endotherm
First Officer
Flight distance : 503241 ft
Australia
Offline
|
Genghis9 Posted at 2018-3-14 20:16
Aardvark and I have had this discussion before in another thread and it has come down to poorly written manuals that are either in error, not updated, or incomplete OR all the above and what the drone actually does or does not do.
Unfortunately, the only way to truly test this stuff is when the drone thinks it is actually airborne. In my early tests I did a ground test using left stick down, CSC, and the emergency shutdown stick RTH button combo methods. In each case shutdowns were near instantaneous and I agree left stick down even shut down a bit quicker than the full 3 seconds but it took the longest before shutdown. As I noted the manual only states that the left stick down will result in a delay and does not state that the CSC or combo methods will have a delay. Yet, as Aardvark's video shows it appears shutdown using the combo method does not occur for about 3 seconds. The problem with repeating that test is basically risking your bird in a fall that could end far more badly than the one in that vid, I don't have that luxury.
After my ground test were called out for the fact that the drone knew it was landed/on the ground and engines in idle, I then tried fooling it in to thinking it was airborne. The results were the same as on the ground, however, the problem was the engines were in idle, so the assertion was the bird was again assuming it was landing just at an altitude higher than where it took off. I next attempted the same thing two stories up with the same shut down timing but maybe I did not factor in something else. The problem here is the bird will spiral down at least two stories during that 3 seconds before shutdown and obviously I cannot afford that either or the end result would end up being rather expensive on my part. Short of another "safe" testing method or a definitive update to the manual that clearly states what the true response times are using which method under what circumstances and conditions, then you are going to need a couple of hundred feet of altitude and the use of drone that you can afford to lose.
I don't think you need to perform any sort of hazardous test to prove or disprove the theories.
With a left-stick-down stop we are instructed to hold it for three seconds, but it doesn't necessarily take that long to determine the aircraft is on the ground, and therefore proceed to disarm the motors. It could be in flight or idling on the ground, so the motor speed is irrelevant. As we are holding it down, the altimeter reading is examined and compared to previous readings milliseconds apart. If it is the same value (or really close) for a set time, it decides it is on solid ground and has "landed". This might only take a second to get sufficient identical altitude values, or it might take longer (up to 3 seconds) for the readings to stabilize so it can make a reliable conclusion. This would explain faster-than-3sec shutdown, or random timings between attempts.
A CSC shutdown is not conditional on the altimeter reading as the left stick method is. You can initiate a CSC in the air or on the ground and it will react the same way. I've tested a P3 on the ground with the delayed firmware, and it responds with a rearward move before flipping over, followed by turning the motors off after 3 seconds. This is why we don't CSC on the ground -- it is still responding to flight control for those 3 seconds before it decides to respond to the emergency command! This even applies to early firmware where it is still responding to flight commands for a much briefer 0.3 seconds. This still causes instability, just not as severe as the 3 second version. Of course this change was introduced to placate the doom-and-gloom crowd who feared an accidental shutdown in normal flight due to stick positions that you wouldn't intentionally perform unless you were drunk.
The 3 second delay was introduced in P3 firmware 1.10 and continued in 1.11. I have no idea which P4 firmware it applies to as I don't have one. I also can't comment on the time it takes for the P4 stick+button shutdown procedure. I'd assume DJI made the same changes for the other models at around the same time they introduced it for the P3. |
|