What diameter magnet works best with AS5xxx sensors?

hardware support
1

My guess is that the 6 mm diameter magnet might be better than the 4mm. I’m guessing I have more room for being off centered with a larger magnet.

But maybe it does not matter.

2

I think the actual part of the die which is doing the sensing is quite small, so I’m not sure how much it matters once you exceed that. The field strength can be independent from the size and does matter substantially, the ones I get from aliexpress seem to vary a lot, some work through PCB well (how I usually design my boards) and others not so much.

3

If anyone has some good links to diametric magnets that work well, perhaps they can share. I’ve got some small ones (6mm x 1mm) - they are ok. I wouldn’t go smaller than 6mm and would prefer ones thicker.
Some of the more expensive magnetic sensors come with diametric magnets - from memory they are bigger again (something like 8mm x 3mm).

Stating the obvious, probably, but make sure they are diametric. Although some people have had success puting normal magnets sideways :slight_smile:

4

A square magnet also works well.

encoder

1 Like
5

whoa, I’m not so sure that magnet will work well. It might work if you use the sensor calibration feature, however I think if you don’t you would get a non uniform relationship between reported angle and actual angle. These devices aren’t like a compass, they have sensors arranged in a square. I tried a cylindrical axially magnetized (instead of diametrically magnetize) magnet once and that’s what happened. I think your square one is going to have the same problem, but as I said soluble with the “calibration”

6

I’m not sure what you mean by “calibration”, but I haven’t noticed any difference what magnet is used. It is important that the magnet poles are in the correct direction.

1 Like
7

I ordered this version, it works well, a little expensive …
https://www.amazon.fr/gp/product/B008U9T3L0

8

OK, so it seems that basically “all magnets work”, even square magnets as long as the poles are diametrically oposed and not axial to the sensor.

I’m using the cheap AS5600 for robot joint positioning ad wanted to optimized it’s performance as much as the AS5600 can be optimized. Next I’ll ask about a motor comutation sensor in a new topic.

Thanks, all of you.