New consumer products such as smartphones have helped make touch screen controllers popular, and touch sensors provide a convenient way to control virtually any type of device.
Touch-sensor controllers now offer versatile performance choices and formats such as sliders and proximity sensors -- and advances in touch sensor technology are making sensor-driven interfaces easier to implement, more intuitive, and simpler for end users.
Most touch-sensor controllers work by detecting the change in capacitance that occurs when something or someone approaches or touches the sensor's conductive metal pad (Figure 1 below).
As conductive objects (such as a finger) move in proximity to the sensor, they alter the electrical field lines of the capacitive sensor and change the capacitance that is measured by the control circuitry.
Industrial applications have used this capacitance-detection technique for many years to measure liquid levels, humidity, and material composition. From these applications, the technology was adapted for human-to-machine interfaces.
Touch-sensor interfaces typically detect a capacitance change by measuring the impedance of a circuit connected to the sensor pad. The touch controller periodically measures the impedance of the sensor input channels and uses these values to derive an internal baseline called the calibrated impedance. The controller uses this impedance value as the basis for all touch/no-touch decisions.
This simplified formula shows the major influences on a touch pad's capacitance when a finger approaches. This formula can be used to determine the capacitance and strength of a sensor pad.
Touch strength increased by
= Pressing harder
= Increasing area of touched surface
= Increasing capacitance
When D is decreased
= Capacitance is increased
= Touch strength is increased
As this equation indicates, the overlay cover thickness and its dielectric constant play a large part in determining the "strength" of the touch. The equation also shows that capacitance sensors are inherently sensitive to the surrounding environment and to the characteristics of the touching stimuli — whether the touch is from a finger, vinyl, rubber, cotton, leather, or water (Figure 1, above).
Table 1 below lists the dielectric constants of various common materials used for covers. With these values in mind, it is interesting to look at the behavior of touch sensors in a kitchen, where oil can easily be spilled on the sensors.
Typical kitchen oils such as olive or almond have dielectric constants in the range of 2.8-3.0. Paraffin at 68 degrees F has a value in the range of 2.2-4.7. These dielectric constants are similar to or even less than that of the polycarbonate (2.9 - 3.2) or ABS materials (2.87 - 3.0) typically used to cover the sensors. Thus, oils do not have much effect on sensor operation.
Conversely, glycerin (in liquid) has a dielectric constant in the range of 47-68, while water's dielectric constant is about 80. Even though these values are lower than those of the cover materials, spilling such liquids on a touch sensor having digital touch detection technology (such as that used in the FMA1127 Touch Sensor Controller, developed and owned by ATLab Inc.) does not cause any abnormal behavior because neither the sensor pad nor the spilled liquid is grounded.
Although a touch sensor's detailed operation and interface depend on the application, capacitive sensor interface circuits and detection methods can be either analog or digital, broadly speaking. One analog technique is to measure a frequency or duty cycle that changes due to the introduction of additional capacitance from finger to ground (Figure 2 below).
A high-resolution analog-to-digital converter (ADC) can be used with this technique to convert the sensed analog voltage to a digital code. The latest capacitance-to-digital converters take advantage of advances in mixed-signal technology to integrate high-performance analog front ends with low-power, high-performance ADCs.
A disadvantage of an analog interface circuit is that the capacitive sensor may be affected by subtle noise, crosstalk, and coupling. Additionally, the dynamic range of the sensor output is limited by the supply voltage, which is continuously decreasing as the semiconductor fabrication technology scales down.
The situation becomes more challenging if the sensor circuitry is integrated on the same substrate with complex digital signal processing (DSP) blocks in a deep submicron CMOS technology. To avoid external disturbances, the device may require software workarounds that place burdensome overhead on interfaced microcontroller's memory and performance.
（to be continued）