Turret Rotation Sensor

[Seat Ninja]

I was wondering how teams in the past (or this year) measured the rotation of their turrets, I.E. what sensors they used. I would say encoders, but there's the whole problem with the turret getting turned while the robot's turned off. Any help would be great, thanks.

[AndyB]

We are using a potentiometer. Using a pot should allow you to find a center regardless of if your turret was turned while it was off. Be very careful with side loads though. Pots hate side loads and tend to stop working after experiencing them.

[Seat Ninja]

Thanks for the fast response. Where would you get this potentiometer? I an assuming it would be continuous rotation and mounted on the rotation motor.

[TheOtherGuy]

We're using one of those grayhill encoders (stole this one from our 2007 robot until we buy one)... here's the mount we made yesterday for it:

http://photos.project1726.org/albums/userpics/10011/normal_IMG_0206.JPG

It's slightly spring loaded since the belt isn't exactly sticky. When the robot is turned off, it's very easy to rotate the turret, so before each match we simply turn it to whatever its initial position is going to be and start. We'll also have reset limit switches so if for some reason it randomly slips it'll still be able to "fix" itself.

[Gdeaver]

Their are absolute rotary position sensors that can measure the magnetic field orientation of a magnet. They give an analog output proportional to the angular position. Cherry electronics has a nice automotive grade unit that is better than a continues rotation pot. Mouser carries them.
http://www.mouser.com/Search/ProductDetail.aspx?qs=sGAEpiMZZMvhQj7WZhFIAPb0P%2f L5OpbJhxqh3niR4xw%3d
The nice thing about them is they always give absolute position.

[EricWilliams]

Absolute encoders can be a lot nicer than continuous pots, but a cheaper solution is to just have your turret turn towards a limit switch at the beginning of the math, and zero your encoder when it hits the switch.

[Jared Russell]

Absolute encoders can be a lot nicer than continuous pots, but a cheaper solution is to just have your turret turn towards a limit switch at the beginning of the math, and zero your encoder when it hits the switch.

This is a nice approach because if there is slip between the sensor and turret, the limit switch will correct for it.

Also, a lot of the advantages of absolute position-sensing sensors are lost once you have a reduction between the turret and the sensor (unless you have a multi-turn pot).