I have a DJI Ronin gimbal that I want to modify in order to be able to tilt the frame left or right more than 45 degrees and the camera to keep horizon.
I opened the gimbal motors. The driver boards are installed on the back of the motors. The board contains the magnetic encoder centered on the board and there is a diametral magnet in the motor shaft. Drivers are connected via CAN interface to the controller. So far so good.

My problem is that from that driver board there are another 11 wires connected to the motor body, neatly, via a connector. I managed to figure out that the phase wires are actually doubled to allow high current to pass to the motor (the gimbal is capable of 7kg load) for a total of six wires. That leaves another 5 wires that I do not know what they are for.

So, to experiment, I removed those 5 wires only from the tilt (not roll) motor connector and I powered up the gimbal. To my surprise it is working fine until you tilt the frame forward or backwards more than 15 degrees or so. At which point the tilt axis suddenly moves about 15 degrees. For instance, if I tilt the frame forward, after passing that 15 or so degrees, the tilt axis suddenly tilts up about 15 degrees.

This happens regardless of the follow settings or smooth settings as DJI calls it.

My questions to you:
1. Is it possible to have a brushless motor that have BOTH rotary encoder AND hall sensors mounted in the motor stator?
2. I am just assuming those wires going inside the motor are for hall sensors but they may well be temperature sensors or other type of sensors but in this case why would the stabilization be influenced by me connecting those wires or not since I already have a rotary encoder?
I was under the impression that absolute rotary magnetic encoder and hall sensors installed on the stator actually do the same thing.
3. Will it be possible to control such motors and drivers using a Basecam CAN_MCU for instance?

Please shed some light or direct me to a good resource to learn this.
Thanks,
Adrian