Welcome to PhantomPilots.com

Sign up for a weekly email of the latest drone news & information

DJI Double Defects Crash! TBE & Low Voltage Forced Landing!

Discussion in 'Phantom 2 Discussion' started by stmwise, Jul 24, 2014.

  1. stmwise

    Joined:
    Jul 23, 2014
    Messages:
    12
    Likes Received:
    0
    My DJI Phantom 2 crashed into the Caribbean sea due to two known DJI defects. First it was the Toilet Bowl effect moved it 170 meters away from home point above the water and then low voltage battery effect forced it to auto land straight down into the water. All were done without any pilot input! This disaster unfolded right after the quadcopter was being lifted from ground. It has been calibrated and was flying in GPS mode when the disaster hits.

    I have been flying the Phantom 2 for the past 4 months. This video was recorded on my PC as I have been very careful with every flight and record each flight i took(Both on PC feed through FPV video transmitter from Phantom and also onboard GoPro camera recording in HD resolution). Onboard the quadcopter was the GoPro Hero 3 Black camera that was also doing recording when it crashed, I have not yet able to recover the the HD video from the P2. There were no interference at this location as I have been flying there for the past few months. The battery inside the Phantom has only been discharged 45 times according to the Phantom Assistant Software but it has had issues with early auto land at 40% for the past two weeks prior to yesterday's 90% forced auto land. Totally unexpected as you can see. just never knew it would auto land at over 90% battery life this time.

    The toilet bowl effect is known by the Phantom 2 owners community and is still being worked out by DJI.

    So here we have a scenario where two defects happened at once and at the same time! If only had either one happened, it would still have survived with minor damages. But this is now totaled due to full submersion. Over USD$1500 worth of equipment loss. I think DJI needs to be held accountable for this! :evil:

    [youtube]http://www.youtube.com/watch?v=hZNxnwQtFuo[/youtube]
     
  2. LuvMyTJ

    LuvMyTJ ADMINISTRATOR
    Staff Member

    Joined:
    Feb 9, 2014
    Messages:
    6,811
    Likes Received:
    4,376
    Location:
    Live! From New York!
    Re: DJI Double Defects Crash! TBE & Low Voltage Forced Landi

    Welcome to the forum, sorry it had to be such bad circumstances. I noticed your satellite numbers were up & down. I have found if there is cloud cover where I usually fly that the satellite number will vary compared to a clear day in the same location. I also see you knew you were having low power landing issues. Continuing to fly with this known issue, especially near salt water is a risky proposition for sure. The TB effect can happen on other quads too, with or without GPS. I have a scratch build X525 that sometimes has issues with the TB effect. Did you try switching to ATTI mode or the return home function when you noticed it was out of control? Did you lose stick input? Were you able to recover the quad?
     
  3. stmwise

    Joined:
    Jul 23, 2014
    Messages:
    12
    Likes Received:
    0
    Re: DJI Double Defects Crash! TBE & Low Voltage Forced Landi

    I consider myself a competent pilot as i have months of experience flying the phantom 2. This disaster strikes totally unexpected and I was still pushing the left throttle stick in my attempt to push it higher to avoid the tree branches. I didn't even touch the right side pitch & roll stick at all. It flew on its on own due to the Toilet Bowl effect and what i was focusing on was the phantom's altitude and the surrounding obstructions from my screen. It went quickly out to the sea and there came the low voltage forced landing. All the while when i was still trying to push it higher on my left throttle. But from my experience before, the low voltage auto land will overpower any of my stick commands making the phantom not responding to any of my inputs. As you can see from the iosd shown in the video of the crash, the voltage dropped below 11.7V and went down as low as 11.4V just about the time it hit the water.

    I put down my remote controller and jumped into the water swimming towards the location where it went down but i couldnt find it. Luckily there was someone who had a snorkel set swimming and saw what happened. He came to me and helped me to search for the phantom and eventually he found it back. By the time when we got back onshore the battery started to burn and we saw the smoke coming from inside so we had to cover it with sand and eventually got rid of it afterwards.

    I can see the phantom is already rusted from salt water due to the full submersion yesterday. It is totally damaged beyond any possibility to repair. I've lost the Phantom 2, the 2 axis gimbal, the GoPro Hero 3 black camera, an ImmersionRC 600mw video transmitter and the 5.8Ghz spironet cloverleaf antenna.
     
  4. ianwood

    ianwood Taco Wrangler
    Staff Member

    Joined:
    Jan 7, 2014
    Messages:
    4,929
    Likes Received:
    1,800
    Location:
    Lost Angeles
    Re: DJI Double Defects Crash! TBE & Low Voltage Forced Landi

    This flight was troubled from the start. Before takeoff, is the gimbal twitching on its own or is the Phantom being moved/jostled repeatedly? Did you power up the Phantom and then place it on the ground? Both will be bad for IMU initialization. Also, the gimbal horizon was way off at start up. That's usually an indication that something is wrong.

    How windy was it? Those clouds are moving at a pretty steady clip and the trees swaying. There is lots of violent pitching and rolling, rates of >15 degrees/s. At one point, you were at 45 degree roll and 35 degrees pitch. At another, you hit 60 degrees roll, practically sideways.

    :shock:

    I can't tell if there was any TBE as the craft was never stable enough. Voltage looked good all the way through. Autoland is at 10.65V. I'd guess the wind was very blustery and well in excess of safe flying conditions or there was an IMU issue.
     
  5. stmwise

    Joined:
    Jul 23, 2014
    Messages:
    12
    Likes Received:
    0
    Re: DJI Double Defects Crash! TBE & Low Voltage Forced Landi

    The gimbal was twitching on its own and that shouldn't be a reason leading to the crash as it had happened every flight i had before, it would usually stabilize the camera within minute or two. The phantom 2 itself was staying in same spot with no noticeably abnormal behaviour before the takeoff. But yes, the phantom was powered on and being put on the ground afterward. I have always waited until the GPS home lock is initialized before taking off and that was the case too before the crash.

    I have not touched either the pitch or roll stick myself but i do see the movement was very unusual when it started flying on its own. It touched the water after completing the arc resembling a 1/4 of a circle as in a TBE. Wind condition was about 15mph yesterday.

    The Phantom 2 always lock its altitude even when experiencing the Toilet Bowl effect before, I just waited for it to correct itself before i start to fly. But instead of locking its altitude at approximately 25 meters height, it AUTO LANDED this time ignoring my full throttle stick upward position. It was a painful experience.
     
  6. ianwood

    ianwood Taco Wrangler
    Staff Member

    Joined:
    Jan 7, 2014
    Messages:
    4,929
    Likes Received:
    1,800
    Location:
    Lost Angeles
    Re: DJI Double Defects Crash! TBE & Low Voltage Forced Landi

    This is bad. The Phantom needs to be stable/still for a good 10-20 seconds after startup so it can initialize the IMU properly. It may not always cause a problem but all it takes is one time. This is most likely the cause of your crash. There is no evidence in the video of autoland. The last viewable part of the video shows extreme changes in attitude likely caused by a bad IMU initialization which resulted in loss of height and the subsequent crash.
     
  7. tom3holer

    Joined:
    Mar 20, 2014
    Messages:
    324
    Likes Received:
    2
    Location:
    Cape Cod, Ma.
    Re: DJI Double Defects Crash! TBE & Low Voltage Forced Landi

    I do realize that when these things happen it can catch you off guard. I am surprised that you didn't use the right stick to try and get it back to your location.

    Tom
     
  8. Buk

    Buk

    Joined:
    Feb 28, 2013
    Messages:
    654
    Likes Received:
    29
    Location:
    Iowa, United States
    Re: DJI Double Defects Crash! TBE & Low Voltage Forced Landi

    Ian,
    Thanks for your evaluation, you bring up some interesting points. In the video above, it shows 117 degrees as the Phantom is set on the ground. Does this mean home and course were established before setting the Phantom on the ground? If home and course are not set does the display show a blank until those are established? If it shows something before home and course are established, what is it showing?

    Thanks for pointing out the importance of using the tools available to you before flying.

    Must agree, the wind was really blowing. Look at the treetops at 49 seconds. Treetop really whipping about.
     
  9. ianwood

    ianwood Taco Wrangler
    Staff Member

    Joined:
    Jan 7, 2014
    Messages:
    4,929
    Likes Received:
    1,800
    Location:
    Lost Angeles
    Re: DJI Double Defects Crash! TBE & Low Voltage Forced Landi

    You are correct. Good catch. It had 9 satellites and the heading and home position were recorded. So it would appear that maybe it was held in hand during the entirety of this part of the startup which is a really bad idea.

    Having it on a stable and relatively level surface is important for the accuracy of the IMU which measures relative movement. When you switch it on, the IMU needs to initialize itself. If it is being moved around or jostled, the initialization will not be accurate which could result in unpredictable behavior.

    After about 20 seconds from power up, you can move the Phantom and it should be OK even if it hasn't recorded the home position or heading yet. It's just those 10-20 seconds where it needs to be stable.
     
  10. noiseboy72

    Joined:
    Jul 13, 2014
    Messages:
    575
    Likes Received:
    17
    Location:
    Lincs, UK
    Re: DJI Double Defects Crash! TBE & Low Voltage Forced Landi

    I would not consider 170M to be TBE. This is usually within 10M and is normally seen in calm conditions.

    To my eyes that flight has all the hallmarks of a sick battery and un-calibrated IMU.
    The altitude on take off is recorded at -2M, and with a pitch of 8 degrees at take off, the bird is trying to compensate and failing.

    The battery voltage drops very quickly on spool up and then continues to drop. This suggests that the current delivery of the battery is below par, so the motors are unable to respond quickly. Also the large amount of horizontal movement on take off would have had me landing straight away as well. It might not be the battery, but arcing on the connectors due to the battery not being pushed fully home. I have seen this on a friends conventional helicopter where after a flight we could not separate the battery connectors as they had welded themselves together!

    Do not confuse good GPS lock with good IMU data. You should - actually must have 0 degrees of pitch and roll and an altitude of 0 for take off. Anything else and the IMU will not know where it is starting from.

    Sorry for your loss and I don't discount a failure on your P2, but I don't think you can totally pin this one DJI.

    From the looks of the gimbal, you may not have the latest firmware loaded into it. Although this would not have caused the crash, it might suggest that other firmware was not up to date either.
     
  11. stmwise

    Joined:
    Jul 23, 2014
    Messages:
    12
    Likes Received:
    0
    Re: DJI Double Defects Crash! TBE & Low Voltage Forced Landi

    I have been trying my best to probe the probable cause of this crash for the last few days and I still can't determine the exact cause of this crash other than blaming it on the battery itself. As i have already explained before, the crash happened while being lifted off the ground and the right side pitch & roll stick was not even used throughout the entire flight. It moved on its own course in a circular shape while i was still applying the full throttle stick upwards.

    Now back to the video, as everyone can see it stopped going any higher after reaching altitude of 25/26 meters even though the left throttle stick was still in full upward position. What could have been stopping it from responding to my stick commands other than the battery defect and subsequent auto land? Wind can't possibly stop it from going upwards, it could push it sideways instead. I have took off from this same spot before and despite windy conditions i was still able to fly it over 700 meters above the ground but the wind has pushed it far out by more than 200 meters away. I have videos of my prior flight from this location to prove it.

    Another point discussed here is the IMU calibration that some suggested as the factor in the crash. I have done both the simple IMU calibration as well as advanced IMU calibration in my Phantom Assistant Software just days ago and have had two successful flights after that prior to the crash.

    Now on to the firmwares in my Phantom and gimbal. I am still flying in firmware 1.08 as that is the only firmware I have been using all along and feel satisfied with. I have had good flight experience with it so i didn't want to upgrade to the latest version 3.04 with all the added limitations and restrictions. What's not broken don't fix.
    My gimbal is the Tarot T-2D and I have the latest version firmware on it, which is version 1.5 downloaded from Helipal website right here: http://www.helipal.com/tarot-t-2d-brush ... ero-3.html

    I can only point towards the battery as the main culprit in this case. As I had early auto lands in the past two weeks at 40% battery life instead of 15% that is programmed in the Assistant software. The battery is also having noticeable swelling although not very severe.
     
  12. noiseboy72

    Joined:
    Jul 13, 2014
    Messages:
    575
    Likes Received:
    17
    Location:
    Lincs, UK
    Re: DJI Double Defects Crash! TBE & Low Voltage Forced Landi

    Any swelling and dump the battery.
    Although it is difficult to ascertain exactly was is going on inside the battery, swelling indicates the battery is either overheating or one cell or more is failing internally.

    As you are finding, it simply not worth flying with a dodgy battery.

    Just a shame they cost so much!
     
  13. Fyod

    Joined:
    May 21, 2014
    Messages:
    684
    Likes Received:
    61
    Location:
    Central EU
    Re: DJI Double Defects Crash! TBE & Low Voltage Forced Landi

    My 2 cents:
    I have a Zenmuse and it takes 10s to start up and level out. If I saw awkward behavior like this, I would not fly, because basically the gimbal is showing some kind of stability problem, which could be the result of a faulty gimbal or bad info from the IMU. If it had been doing this for a longer time, I would certainly look into why and find a fix.
    I could be wrong, but the extreme angles remind me of when I had prop wash problems. High (enough) wind could be a culprit or just made it much worse than on a calm day.

    Sorry for your loss.
     
  14. ianwood

    ianwood Taco Wrangler
    Staff Member

    Joined:
    Jan 7, 2014
    Messages:
    4,929
    Likes Received:
    1,800
    Location:
    Lost Angeles
    Re: DJI Double Defects Crash! TBE & Low Voltage Forced Landi

    Lot's of misinformation here. The battery has nothing to do with this crash. Zero. The voltage never dropped below 11.5 which means there was ample power being supplied to the craft at all times.

    This is an IMU initialization failure as result of the operator holding it in their hands while it started up. The problem is further exacerbated by windy, turbulent conditions.

    And fyod is correct to assess that there was an indication of a problem in the way the gimbal did not orient itself properly before the flight even started.
     
  15. stmwise

    Joined:
    Jul 23, 2014
    Messages:
    12
    Likes Received:
    0
    Re: DJI Double Defects Crash! TBE & Low Voltage Forced Landi

    Hello Ian, if the battery has nothing to do with this crash, would you please explain why was my phantom not responding to my up throttle stick? The only times when i experienced the Phantom not responding to my stick commands were the times when it goes into auto land due to battery issues.

    According to you, this crash was due to IMU initialization failure. Could this IMU initialization failure have stopped it from going higher than 25/26 meters as shown in the video? If thats the case, then why did it respond to my full throttle up to 25 meters and not higher?

    This crash certainly left me with more questions than answers. Can anyone provide a clue?
     
  16. ianwood

    ianwood Taco Wrangler
    Staff Member

    Joined:
    Jan 7, 2014
    Messages:
    4,929
    Likes Received:
    1,800
    Location:
    Lost Angeles
    Re: DJI Double Defects Crash! TBE & Low Voltage Forced Landi

    Autoland is at 10.65V. You did not hit 10.65V. Not even close. You did not have an autoland event.

    I cannot explain why it wouldn't go higher. But I do know from the video, it could barely maintain steady flight. There were violent changes in attitude during which it would be expected to lose altitude. This could have cancelled out the command to go higher.
     
  17. Buk

    Buk

    Joined:
    Feb 28, 2013
    Messages:
    654
    Likes Received:
    29
    Location:
    Iowa, United States
    Re: DJI Double Defects Crash! TBE & Low Voltage Forced Landi

    Software limit?
     
  18. Mike

    Joined:
    Jul 6, 2014
    Messages:
    163
    Likes Received:
    2
    Re: DJI Double Defects Crash! TBE & Low Voltage Forced Landi

    I don't know what caused the crash, but just because the osd showed the voltage at 11.5 didn't mean it wasn't battery related. I have four P2 batteries and experienced 1 with a defect. I had 6 flights on it when it started acting up. It would force an auto-land even though the osd showed 12.4v and over 95%. It was only this one particular battery that would cause the problem. It would fly for less than a minute then the red, fast flashes would start and it would start to land. I could still control lateral movement but couldn't gain any more altitude. I had that defective battery replaced. I'm not suggesting that the OP experienced the same problem, but I must respectfully disagree that the battery couldn't be a contributing factor.
     
  19. noiseboy72

    Joined:
    Jul 13, 2014
    Messages:
    575
    Likes Received:
    17
    Location:
    Lincs, UK
    Re: DJI Double Defects Crash! TBE & Low Voltage Forced Landi

    The battery issue comes down to current delivery and this is limited by the battery source impedance. The motors are not receiving the current level they require, so the quad does not gain height. This is compounded by the incorrectly initialised IIMU, which has bad data AND insufficient current delivery available to allow it to fly.

    The OP says the battery shows signs of swelling. This alone is enough to suggest that not flying with that battery again would be a very good idea.
     
  20. stmwise

    Joined:
    Jul 23, 2014
    Messages:
    12
    Likes Received:
    0
    Re: DJI Double Defects Crash! TBE & Low Voltage Forced Landi

    I will post more videos in the coming days with iOSD flight data to show that IMU initialization failure is NOT the cause of this crash as suggested by Ian. Hope this helps with members of the Phantom 2 owners community. Everyone can learn from my loss and hopefully this will not happen to anyone anymore.

    SWOLLEN Battery? Early Auto land? Please discard your Phantom 2 intelligent battery!