Indexing a wheel to a detent location.

[cmmguy]

I am building a device that rotates to three positions.
The positions must be very accurate and not move once in position.
The idea is to use a rotating disk with three "v" detents located around the perimeter of the disk which are 120 degrees apart. See the example model.
A motor of some sort will be used to rotate the disk to each detent and a coded limit switch setup will feedback which position is locked in.

My question is what type of motor system can I use to rotate the disc that will allow me to turn off the motor to allow the motor drive system to "float" or free-wheel while the detent takes over the precision locating function? My experience with servos or steppers is that they hold a position and dont allow floating. Will I have to turn off the power to the motor for this to work?

Any ideas?

Thanks

[Adrenalynn]

Unless you have a clutch, yes, you'll have to power down. Any gear motor will introduce a lot of drag.

[jes1510]

A simple gearmotor will work since you will be getting feedback from the limit switch. Adrenalynn is right as well. If you add a gearmotor then the gear train will induce a lot of drag while turning the mechanism by hand. Also keep in mid you will be genrating back-emf when you turn the motor by hand since it will act like a generator.

[robologist]

I tried to find a 3 sided Geneva mechanism, but found 4 and 6 stop versions (http://www.meadinfo.org/2009/05/geneva-wheel-design.html). I'd also wondered about a pushrod advancing a cogged wheel, or one with detentes similar to what you have above. Latching mechanism could be the ball you have, spring loaded, or a catch of some sort, that detaches when the pushrod advances.

[cmmguy]

I tried to find a 3 sided Geneva mechanism, but found 4 and 6 stop versions (http://www.meadinfo.org/2009/05/geneva-wheel-design.html). I'd also wondered about a pushrod advancing a cogged wheel, or one with detentes similar to what you have above. Latching mechanism could be the ball you have, spring loaded, or a catch of some sort, that detaches when the pushrod advances.

that might be a better solution than the detent method. The detent method has a small about of impact associated with the ball sliding into the vee but this geneva mechanism has not violence in the system which is much better. Now I only have to make the motor stop at a specific point in rotation and it seems that it does not have to be that accurate to achieve the index accuracy.

Thanks to everyone for their answers.