I bought myself another brand new Spark. I tried to take off twice using palm launch, but it just dropped to the ground after leaving my palm (so did my heart seeing the new Spark just drop straight onto the ground). I know how to do a proper palm launch because I had done that over 100 times with my first Spark. Hence I just couldn't understand why this new Spark failed to do so.

After that, instead of a palm launch, I decided to launch from the ground. The Spark flew up a foot up and start to flip and crashed. I powered it off, and on, and tried to do a screen launch and it flew up and start to fall again. This time I tried to push the stick up but after rising slightly it crashed to the ground again. This is a brand new Spark and I did not expect the first 4 launches to crashed like that.

I changed the battery, and everything is working fine. And when I changed back to the first battery, the Spark also behaves normally, and also for the third and fourth battery. For the 4 batteries, I tried flying around near me and everything seems fine. My question is why did the Spark failed the first 4 launches? I had a decent amount of faith in the Spark, which is the reason why I bought a second one. But this time the experience I went through dented my confidence in the Spark.

How do I extract the log for some experts here to analyse what went wrong? Btw is there a limited number of flights recorded in the Spark? I forgotten I can extract the log data at that moment and tried to fly many times to ascertain the functionality of the Spark. Only when I reached home then I realised I might have overwritten the first 4 initial launches that crashed.

Any advice is appreciated.

Flight Data can be downloaded using Assistant 2.  There are dozens of posts here describing how use Assistant 2.  While you're there, update to the latest firmware.  What phone/monitor are you using?  What operating system are you using?  Did you update the firmware before your problems?  Finally, after you have gathered the data file, updated the software(s), start an online Support Chat session.  The Chatter will want all of the information you've collected.  You will probably get a Case# if the advisor feels there is a problem with the aircraft.

What WebParrot said is correct, download the .DAT files from the drone or the txt logs from your device and someone might be able to shine some light on your issue. It's strange that your issue resolved itself though.
3 Questions.
1. Did a propeller fly off at all on any of the crashes ?
2. Was a controller involved in any of the crashes or were you using the spark and your phone/tablet only ?
3. Did you do any troubleshooting that might have resolved the issue or did you simply restart the drone(by changing batteries)?

Hi, we're very sorry to hear about the unpleasant experience that you had with your spark. If the drone is having a constant issue and crashed, I suggest you send it in for repair and data analysis, we have professional analysis team who will help to figure out the reason. Should you have further question, please let us know. Thank you.

In my experience that type of behavior is due to the props being installed incorrectly or just bad/damaged props.

Had such problem with other drone (not dji) in 2 cases:
1) props installed not on correct places
2) damaged contacts on battery

This is one of those instances where, like DJI Natalia mentioned, sending it in to be checked out might be the best option.

Send it back and let them check the black box logs

Thank you for taking time to give your valuable advises. I will test flight a few more times and evaluate the options suggested

To answer some of the questions, all batteries, AC, RC, Go 4 app are updated to the latest version and I am using iPhone 6 Plus dedicated for droning.

If someone can help to analyse, this is what I have extracted from the Spark, hopefully the first 4 crashes are recorded.  https://1drv.ms/u/s!ArdwMz9L9jtGgd5nIDhOix7sG8uioA

G'day, I'm checking your flight records with CsvView & DatCon software. to 12.43
Extracted  5 files,  FLY040.DAT - FLY045.DAT, with the "ExtractDJI.exe" app from CsvView or DatCon.

So far FLY040.DAT shows a short flight up to 7.3 m. The battery current increased to 6 amps whilst  the voltage dropped from 12.5 v to 11.7 v. Then the voltage recovered to 12.3 v.  I don't know if this is normal with Spark, but maybe a bad battery connection caused the problem. I'll check your other *.DAT files & compare the results, in spare time.

Disclaimer: I am using 3rd party software to decode *.DAT files - not all information provided may be accurate.

FLY045.DAT has errors via CsvView. Something is wrong as the other files decoded without this error. "Left front motor has a signal MotorCtrlWM that isn't defined" . The current & volts seem low. 11 volts down to 8 volts with 0.9 to 1.2 amps.  

Check the contacts are clean & replace your batteries if required. Only 45 flights with problems?
FLY040.DAT shows 12.5+ volts at 1 amp. Up to 6.5 amps @ 11.8 volts.

Maybe you have a bad battery or dirty contacts? Clean the contacts with an eraser from a pencil - the soft rubber part on top. If that fails, return or replace the battery. If several batteries fail, return Spark.
Again - disclaimer - 3rd party software used to decode  - information may not be accurate.
I wish I could help further.

Sparkz71 Posted at 2018-10-20 21:11
FLY045.DAT has errors via CsvView. Something is wrong as the other files decoded without this error. "Left front motor has a signal MotorCtrlWM that isn't defined" . The current & volts seem low. 11 volts down to 8 volts with 0.9 to 1.2 amps.  

Check the contacts are clean & replace your batteries if required. Only 45 flights with problems?

Hi Sparkz71, thanks for taking time to analyze for me. Greatly appreciated that. It confirms my wild guess that probably one or more of the motors were not performing normally.

Yes, only the first 4 flights of the Spark had errors and crashed. Subsequently I can launch and fly without it falling after changing batteries. Even later when I swapped back to the first battery, there were not problems too.

Now I have other questions in  my mind, since it is a new Spark, how come the dat are recorded as Fly040 - Fly045.dat. A refurbished Spark?

Anyhow thanks mate, and others who have given their valuable suggestions and opinions.

Ronnie.Choy Posted at 2018-10-21 17:57
Hi Sparkz71, thanks for taking time to analyze for me. Greatly appreciated that. It confirms my wild guess that probably one or more of the motors were not performing normally.

Yes, only the first 4 flights of the Spark had errors and crashed. Subsequently I can launch and fly without it falling after changing batteries. Even later when I swapped back to the first battery, there were not problems too.

As far as I have noticed, the number of the dat file increases after each power cycle or reboot.
I don't know why the dat files start at 40 - older dat files are deleted due to limited storage within Spark.

A few older dat files may have existed if you updated the firmware of Spark & the batteries, plus the IMU & compass calibration process, increasing the number of the dat file with each reboot (??)

My first 'recorded' dat file is FLY087.DAT in Spark, 5 days since purchasing. I flew several flights (at least 35 according to the video file numbers). I had no reason to use DJI Assistant 2 apart from upgrading the Spark firmware. Since learning further I downloaded the *.dat file & 'lost' the first files. I can't confirm the initial number recorded.  

Maybe somebody with a virgin Spark & a charged battery can share their initial *.dat file from  DJI Assistant 2 .

Comparing data ... see the voltage problems? I had no drama up to max height.

Hi, I don't believe any of those DATs, are referring to your crash flights, straight away i can see the oldest file 040.DAT is dated 19/10/2018 - 10:10 GMT, your post is 19/10/18 - 04:13. Also i don't see any serious crashes in the flights, you have some interesting landings though, like your picking up the drone very quickly, i guess your doing palm landings ?

DAT files are created regardless of flights, simply powering the drone on creates a DAT file, that DAT file ends when you either power off the drone or the file reaches its max size(300mb i believe), in which case it creates another DAT file. If your flight recorder storage gets full(spark has the lowest storage)it overwrites the oldest file. In some cases the flight recorder can save a flight record(mark it so it wont get overwritten), i have no idea what the criteria is for this or why, but i have seen it happen.
A new drone should have 0 DAT files, its not impossible for a new drone to come with some DATs but i assume these would be extra QA checks by DJI and there would not be 40 of them.

Best thing to do would be to get your .log files from the phone(if any) you used during the crash, DJI will probably request these if you log a case with them. These should contain every flight you have done with that phone.
Also be sure to check your controller on the calibration screen and make sure all the sticks do what they should and aren't giving commands when not touched, a faulty controller can give ghost commands(you press nothing but a command is output) and give no command when you clearly press the sticks/buttons, It's probably fine but better to check all the same.

Ben Mason Posted at 2018-10-22 00:33
Hi, I don't believe any of those DATs, are referring to your crash flights, straight away i can see the oldest file 040.DAT is dated 19/10/2018 - 10:10 GMT, your post is 19/10/18 - 04:13. Also i don't see any serious crashes in the flights, you have some interesting landings though, like your picking up the drone very quickly, i guess your doing palm landings ?

DAT files are created regardless of flights, simply powering the drone on creates a DAT file, that DAT file ends when you either power off the drone or the file reaches its max size(300mb i believe), in which case it creates another DAT file. If your flight recorder storage gets full(spark has the lowest storage)it overwrites the oldest file. In some cases the flight recorder can save a flight record(mark it so it wont get overwritten), i have no idea what the criteria is for this or why, but i have seen it happen.

Those DATs are all I can retrieve from the Spark. Throughout that afternoon I powered on and off the Spark probably about 8 times and if I include the earlier day about 2 times for updating the Spark, the total should be about 10 times. I am not sure why only 5 DATs are exported.

You are absolutely right about the palm landing. I do all my landings using my palm, except the first 4 when the Spark just dropped immediately after palm launches (twice) and land launches (twice).

Thank you for taking your time to analyse the data.

Ronnie.Choy Posted at 2018-10-22 07:47
Those DATs are all I can retrieve from the Spark. Throughout that afternoon I powered on and off the Spark probably about 8 times and if I include the earlier day about 2 times for updating the Spark, the total should be about 10 times. I am not sure why only 5 DATs are exported.

You are absolutely right about the palm landing. I do all my landings using my palm, except the first 4 when the Spark just dropped immediately after palm launches (twice) and land launches (twice).

Well you can still get the log files from your phone, these don't really delete unless you delete them yourself, they don't give as much information as a .DAT but they could still be useful.

Ben Mason Posted at 2018-10-22 07:53
Well you can still get the log files from your phone, these don't really delete unless you delete them yourself, they don't give as much information as a .DAT but they could still be useful.

Funny thing that the first palm launch is not recorded in the phone flight record, but I manage to find and export the other 3 failed take-offs.

Here is the link to the CSVs https://1drv.ms/f/s!ArdwMz9L9jtGgd5o0FH6yDdJ8CEKbA

Ronnie.Choy Posted at 2018-10-22 08:46
Funny thing that the first palm launch is not recorded in the phone flight record, but I manage to find and export the other 3 failed take-offs.

Here is the link to the CSVs https://1drv.ms/f/s!ArdwMz9L9jtGgd5o0FH6yDdJ8CEKbA

Maybe you didn't use your phone for that flight ? palm launch and land. Also that is the incorrect file, see link for more info.

https://www.phantomhelp.com/logviewer/Upload/

Ben Mason Posted at 2018-10-22 23:02
Maybe you didn't use your phone for that flight ? palm launch and land. Also that is the incorrect file, see link for more info.

https://www.phantomhelp.com/logviewer/Upload/

I am pretty sure the first file is palm launch, the second file is using the DJI Go 4 app on the phone to launch, and the third file is using the joysticks to start the motors and trying to launch upwards.

I will use your link to explore later