The story of magnetic levitation haptics

Conventional haptics

Conventional 3-DOF haptic interfaces are essentially back-driven robot arms, with either serial or parallel kinematics using motors, encoders, and mechanical linkages or transmissions. To provide good haptic interaction, they must have low inertia, low static friction, minimal backlash or lost motion, and minimal cogging effects from motor commutation. To provide the highest degree of haptic transparency, they must have high position resolution and high mechanical bandwidths, free of resonances. In closed-loop operation with a virtual environment, a device should approach perfect freedom moving through empty space, and approach perfect rigidity when emulating contact with a rigid object. Providing all these attributes challenges the art of mechanical design. Further requirements arise in 6-DOF devices requiring the addition of three additional bulky motors and encoders at the user's handle.

Maglev haptics

Butterfly Haptics has taken a different approach to providing a 6-DOF user interface. We have eliminated all the mechanical complexity in favor of a single moving part levitated by magnetic fields. The user's handle is rigidly attached to a lightweight "flotor" that floats between stators with permanent magnets providing strong magnetic fields. The position and orientation of the flotor is tracked by optical sensors. As the user moves the handle through its motion range in 6 DOFs, position information is sent to the user's application. Conversely, forces and torques are sent to the handle from the user's application.

Maglev haptics provides the highest resolution and highest position and force bandwidths of any known method. There is an essentially direct electrodynamic connection between the computer and the hand, conveying gross force and torque effects to the proprioceptive sensors as well as subtle vibratory effects to the skin sensors. The high performance comes at the expense of a small workspace. For many haptic applications, scaling, indexing, and rate control methods can be used to effectively overcome this limitation.


Magnetic levitation of a rigid body using Lorentz forces was invented by Dr. Ralph Hollis, a founder of Butterfly Haptics, in 1984. The first application was a precision robot wrist. Its use as a haptic technology was invented several years later by Dr. Tim Salcudean and Ralph Hollis while at IBM Research. Subsequently, Salcudean and Hollis pursued further developments at the University of British Columbia and at Carnegie Mellon University, respectively. Its use as a vibration isolation method for microgravity experiments was developed by Salcudean and the Canadian Space Agency, where systems flew in the Mir space station for three years, and on the STS-85 shuttle mission.

Professor Hollis and his Ph.D. student Peter Berkelman developed the first relatively large workspace magnetic levitation haptic device at Carnegie Mellon in 1997 with funding from the National Science Foundation. This prototype device has been used for many psychophysical investigations of hard contact, texture synthesis and perception, and deformable object perception studies and is still in use today. This work has been carried out by Hollis, Ph.D. student Bertram Unger, M.S. student Vinithra Varadharajan, and Professor Roberta Klatzky.

In 2004, an NSF Major Research Instrumentation grant allowed Hollis and his group to begin developing an improved version with higher performance, greater useability, and lower cost. As part of the grant, a consortium of researchers was formed and a number of the new Beta maglev haptic systems were built, with six systems distributed to members of the consortium.

Butterfly Haptics was spun out in December, 2007 to commercialize the technology developed under the Major Research Instrumentation grant with an exclusive world-wide license from Carnegie Mellon University.