Erratic braking on P3A

[idahoflyer]

Does anybody know about how much distance the P3A is supposed to cover while auto-braking from full speed? I was flying yesterday on a perfectly calm day (wind @ 0 MPH) well away from any interference with the drone at relatively close range and noticed really erratic braking behavior. A couple of times it seemed like it was coming to a stop much more aggressively than usual, and a couple of times it coasted for an alarmingly long distance before coming to a stop.

On my last flight last night I actually had my first crash when it went coasting in to a tree. It was in P-GPS mode the whole time (confirmed looking at the flight record that it was not in Atti) and if I'm reading correctly off the flight record it actually coasted about 125 feet after I released the forward throttle before it hit the tree. I actually went full reverse for the last 40 feet and it just kept coasting. Luckily the damage seems fairly minimal (thought I haven't taken it back up in the air yet) - the shell popped open a bit on one arm and won't fully snap back closed, one of the sonar sensors has the aluminum ring around the outside bent in a bit (not sure if that hampers functionality), and there's just a bit of scuffing on the gimbal, shell, and battery. All in all it could have been much worse but I've never seen my drone take so long to come to a stop, especially with giving a full reverse throttle for the last 40 feet.

Has anybody else experienced anything like this? I'm wondering if this is a scenario where DJI might foot the tab for repairs since the auto-braking didn't seem to work. I know others have gotten crash repairs under warranty when the flight record showed the equipment didn't act as it was designed to, but not sure how outside of the normal stopping distance this was.

Edited: Note that I've never changed any of the auto braking settings - all of those settings should still be at factory defaults.

 

[msinger]

If you upload your flight log here and post a link back here, we can check it out and try to explain what happened.

 

[idahoflyer]

Here it is: Phantom 3 Log Viewer - PhantomHelp.com

 

[msinger]

Did you really take off at 38% battery or is the first part of your log missing?

 

[idahoflyer]

I started with a full battery but briefly landed and took off again a few times during the flight including at around 38% (maybe it starts a new log at each takeoff rather than each time it turns on?) That last flight was just a short trip to get a couple of stills and then zip back. Got it almost back to me and still had 23% battery when the crash happened. My critically low battery warning is set at the default 10% and I hadn't gotten low enough to trigger auto-RTH.

 

[msinger]

When your Phantom was out at the furthest point, it started to descend straight down -- nearly hitting the water. Did you do that?

 

[idahoflyer]

Yes, the altitude reading is a bit misleading - it's a reservoir with a low water level right now, I probably had another twenty feet or so before it would have hit the water. I was up on a bank that was probably 30 feet or so above water level when I took off.

 

[msinger]

After that point, your flight looks very odd. As your Phantom was coming back, there are some points where it seems like it came to a dead stop, backed up a bit, and then continued full throttle forward again. The time span is very short in between those points, so I cannot imagine it was actually coming to a dead stop. Can you explain what was happening there?

Also, did you change the brake setting in the DJI GO app? If you greatly reduced it, your Phantom would coast quite a bit before coming to a stop.

 

[idahoflyer]

After that point I just zipped back to me to finish my flight. Definitely no stopping or backing up, just full speed towards myself until the events described in my original post. The brake setting is at default, I have never modified it.

 

[msinger]

@BudWalker, have you ever seen anything like this?

 

[BudWalker]

For some reason I didn't get the notification until just now. This is very strange.

This looks like the P3 abruptly speed up for some reason, detected the error, over corrected (to 0 mph), and then resumed normal speed. Multiple times. No pitch input from the RC. Very strange. @Luap needs to look at this. The motor commanded speed might be interesting. If it were a motor speed sense problem, like the P3 thought the back motors were going slow, then you'd expect there to be stability issues.........

 

[Luap]

idahoflyer, if you could post the Dat files of that day to check the motor speed amongst other.

 

[BudWalker]

I think it's significant that the P3 isn't just braking. It is actually speeding up and then braking to recover.

I was supposing that there could be faulty indications coming from the gyro (specifically gyroY). Specifically, gyroY producing temporary erroneous data indicating it was pitching up. The P3 would temporarily increasing the back to front motor speed difference to arrest the pitching up. Then when the gyroY was producing correct data it would recover by doing the opposite action (i.e. braking).

But, I don't think this is very compelling. The speed is always reduced to exactly 0 and the P3 never "backs up".

This plot shows this. Since the P3 is moving in a westerly direction it's longitude is decreasing. When it speeds up there is a dip in the longitude's trend. But that isn't followed by a bump (i.e. momentary increase in the trend) in the longitude. I.e. the P3 doesn't "back up"

@Luap is right. We need the .DAT. To see how to do that go here.

 

[idahoflyer]

What's the best way to get my .DAT file on here? I've got it downloaded off the Phantom but the file is 255 MB.

 

[BudWalker]

What's the best way to get my .DAT file on here? I've got it downloaded off the Phantom but the file is 255 MB.

Can you use Dropbox?

 

[idahoflyer]

How about this (Google Drive): FLY064.DAT

 

[BudWalker]

That's the right .DAT. For anybody using DatCon to look at this the incident happens after the pilot stopped and then started the motors. The default "Time Axis" settings set the upper end at the first motor stop. You'll need to override the upper with something like

 

[Luap]

haven^t had much time to look at it but here is the Dashware representation of the P3 Pitch vs Speed vs RC pitch of the last 2.4 minutes of the flight

 

[msinger]

Luap, do you have any idea what made it come back full throttle like that?

 

[BudWalker]

In my previous posts I was looking at the abrupt stopping and starting. I don't think this actually happened. There must be some error in how the data was extracted from the .txt and/or how the speed was then computed.

Extracting the data from the .DAT we have

It's not the throttle, as referenced above, that controls the speed. It's the elevator. The throttle was reduced but the elevator was kept at the max until 3 seconds before impact at 1167.

At 1147 the right stick (elevator) is pushed forward and the speed starts increasing. I'm not sure what's normal, but the rate of increase does seem sluggish. At 1164 the right stick (elevator) is brought back and the pitch starts increasing at the same time at a rate of 35 degrees/sec. Again, I'm not sure what's normal but 35 degrees/sec seems to be pretty aggressive braking. Impact was at 1167 leaving 3 seconds to slow from 16 M/S to 3 M/S. That's a delta of 30 MPH in 3 seconds. The distance traveled was 30 meters