi locations. For mobile navigation, land sea or air GPS is the only accurate system for use from sea level and upward. When law enforcement agencies use cellular data to determine a suspects location this only works to a realistic accuracy of 100metres.

What is your question? Are you saying that gps in your phone or car is the same as your drone well it’s not.
If your looking for more proof it’s all out there.

If you think about it I can receive gps signal on my phone in my office I can’t get gps on my drone.because my GPS receiver needs to be in direct contact with satellites.

hallmark007 Posted at 2018-1-3 13:39
What is your question? Are you saying that gps in your phone or car is the same as your drone well it’s not.
If your looking for more proof it’s all out there.

I second this information.
There is a difference between GPS on the car systems and the Drone . The theory that it uses satellites for connection is absolutely true, the method used to produce directional indication may be different.

What DJI could have done between a positional and orientation conflict is to have the user choose  how SPARK moves from that moment in the Sky rather than using compass information and rejecting GPS  to run the AC and then report of  flyaway occurs.

hallmark007 Posted at 2018-1-3 13:39
What is your question? Are you saying that gps in your phone or car is the same as your drone well it’s not.
If your looking for more proof it’s all out there.

I think you're mistaken if you believe car gps system use cell towers for navigation. Car GPS, tablet GPS (without phone capability), golf range finder GPS, health tracking GPS, and many industrial applications all use the satellites without the aid of cell towers and all are very close to the ground. It takes a little longer, but my Spark will pick up more than 8 satellites in my home, as do my golfing range finder and wi-fi tablet.

Cell phones do use tower data to enhance their location ability - generally to speed up acquisition as opposed to increased accuracy.

Both GPS and compass information is required. Sure, the Spark could easily determine heading simply via GPS but it would have no clue about orientation so it would have no clue which rotors to spin up in order to move in a specific direction. Now, it could use trial and error, but without the compass that would be one heck of a drunken stroll back to the home point for the little guy.

If anything compass data being inaccurate should stop the drone it its tracks and maintain GPS lock and hover. Send an error to the pilot who can then determine orientation (should still be in line of sight, if you're a behaving drone pilot) and allow them to pilot the drone home. Tossing the drone into ATTI mode just leads to frustrated users.

With all that said I am still unsure if the OP case has anything to do with the behavior I described above nor am I aware of how the Spark firmware handles loss of compass accuracy.

Sliderul Posted at 2018-1-3 20:42
I think you're mistaken if you believe car gps system use cell towers for navigation. Car GPS, tablet GPS (without phone capability), golf range finder GPS, health tracking GPS, and many industrial applications all use the satellites without the aid of cell towers and all are very close to the ground. It takes a little longer, but my Spark will pick up more than 8 satellites in my home, as do my golfing range finder and wi-fi tablet.

Cell phones do use tower data to enhance their location ability - generally to speed up acquisition as opposed to increased accuracy.

I think you will find gps systems in cars phones etc use many different sources of information ie google maps government data base aerial photography measurements , drones don’t use this data .
Be super careful to study the differences between the map database's actual data resolution, and the current situation around you. Buildings come, hillsides get bulldozed away, rivers dry up, etc. Tides also ebb and flow, wax and wane, so the morning's sea level isn't even the same as the afternoon sea level.
So what we get here is many think because they can walk through high rise buildings and still see gps working on their phones that somehow this will be the same with their drones, and while you pick up 8 satellites in your home try moving your drone in your home with your hand over the top where gps is located and you will see a change.
My point is that receiving gps in a drone is different than car phone tablet.

asourceunknown Posted at 2018-1-4 01:57
Both GPS and compass information is required. Sure, the Spark could easily determine heading simply via GPS but it would have no clue about orientation so it would have no clue which rotors to spin up in order to move in a specific direction. Now, it could use trial and error, but without the compass that would be one heck of a drunken stroll back to the home point for the little guy.

If anything compass data being inaccurate should stop the drone it its tracks and maintain GPS lock and hover. Send an error to the pilot who can then determine orientation (should still be in line of sight, if you're a behaving drone pilot) and allow them to pilot the drone home. Tossing the drone into ATTI mode just leads to frustrated users.

This is how it is handled at present, and this is the way most if not all handle compass problems, many drones carry more than one compass and IMU as redundancy , in these drones we should see less of this happening as a result. Spark has only one.

Orientation (compass data) and position (GPS data) are different but related.
When the Spark is moving and getting position information from the GPS that tells the Spark that it is travelling in a different direction from what it's getting from compass data, The Spark programming can't work with the data conflict.
The solution is to drop data from one source and since the Spark can't fly without a compass but can without GPS, it's the GPS data that gets dropped.

hallmark007 Posted at 2018-1-4 02:49
This is how it is handled at present, and this is the way most if not all handle compass problems, many drones carry more than one compass and IMU as redundancy , in these drones we should see less of this happening as a result. Spark has only one.

Orientation (compass data) and position (GPS data) are different but related.

I highly doubt this is how the drone is operating. It seems more likely to me that the drone really doesn't care or calculate directional data from the the last and current GPS location and there is really zero reasons for it to do so. All that calculating that information would get you is if the compass and GPS data do not match, which may or may not be reported as compass errors but also could mean that the drone is facing a strong wind in any direction.

For example aircraft of all types need to face the nose into the wind to ensure they're not drifting off course when flying in a straight line trajectory and that angle into the wind is directly related to the speed of the wind. I'd say that DJI might be calculating this data to throw high wind warnings but that can be calculated by the gyro inside the drone without the need for storing past GPS data.

asourceunknown Posted at 2018-1-4 04:42
I highly doubt this is how the drone is operating. It seems more likely to me that the drone really doesn't care or calculate directional data from the the last and current GPS location and there is really zero reasons for it to do so. All that calculating that information would get you is if the compass and GPS data do not match, which may or may not be reported as compass errors but also could mean that the drone is facing a strong wind in any direction.

For example aircraft of all types need to face the nose into the wind to ensure they're not drifting off course when flying in a straight line trajectory and that angle into the wind is directly related to the speed of the wind. I'd say that DJI might be calculating this data to throw high wind warnings but that can be calculated by the gyro inside the drone without the need for storing past GPS data.

If it doesn’t collect directional data how does it find its way home?

S.J Posted at 2018-1-3 04:00
YOU SHOULD NOT USE THE INTERNET WHILE FLYING THE spark.  Also you were in a warning Zone area ... Yaw errors are caused by many factors . From what i deduce from  the flight log, you were using the internet or a second GPS for the flight ..Also you were flying very low where GPS was weak.

Just curious....what information drew you to the conclusion of "using the internet connection" .... or "a second GPS for the flight" ?  Similarly, his altitude was showing higher that that when all was going well.  

I suggest  if compass is ditched, the spark can't just hover since it's bound to drift and need compass to maintain gps position, which is where we came in.

Also in what way do 2 compasses which are subject to identical field improve matters?

hallmark007 Posted at 2018-1-3 12:40
Gps in cars and phones receive signal in much different ways than your gps in your aircraft , gps in drones takes signal directly from satellites,so yes if you are above obstacles in the clear sky you have much better chance of receiving higher gps sat count.

Cars and phones receive signal from antenna masts signals being bounced around from strategically placed very large GPS receivers and that’s why they don’t have problems driving through cities etc.

"Gps in cars and phones receive signal in much different ways than your gps in your aircraft , gps in drones takes signal directly from satellites"


No....NO you are absolutely WRONG.


Sure phones use cell information, A-GPS and GLONASS , but cars DO NOT. I never mentioned phones, I said cars.


Cars receive GPS signals EXACTLY the same way, from the global posistioning satellites to thier in built microwave antennae .

TomTom

Spark

ped078 Posted at 2018-1-4 10:28
"Gps in cars and phones receive signal in much different ways than your gps in your aircraft , gps in drones takes signal directly from satellites"

The fact of the matter here is gps in cars receive information in more ways than drones so if your driving through streets flanked by high buildings you will continue to get information from your gps in car , unlike flying a drone low in same situation you are likely to loose gps or get weak gps signal.

I think you will find gps systems in cars phones etc use many different sources of information ie google maps government data base aerial photography measurements , drones don’t use this data .
Be super careful to study the differences between the map database's actual data resolution, and the current situation around you. Buildings come, hillsides get bulldozed away, rivers dry up, etc. Tides also ebb and flow, wax and wane, so the morning's sea level isn't even the same as the afternoon sea level.
So what we get here is many think because they can walk through high rise buildings and still see gps working on their phones that somehow this will be the same with their drones, and while you pick up 8 satellites in your home try moving your drone in your home with your hand over the top where gps is located and you will see a change.
My point is that receiving gps in a drone is different than car phone tablet.

Bright Spark Posted at 2018-1-4 08:22
I suggest  if compass is ditched, the spark can't just hover since it's bound to drift and need compass to maintain gps position, which is where we came in.

Using gps to maintain position when you don't know your own orientation seems problematic.  i.e. The GPS tells you that you've drifted west and you need to correct by moving east - how do you do this when you don't know where east is?  It again goes back to drunken salior technique of trial and error.



  

web_parrot Posted at 2018-1-4 08:10
Just curious....what information drew you to the conclusion of "using the internet connection" .... or "a second GPS for the flight" ?  Similarly, his altitude was showing higher that that when all was going well.

it was only a comment.  Also if the SPARK has to give you a fly zone warning, it must be linked to the internet constantly about updating NFZ  list.

During my flights , rc /ac conflicts happen while I am linked to the internet.

hallmark007 Posted at 2018-1-4 10:42
The fact of the matter here is gps in cars receive information in more ways than drones so if your driving through streets flanked by high buildings you will continue to get information from your gps in car , unlike flying a drone low in same situation you are likely to loose gps or get weak gps signal.

I think you will find gps systems in cars phones etc use many different sources of information ie google maps government data base aerial photography measurements , drones don’t use this data .

No.

Just no.

Cars receive their GPS signal EXACTLY the same way the Spark does...through  their GPS antenna.

Phones receive their position data from different sources, Cell information, wifi ect, cars don't...I say again, I didn't mention phones, YOU did...I said cars.

ped078 Posted at 2018-1-20 14:59
No.

Just no.

Cars receive their data in many ways just like phones and there is a difference between information received by drones and information received by cars.

http://www.aisin-aw.co.jp/en/pro ... tructure/index.html

hallmark007 Posted at 2018-1-20 15:21
Cars receive their data in many ways just like phones and there is a difference between information received by drones and information received by cars.

http://www.aisin-aw.co.jp/en/products/information/structure/index.html

Nope.

You're wrong. Cars receive their GPS data through their GPS antenna.
True some connect through a cell phone...but thats only for additional data and if that phone is unavailable it's primarily down to the GPS antenna.

ped078 Posted at 2018-3-17 14:37
Nope.

You're wrong. Cars receive their GPS data through their GPS antenna.

One minute you say I’m wrong next I’m right whatever you think yourself.

hallmark007 Posted at 2018-3-17 15:01
One minute you say I’m wrong next I’m right whatever you think yourself.

I don't know about commercial systems, but I was managing a fleet of taxicabs with GPS-based onboard computers, and these systems used a normal GPS antenna, either mounted inside or outside the car (depending on the driver's preference, most of the time on the dashboard, behind the windshield). I had never had complaints of drivers not being positioned, yet I'm sure that this system didn't use any other data, like cellphone triangulation. The only exception is, of course, when the cars were parked in a garage. As long as the antenna can "see" the sky, it should be able to pick up a few satellites.