Thanks JJB. I assue the HP was set based on the 11 sats available at motor start - but I have no clue how many are used.
I suspected the low light may have hampered the OA in this case. The only thing I could link in the guide was “The aircraft cannot avoid obstacles during Failsafe RTH if the Forward and Backward Vision Systems are unavailable.” but I now also read “The Vision Systems cannot work properly over surfaces that do not have clear pattern variations.- Flying over surfaces with repeating identical patterns or textures” and the wall it hit is covered with dark grey slates.
I would expect the RTH sequence to abort if a vision sensor is inactivated, and trigger an exception call - at least when it is within the 5-20m radius (maintain current altitude mode). Or simply reject an RTH call when within a certain radius if this can introduce risk or damage.
Hi, 11 sats looks good, but it depends where the 11 are in the sky to get a good HP.
Signal reflections of the surrounding may disturb a good position as well.
True, DJI should improve on many things regarding to inform users. But the manual is clear about low vision etc.
Rejecting a RTH ; well mayby it helps if the user get a choice on the screen, continue RTH yes/no.
But having said that, there are so many exceptions on the ‘normal’…DJI won`t do that.
So imo ‘pilots’ should use always common sense and knowlegde and don`t fly in situations where exceptions are to be expected.
cheers
JJB
Apoligies for the delay, was an intense work week. Wanted to say thanks to you chaps for the diagnostics, sound advice and great responses. I will definitely be more circumspect wrt gps accuracy and vision systems.
The MA2 is an amazing drone (not difficult to impress as this is my first decent uas), and the firmware releases are exciting.
i guess we will see more predictability as the frameworks for levels of autonomy for UAS mature (like the SAE levels for automotive) to be less deterministic and more adaptive, as avoiding injury is paramount. But for the time being the ‘grain of salt’ principle will prevail 
All the best.
Sounds like a pretty good crash. I have no solution for the error that caused the crash but I had a head-on due to processor delay. Several crashes usually damage the gimbal and chip props. I’m going to install prop guards which should also provide some degree of gimbal protection as well as possibly protecting my fingers.