18.1 Tuning for Noise Performance

In any touch sensing application, the system designer must consider how electrical interference in the target environment may affect the performance of the sensors.

Touch sensors with insufficient tuning can show failures in tests of either radiated or conducted noise, which can occur in the environment or power domain of the application or may be generated by the application itself during normal operation.

In many applications, there are quality standards that must be met where EMC performance criteria are clearly defined. However, meeting the standards cannot be considered as proof that the system will never show EMC problems, as the standards include only the most commonly occurring types and sources of noise.

Noise immunity comes at the cost of increased touch response time and power consumption. The system designer must carry out a proper tuning of the touch sensors to ensure the least power consumption. The QTouch modular library has several user-configurable features that can be tuned to give the best balance between touch response time, noise immunity, and power consumption.

Noise Sources

Noise sources that affect touch sensor performance can occur in a wide variety of applications like:
  • Motors
  • Piezo buzzers
  • PWM controls
  • Fluorescent lamps
  • Radio transmissions
  • Inductive cook tops
  • Power supply/chargers
  • Mains supply

Applicable EMC Standards

  • Conducted Immunity EN61000-4-6
  • Electrostatic Discharge (ESD) EN61000-4-2
  • Electrical Fast Transient (EFT) EN61000-4-4

Noise Counter Measures

The effects of noise are highly dependent on the amplitude of the noise signal induced or injected onto the sensors, and the frequency profile of that noise signal.

Generally, this noise can be classified as:
  • Broadband noise
or
  • Narrow band noise

Broadband Noise Counter Measures

In broadband noise, most of the noise spectrum lies outside the sampling frequency. Provided that the maximum and minimum voltage levels of the acquisition signal combined with noise signals are within the input range of the measurement system and a sufficiently large number of samples are taken, broadband noise interference can be averaged out by setting a high value of oversampling.

Excessive noise amplitude can saturate the analog front end. In this case, the acquisition signals combined with the noise signals are outside the input range of the measurement circuit, which results in clipping of the measurements.

Often the clipping is not observable in the resolved measurement, as it occurs only on a portion of the measurement samples, but the presence of clipped samples prevents effective averaging of the sample points.

In this case, averaging of samples will not result in a noise-free measurement, even with large rates of oversampling. The resolved signal will show a shift from its correct level due to the asymmetry of signal clipping.