Self-cap works by sensing one sensor’s capacitance to circuit ground. This is done by applying an excitation signal at one sensor (TX) and sensing the amount of charge or current it takes to fully charge this sensor with a receiver (RX). There are two possible loops for current to flow in this model:
- Direct coupling from the sensor through the human body (I2).
- Fringe field coupling from the sensor to an adjacent sensor (I1).
For self-cap the dominant source of signal is I2. This signal is largely governed by the direct capacitive coupling between the finger and sensor (shown as blue field lines in Figure 1).
The direct capacitive coupling can be estimated by the parallel plate capacitor equation C = E0*Er*A/d, where E0 is the permittivity of free space, Er is the relative dielectric of the touchscreen overlay, A is the area covered by the finger, and d is the distance between the finger and the sensor (separated by the touchscreen overlay material). C1 and C2 in Figure 1 are the capacitances of the mobile device and human body to earth ground, respectively. These capacitances are normally assumed to be much greater than the direct coupling capacitance; therefore, the series combination of all three capacitances reduces to the direct coupling capacitance. However, C1 is often small enough to reduce the overall direct capacitive coupling, especially when the mobile device is completely battery powered with no charger connected.
The fringe field signal I1 adds some touch signal during a touch because the finger steals this signal away and diverts it through the human body to ground (adding to I2). Water that is untouched on the touchscreen has the largest effect on I1 and is the main source of error for capacitive touchscreens. Water strengthens the fringe field between adjacent sensors and increases capacitance. Depending on the touchscreen overlay thickness and dielectric, this may cause the capacitance to change enough to look like a light finger touch to the sensing circuit and cause false touches. One solution to this problem is to use a driven shield (sometimes called “guard”). See Figure 2.
By driving the adjacent sensor with an exact replica of the TX, I1 is eliminated and no capacitance is detected by the sensing circuit. For this solution to be practical, the touchscreen controller must be able to dynamically switch its sensing pin function from TX, RX and Shield on the fly so that the entire touchscreen can be sensed. The shield works equally well with traditional CapSense buttons.
Figure 3 provides an alternate graphical view to better understand how I1, I2 and sensed current IRX change with touch, water and shield versus no shield.
Mutual capacitance works by sensing the capacitance between two sensors. See Figure 4.
TX is applied to one sensor and RX to an adjacent sensor. The physics for mutual-cap sensing are the exact same as in self-cap, but now the dominant source of finger signal is due to the fringe field instead of direct coupling. The finger effectively steals charge in the form of current that would normally flow through the fringe field (I1) and diverts it to ground through the human body (I2). The net effect is a reduction in mutual capacitance between the two sensors. Water on the touchscreen that is not touched by a finger produces the same behavior as in self-capacitance; it adds capacitive signal by increasing the fringe field strength and increasing current flow to RX. Figure 5 provides another graphical view of mutual capacitance sensing.