First flight, instant crash after takeoff, FLYxxx.DAT?

[Pac]

Hi all!

I finally received my Spark (don't ask me as a Mavic owner WHY I bought it). After charging, activating and connecting I finally wanted to fly. So I put it down on a parking lot, turned it on and waited for the first GPS fix.
I recognized a compass error and thought it might be due to the proximity to a car, so I moved the spark to a slightly different location, no more compass error.
At 13 sats I thought it would be safe to fly so I turned on the rotors and wanted to take off - nope, stick configuration set to mode 2 so no take off with the right stick (I am left-handed and using mode 1 since ever on Mavic). Engines off, DJI Go App to change mode. Everything green (no errors, ready to take off). So right after taking off the Spark went crazy, drifting to the left at a height of 10 cm/4 inches so fast I couldnt to anything but trying to CSC - as this takes a second it already hit the wall 2 meters / 6 ft away. Luckily it flipped over and turned of the props.

Besides some sratches and one prop needing to be changed, nothing happened. However, I am curious why this happened. So first thing was to upload the flight to AirData UAV to check GPS and Sensors. And guess what? I can see no errors, however data is limited as it was so short (less then 10 seconds). It showed me that the Spark hit the will at 11 km/h / 7mph.

So I thought firing up DJI Assistant 2 will help me getting the FLYxxx.DAT which might include more information. Needed to Update from 1.1.0 to 1.1.2 to recognize my Spark.
However, the menu item I am used to from Mavic is missing here! I can access my memory card and I can see another drive letter, but I can't access it! I can copy the Blackbox and I can upload Flight Data to DJI.

Any suggestions on how to do this?
Thanks!!


Here are some AirData Screenshots:

 

[Stickboy]

What firmware/ app version are you using?

 

[Pac]

Stock FW it came with .400 (I know .500 is out but I read about Sparks flying out of the sky so I thought .400 is fine for now) and DJI Go App iOS 4.1.4 (4.1.5 shows me same changelog).

 

[Dronason]

Yes, the DAT file has more info than the app .TXT flight record.
To get your .DAT flight record, you were right, ASSISTANT2, DATA UPLOAD (strange name), press CONFIRM, select the flight you want to check and press SAVE TO LOCAL.
Then you select the directory where it should be saved.


Then you can use CsvView CsvView Downloads to plot selected flight data versus time.
DatCom DatCon Downloads can be used to convert the files to other formats.
Minimum reading is the CsvView manual otherwise you will be lost.

Regarding your issue, maybe the compass is not well calibrated (whatever says the app) or you have a lot of steel in the ground and your height was not enough that its influence is negligible. Since a first crash with a Phantom similar you yours I always lift it up a few meters (above my head) and hover for a few seconds to see if it is stable before going further away.
Maybe an IMU calibration could be good, at least check that their values are really close to zero. Normally you should have got some error messages in case the sensors output are conflicting. See DJI Spark Accessories & Tips - Spark Help section How-Tos, Calibration.
Also a compass calibration in a field where any metal is fare away could be recommended. See Compass Calibration Guide
Most probably you already know all of this from your Mavic, sorry in advance.

 

[msinger]

@Pac, if you'd like others to review your flight log, then please post the TXT and/or DAT flight log here. If posting the DAT flight log, you'll likely need to upload it to a file sharing website (e.g. Dropbox.com) since those files are usually quite large.

 

[Blade Strike]

More than likely this is due to the rebar and other metal in the parking area.

The spark compass is very close to the ground. Once you lifted off you broke the influence from the metal and compass started to throw errors. Craft disengages from gps mode when this happens.

 

[Jeff7577]

More than likely this is due to the rebar and other metal in the parking area.

The spark compass is very close to the ground. Once you lifted off you broke the influence from the metal and compass started to through errors. Craft disengages from gps mode when this happens.

This is backwards. Ferrous metals will throw off the compass and set errors. Moving away from them will allow normal compass operation. If rebar had any effect it was because it was still within proximity to skew the compass. I keep reading about drones flying away at 200 feet and people blame it on taking off from an area with some type of magnetic interference. WRONG. Unless that interference was a giant piece of iron or steel floating in the sky it had no effect.

 

[Blade Strike]

Actually it's not wrong... a minor influence like rebar can creat just enough bias in the compass for it not to through an error. Once that influence is broken by taking off thats when all hell can break loose for an inexperienced pilot that doesn't know how to fly in ATTI.

I work and do testing for DJI , I know just a little about these things!

 

[Dronason]

Usually the story is (own experience as noob):

• Poor location with rebar in the ground
• App says compass error, please calibrate, so let's do it
• Calibration is done at the worst possible place
• App says all fine
• Take off and the mess starts as soon the influence of rebar is reduced

The right way is:

• Poor location with rebar in the ground
• App says compass error, please calibrate
• Forget flying now at that place
• Found and open area for compass calibration and do it
• When at a place with rebar in ground, the app says compass error, please calibrate
• DO NOT calibrate but put the Spark on a box (plastic) so it is a few feet up the ground
• App says all fine
• Take off and fly
• For landing, either palm landing or on the box

Practice of ATTI mode is very important and give a lot of confidence, use a cheap drone for that (40$). It can be done indoor during winter or rainy saison.
That is what worked for me since Phantom 3, the quality of the compass calibration is the most important. As soon the Spark is in the air with some altitude, any influence is gone.

 

[Blade Strike]

Usually the story is (own experience as noob):

• Poor location with rebar in the ground
• App says compass error, please calibrate, so let's do it
• Calibration is done at the worst possible place
• App says all fine
• Take off and the mess starts as soon the influence of rebar is reduced

The right way is:

• Poor location with rebar in the ground
• App says compass error, please calibrate
• Forget flying now at that place
• Found and open area for compass calibration and do it
• When at a place with rebar in ground, the app says compass error, please calibrate
• DO NOT calibrate but put the Spark on a box (plastic) so it is a few feet up the ground
• App says all fine
• Take off and fly
• For landing, either palm landing or on the box

Practice of ATTI mode is very important and give a lot of confidence, use a cheap drone for that (40$). It can be done indoor during winter or rainy saison.
That is what worked for me since Phantom 3, the quality of the compass calibration is the most important. As soon the Spark is in the air with some altitude, any influence is gone.

Sound and very accurate advice if there's enough in the ground to cause error.. the one that gets people is when there's only enough to influence the bias at arm..

What really gets people are those old fiberglass tables at parks and beaches.. not enough to error the craft but enough to make the craft drop out of atti once it's about 5' away from it.

 

[Pac]

Thanks for all the comments. I flew my Mavic from this spot quite often without any issue. And of course I know NOT to calibrate the compass when it throws errors as you will do a false calibration and have fun once in the air.
Things went so fast, no chance to manually correct it.

I will now get the DAT files (thanks for that comprehensive guide!) and post it here!

 

[Sparkbart]

I'm going to apologize up-front, while I'm going to come off sounding like a "prick", its not my intent. Its just too easy for people to claim to be experts without any proof of any kind...

Well you may work for DJI, but are you an educated engineer? Are you a schooled Mechanical Engineer or Electrical Engineer? Are you paid by DJI to test for them? What is your background? Is it technical/engineering?

I'm not picking on you, but just because DJI pays you to test stuff is it testing from a user's experience or is it from a technical, investigative, engineering type testing? Do you use special DJI equipment and software to fly in a "debug" mode?

Again, I'm not picking on you but its too easy in this day and age for someone to claim something but not really have the qualifications or background to make any real comments on that area... I'll use myself as an example. I'm a Software Engineer specializing in a very specific field and I don't work with Windows based systems. Ask me anything about Windows and I couldn't tell you jack about them. People think because I spend my days working on "computers" I know EVERYTHING about every computer/smart device.. I don't....

Actually it's not wrong... a minor influence like rebar can creat just enough bias in the compass for it not to through an error. Once that influence is broken by taking off thats when all hell can break loose for an inexperienced pilot that doesn't know how to fly in ATTI.

I work and do testing for DJI , I know just a little about these things!

 

[Sparkbart]

+1

I couldn't agree more.

I've got all kinds of DJI drones and even had a Yuneec Typhoon H. If I received a "compass" warning. First thing I'd do is move the drone to a different location and check. If I received the error again, then I'd recalibrate. If not I'd fly. I've never had an issue doing this with my Inspire, Phantoms, Mavic, Spark, and Typhoon H.

Once you move away from a magnetic interference source, and you don't get an error its fine. I think people that lose GPS signal or something that causes a loss of control when the drone is away from them is due to other reasons. OR that the first "compass" error may have been a harbinger of other issues.

This is backwards. Ferrous metals will throw off the compass and set errors. Moving away from them will allow normal compass operation. If rebar had any effect it was because it was still within proximity to skew the compass. I keep reading about drones flying away at 200 feet and people blame it on taking off from an area with some type of magnetic interference. WRONG. Unless that interference was a giant piece of iron or steel floating in the sky it had no effect.

 

[Blade Strike]

I'm going to apologize up-front, while I'm going to come off sounding like a "prick", its not my intent. Its just too easy for people to claim to be experts without any proof of any kind...

Well you may work for DJI, but are you an educated engineer? Are you a schooled Mechanical Engineer or Electrical Engineer? Are you paid by DJI to test for them? What is your background? Is it technical/engineering?

I'm not picking on you, but just because DJI pays you to test stuff is it testing from a user's experience or is it from a technical, investigative, engineering type testing? Do you use special DJI equipment and software to fly in a "debug" mode?

Again, I'm not picking on you but its too easy in this day and age for someone to claim something but not really have the qualifications or background to make any real comments on that area... I'll use myself as an example. I'm a Software Engineer specializing in a very specific field and I don't work with Windows based systems. Ask me anything about Windows and I couldn't tell you jack about them. People think because I spend my days working on "computers" I know EVERYTHING about every computer/smart device.. I don't....

Well since you asked, in my day job I'm an network and systems engineer. On top of that I'm been building and designing custom rigs for many years .

Anyone that has been doing this for a while knows the above can happen.. a small compass bias turns into a large one once the field breaks.

Yes we do have custom tools as we test.. below is an example of just one of the screens.. I was testing BT interference..

BT ON


BT OFF

 

[Blade Strike]

+1

I couldn't agree more.

I've got all kinds of DJI drones and even had a Yuneec Typhoon H. If I received a "compass" warning. First thing I'd do is move the drone to a different location and check. If I received the error again, then I'd recalibrate. If not I'd fly. I've never had an issue doing this with my Inspire, Phantoms, Mavic, Spark, and Typhoon H.

Once you move away from a magnetic interference source, and you don't get an error its fine. I think people that lose GPS signal or something that causes a loss of control when the drone is away from them is due to other reasons. OR that the first "compass" error may have been a harbinger of other issues.

Once again you are correct if the interference is big enough to throw an error.

Guys my main role is enterprise equipment.. I work with things like RTK.. if this wasn't the case why would I even be making the statement!?

I'm only hear to help out and give some support. It's your choice to take it or leave it.. Not hear to argue

 

[Pac]

+1

If I received a "compass" warning. First thing I'd do is move the drone to a different location and check. If I received the error again, then I'd recalibrate. If not I'd fly. I've never had an issue doing this with my Inspire, Phantoms, Mavic, Spark, and Typhoon H.

Once you move away from a magnetic interference source, and you don't get an error its fine. I think people that lose GPS signal or something that causes a loss of control when the drone is away from them is due to other reasons. OR that the first "compass" error may have been a harbinger of other issues.

Exactly what I did. "Red" compass error (where I dont have any issues with Mavic), moved Spark to different location and compass error went away. 13 Sats also seemed fine. But right after takeoff (it was in the air for 0.5 seconds) it drifted at full speed (no sport mode) into the wall.

Whatever, here is the .DAT file of the flight. Looking forward to some info what happened. I am pretty sure it is compass related: DJI_ASSISTANT_EXPORT_FILE_8.5.2017 1744.DAT

Thanks for all the helping answers, very kind.

 

[PascalG]

There are two things which I find odd in those reports of compass errors and erratic behavior

- spark seems to be more sensitive to problems than other drones like mavic. While I haven't experienced a compass error the number of such problems

- more worrying is why would the drone move in any direction at speed with a compass error? I can understand the drone switching to atti mode and letting the pilot fly manually and being confused as to which way to go on its own. Drifting in the wind or maybe pivoting on its axis but why would it suddenly head for the nearest wall, tree or body of water.

 

[Pac]

True PascalG.Maybe the error level is just little below threshold that throws an error - but faulty enough to make it not only drift but head to the wall at full speed. I have flown quads that dont have GPS, Sensors etc. so I think I could be able to land it instead of crashing it. But only thing I tried was to shut down motors through CSC as it was close to the ground.
BTW, does CSC work on Spark?

Waiting for someone to read my .DAT file posted in my comment before

 

[Dronason]

There are two things which I find odd in those reports of compass errors and erratic behavior

- spark seems to be more sensitive to problems than other drones like mavic. While I haven't experienced a compass error the number of such problems

- more worrying is why would the drone move in any direction at speed with a compass error? I can understand the drone switching to atti mode and letting the pilot fly manually and being confused as to which way to go on its own. Drifting in the wind or maybe pivoting on its axis but why would it suddenly head for the nearest wall, tree or body of water.

It will switch to ATTI mode after compass error where only the accelerometer and gyro sensors are used. If these one have some zero error, they will induce a drift in addition to the wind. It can be from negligible to significant. The accelerometer error will tend to make the Spark not flat and so it will have an angle and start to drift. The gyro error will create some rotation that will amplify the drift and add some rotation at same time. It is there that IMU calibration is critical. When there is GPS and compass they can compensate for small IMU error.

 

[Dronason]

Exactly what I did. "Red" compass error (where I dont have any issues with Mavic), moved Spark to different location and compass error went away. 13 Sats also seemed fine. But right after takeoff (it was in the air for 0.5 seconds) it drifted at full speed (no sport mode) into the wall.

Whatever, here is the .DAT file of the flight. Looking forward to some info what happened. I am pretty sure it is compass related: DJI_ASSISTANT_EXPORT_FILE_8.5.2017 1744.DAT

Thanks for all the helping answers, very kind.

Thanks for sharing your DAT file. First of all I am not THE expert on such analysis but I am on same boat to learn from mistake.

At first I don't see the motors activated, the power is like 10watt so it is more standby mode. When motors are on it is typically around 60W on mine. There is also no mark in status of take off or any status change. So I have a first doubt if the DAT file was from a flight. As said I am also on learning curve. In DJI assistant, it is not that simple to found the right flight log, it can be confusing.

This is the full log, In RED marked near the time scale is some activity and is zoomed in next picture, BLACK is some standby time and BLUE is power-up.
In general, GPS satellites was not that much 5 to 9.
Magnetometer reported too high value ~2000, it is typically 1600 on mine which is the value you get too at end of record. This mean there was some strong magnetic interference before it moves. The value changed at start of RED time area but the final value was still high.


The RED area, zoomed:

There was only a throttle stick up action on the RC. But there could be some glitches as Spark is new and the program CsvView was recently adapted to support the Spark.
The RED area (16 seconds) is maybe the move and crash. Then in the blue area you pick it up and switched off?
As said at start, some elements make me thinking that it maybe not the DAT of your crash.

Here is a normal flight with mine plotting the same parameters.