T16000M Hall Sensor video

[AngleOff66]

Just stumbled across this by accident.

 

https://www.youtube.com/watch?v=jBo_dAvgZ1k

 

I wonder if something like wire glue is good enough?

 

He mentions 5mm x 1mm, though someone else mentioned N52 8mm x 1mm in comments. Another commenter mentions using 15 magnets.

 

Maybe this will help someone?

[Sokol1]

In the video the hot glue is just for keep the Hall sensor in the place and for this is OK.

The Hall sensor is soldered in the wires, replacing the fault twist rudder potentiometer.

 

The Neodymium magnet need have a size that fit in the place, so both 5x1 or 8x1 mm is OK.

Can use include a ring magnet, better if diametrically magnetized.

 

No reason for use 15 magnets for one sensor, that guy post is not clear about what he did or are talking.

 

For Hall sensor use Honeyweel SS95A1 or SS49E, or Allegro A1324LUAT.

 

The biggest difficulty of this mod is adjust (place) the Hall sensor to be in the center of magnet field, what is done bending their legs, but the adjust is in fractions of milimeters... So, patience Padawan.  

 

 

Edited November 5, 2021 by Sokol1

[Cynic_Al]

I agree this mod can and should be done using only two magnets mounted on opposite sides of the sensor, with like magnetic poles facing each other.  In this configuration the distance between the magnets sets the correct centre point and rotation angle for calibration.  If you find that twisting the grip left and right of centre results in movement in one direction only, flip-over one of the magnets.  If rudder direction turns-out to be reversed relative to the way it was prior to the mod, flip both magnets.

 

When performing calibration, don't use the test screen shown in the video. Instead display the calibration screen for the Z-rotation axis, so that the percentage displacement values can be seen.

 

Having said all that, this joystick (at least the earlier model with the green base) has a nasty built-in deadzone around the centre position of the twist axis, which this mod will not remove, nor will any software.  This deadzone accounts for about 11% of the total twist movement, which I consider to render it useless for precision aiming.  I cannot say whether the later models retain this deadzone.