Please select Into the mobile phone version | Continue to access the computer ver.
Spark falling out of the sky - Findings, opinion and my conclusion
4003 28 2017-7-20
Uploading and Loding Picture ...(0/1)
o(^-^)o
Phuong Do
Second Officer
Flight distance : 29800 ft
Vietnam
Offline

Hi guys,
Thank you for all of you guys who have been following, supporting and giving comments on my threads about this matter: Spark got power cut-off during flight and fell out of the sky. Please note, I don't talk about pilot error cases here where people unintentinally do the CSC conbination or not reading the manual carefully.

1. Findings:
-I have been trying to separated between Software - Hardware issue, and for some reasons people do not want to accept my logic. So I will just say again like this: I absolutely do not think this is software issue. Why? It is just a simple logic: All Sparks use the same code, if there is problem, all Sparks would fall from the sky by now, not just a few cases. And, put yourself into DJI shoes, if it is a software issue, will you do nothing for nearly a month or you will release a firmware update as soon as you can? Also, some veterans pilots said that don't worry, when Mavic and Inspire first released, they also fell from the sky, but DJI fix after some update... However, I think you are looking inside the box, how about think outside the box and consider this scenario: is it really true that DJI fix the fall out cases of Mavic, Inspire, Phantom,... "by software update"/ Or they just waiting for cases to report, send in and fix. If it is the quality issue and just happen to a few units, of course after all those "faulty" unit sent back and fixed, the problem gone. That made you think the problem fixed by update but that is the hardware faulty. I got a feeling this is the case happening here.
       1A-Talking about hardward, we (me and some other users in the forum) got an assumption that it is overheat problem that the temperature sensor thought Spark is overheat so it cut off the power mid-flight... I would like to reject this assumption. First we don't know if Spark has that kind of function or not (which let it cut-off power when overheat), I highly doubt that, especially for a flying toy like this, DJI would not program such ability for Spark. Also, if it is really overheat, it will not let you power-up again right after falling (because still being overheat), but guys reported that they still can power-on right after the fall.
       1B-Talking about battery issue: I highly suggest it is the issue with battery, no issue Spark itself. Anh we got 2 assumptions:
            1B1-The contact between Spark and battery not tight enough (aka the Gopro Karma similar issue). However, i did a small test by turning on the battery (without connecting it to the aircraft), it can maintain the power itself and only turn off after I double tab the power button again. So, if it is the contact issue, when Spark and battery lost contact, the battery should be still power on when it landed on the ground, because it should be able to maintain the power itself like in the experiment. However, cases reported that battery was off when they pick it up. And those users also confirmed they have check carefully the latch between the battery and Spark --> So reject this assumption.
            1B2-This is the most likely root of issue in my opinion, the battery itself has problem. Now, software or hardware I don't know. If software, could be some glitch that trigger the power off randomly/ or under some combinations of outside elements. Or if hardware, some mistake during manufacturing make it work unpredictable and random shut-down.

2.Opinion:
-I made a poll to see how many people got the fall out issue, although not many vote, and I assume that meant they do not have power issue. The number of battery issue is only 11/44+, considering thousand of Spark users on this forum. I guess this is not a critical issue.
-Now, I do not saying that DJI is doing it this way, I just put myself into DJI shoes and say what I would do: First, we all accept that faulty cases are real, I did not make them up --> which tells that there are random error with Spark. However, I believe DJI will not official announce there is issue with the hardware. Why, you ask? Do a simple maths: if DJI announced globally that there are some faulty units out there, all users will request to send the product back for checking/ fixing. And not only Spark, but Mavic, Phantom, Inspire as well. That will cost them a fortune. And please note, seems like there are only some defected items only, I also think that it is not necessary. Rather than that, DJI just need to focus on develop the support team, and when there are defected cases, they will fix for those individual only (and they can even charge back customers IF customers cannot prove it is not pilot error). The picture is quite clear there.

3.My conclusion:
So you will think: oh no, I am so scared, I want to stop flying, I want to return Spark, I don't like DJI anymore. PLEASE DON'T. As I said earlier, we all think that Phantom, Mavic, and other are fixed after a few updates, but who knows, maybe that was just because DJI have successfully replaced those faulty units already, and see how happy people are with DJI all around the world. Same case here for Spark: If you don't have any issue that tells you are having a good product -> Keep flying, because there is nothing else you could do anyway, nothing happen yet means it is a good product. If your Spark is falling from the sky or not working, that tells you are having a bad quality product and should definitely go through DJI customer care (and hopefully DJI will be fair in compensation for customer).


So...guys, thank you for all of your inputs, comments and support until now for this falling matter, but I think I will drop this topic for now because I do not have access to internal code for reviewing more. And you can see that not many users are having issues. If the cases of falling keep increasing, then surely DJI cannot stay quiet (I believe need to be at least the number is around 2,500 units like Gopro Karma incidents for DJI to actually do something).


My advice for new users are:
-Do not fly recklessly.
-Follow the pre-flight checklist.
-Be aware with the battery, I would suggest to a Spark firmware refresh with every new battery (Connect to PC, use DJI Assistant 2 to re-run the newest firmware for Spark)

I am new to DJI, Spark is my first product with them, and I start falling in love with drone. As long as DJI focus more on customer assistant, and be fair in solving the issue (For example, should not charge customer for repairing if that is not pilot error), I will surely keep buying more products form DJI.

OK guys, hope this post helps to release your worry about the issue and start enjoying Spark.

By the way, very sad case of Chester Bennington (Linkin Park's vocalist) today. I hope he Rested in Peace.

2017-7-20
Use props
cosmic68
Captain
Flight distance : 89042 ft
United Kingdom
Offline

Hey thanks for this-do you know how to update the batteries on a Mac? I downloaded DJI Assistant 2-run it on iMac, connect batteries via USB and...nothing happens? Batteries are connected to Mac on charging hub which is also plugged in to wall outlet. Thanks.
2017-7-20
Use props
Phuong Do
Second Officer
Flight distance : 29800 ft
Vietnam
Offline

cosmic68 Posted at 2017-7-20 23:26
Hey thanks for this-do you know how to update the batteries on a Mac? I downloaded DJI Assistant 2-run it on iMac, connect batteries via USB and...nothing happens? Batteries are connected to Mac on charging hub which is also plugged in to wall outlet. Thanks.

Hi, it is not the firmware for battery, it is the firmware for Spark. So I meant you re-run Spark firmware (and with the new battery on). I found that helps with my case.
2017-7-20
Use props
Prosto
lvl.1
Russia
Offline

At last -- a voice of reason among the hysterical outcry! Well put, Do!
2017-7-20
Use props
hallmark007
Captain
Flight distance : 9812789 ft
  • >>>
Ireland
Offline

I think the problem here is we don't or probably won't know the answer.
One day everyone is convinced it's battery , next day it's firmware then it's hardware, yet we have not see enough evidence in any of these to sat conclusively what it is, one thing seems common in the small few cases here is.
Most if not all were hovering, most flipped over power cut and dropped, why when power cut happened did these AC not fall straight down always flipping.
So here's my theory , we know when Spark is hovering we can catch it turn it just over 90 degrees and motors will cut. I.e. Flip it. This is obviously caused by some sensor , could the setting for this sensor be slightly off in some sparks causing it to shut down motors before AC reaches beyond 90 degrees.
Was wind a factor tipping spark to set off this sensor.

Just another theory to add to the mix.

One thing I will say this is a big problem in a very small amount of sparks.
2017-7-21
Use props
Phuong Do
Second Officer
Flight distance : 29800 ft
Vietnam
Offline

hallmark007 Posted at 2017-7-21 01:18
I think the problem here is we don't or probably won't know the answer.
One day everyone is convinced it's battery , next day it's firmware then it's hardware, yet we have not see enough evidence in any of these to sat conclusively what it is, one thing seems common in the small few cases here is.
Most if not all were hovering, most flipped over power cut and dropped, why when power cut happened did these AC not fall straight down always flipping.

That could be another cause Mr. Hallmark, but I am afraid I have to reject that assumption. Because:
-Yes, I agree that motors will shut off when you flip AC over, but please remember, that will only shut-off motors but battery is still on, you can try that and see the battery LED still on, aicraft LED still on.
However, the falling cases reported that the battery was shut-off when it landed on the ground.

I understand your point in explaining the confusion between hardware and software, but if you read the logic in my thread above again you can see that it not likely be the aircraft software here.
2017-7-21
Use props
BriRi
lvl.3
Flight distance : 1538845 ft
United States
Offline

Phuong Do Posted at 2017-7-21 01:41
That could be another cause Mr. Hallmark, but I am afraid I have to reject that assumption. Because:
-Yes, I agree that motors will shut off when you flip AC over, but please remember, that will only shut-off motors but battery is still on, you can try that and see the battery LED still on, aicraft LED still on.
However, the falling cases reported that the battery was shut-off when it landed on the ground.

You are correct Phuong Do.

The shut offs that concern me seem to involve complete shut offs, not just the motors, but the lights and entire craft. Those cases may be a small fraction of SPARKS on this forum but only a tiny fraction of people who own a SPARK ever use this forum. I don't personally know anyone else who has a spark, but I know at least 10 people with Phantoms and Mavics and NOT ONE of them ever uses DJI forums or PHANTOM/Mavic Pilots forums.

Its too bad DJI removed the list of reported shut offs from this forum. I bet it would be close to 50 by now.

I hope they get the problem fixed. It appears the fly away problem may be real in a small percentage of cases as well.
2017-7-21
Use props
userc2f3482696
lvl.3
Flight distance : 3529009 ft
  • >>>
Cyprus
Offline

The fan of my spark was acting strange for 1-2 hours..it was 30-33 degrees celcius,the spark was already hot enough,and the fan was switching randomly off and back on after 2 seconds while the spark was too hot..then i finished my flying and went to the hotel to download the videos..so i left the spark on for 15 minutes,also to see how the fan will act..while i downloading the videos,the fan did it again,for 3 times,then went into full speed..for that moment and later on,and it's been 10 days from that day,i never had any problem..the fan working perfectly,(when it's not too hot is working medium speed and when it's hot at full speed)..so I don't know if other sparks have this problem causing them to overheat.. also that day after finishing 1 battery i was wait 10 minutes to fly again..
2017-7-21
Use props
SparkDad3
New
Flight distance : 207188 ft
United States
Offline

Thank you so much for this!
2017-7-21
Use props
Phuong Do
Second Officer
Flight distance : 29800 ft
Vietnam
Offline

Guys, guys, I just linked up 1 more thing and that also strengthens my conclusion: it is nothing to do with the software, and we are dealing with quality control here. You can ask Mr. SparkDad3 above, he got his Spark fell twice, already sent back to DJI for fixing and also on the way back to him, expecting next Monday. (Mr. SparkDad3, if you can please help to confirm the information above, thanks):
--> So, logic, logic again: By fixing and sending back to Mr. SparkDad 3, I believe DJI must be confident in the fix (Otherwise, they should keep it until they found out the reason and fix), which means they know the issue, and they can fix it. Correct?
So, let say again if it is the software matter: seeing that the cases of falling Spark keep increasing recently, it does not make sense at all that DJI does not release the firmware right away (But wait until next week). The only explanation that make sense is: there are some bad hardware products were released. And firmware release maybe will fix some minor glitches and behavior but I think it has nothing to do with the falling cases.
So, hope this makes sense again, and nice weekend guys. Let's wait to see what the new firmware can do.


2017-7-22
Use props
Didiko
First Officer
Flight distance : 18399 ft
Belgium
Offline

Pretty worrying as the limited flight time of the Spark encourages the purchase of several batteries.

I am currently using 3 of them to try to maximise my flight time per session.

Let's hope it happens on a very small amount of batteries...
2017-7-22
Use props
Phuong Do
Second Officer
Flight distance : 29800 ft
Vietnam
Offline

Didiko Posted at 2017-7-22 05:46
Pretty worrying as the limited flight time of the Spark encourages the purchase of several batteries.

I am currently using 3 of them to try to maximise my flight time per session.

I can only give you advice for now that to minimize the error, be cautious:
-It is advisable to refresh the Spark firmware with the new battery. (No one prove this, but this helped me, so...)
-After all the refresh, do the IMU and compass calibration, in the open area with no metal objects and interference around.
-On the phone, goes to airplane mode, close all the background apps.
-For Android only: after the RC and AC connected, plug in the OTG, do not open DJI Go 4 yourself, let Android open automatically. If it does not, plug out and replug. I found my signal is solid this way (Note: This is a tip from Mavic User as well, it works for them).
-Let the aircraft cool down about 5-10mins between each battery. (Seems like it is not heat issue, but it doesn't harm to be cautious)
-If you notice any thing strange, it is always advisable to reset every thing such as: reinstall DJI Go 4, Refresh firmware and recalibrate IMU, compass, Remote. (Please note some guy complained after follow my advice to refresh firmware, his connection getting bad. I strongly believe it is not the case, reset the app and firmware will not cause any problem unless the firmware from DJI has problem itself to begin with).
-If you travel to another country, strongly do all the reset and calibration before flying.
-After all if fail, maybe you are unlucky to get the bad quality product, provide all the information to DJI for repair.
2017-7-22
Use props
Didiko
First Officer
Flight distance : 18399 ft
Belgium
Offline

Phuong Do Posted at 2017-7-22 06:11
I can only give you advice for now that to minimize the error, be cautious:
-It is advisable to refresh the Spark firmware with the new battery. (No one prove this, but this helped me, so...)
-After all the refresh, do the IMU and compass calibration, in the open area with no metal objects and interference around.

Many thanks for your advice.

Fingers crossed
2017-7-23
Use props
moh_rafiudeen
lvl.4
Flight distance : 328957 ft
United States
Offline

Phuong Do Posted at 2017-7-22 06:11
I can only give you advice for now that to minimize the error, be cautious:
-It is advisable to refresh the Spark firmware with the new battery. (No one prove this, but this helped me, so...)
-After all the refresh, do the IMU and compass calibration, in the open area with no metal objects and interference around.

You really do a good service. Thanks a lot for your tips.
2017-7-25
Use props
Phuong Do
Second Officer
Flight distance : 29800 ft
Vietnam
Offline

moh_rafiudeen Posted at 2017-7-25 04:02
You really do a good service. Thanks a lot for your tips.

Hi, No problem, actually I am not working for DJI. I am also just a user like you.
All those advice based on my experience with Spark so far.
2017-7-25
Use props
R&L Aerial
Captain
Flight distance : 298100 ft
  • >>>
United States
Offline

Hardware vs software, when I first received my mavic I had a common VPS issue that would cause the mavic to go into landing mode every time I pulled back on the throttle. DJI issued a firmware update to fix the issue and for some mavic pilots, it worked. For others with the exact same issue, it did not and the were forced to send there mavics in for core board replacement...
2017-7-25
Use props
Phuong Do
Second Officer
Flight distance : 29800 ft
Vietnam
Offline

Not sure if my conclusion is true or not, but you can see in the new update, they asked you to update with each battery. After I ran update, I plug in another battery and it said something about conflict software and ask you to update with the new battery again...So, make sure you run the firmware with each battery.
2017-7-26
Use props
OneMatt
lvl.4
Flight distance : 155984 ft
Canada
Offline

hallmark007 Posted at 2017-7-21 01:18
I think the problem here is we don't or probably won't know the answer.
One day everyone is convinced it's battery , next day it's firmware then it's hardware, yet we have not see enough evidence in any of these to sat conclusively what it is, one thing seems common in the small few cases here is.
Most if not all were hovering, most flipped over power cut and dropped, why when power cut happened did these AC not fall straight down always flipping.

You may be on to something, an IMU issue. Perhaps not the same as the full shut down, but something.

If the IMU thinks it is being tilted, it will fight it (like when you grab it and flip it). When it reaches 90, motors shut. So, if the IMU thinks it is being rolled, it will flip itself over and shut off. This explains the flip death.

IMU issues would also explain the flyaways, as if it believes it is off level, it will try to self level, which leads to a banking causing the Spark to fly away. This is why the Phantom and Mavic have dual redundant IMUs, to correct for errors. Surprising DJI wouldnt have this, knowing the issue from the past.

OP: i know you rejected this idea, but is it possible that this is the cause of the shutdowns, however since the aircraft fell some distance that power was knocked out on impact ?
2017-7-26
Use props
ImHereToCrash
Captain
Flight distance : 5381368 ft
  • >>>
United States
Offline

OneMatt Posted at 2017-7-26 05:56
You may be on to something, an IMU issue. Perhaps not the same as the full shut down, but something.

If the IMU thinks it is being tilted, it will fight it (like when you grab it and flip it). When it reaches 90, motors shut. So, if the IMU thinks it is being rolled, it will flip itself over and shut off. This explains the flip death.

i believe the IMU issue for fly aways, i have seen it fully banking forward...on a few of mine, explains why it goes full speed away..     thats a good catch.  i didn't think of that..

DJI, for $500/unit price, could have have given us a few extra redundancies.   i can  buy an argument why occusync was skipped... being a selfie drone nd smaller..semi fair.  but no dual redundant IMUs or compasses?  come on!
2017-7-26
Use props
Kyokushin
Captain
Flight distance : 296381 ft
Poland
Offline

You will never know what was the cause as it was in a case of P3P stress cracks and RC changes.

DJI never admitted it was fautl (cracks), someone got after 2 flights, someone as me never, someone got replaced, but it was never explained. And as now it touched a small number of copters.

In my opinion - as i readed what You all wrote in few topics, and what i saw in 2 years on DJI forum - its rather hardware malfunction - its result of quality control.

What is happening probably:
- something is overheating and fuse is turning off all
- short circuit (may be caused by temperature in some board)
- no contact - also may be caused by a temp.
- vibration causes power loss

first two we cant fix, and nobody can, it can be only replaced, but DJI never admit it was faulty. Even if they find solution they will silenty improve production process, nothing more. Maybe they even did it already.

the third - i had few times a problem in my P3P with connectors - the fix was simple - a scratch a metal surface of connector by a something. The connectors was covered by i do not know what and caused no electric pass. After cleaning/scratching it backed to normal. Maybe similar situation is here and you just need to keep clean battery connectors, or even scratch them a little

the fourth - you may use some strip to keep battery on place. Vibrations may cause a loose of contact for a moment.

A person whose Spark fallen may make a test - put battery on bird, turn it on without props then use a little force on battery to make vibrations then observe it is possible to make quick power off by that action.
In other test - a person whose Spark fallen can clean battery connectors then try to hover by few battery charges on that same battery.

Even it is software failure, then you will never see it on changelog. DJI will never admit mistake because it will cause returns, costs, sues.

If You not figure out by Yourself then You will NEVER know what is going on.

Spark is a product with a lifecycle, probably one year. After one year nobody will care of you and your fear bout falling spark. There will be Spark 2.
You can think 'nooo, DJI will never do that!' Yes they will, and they did in a case of P3P stress cracks and few other things.

For testing it you do not need to fly and risk. You may just start engines without props.
2017-7-26
Use props
ImHereToCrash
Captain
Flight distance : 5381368 ft
  • >>>
United States
Offline

Kyokushin Posted at 2017-7-26 13:12
You will never know what was the cause as it was in a case of P3P stress cracks and RC changes.

DJI never admitted it was fautl (cracks), someone got after 2 flights, someone as me never, someone got replaced, but it was never explained. And as now it touched a small number of copters.

sadly how most companies operate.

that being said it would be better if DJI admitted outright to a potential problem that is being investigated.. even if they had to limit our sparkies a bit as a precaution until they find out what is causing it and what to do about it.  

the way i see... its worth it for a company to do the embarrassing and admit a problem.  to many companies dont... and consumer trust is usually a big selling point..   people want to return to what they are familiar with and who they trust.   similar to samsung note 7.. not that many people actually  mock samsung for it, few that occasionally joke about... but compared to the iPhone 4 antenna gate.. people still today, like 7 years later use the "your holding it wrong" meme still...

even if it is a small fraction of Sparks...  they are only causing more problems for themselves potentially further down the road if they cannot fix this problem in a more timely manner.  it starts straying off more from "maybe  something wrong here"  to "company trust issues" to "legal issues" the longer they wait.      
2017-7-26
Use props
Kyokushin
Captain
Flight distance : 296381 ft
Poland
Offline

Few will fall, they will correct hardware or software, then finally they will stop falling, problem solved, and at the end - pilot error - someone will say, then dji will stop answer these topics or will do extremely rare or offtopic.
No proofs, just few folks with crashed sparks (some will get replacement to shut them up) against a corporation with no spark fallen since few months. After a year no one will remember.
Its sad, but it will happen.
2017-7-26
Use props
tompiotrowski
lvl.3
Flight distance : 2814265 ft
  • >>>
Australia
Offline

"Cross your heart and hope to die"... How many of you who experienced Spark falling of the sky, crashed this very drone before?  Did your Spark battery get disengaged?  Were the tiny pins above the latches left intact? When my Spark crashed few weeks ago after going "haywire", the battery upon impact got dislodged and one of the pins broke off.  I have attempted to reinsert the battery, but noticed immediately how insecure fit was from then on.  I suspect that this could be the primary cause for Sparks loosing power and falling down.  Whilst Spark's unibody seems to be very robust, its battery attachment is very delicate, good enough when two latches are engaged, but not necessarily when only one of them is operational.
2017-7-26
Use props
harrismix
lvl.1
United States
Offline

OneMatt Posted at 2017-7-26 05:56
You may be on to something, an IMU issue. Perhaps not the same as the full shut down, but something.

If the IMU thinks it is being tilted, it will fight it (like when you grab it and flip it). When it reaches 90, motors shut. So, if the IMU thinks it is being rolled, it will flip itself over and shut off. This explains the flip death.

but it's only happening when the Spark is used with a controller - no crashes when used with a phone
2017-7-27
Use props
PaTaRnAk
lvl.4
Flight distance : 83383 ft
United States
Offline

tompiotrowski Posted at 2017-7-26 20:12
"Cross your heart and hope to die"... How many of you who experienced Spark falling of the sky, crashed this very drone before?  Did your Spark battery get disengaged?  Were the tiny pins above the latches left intact? When my Spark crashed few weeks ago after going "haywire", the battery upon impact got dislodged and one of the pins broke off.  I have attempted to reinsert the battery, but noticed immediately how insecure fit was from then on.  I suspect that this could be the primary cause for Sparks loosing power and falling down.  Whilst Spark's unibody seems to be very robust, its battery attachment is very delicate, good enough when two latches are engaged, but not necessarily when only one of them is operational.

Never crashed mine before it fell... And I always double check before takeoff if the battery is well seated....
2017-7-27
Use props
fans5d9349b7
lvl.3
United Kingdom
Offline

harrismix Posted at 2017-7-27 00:49
but it's only happening when the Spark is used with a controller - no crashes when used with a phone

Mine died in the air only connected to an iPhone 7. I don't have an RC.
2017-7-27
Use props
OneMatt
lvl.4
Flight distance : 155984 ft
Canada
Offline

harrismix Posted at 2017-7-27 00:49
but it's only happening when the Spark is used with a controller - no crashes when used with a phone

Not true at all. Mine crashed with neither pbone or controller connected.
2017-7-27
Use props
fans457818dd
New
Flight distance : 22346 ft
United Kingdom
Offline


I also had a fault.

This is what happened, I was testing active track. The spark went into return home mode at 20% battery, I over rid it and was flying it in sport mode. I made a sharp turn and the SPARK just fell out the sky.

It did not seem to suffer any damage upon inspection, but I now have a little less confidence in the unit as I don't think it would fair so well if it fell from 100s of metres.

DJI didn't seem to offer any answers or show great concern when I reported it to them.

2017-7-27
Use props
rtanner
lvl.1
Flight distance : 410453 ft
United States
Offline

I purchased two Sparks and returned them to DJI I feel they are a piece of junk and the gesture mode did not work consistantly
2017-7-27
Use props
Advanced
You need to log in before you can reply Login | Register now

Credit Rules