Capacitive sensors can replace mechanical switches for touch control

Capacitive touch-sensor interfaces typically consist of a capacitive sensor, a capacitance-to-digital converter (CDC), and a host processor (Fig. 1). The sensors are manufactured using traces on a standard two- or four-layer pcb or flex circuit, and thus don't require any external components or materials.

Reliable sensors must be immune to the effects of environmental changes, and must maintain accurate sensitivity levels under all operating conditions. As the temperature or humidity changes, the characteristics of the pcb material will change. As a result, the output level of a printed-circuit capacitive sensor will drift. This could occur when a user goes from an air-conditioned automobile to a hot, humid environment, for example. To avoid intermittent contact errors, real-time drift compensation must be included in the CDC (Fig. 2).

The sensor's ambient value, measured by the CDC during the period when the user isn't making contact with the sensor, drifts as the environmental conditions change (e.g., increasing temperature or humidity). To compensate, the high and low threshold levels are dynamically changed to determine a valid sensor contact. The reference levels are rescaled at positions 2, 3, 5 and 6 to maintain optimal threshold reference levels, automatically tracking and compensating for drift errors.

Printed circuit boards can be plagued by parasitic capacitances as large as 20 pF. This capacitance would offset the threshold at which the capacitive touch sensor is seen as pressed, thus changing its sensitivity. To compensate for the parasitic capacitance, DACs could be programmed to offset the input to the CDC. The parasitic capacitance is uniform from board to board, so a simple adjustment can be made at the time the boards are fabricated. This eliminates the need for external RC tuning components, thus minimizing the costs associated with material, assembly, and test. Individually adjusting the offset for each sensor lets designers take full advantage of the converter's resolution (Fig. 3).

Traditional mechanical switches have a familiar sensitivity and tactile feedback. These parameters must also be considered and optimized for capacitive sensors. Different sensors will likely require unique sensitivities depending on the switch function or the switch's physical location on the product. Furthermore, one set of sensitivity settings may not be suitable for every user, so a feature that would allow the user to select different sensitivity settings—from a sensitivity control menu, for instance—would be ideal. The AD7142, for example, supports these sensitivity requirements by allowing an individual 16-bit sensitivity control register to be programmed for each sensor. These registers could also be embedded into the host firmware and provided in a menu display, allowing the user to select different sensitivity levels to meet their particular needs.

Battery power would be consumed unnecessarily if each sensor input was sampled during periods when the user wasn't making contact with a sensor. To maximize battery efficiency, the CDC should be able to detect when the user has stopped making contact with the sensors, and automatically switch to a low-power mode. When the sensor is re-contacted, the IC would automatically be reconfigured into the normal operating mode.

For additional power savings, a complete shutdown mode should be included. In this case, the entire IC would be shutdown whenever the sensors where disabled. Disabling the sensor switches is typically accomplished in portable products by setting a mechanical switch or selecting a blocking mode from a control menu.

An additional benefit of replacing traditional mechanical switches with capacitive sensors is a simpler manufacturing and assembly process. While traditional mechanical switches require manual insertion of each switch into dedicated holes in the plastic housing, a single capacitive sensor board consisting of all the switches can be placed under the plastic casing one step. A sensor board mounting cavity containing an alignment notch and some glue is all that's needed for the sensor board installation and alignment.

The host processor board will emit electromagnetic noise that can couple into the capacitive sensor and sensor traces, resulting in unpredictable sensor behavior. This could degrade performance, but simple methods help minimize the electromagnetic interference (EMI) effects on the sensors. First, the CDC should be mounted on the sensor board. This will minimize the sensor trace lengths, which reduces the chance of EMI being coupled onto the traces. Using a four-layer sensor board with a solid ground plane will provide additional EMI shielding to the sensors. If these two methods don't effectively isolate the EMI noise from the sensors, a grounded metal shield could be placed over the sensor board cavity (Fig. 4).

Capacitive sensor fields can couple to conductive surfaces such as the product's metal casing or conductive metal painted coatings, resulting in unpredictable sensor behavior. This places a mechanical constraint on how close the edge of a capacitive sensor can be located from the edge of a metal surface. Furthermore, the sensitivity of a capacitive sensor is also related to the thickness of the plastic located directly above the sensors. If the plastic is too thick, the flux field lines won't effectively pass through the plastic, making sensor performance unreliable. Typically, the distance from case to sensor should be greater than 1.0 mm, and the plastic thickness should be less than 4.0 mm to maintain proper sensitivity ranges.

About the author
Wayne Palmer is an applications engineer for the sensor products group at Analog Devices. Wayne received a BSEE from Northeastern University, Boston, MA. He can be reached at wayne.palmer@analog.com.