I use a hall effect sensor instead of a reed sensor to measure my gas ( heating not car ) usage.
It's not a sensor that's connected to an analog input but one with a cutoff potentiometer on that is connected to a digital pin.
looks like this:
The thing gives a constant LOW when the dial is between 1 and 7 and a constant HIGH when on 9. when it's between 7 and 9 it flutters between 1 and 0 . also when the dial is between 9 and 1.
I add up a few hundred reads then process the result. when it's above 95% I put 1 in a secondary counter which gets subtracted when not above 95% . I do the same with a second secondary counter when values of the first are 5% or lower. When that HIGH secondary counter secondary counter gets to 12 there is a boolean that gets set to true. Call it switchingAllowed. When that secondary LOW counter gets to 12 and switchingAllowed is true I set switchingAllowed to false and add 1 to my rotation counter. The 12 are like a moving window.
A full rotation looks something like this..
Pin Reads: (lot more of course)
1111111111111111110111010101101010101010010100000000000000000001010101010101111111111111111111111111111111111111
Let's assume I read 200 states in a row my results after that would be:
200,200,200,200,200,198,187,192,140,123,90,93,82,76,47,23,4,2,6,5,1,2,3,0,0,0,0,0,0,0,0,0,1,0,2,4,17,28,45,78,95,123,146,168,184,189,192,198,200,200,200,200,200,200,200
The secondary counters would look something like this over time after the first turn.
low:
0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,2,1,2,3,4,5,6,7,8,9,10,11,12,12,12,12,12,12,12,11,10,9,8 ....
high:
12,12,12,12,12,12,12,12,12,12,12,12,12,11,10,9,8,7,6,5,4,3,2,1,0,0,0,0,0,0,0,0,0,0,0,0,0,1,2,3,4,5,6,7,8,9 ......
When explained like this I'm sure it can seem a bit bulky or weird but in code I'm sure you can see it's just a few lines.
In short , a turn counter isn't allowed to be added unless I'm sure the dial was on the other side first.
This is running on a PI. If I had an arduino for it at the time I would probably have gone analog for the reads.
Maybe your "jitter" issue can be solved in the same way?
I also believe, like others replies going optical is the better solution.
regards.
PS: 4 coffee measuring spoons, a few bits of lego technic, a tracking or "obstacle avoidance" sensor and a piece of tin foil should make a great test setup before you buy anything expensive.
EDIT
PPS: Thinking about it some more I'm sure those things turn backwards or swing in certain low wind
situations too no ? This could probably be detected by adding a second optical sensor.
or if you want to be absolutely sure, use 3 sensors and only count a turn when sensor 1, 2, 3 registered a turn in that order. If 1 triggered followed by 3 or 2 followed by 1 it was going backwards..
Or 1, 1 ,1 in that case the piece of tin foil was swinging over a sensor.
In this scenario if this is important, your precision has also augmented to 1/3 turn instead of a full turn. Every time the sequence of the last 3 triggers was 123, 231 or 312 you can be pretty sure 1/3 turn went ok.
EDIT2:
hanging a ping pong ball on 4 springs or elastic bands and point a few distance measuring lasers at the thing ?
idk what precision/price ratio that would be though..
Or maybe a ping pong ball wired to 4 load cell sensors measuring the force on the ball from any direction.
Not sure about the precision of that one.
If you make a 3D grid of those things you should even be able to measure local gusts and quick wind turns.
Or you can also use an antenna to sense the movement of an object close to it like again a ping pong ball :P the same way a theremin works..
All this to say, there's a lot you can do yourself that doesn't have to look like or work like commercial windspeed meters either.
EDIT3:
The "flex and stretch" bit here contains some useful info:
http://hackaday.com/2016/09/16/resistance-in-motion-everything-you-should-know-about-variable-resistors/
Idk if this particular item is precise enough at that diameter for suspending a ping pong ball and measuring wind forces on it but still an interesting read. There's probably finer elastic resistors being developed as we speak, maybe even available idk.
