Hey There Studi Posted at 2016-11-7 15:02
Thanks Tahoe Ed! I filled out the DJI Analysis Form and submitted it. They said they will investigate if it is Pilot error or product error. Even better, they will try to find it for me. Fingers crossed.

Sorry for your loss, lived in SH for over a decade, from what I can see, if you are upwards to 400 ft, you should be ok.  Like alexcabrils says, unlike Pudong,  that area should be clear of sky scrappers.  You might have encountered interference from radar monitoring traffic on the river.  I once flew at a height around 50 meters may be 3 miles further south west from where you were and was interfered by a marine radar hidden in a building.

Hey There Studi Posted at 2016-11-5 04:53
Hey Zatx,

Thats what exactly prompted me to write in the forum; seeing if I missed something. I'm usually a very conservative pilot and don't take unnecessary risk (for the sake of people around, not necessarily my drone).

"Studi": Thanks for sharing all your detail.  I'd like to share with you my theory and understanding of your issue.  I am just only beginning to experiment with my Mavic and have been scratching my head with some things.  I've been in the RC hobby for about 20 years; helicopters before drones. Please bear with me - I'm going to (try) and be brief; I love these types of issues so I love to be analytical which often turns into being long-winded.  I'll start off by stating that I am an engineer with a good deal of understanding with signal technology; having said that I am at risk for over-simplifying other factors that contribute to this issue as the factors in play here with the Mavic, it's RF transmission characteristics and in general, can be a bit of an 'art' rather than science.  I say 'art' because there is only so much that can be understood from the pure telemetry data itself.  It was your mapping data which provided me my conclusion.   I was curious about your flight map data because of the environment: specifically the river opening in the middle of the urban surrounding - this created a natural (or unnatural!) boundary around your Mavic.  From your Home position you were on the building side, presumably facing waypoint A which was located across the river, then you flew more or less 90 degrees along the buildings across from you.  Looking at your telemetry, and the Mavics positioning; the majority of your flight maintained a direct line-of-sight with the Mavic's antenna.  The buildings in the direct background provide for strong positive signal reflection.  Because you maintained both line-of-sight and a reasonable angle to the Mavic's antenna (think of your RC as two goal posts facing the Mavic's antenna as opposing soccer team's goal posts are parallel to each other) there is very little chance off any signal interference or signal filtering causing errors in send/receiving. However....the discussion I read wrt whether your Mavic took a right or left turn clued me in.  At point H, where you lost signal, the Mavic's antenna abruptly turned 90 deg to you and ended up perpendicular to your home position/RC (your maintained contact for a short while until the angle between the RC and the Mavic's antenna closed).  The building directly behind your Mavic likely contributed by providing a strong signal reflection that confused the Mavic.  The Mavic works best by receiving a DIRECT
signal versus a secondary/reflection.  This technically would not be 'interference' but rather a phenomenon known as comb filtering.  The signal strength from this reflection combined with the attenuation (weaker) of the primary signal (which is primary is debateable as it is 2 way) is what caused the Mavic to not have an established connection.  I would have hoped that signals reflecting off the buildings on the far side of the channel would have been sufficient (appears the comb filtering was 'louder' than these signals).  FYI your Mavic did NOT hit the building as a root cause; although it is probable that it did shortly after losing communication.  Although not much can be done about the positioning of the RC wrt the Mavic's antenna (your flew it quite naturally/as expected) I personally think DJI's engineers need to re-think the logic of how the Mavic behaves once / if GPS signal is lost (or rather the inability of the Mavic to transmit it's GPS position to the RC; as I highly doubt the Mavic is actually losing satellites; I could be wrong in this regard as I am still myself trying to understand the quality of its GPS sensor and logic).  The firmware and software could be tuned to consider the Mavic's signal processing/timing wrt the relative angle of the RC vs Mavic's antenna.  A bit of R&D to do.  I believe they don't have this issue in the Inspire design and also I believe the Inspire is aided via backup GPS data via optional GSM network.  Some 3rd Party could vastly improve the Mavic's abilities in this regard by pairing an external GPS afixed to the Mavic (this would also double as a cheap locator).  I've an idea to do this as a locator by fixing a $50 smartwatch with a sim card as insurance against fly-aways   A military-grade drone would of course have rotating antenna (my wish list for the Mavic 2 ) or omni-directional like an AWAX plane.  IMO I'd prefer a 3-4 km range with omni vs 7km without....

To prevent this all I can think of is to adjust flying to maintain a 'goal post to goal' post orientation of the antenna as best possible (using more roll vs yaw on straight-aways when at far distances).  See I told ya I'd be short....