Welcome to PhantomPilots.com

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

Flyaway today

Discussion in 'Other DJI Multi-Rotors' started by syotr, Oct 16, 2014.

  1. syotr

    Joined:
    Jul 30, 2013
    Messages:
    898
    Likes Received:
    136
    Had a flyaway with my quad today. This is a scratch built model with NAZA Lite, GPS and DJI E300 propulsion kit. I have had dozens of successful flights with this quad and hundreds with my Phantom 1 before it.
    I took it to a location where I have flown the Phantom a few times before. Went through my normal startup checks. Quickly got 6+ satellites and set home position. Manually reset home position to be sure. Lifted off and right away it seemed to be drifting and not holding position.
    I brought it down and did a compass calibration. Launched again and again it started drifting. Tried to land but it took off like a rocket about 50 feet away and dove at an angle into the ground. Fortunately it did not hit anyone or damage any property. I had to climb under a high chainlink fence to retrieve it.
    Damage to the quad was minimal. It broke 3 of the 4 legs which were 6mm cf tubes and snapped the mount for the GPS tower. Brought it home and cleaned up the damage. Hooked it up to the NAZA assistant and checked everything. All looked normal. Recalibrated IMU and transmitter anyway. Now it seems to fly fine but I am hesitant to trust it.
     
  2. SanCap

    Joined:
    Sep 18, 2013
    Messages:
    715
    Likes Received:
    61
    Location:
    South West Florida
    Is this your Dex quad? I had similar things happen with my dex quad, I attributed it to heat build up or pressure build up inside the waterproof housing. I know you have some sort of pressure relief and heat sinks. Mine almost hit one of my friends on its uncommanded dive towards the ground. I never trusted it after that and disassembled it.

    My post in no way blames this towards the Dex quad but rather to the way I built the quad.
     
  3. syotr

    Joined:
    Jul 30, 2013
    Messages:
    898
    Likes Received:
    136
    Yes, it was my Dex quad. Don't think it was a heat or pressure issue. This happened within seconds of launch so no time for heat or pressure to build. The only thing that was out of the ordinary was some dark clouds to the South which may have affected GPS but the sky overhead was clear and I got GPS lock very quickly. I did not have time to switch to ATTI mode which might have saved it.
    This is the classic NAZA flyaway where the bird drifts a bit and then takes off like a rocket in some random direction, to disappear forever or crash to the ground. I have had two similar incidents with my Phantom.
    I am thinking of switching to a Vector FC. I have one that I ordered when they were first announced. While I was waiting for it to be delivered, I got impatient and bought the NAZA Lite just to get the bird in the air. Once I got it tuned, it has been flying great until yesterday. I suppose it may have been some sort of interference but I no longer trust the NAZA.
     
  4. SanCap

    Joined:
    Sep 18, 2013
    Messages:
    715
    Likes Received:
    61
    Location:
    South West Florida
    Did you buy the carbon fiber or fiber glass Dex shell?
     
  5. syotr

    Joined:
    Jul 30, 2013
    Messages:
    898
    Likes Received:
    136
    Fiberglass. I discovered something this evening. My s2 (left hand) switch has failed. Not sure if that had anything to do with the crash but the switch is not making contact in the lowest position (homelock.) I had noticed that lately I was having difficulty resetting home position by switching between CL and HL positions.
     
  6. thumpinhard

    Joined:
    Aug 12, 2014
    Messages:
    96
    Likes Received:
    0
    sorry to hear about the fly away..sounds exactly like mine..at least yours didn't hit water or people etc.
    i have noticed in my area there is a big difference when overcast regarding g.p.s. lock.
    i also notice my area that the morning around 8-9 am or so there is a radically reduced signal.
    i know some people insulate the shells with foil lined with paper etc. and it seems to make a pretty big difference.
    i am planning on doing that when i get my triumph and will be using a gps app to check signals as well.
    i know dji seems to think the fly aways now are pilot error..and i am guessing they are mainly correct,
    however i have also heard that the g.p.s. signal reduces by quite a bit just by turning on the camera..but i am not sure if that is one persons experience or if it is common knowledge now.
    in my case i am planning on flying atto mode mainly i think and to just be prepared to put it in home lock or turn off the transmitter if i need to maybe.
    i did some atti flying yesterday and found it to be no problem strong wind might be a problem but in light wind it was easy.
    hopefully someone will verify or correct me on this.
     
  7. syotr

    Joined:
    Jul 30, 2013
    Messages:
    898
    Likes Received:
    136
    I have practiced quite a bit flying in ATTI mode and it is nice for shooting smooth video. It takes a lot of practice to fly around trees and other obstacles as you have to anticipate the drift and correct for it.
    I think there are a lot of flyaways that are not pilot error. they may be caused by outside interference, a stuck sensor or some bug in the program. When the quad takes off full speed in some random direction and does not respond to any command, this is not pilot error. Turning off the transmitter or switching to homelock will not help. If the problem is some confusion of GPS or compass, switching to ATTI mode may recover it.
     
  8. thumpinhard

    Joined:
    Aug 12, 2014
    Messages:
    96
    Likes Received:
    0
    i am sure you are right. my guess as well as that type of fly away id due to g.p.s. signal loss and home lock or turning off the transmitter would only help it it was maybe some bug maybe.
    i m sure my fly aways have been due to gps signal loss..even though i had good lock to begin with.
    the solar flares are a factor as well from what i hear.i guess manual or atto mode would be the emergency back up plan. i could not do manual well enough..but atti would be smart.
     
  9. Browneee

    Joined:
    Jun 20, 2014
    Messages:
    54
    Likes Received:
    1
    Just reading your post... Sorry to hear about flyaway... Just giving some thought to the whole GPS errors... Now I'm an airline pilot and we use some pretty fancy GPS avionics for navigation ( don't claim to know the ins and out of how its work exactly, been flying for 20 years ). However when we shoot approaches we used to have to check for what was called "RAIM" with the old GPS systems. New ones do it for us now :).... And if it's doesn't that what first officers are for :). Any ways seriously I'm wondering if the little GPS's we all have in our copters suffer a similar thing ?.

    Now I'm going to cut and paste something which if anyone can be bothered finding out, it may solve/aid in the mystery...

    "RAIM detects faults with redundant GPS pseudorange measurements. That is, when more satellites are available than needed to produce a position fix, the extra pseudoranges should all be consistent with the computed position. A pseudorange that differs significantly from the expected value (i.e., an outlier) may indicate a fault of the associated satellite or another signal integrity problem (e.g., ionospheric dispersion). Traditional RAIM uses fault detection (FD) only, however newer GPS receivers incorporate fault detection and exclusion (FDE) which enables them to continue to operate in the presence of a GPS failure.

    The test statistic used is a function of the pseudorange measurement residual (the difference between the expected measurement and the observed measurement) and the amount of redundancy. The test statistic is compared with a threshold value, which is determined based on the requirement probability of false alarm (Pfa).

    RAIM
    Receiver autonomous integrity monitoring (RAIM) provides integrity monitoring of GPS for aviation applications. In order for a GPS receiver to perform RAIM or fault detection (FD) function, a minimum of five visible satellites with satisfactory geometry must be visible to it. RAIM has various kind of implementations; one of them performs consistency checks between all position solutions obtained with various subsets of the visible satellites. The receiver provides an alert to the pilot if the consistency checks fail.

    RAIM availability is an important issue when using such kind of algorithm in safety-critical applications (as the aeronautical ones); in fact, because of geometry and satellite service maintenance, RAIM is not always available at all, meaning that the receiver's antenna could have sometimes fewer than five satellites in view.

    Availability is also a performance indicator of the RAIM algorithm. Availability is a function of the geometry of the constellation which is in view and of other environmental conditions. If availability is seen in this way it is clear that it is not an on–off feature meaning that the algorithm could be available but not with the required performance of detecting a failure when it happens. So availability is a performance factor of the algorithm and characterizes each one of the different kinds of RAIM algorithms and methodologies.

    Fault detection and exclusion
    An enhanced version of RAIM employed in some receivers is known as fault detection and exclusion (FDE). It uses a minimum of six satellites to not only detect a possible faulty satellite, but to exclude it from the navigation solution so the navigation function can continue without interruption. The goal of fault detection is to detect the presence of a positioning failure. Upon detection, proper fault exclusion determines and excludes the source of the failure (without necessarily identifying the individual source causing the problem), thereby allowing GNSS navigation to continue without interruption. The availability of RAIM and FDE will be slightly lower for mid-latitude operations and slightly higher for equatorial and high-latitude regions due to the nature of the orbits. The use of satellites from multiple GNSS constellations or the use of SBAS satellites as additional ranging sources can improve the availability of RAIM and FDE.

    RAIM prediction
    GNSS differs from traditional navigation systems because the satellites and areas of degraded coverage are in constant motion. Therefore, if a satellite fails or is taken out of service for maintenance, it is not immediately clear which areas of the airspace will be affected, if any. The location and duration of these outages can be predicted with the aid of computer analysis and reported to pilots during the pre-flight planning process. This prediction process is, however, not fully representative of all RAIM implementations in the different models of receivers. Prediction tools are usually conservative and thus predict lower availability than that actually encountered in flight to provide protection for the lowest end receiver models.

    Because RAIM operates autonomously, that is without the assistance of external signals, it requires redundant pseudorange measurements. To obtain a 3D position solution, at least four measurements are required. To detect a fault, at least 5 measurements are required, and to isolate and exclude a fault, at least six measurements are required, however often more measurements are needed depending on the satellite geometry. Typically there are seven to 12 satellites in view.

    The test statistic used is a function of the pseudorange measurement residual (the difference between the expected measurement and the observed measurement) and the amount of redundancy. The test statistic is compared with a threshold value, which is determined based on the requirements for the probability of false alarm (Pfa) and the expected measurement noise. In aviation systems, the Pfa is fixed at 1/15000.

    The horizontal integrity limit (HIL) or horizontal protection limit (HPL) is a figure which represents the radius of a circle which is centered on the GPS position solution and is guaranteed to contain the true position of the receiver to within the specifications of the RAIM scheme (i.e. which meets the Pfa and Pmd). The HPL is calculated as a function of the RAIM threshold and the satellite geometry at the time of the measurements. The HPL is compared with the horizontal alarm limit (HAL) to determine if RAIM is available."
     
  10. Gizmo3000

    Joined:
    Jan 25, 2013
    Messages:
    1,859
    Likes Received:
    2
    Location:
    Sherman Oaks, CA
    Sounds like a classic true flyaway.. same thing happened to me
    just be glad you were able to retrieve yours .. I'm still wondering just where my F450 flew off too (in the middle of Hollywood), someone surely must have found it by now.
    I've started flying with a GPS locator (tagg) ,.makes me feel much more secure flying.
     
  11. arizeda

    Joined:
    Sep 2, 2014
    Messages:
    175
    Likes Received:
    23
    Which tag are you using?
     
  12. sergekouper

    Joined:
    Apr 20, 2014
    Messages:
    752
    Likes Received:
    128
    Location:
    London UK
    Very useful information to take on board, Thanks! A shame we can't know if our GPS is FD or FDE... I was personally thinking to a fratricide frequency in use at the time it took off... I use a spectrum Analyser to sweep the area before taking off, just in case...
     
  13. syotr

    Joined:
    Jul 30, 2013
    Messages:
    898
    Likes Received:
    136
    Many of these locators require phone service to work. Lots of the places I fly have no cell service at all. I am lucky to get even 5 GPS satellites. I often have to launch with 5 satellites and get more once I am in the air.