Well, that was a short love affair with my Spark.

I was flying while on vacation in Durango, CO. When I hit RTH (yes, homepoint was recorded) it flew VERY high. Much higher than the 20M I have it set for RTH.

Then it just hovered there, as I could see from the camera. I trid to take control and bring it home, but it was only intermittently responding.

Next, they battery alarms went off, with several consecutive warnings... It showed it altitude as -274 ft??

Anyway, we have SCOURED the field near its last location for two days.

I've read enough on the forums to know that RTH says it flies straiught up, the to the HomePt and then straight down. It only flew upward to am amazingly high distance.

I have no idea what to do now, its battery is dead, Find My Drone shows its last point but we have had five people searching for two days. Cannot find it. Very Sad.

Cautionary tale, I guess.

Check out your TXT flight log to see if it explains what happened. You can upload and view it online here. If you'd like other people to review and comment on your flight log, then please post a link back here after you upload it.

Someone will be on here any minute to tell you how to upload the log files to be examined.   When I lost mine once I had not realized that it had been in record when it crashed into a tree.. it wasn't until I later looked at the video that I was able to find exactly where it crashed and I found it...   No help now if you were not recording... but maybe.   It sounds like you had connection to it until the end.. so the last recorded point in the log should be pretty darned close.... Did it have any errors that would have prevented it from returning to home... compass, yaw, atti mode????  If it was in atti mode it would have drifted on the wind until it lost connection with the remote, then it would have landed.. so  look in the direction that the wind was blowing.. but, if as you say you saw it start to land.. then that would not have been the case.. again.. the log file will have a lot to tell.

LouisP Posted at 2018-3-29 13:22
Someone will be on here any minute to tell you how to upload the log files to be examined.   When I lost mine once I had not realized that it had been in record when it crashed into a tree.. it wasn't until I later looked at the video that I was able to find exactly where it crashed and I found it...   No help now if you were not recording... but maybe.   It sounds like you had connection to it until the end.. so the last recorded point in the log should be pretty darned close.... Did it have any errors that would have prevented it from returning to home... compass, yaw, atti mode????  If it was in atti mode it would have drifted on the wind until it lost connection with the remote, then it would have landed.. so  look in the direction that the wind was blowing.. but, if as you say you saw it start to land.. then that would not have been the case.. again.. the log file will have a lot to tell.

Thanks for the encouragement!

This is the end of looking on the second day, so all 5 of us are kinda tired and kinda bummed.

How do I upload the flight data from the iPhone I was using?

See the instructions at the link I posted above.

msinger Posted at 2018-3-29 13:21
Check out your TXT flight log to see if it explains what happened. You can upload and view it online here. If you'd like other people to review and comment on your flight log, then please post a link back here after you upload it.

Sorry, I had missed your post earlier.

Here is the log: http://www.phantomhelp.com/LogViewer/6HFJ3AN42T98207VNQWN/

EDIT:
I notice on the last entry the IMU Alt is: -272.6ft
While the VPS Alt is: 7.9ft

What does that mean?

Thanks for any help!

Did you have any kind of accessories installed on the arms and/or beneath your Spark?

msinger Posted at 2018-3-29 14:13
Did you have any kind of accessories installed on the arms and/or beneath your Spark?

Nothing. It was naked .   : )

Here are some things I can tell you after reviewing your flight log:

• Your RTH altitude was set to the default setting (20 meters or 65 feet).
• The battery was at 57% on takeoff (which is too low to be starting a new flight).
• When RTH was initiated, your Spark ascended to 65 feet (as expected) and then ascended another 20 feet for some odd reason. It appears it hit a very strong gust of wind at 65 feet and I'm assuming that extra altitude was related to what it was doing to attempt to stabilize itself.
• After you cancelled RTH, you had the right stick in the full down position. That made your Spark fly backward away from you at full speed.
• After that point, there were periods where you flew your Spark toward the home point, then away from the home point, and then toward the home point again (it appears you were a little confused).
• The battery reached the critically low level and your Spark started landing at its current location (since it wasn't safe for it to continue to fly). You continued to attempt to fly your Spark at full stick while it was trying to land.
• The negative altitude during the auto landing procedure is a bit puzzling. I'm not sure why/how that occurred.
• According to your log, your Spark landing at 37.40516409, -107.8140977. It should be close to that location.
  

FYI, Weather Underground shows the wind in your area was blowing at 30 MPH with gusts up to around 40 MPH (at ground level). See the graph below. The Spark is not designed to fly in high winds like this. You probably should have packed everything up and flew a different day.

Thanks for the thorough analysis.

The only thing I can add was that it was VERY calm when we started the flight. Once it was across the road the wind kicked up pretty quickly and I tried to bring it home.

Ah well, I guess I should sell off the batteries, controller and all the other gear.
Much appreciate the help.

Slow speed of RTH couldn't beat the winds?  Needed Sport mode and Maverick (or Iceman) working the sticks to get it back to the home carrier safely.

Curious...  what firmware was loaded on your Spark?   Was it .600 by chance.   I lost my Spark too.  Was in the sky and just completely disappeared.  DJI will NOT DO ANYTHING for you.  EVEN with an extended warranty/DJI "care".   I hate tell you this, but after you made your purchase, DJI has ABSOLUTELY ZERO intention of helping you.   DJI is primarily a company that ONLY cares about its CORPORATE customers.   DJI will only support you if your buying one of their industry products, like the Inspire or the Ronin.   DJI simply doesn't care about their consumer market.   FACT!  

@Wrathpanda, did you plug in the 37.40516409, -107.8140977 coordinates into an app like Google Maps and go right to that spot?

msinger Posted at 2018-3-29 17:37
@Wrathpanda, did you plug in the 37.40516409, -107.8140977 coordinates into an app like Google Maps and go right to that spot?

Yes, we did. One used Google, I just downloaded a specific GPS app and plugged in the coordinates to take me there.

We had five people search in an outward spiral from that point. Then we did a sweep and then a criss-cross. This was all day today and much of yesterday afternoon, until the sun set.

Kudos to my family for spending their vaca helping me look. I am certain it is not there.

NerdyGuy Posted at 2018-3-29 17:15
Curious...  what firmware was loaded on your Spark?   Was it .600 by chance.   I lost my Spark too.  Was in the sky and just completely disappeared.  DJI will NOT DO ANYTHING for you.  EVEN with an extended warranty/DJI "care".   I hate tell you this, but after you made your purchase, DJI has ABSOLUTELY ZERO intention of helping you.   DJI is primarily a company that ONLY cares about its CORPORATE customers.   DJI will only support you if your buying one of their industry products, like the Inspire or the Ronin.   DJI simply doesn't care about their consumer market.   FACT!

It was whatever is current. I updated the spark and batteries just before we left a week ago.

Terribly sorry for your loss, have you contacted our support to start a case? If you haven't, please contact our support on the following page: http://www.dji.com/support

•The battery was at 57% on takeoff (which is too low to be starting a new flight).


Why is 57% too low to start a new flight.  Several times I have flown a short flight on a new battery, landed, powered off, changed a filter, and took back off again.  Also, when filming a sunset, I will takeoff and get a shot or pano and then land and power off for 5 minutes...then takeoff again and do the same thing over and over again.  What is significant about taking off with a fully charged battery???

Wrathpanda Posted at 2018-3-29 19:06
Yes, we did. One used Google, I just downloaded a specific GPS app and plugged in the coordinates to take me there.

We had five people search in an outward spiral from that point. Then we did a sweep and then a criss-cross. This was all day today and much of yesterday afternoon, until the sun set.

Just to confirm without a doubt, here is where your flight log shows your Spark landed:



Due to the odd altitude displayed in DJI GO, I'm not sure if that's the same location you're seeing in Find My Drone.

Rustic17 Posted at 2018-3-29 22:03
•The battery was at 57% on takeoff (which is too low to be starting a new flight).

Starting a new flight that includes a maneuver that involves purposely flying over 900 feet away from the takeoff point probably should never be done on a battery that is only 57% charged. In this case though, you might be onto something. The OP seemed to be unaware of the weather conditions, the low battery warning, and the critically low battery warning. So, more battery power might not have been of much help.

This case can be summed up as a pilot error.

95% of the  'fly away'cases in this forum were solved as pilot errors  and especially while predicting  the wind / gust during the time of flight.
still dont understand why you were toying with the joysticks while the SPARK was in RTH mode.

S.J Posted at 2018-3-30 05:26
This case can be summed up as a pilot error.

95% of the  'fly away'cases in this forum were solved as pilot errors  and especially while predicting  the wind / gust during the time of flight.

Glad you ‘summed up’ this one... 95%? Where are you getting your numbers?

Rustic17 Posted at 2018-3-29 22:03
•The battery was at 57% on takeoff (which is too low to be starting a new flight).

Didn't John Denver take off with a half fuel tank?  Doing some low level work when he ran dry and didn't successfully switch over to reserve/new tank.  Why take off already having a problem?

I would say 50% is enough if you keep the flight short and are current on your Spark emergency procedures (can fly it home yourself in high winds without RTH).

Did this drone land downhill from it's take off point?

Still this is better than the mavic owner who flew outbound over the ocean with a 20mph tail wind until he had less than 50% battery.  He thought it was a software issue that it didn't return.  He thought he should get a free replacement.

First off, I'm very sorry to hear of the loss of your aircraft.  I know it can be rough on the emotions and wallet when that happens.  Below is my attempt to review your logs to try to understand what may have happened.  It is done so in an effort to help to see if I understand and interpret the logs correctly, but also to serve as a learning point for myself, others, and for the pilot for future flights.  Having said that, below is my interpretation.  Let me know if anybody sees something I'm interpreting incorrectly or missing.

After a quick look at the logs, it looks to me like this is a continuation of another flight on the same battery.  Firstly, the first line of the log starts with a time stamp of "4m 49.4s" and the home point is 29.5ft away immediately as soon as the flight is started.  So my assumption is that the pilot took off originally (log would be in a different log file), flew around a bit, and then landed about 29.5ft away from their original take off point.  


Wrathpanda, can you confirm that and do you have any details about that flight that immediately preceded this one?  Did it land normally from the first flight or was it a crash landing?  On this flight did it land into the dirt, snow, bushes, or something like that?

At first glance at the Altitude results recorded in the log files, there are a couple of things that stand out.  As mentioned above by others, the -275ft values are a bit abnormal, but could be explained by elevation changes in the landscape.  I had a quick look via maps and Google Earth, but in the area of the flight, from a digital view via the internet, that doesn't seem likely.   There are terrain changes in the area, but in the direct area of the flight, it doesn't seem like it would be almost 300ft difference.  However hard to tell without being there.

Wrathpanda, any thoughts on elevation changes of the landscape?

Also, the other thing that is odd is that the VPS elevation sensors were registering a value of somewhere around 6-12ft for the entirety of the flight even with the altitude coming from the IMU (i.e. barometer) changing quite significantly.  When I've seen this in the past, it had quite often been a 3rd party accessory (i.e. sunshade, legs, etc) attached to the aircraft.  Other times it has been due to moisture or dirt on the VPS sensors.  

Wrathpanda, is it possible that in the first landing that immediately preceded this flight that the sensors could have gotten something on them?

As MSINGER mentioned, when the RTH procedure was invoked (7min 56s mark), it rose to an altitude higher than the 20m (65ft) value set.  However, my theory for why it did this is not due to wind catching it, but instead because the VPS sensor was registering a value of around 7 or 8ft high and it was using that instead of the altitude being registered by the IMU (i.e. barometer).  So it was trying to continue up to a height of 20m.  It would have continued to go higher and higher however the pilot cancelled the RTH procedure and it went back to GPS mode so it stopped climbing.

The interesting thing is that the aircraft was only about 18ft up and 95ft away with 34% battery left with the pilot in full control of the aircraft when the RTH procedure was invoked.  As a potential learning point, unless there are extenuating circumstances that aren't apparent in the logs, I would suggest the pilot to manually fly the aircraft home in this situation rather than flipping to RTH and asking the computer to do it.

During the RTH procedure, as it was rotating to the correct heading and then trying to rise to the 20m RTH altitude, it appears the wind may have been pushing it a bit further away from home.  My guess is that it was this extra distance and extra height that caused the pilot to choose to exit RTH mode and try to take control again.

Wrathpanda, any comments to add at this point about what you saw in real time?

Once the pilot took control again after the RTH procedure, it looks to me like the aircraft was responding to the pilot's stick inputs as expected.  When the pilot gave stick inputs to move forward, it moved forward.  When the stick inputs were to go backwards, it went backwards.  When stick inputs said to climb, it climbed, etc, etc.

At this point the aircraft was approximately 118ft away at around the 9m 15s marker and with 17% battery left.  From then, the pilot provided stick inputs that led the aircraft much further away from the home point with it eventually getting out to just over 900ft away at the 11m 1s marker with 3% battery left.  

During this trip away from the home point, the battery got to a critical level (around 10%) at the 10m 13s of the flight.  At that point, the aircraft went into AUTO LANDING mode and started to descend.  Even though the battery was only at 10% and was auto landing, the pilot was still providing stick inputs for almost another 50 seconds that continued taking the aircraft further away.

At about the 11m marker with 3% battery left and being about 900ft away, the pilot's inputs changed to start bringing the aircraft back closer to the home point.  The aircraft is still in auto landing mode and descending during this period.

For the last couple minutes of the flight, the pilot was giving the aircraft conflicting inputs at different times.  For a bit, it was coming back closer to home, then the pilot would change the inputs and send it further away.  This continue back and forth for bit until the last 20 seconds when there was very little input into the sticks. while the aircraft continued to descend.  My guess is that it eventually ran out of battery and it was at that point the log ended.  With the way the elevations were a bit jacked up, its a bit hard to know for sure how high exactly it was when the battery finally died.  However, for the last several seconds, it looks like its location was quite steady.  Its GPS coordinates aren't moving hardly at all and its speed is pretty much at 0.  

I would think it should be quite near where the GPS coordinates mention.

Okay, so that is my interpretation.  What have I missed or interpreted incorrectly?

95% of Flyaways are pilot errors? That is a rather a sweeping statement.

Gunship9 Posted at 2018-3-29 17:10
Slow speed of RTH couldn't beat the winds?  Needed Sport mode and Maverick (or Iceman) working the sticks to get it back to the home carrier safely.

adjust the RTH speed to 13 (google is your friend) to beat the wind in RTH
#RTH Speed increase
g_config_go_home_gohome_idle_vel = 13

True there is a bunch of pilot error in this one.. .but, that IMU altitude and vision altitude is not right.. poor spark didn't know where it was in the sky.   It does look like the IMU was faulty during this flight.  I would suggest that you open a case with DJI and let them take a look at the logs.   They might agree that there was a hardware problem and give you  a discount on getting a new spark.  Overall this was one of the more interesting logs to examine... no real errors at all from the spark.

DJI Thor Posted at 2018-3-29 21:28
Terribly sorry for your loss, have you contacted our support to start a case? If you haven't, please contact our support on the following page: http://www.dji.com/support

Thank you. Yes, filing now.

S.J Posted at 2018-3-30 05:26
This case can be summed up as a pilot error.

95% of the  'fly away'cases in this forum were solved as pilot errors  and especially while predicting  the wind / gust during the time of flight.

@SJ -
Don't be too eager to proclaim pilot error so quickly.

You were not there, so you wouldn't know that it was acting erratically. The altitude readings were unintelligible, so I used the down stick to try to see if I could actually hear it as it got closer to the field.

Yes, at four months in,  I am a noob at this, but I also love everything about it. Were you born an excellent RC pilot and have never made an error? Maybe compassion would have been the right posture. Other people's posts are very encouraging or consoling.

Not sure what value you have added to my situation?

KlooGee Posted at 2018-3-30 09:58
First off, I'm very sorry to hear of the loss of your aircraft.  I know it can be rough on the emotions and wallet when that happens.  Below is my attempt to review your logs to try to understand what may have happened.  It is done so in an effort to help to see if I understand and interpret the logs correctly, but also to serve as a learning point for myself, others, and for the pilot for future flights.  Having said that, below is my interpretation.  Let me know if anybody sees something I'm interpreting incorrectly or missing.

After a quick look at the logs, it looks to me like this is a continuation of another flight on the same battery.  Firstly, the first line of the log starts with a time stamp of "4m 49.4s" and the home point is 29.5ft away immediately as soon as the flight is started.  So my assumption is that the pilot took off originally (log would be in a different log file), flew around a bit, and then landed about 29.5ft away from their original take off point.  

Wow -- you also did an amazing analysis. Thanks for spending the time.
I will try to answer each question below:

Wrathpanda, can you confirm that and do you have any details about that flight that immediately preceded this one?  Did it land normally from the first flight or was it a crash landing?  On this flight did it land into the dirt, snow, bushes, or something like that?

Prior flight took off and landed on a cement patio. After it landed we took off again from the same spot. After the first flight, it landed just fine (no crashes).

Here is the log from the previous flight:  http://www.phantomhelp.com/LogViewer/X6PQUZXBSHIGSPH2EE6U/


Wrathpanda, any thoughts on elevation changes of the landscape?
The field was gently sloping to the left (or West) but nothing like a 300 ft drop. Change in elevation was maybe 20 feet all the way to the road you see to the west.


Wrathpanda, is it possible that in the first landing that immediately preceded this flight that the sensors could have gotten something on them?
Nothing was on the Spark, not even the blade guards. It was just stock, if you will.

During the RTH procedure, as it was rotating to the correct heading and then trying to rise to the 20m RTH altitude, it appears the wind may have been pushing it a bit further away from home.  My guess is that it was this extra distance and extra height that caused the pilot to choose to exit RTH mode and try to take control again.
Exactly. After we hit RTH, it actually ascended quite far (judging by the view from the camera). The house was looking tiny. It was also just hovering, no movement. One thing that is very disconcerting is the persistent alarms that sound when the battery is low. Yes, I know it needs to come home now... its gut wrenching to see it just hover very high and watch the battery drain. So I tried to take control. I was able to bring it down, but it was reacting "in spurts" almost like a very laggy video game... Nothing would happen and then it would exectute a bunch of commands. Note at this point the camera was also only updating every few (3-5) seconds. So for a noob, this was very challenging.

As many people have said, it should be somewhere near the final GPS point. It simply isn't. We looked for two days. I don't know where it went, but we looked all over the area. Nothing.

As I said, many thanks for the thoughtful analysis and the time spent helping me.

Wow, this community rocks!

I am amazed at how much effort and encouragement you provide each other. Having been on a LOT of other R/C forums, what you folks have here is truly unique and impressive.

We have left Durango CO, and maybe someone will find the Spark and call my phone number written on the drone.

I will contact DJI and send them my logs for analysis.

I hope to get better at flying an learn more in the coming months.
If any of you live near Boulder, CO -- I'd like to buy you a beer for helping!

Many thanks to all for the help.

Wrathpanda Posted at 2018-3-30 16:14
@SJ -
Don't be too eager to proclaim pilot error so quickly.

Hi panda,
Just leave your emotions behind.
I was not pointing at you but replying in general to all the posts that i have read in this forum so far
You may have a point but the conclusion from my side is a pure pilot error which could have been avoided .

I agree with S.J here pilot error but us seasoned pilots need to remember the Spark is marketed as a safe smart drone and we all know it isn't , DJI really fooled a lot of us into spend big sums of money for a safe smart advanced drone when in fact most of the safety systems go wrong too often and make the drone dangerous to fly and fly away, i suffered battery drop out in the first three weeks of ownership in august last year, now on my third Spark because of issues all relating to DJI.

We need to advise the community of new members that the Spark is still in beta phase , always will be.

battery drops, imu and yaw errors, your new Spark is bound to crash or fly away unless you have a lot of prior experience fly ATT mode.

See if you can run a Diagnostic DJI Assistant 2 on your last flight to see what the trend was or where in the flight something went wrong?

davidmartingraf Posted at 2018-3-31 03:06
See if you can run a Diagnostic DJI Assistant 2 on your last flight to see what the trend was or where in the flight something went wrong?

You should also open up a case with DJI Support and send them the information you have from your last few flights saved from the Spark to see if they notice any trends from past flights that perhaps showed something in the upcoming, last flight?

@xtro, my flight log viewer does not allow the data columns to be modified if that's what you're wondering.

KlooGee Posted at 2018-3-30 09:58
First off, I'm very sorry to hear of the loss of your aircraft.  I know it can be rough on the emotions and wallet when that happens.  Below is my attempt to review your logs to try to understand what may have happened.  It is done so in an effort to help to see if I understand and interpret the logs correctly, but also to serve as a learning point for myself, others, and for the pilot for future flights.  Having said that, below is my interpretation.  Let me know if anybody sees something I'm interpreting incorrectly or missing.

After a quick look at the logs, it looks to me like this is a continuation of another flight on the same battery.  Firstly, the first line of the log starts with a time stamp of "4m 49.4s" and the home point is 29.5ft away immediately as soon as the flight is started.  So my assumption is that the pilot took off originally (log would be in a different log file), flew around a bit, and then landed about 29.5ft away from their original take off point.  

Hi,

Great analysis! Your story can be seen on this chart.
See the lower chart ; moving heading is backwards flying away form homepoint.

cheers JJB

m1n1s Posted at 2018-3-31 00:23
I agree with S.J here pilot error but us seasoned pilots need to remember the Spark is marketed as a safe smart drone and we all know it isn't , DJI really fooled a lot of us into spend big sums of money for a safe smart advanced drone when in fact most of the safety systems go wrong too often and make the drone dangerous to fly and fly away, i suffered battery drop out in the first three weeks of ownership in august last year, now on my third Spark because of issues all relating to DJI.

We need to advise the community of new members that the Spark is still in beta phase , always will be.

Define safe.  Safe smart drone like my unstabilized helicopter is, or my model piper cub?  Even at the Spark's worst, it will keep itself upright in ATTI mode.  You still need to be able to fly a model in that you know left from right and forwards and backwards.

It is debatable whether DJI was advertising the Spark's ability to fly over the horizon, out of sight of the pilot, and still always return home if it runs into problems.  It is a $400 drone and no cheap autopilot is that good.  Mavic's are lost too and they have dual compasses and IMU's.  Even Predators are lost on occasion.  What does a predator have?  Dual or triple autopilot?  So, radio control flight, beyond the pilot's sight is still iffy and always will be.   

Flying a Spark in eyesight and close enough to maintain orientation with it, is safe, once you learn to fly a drone.

Gunship9 Posted at 2018-3-31 05:33
Define safe.  Safe smart drone like my unstabilized helicopter is, or my model piper cub?  Even at the Spark's worst, it will keep itself upright in ATTI mode.  You still need to be able to fly a model in that you know left from right and forwards and backwards.

It is debatable whether DJI was advertising the Spark's ability to fly over the horizon, out of sight of the pilot, and still always return home if it runs into problems.  It is a $400 drone and no cheap autopilot is that good.  Mavic's are lost too and they have dual compasses and IMU's.  Even Predators are lost on occasion.  What does a predator have?  Dual or triple autopilot?  So, radio control flight, beyond the pilot's sight is still iffy and always will be.   

DJi market the drone for first time pilots , with all its safety systems.

Although as we have seen the Spark can fly off even in guesture mode and usually within VLS , so i'm saying its not as safe as advertisted , you really need to have pervious experience flying drones non-attitude hold or gps and the like.

There seems to be three problems with the Spark , non-FCC mode , the RC so you can fly it out of VLS and the poor design on the battery.

Spark was design for the USA market , DJI sell off your images so rich chinese can buy up your land

JJBspark Posted at 2018-3-31 05:28
Hi,

Great analysis! Your story can be seen on this chart.

Thanks JJB!  Your log analysis tool is looking awesome!