Hall Effect Sensor direction sensing

[deepakomanna]

Dear Sir,
In my application I am using Hall Sensor "ATS665LSG" ( http://www.allegromicro.com/en/Products/Part_Numbers/0665/ )
& On the rotor there is one magnetic slot. So whenever it rotates in clockwise direction this gives high-to-low pulses & when it rotates anti-clockwise it gives low-to-high pulses .
So in my application the rotor is running in clockwise but sometimes it rotates anti-clockwise. So please tell me how can I sense the direction i.e. rotor is running in clockwise or anti-clockwise.
_________________
Thank You,
With Best Regards,
Deepak.

[FvM]

According to the datasheet, the said sensor is not providing any kind of directional information. The only option would be in an asymmetricall structure in your hardware, e.g. two slots of different width.

Two sensors with a mechanical offset is another solution.

[Ttelmah]

Spot on.
If you read the data for these, they are _speed sensors_, _not_ _direction sensors_. This is the whole difference between devices like these, and units like quadrature encoders.
There are ways of adding directional sensing, but all are going to involve some extra hardware, or changes to your hardware.
If (for instance), you are reading off a toothed plate, rather than a gear, then you can fractionally stagger a tooth. Setting one a little 'early' as you rotate in one direction. This only gives the ability to detect direction once per rev, and if the speeds are reasonably slow changing, but what you then do, is have a PLL, running off the pulse train, and look for the one pulse arriving signifcantly early/late.
The second alternative is to add a second sensor, offset by a fraction of a tooth. This is how quadrature encoders are done. This is nice, since it not only gives direction information for every tooth, but also improves the potential resolution as well.
There are then other routes with different sensors and alternative hardware.
As FvM says, What you have, just does not provide directional information.

Best Wishes