Hall Effect Sensing Input and Like Polarity
Haptic Feedback in the Liquid Interface System
Kasun Karunanayaka, Jeffrey Tzu Kwan Valino Koh,
Eishem Bilal Naik, and Adrian David Cheok
Keio-NUS CUTE Center, NGS, National University of Singapore
{kasun,jeffrey,adriancheok}@mixedrealitylab.org
http://mixedrealitylab.org
Abstract. Liquid Interface is an organic user interface that utilizes
ferrofluid as an output display and input button embodiment. Using
a matrix of Hall effect sensors, magnetic fields generated by rare-earth
magnets worn on the fingertips are measured and are then converted into
signals that provide input capability. This input actuates an array of elec-
tromagnets. Both Hall effect sensors and electromagnets are contained
beneath the surface of the ferrofluid. By matching like polarities between
the electromagnets and the rare-earth magnets, haptic force feedback by
means of magnetic field repulsion can be achieved.
Keywords: Organic User Interface, Ferrofluid, Magnetic, Hall Effect.
1 Introduction
Building on the idea of previous ferrofluid artworks [2], and adhering to the
characteristics of organic user interfaces (OUI) [1], Liquid Interface (LI) provides
an input/output solution based on ferrofluid.
The system is composed of a pool of ferromagnetic liquid combined with a
sensing and actuation mechanism. The sensing is achieved through the use of an
array of Hall effect sensors. Actuation is produced by an array of electromagnets.
Users can interact with the system by wearing magnetic rings. The magnetic ring
position is detected by the array of Hall effect sensors, which in turn actuates
the electromagnets and the audio server. The magnetic field of the active elec-
tromagnets morphs the ferrofluid to create ”buttons”. When a button is pressed
the system generates a sound. The electromagnetic fields produced by the array
repel the rare-earth magnets worn on the fingertips, giving the user a haptic
response. Our previous work includes a detailed system description as well as
describes a series of experiments to measure spike height versus current, distance
of two adjacent spikes, transient state of the system and the static linearity of
the system [4].
In order to discern the parameters in which Hall effect sensing would be most
effective, a new series of experiments were conducted. These include experi-
ments for understanding the vertical distance of rare-earth magnets from the
surface embedded with Hall effect sensors, the horizontal sensing effectiveness
D. Keyson et al. (Eds.): AmI 2011, LNCS 7040, pp. 141–145, 2011.
c
© Springer-Verlag Berlin Heidelberg 2011

142 K. Karunanayaka et al.
to understand the quality of precision for cartesian coordination, and finally an
experiment to characterize the effectiveness of the Hall effect sensors employed
under the influence of multiple magnetic fields.
2 Experiments and Results
2.1 Experiment 1: Hall Effect Sensor Reading versus Vertical
Distance
This experiment has been conducted using a Hall effect sensor and an electro-
magnet that generates an average flux density on the surface from 450 to 1950
Gauss for the range of 6V to 24V with 1.9 to 7.5A of electrical current. In the
experiment we kept the power of the electromagnet at a constant voltage of 10V
and a driven current of 2.44A, with the sensor on the vertical axis on top of the
electromagnet. The sensor reading is measured versus the distance to the elec-
tromagnet. The value of the sensor output voltage taken is the mean value in
one second. This plot shows that the sensor is most sensitive with respect to the
vertical distance from 0cm to 3cm. When the distance is greater than 3cm, the
change in output is much smaller. At larger distances, for example the values of
6cm and 7cm, the difference in voltage is only 0.011 volts. However such a small
voltage difference is not detected by the micro-controller used for this iteration
of the system.
Fig. 1. Sensor output versus vertical distance
2.2 Experiment 2: Hall Effect Sensor Reading versus Horizontal
Distance
Once more keeping the power of the electromagnet constant, the sensor is placed
on the vertical axis of the electromagnet, at 2cm, since at this distance the
sensor is most sensitive, registering the largest change in values with respect
to distance moved. The plot shows that the sensor voltage is very close to 2.5
volt (zero field voltage) after 3.5cm displacement. Experiment 1 shows that the

Hall Effect Sensing Input and Like Polarity Haptic Feedback 143
Fig. 2. Sensor output versus horizontal displacement
resolution of the system cannot distinguish any smaller change that within 0.02
volts, the magnetic field at 3.5cm and beyond are too small to cause a change
in the microprocessor. This experiment shows that the magnetic field out of the
horizontal area of the magnet is too small to be detected at the optimal vertical
distance.
2.3 Experiment 3: Characterization of Hall Effect Sensor Readings
under the Influence of Multiple Magnetic Fields
In this experiment the readings of the Hall effect sensor are measured to deter-
mine the influence of the magnetic fields generated by the electromagnets and
neodymium magnets. The goal of this experiment is to determine which combi-
nations of the two magnetic fields (electromagnet and neodymium) cancel one
another.
The sensor is supplied the rated of 5V and is positioned such that it is in level
with the top of the electromagnet and directly next to it. Its output is connected
to an oscilloscope. A non-magnetic material at varying heights directly above the
Fig. 3. Sensor reading values obtained for different distances versus PWM

144 K. Karunanayaka et al.
sensor holds the neodymium magnet. Its pole direction is fixed, with the South
Pole facing downwards. The reading of the steady-state output voltage of the
sensor is recorded using the oscilloscope, while varying the height and direction
of the neodymium magnet and the PWM input to the electromagnet.
First the default sensor value is taken without the neodymium magnet or
electromagnet influence. Next, with the neodymium magnet pole at South Pole
(facing down), the PWM values and distances are measured.Here the strength
of the electromagnets field serves to decrease the reading of the sensor, whereas
the position of the neodymium magnet field serves to increase the reading of the
sensor. This results in a case in which the value of the sensor is unable to detect
the presence of the neodymium magnet due to the electromagnet’s field.
From the data we gathered, this occurs in the case when the distance of the
neodymium magnet is 7.0cm. If the electromagnet is off, the reading is 2.53V,
but if the electromagnet is turned on, the reading falls below the 2.50V neutral
value. To circumvent this problem, we use like poles instead of unlike poles. This
approach has the peripheral advantages of preventing the two magnets from
attracting each other and preventing the neodymium magnet from picking up
the ferrofluid as well as add haptic feedback.
3 Discussion
By using the results obtained in these three experiments, we were able devise an
algorithm that performs accurate sensing of nearby magnetic fields for the Liquid
Interface system. It is possible to track magnets worn on finger tips precisely as
each sensor needs only to be able to detect a given magnet directly above it and
will not be affected or disturbed by other magnets nearby. The sensitivity of our
sensor is very effective in detecting movement within the 3cm range. It is able
to detect even very subtle movements. The sensors are placed directly on top of
each electromagnet while the user’s fingertips carry strong neodymium magnets
with like poles of each magnet/electromagnet pair facing one another. When the
electromagnet is turned on, the sensor’s output becomes fully saturated. If a
neodymium magnet of the same pole is brought near to it, the sensor’s output
drops. This is detected as the presence of the user’s hands.
The micro-controller firmware handles the sensing input, actuation output and
communication with the server to produce music. Upon system initialization, the
system first performs a calibration. This process takes up to 20 seconds. During
the calibration each electromagnet is turned on to maximum power in order to
find the offset value of the sensor. This offset value is then used in determining
if a neodymium magnet (user’s hand) is nearby when the value varies from the
offset. To handle the sensing input, the micro-controller continuously polls the
analog-to-digital converter modules at a frequency of once every 200 milliseconds.
This is accomplished using a timer module and is to ensure that each analog-
to-digital conversion is given sufficient time to complete. Complex gestures are
handled by constantly storing interactions from the previous 2 seconds in the
program memory. The stored interactions can then be interpreted as necessary to

Hall Effect Sensing Input and Like Polarity Haptic Feedback 145
produce any gestures other than simple activation. To handle actuation output
the micro-controller sets the PWM duty cycle for each output if necessary. Each
time an interaction is recorded, the micro-controller sends a unique character to
the server via an RS-232 connection. This is interpreted by the server for use in
music production.
4 Conclusion
We outlined in this paper, three new experiments that enabled us to develop an
input sensing mechanism based on the Hall effect. These include findings reveal
the relationship between perpendicular and horizontal distances of the Hall effect
sensor and the magnetic field generated by the electromagnetic array, the char-
acterization of the magnetic Hall effect sensor readings under the influence of
multiple magnetic fields, and the relationship of distance from the sensor versus
PWM.
We also discussed the addition of haptic feedback facilitated by the repelling
force of rare-earth magnets placed on the fingertips with like polarities matched
to the electromagnets, thus providing an additional modality of feedback. This
input accessory provides a means for users to interact with the LI system without
the need to touch the ferrofluid, and still provides instantaneous tactile response.
Acknowledgement. This research is carried out under CUTE Project No.
WBS R- 7050000-100-279 partially funded by a grant from the National Research
Foundation (NRF) administered by the Media Development Authority (MDA)
of Singapore.
References
1. Holman, D., Vertegaal, R.: Organic user interfaces: designing computers in any way,
shape, or form. Communications of the ACM 51, 48–55 (2008)
2. Kodama, S.: Dynamic ferrofluid sculpture. Communications of the ACM 51, 79
(2008)
3. Koh, J.T.K.V., Karunanayaka, K., Sepulveda, R., Tharakan, M.J., Krishnan, M.,
Cheok, A.D.: Liquid interface: a malleable, transient, direct-touch interface. In: Pro-
ceedings of the 7th International Conference on Advances in Computer Entertain-
ment Technology (ACE 2010), pp. 45–48. ACM, New York (2010)