Non-stable quiescent output voltage hall sensor

[Farad22]

Hi all. At the moment I'm experimenting with a ratiometric hall sensor (TL173C) and a permanent magnet. The problem is my quiescent output voltage doesn't return to the same voltage (0.5 x Vsupply) after removing the magnet. Especially with a change in magnet orientation (N ---S) the quiescent voltage can differ around 20 - 30 mV after removing the magnet. The difference in quiescent voltage after applying the same orientation a few times is smaller (5-10 mV). What could be the issue with this? Thanks in advance.

The sensor is hanging in free air and sensitivity is 1.5 mV/Gauss.

 

[jiripolivka]

How your mount the Hall sensor? Any ferromagnetic body close to it gets magnetized by the permanent magnet and affects the Hall sensor.
Check your screws, nuts, also filings on the magnet.
How stable is magnet's mechanical position when close to the sensor?

 

[Farad22]

The sensor is hanging on its electrical wires and hanging in free air with 30 cm clearance in x,y,z directions. I soldered the wires directly to the hall sensor. The magnet itself I bring by hand to the hall sensor. It's a simple bar magnet (no screws, nuts). One interesting thing I observed is that the hall sensor itself is attracted to the magnet when it is really close. So maybe something get magnetized in the sensor itself and therefore gives non-stable quiescent output voltage?

 

[Farad22]

Anyone know why this ratiometric hall sensor (TL173C) would not return to its quiescent output voltage after changing magnet orientation and moving the magnet away? It's hanging in free air, so no metal surroundings.

I tested with another sensor SS496A1, this one showed a similar behavior although a smaller difference in its output voltage (+/- 5mV).