Magnet placement and hall effect sensors

[Shawn]

I had read a number of tutorials regarding the use of hall effect sensors in flight controls and it was my understanding that the axis of the magnet should be placed on the same axis that the stick rotates on. I struggled for hours trying to get a decent reading on the sensors and all of a sudden the roll axis starting working brilliantly for seemingly no reason. Fast forward another hour or two and I suddenly realize that the magnet on the working sensor had slipped down on the axis and instead or rotating it was following more of an arc, sweeping along the sensors face?? Maybe I'm the only one who didn't understand how this was supposed to work but just in case anyone is having difficulty with their setup I made this short, crappy video to show how the pitch axis on my centre stick is setup.


http://www.youtube.com/watch?v=VH0XI91lA54

This set up is still being worked on as I need to find a better way to mount the sensor but I am very pleased with how the HE sensors are now working, there is less than 6" of total throw at the top of the stick (which is full length) I can make changes to pitch and roll with just slight pressure on the stick.

And a quick shot showing the control stick and FSX running in the back ground.


http://www.youtube.com/watch?v=96q3zomJmdc
Shawn

[Matt Olieman]

Shawn, thanks for the videos and sharing your experience. :)

Matt Olieman

[Shawn]

My pleasure Matt, hopefully it will save someone a few frustrating hours. I have seen a few posts from other people who said they found HE sensors outputs to be jittery and hard to calibrate, that was exactly what I was experiencing until dumb luck found me.

[wledzian]

A linear hall sensor produces a voltage proportional to the field strength normal to the sensing element. If you're rotating your magnet along its N-S axis, the field at the sensor will have no field lines normal to the sensing element, and you will get a 'neutral' signal. However, if the magnet moves off-center, the field will have a normal component that will change as the magnet is rotated around the sensor.

Figure A:
Similar to your original configuration. The field lines do not pass through the sensing element. When the axis is rotated, the sensor measures no change.

Figure B:
Your 'slipped' configuration.
With the magnet no longer centered on the axis, there is a perpendicular component to the field lines at the sensor which changes as the magnet is rotated around the sensor.

Figure C:
Similar to most of the 'tutorial' cases.
The magnet is positioned on the stick's rotational axis so that the field lines are normal to the sensor (parallel to the N-S axis) and change as the axis is rotated.

5773

Hope this helps!

[Shawn]

Thank you Wledzian, that did help explain things and made me take a second look at the photos/tutorials I had been reading. Turns out I was only assuming that both the magnet and the sensor were on the same axis, had I looked closer at the photos, I would have figured it out. In my defence, most of the diagrams that accompanied the photo's were similar to the example C that you gave. That configuration did work but didn't register much change, plus it was extremely sensitive to movement of the sensor?

Cheers
Shawn

[wledzian]

I just took a look at your video; it appears that your sensor is not in line with the stick axis - is this correct?

There are lots of workable configurations of magnet and sensor; the only real important bit is that there is no perpendicular flux component when the stick is neutral, and this component changes in opposite directions as the stick is moved in opposite directions.

As your stick has very little angular movement, I'm not surprised that the figure C configuration didn't provide much signal change. Here's another configuration you may want to try: (in all the configurations, assume the sensing direction is out of the page)
5778

In this configuration, the field is neutral when the stick is centered, and passes through the sensor in a direction based on the stick movement direction. The sensitivity of this configuration can also be adjusted by keeping the original relationship of the magnet and sensor as shown, but moving the axis towards the sensor.

[Shawn]

The sensor is being held on with some black tape at the moment, so it might not be centred. I had made some aluminium tubes that were supposed to protect the sensor and provide some easy adjustment but took these off while I tried to figure things out. The black tape was handy.

This is how things are set up currently, and while it's working better than I expected. I now understand that there are a few other options, I will certainly test these and see how they work out. The steel bearings make a great sticking point for the magnets and I like that the sensors do not move, as it might save some wire "wear and tear" down the road. Thanks again for the information regarding the setup.

http://i1235.photobucket.com/albums/ff438/shawnLund/Image3.jpg