srandall25 said:
I've exhibited this wobble of death a few times since I've been flying the Vision+ when it first came out in April. I've studied many similar reports from others on the same subject matter. Many people jump to the VRS conclusion on this.. which it may very well be... but after my last wobble of death occurred, I noticed a compass error on my phantom after immediately connecting it to the assistance software. I also remembered that I was flying very close to a huge rock side of a mountain... a rock that is very rich in iron... then i remembered another time i had set the phantom on the same type rock as a launching point, and the DJI app telling me of a compass calibration error.. even after doing the compass dance many times... the rock was apparently throwing the compass off. Iron is known to do this. At this point, based on all my experience, I personally believe that a failed compass during flight can indeed cause this wobble of death... and it is also my belief, that a lot of other presumably VRS reports, are actually due to a failed compass or compass interference... I believe what is actually happening is that when the compass is not reporting the correct orientation, the GPS algorithm that keeps the phantom stable at a fixed point, is no longer properly executing the proper flight corrections (with respect to orientation) that is needed to keep the phantom stable at that fixed point. The lack of a proper fixed orientation or compass setting results in these 'jerky' sporadic corrections... causing this rapid oscillation or 'wobble of death' as I call it... I could be dead wrong on this.. but this is my belief based on everything I've personally experienced and read here in the forums.. and trust me.. I've read my share of threads in this forum regarding this subject... Reading your report convinces me even more that compass interference is more than likely the culprit...
That gets interresting. Let me describe what happened in the second, nearly crash. I was descending sloowly from a height of about 5m. No wind whatsoever, I watched the video again and again. Then there was a rhytmic movement, left right,left right with a frequency of about one second. While this happens phantom loses height about 1m per second. In this second case it re stabilised for whatever reason. I think I tried to accellerate upwards.
I know that my ground was a ceiling from an underground parking, with a lot of steel inside for sure. I set up some hundred sat dishes in my life and know how unreliable compasses can be soon as they get close to metal.
BUT: I am quite sure other drones also use a compass, so: Is this a general rule we are not knowing by hard or is it problematic firmware code that makes compass disturbances so dangerous? Are we not more likely talking of gyro problems? Although gps position jumps could be a scenario as well.
I think the programmers from dji should read this because of that frequency of about one second. This shows that we are not talking of a value that changes because of interference and stays for a while until it gets back to correct value. What we see are several unsuccessful attempts of the drone to level itself out. Every attempt leads to the next bad angle, not to a leveled state. Also usually if e.g. entire left side of drone is too low firmware will surely try to raise left side, not to lower right side or am I mistaken? So why is it loosing height when trying tolevel the drone out?
Anyway, I think we are facing a "regelkatastrophe" here. (Hard to translate) This is if you try to level something out and everytime you put far too much or by far not enough.
I will go back to site with my analog and my digitalcompass and check for anomalies.
My client wants me to film a lot over water, so I have to be sure not to make mistakes then. I will stay tuned here about this issue.
I will try to point someone from dji to this thread... lets see if this is possible.