Saturday, December 30, 2017

Reverse Engineering a DJI Phantom 3 Standard Gimbal

My dad owns a DJI Phantom 3 Standard. For some reason (probably the multiple crashes into trees), the fragile gimbal got the worst of it. Something in it is broken. When you turn the drone on, the gimbal stabilizes, goes weird, and then goes limp. So something isn't right. But I'm not going to try and fix this. I'm going to try and reverse engineer it. I'm fairly certain that this is a single board computer. So let's take it apart, power it up, and see if we can communicate with it.



The first step, as mentioned above, is to take this apart. I wanted to get it back together at some point too, so I was careful about taking it apart. Here is what the main circuit board looked like on its top and bottom, respectively.


All of the interesting stuff is under the shield/antenna, but I don't want to remove it just yet. At this point, I had to find where the VCC and GND pins were on the connector in the bottom right of the picture of the top of the board.


With a continuity meter, I was able to determine that the first two pins were ground (it was continuous with the big shield. The next two pins were also continuous, so I guessed this was VCC. I was right. So I soldered and superglued the two power lines to the board.


I read somewhere on the internet that it took 12V power, so I wired it to a PC power supply and turned it on. The gimbal turned on perfectly. Except for the broken behavior. But other than that...

I noticed that on several parts of the board, there were TX and RXs to different chips (not well labeled), so I probed some of them until I found one that produced ASCII text.

None of them did. One of them seemed to put out a debug packet whose header matched what I found on the internet to be standard DJI headers, but nothing of any consequence came from chasing down that rabbit hole. It seemed that despite the markings on the board of TX and RX, nothing was true serial as I had hoped. This is all the time I can spend on this now, but I will post again if I'm able to make any further progress on this. My hope is to get an image from the camera or at least to force the camera to record something again - gimbal or no gimbal.

No comments:

Post a Comment