Design Considerations
AN64846 - Getting Started with CapSense
®
Doc. No. 001-64846 Rev. *X
60
3.4.5 Event-Based Filters
Event-based filters involve a special filtering method, where a pattern observed in the sensor data causes a predefined
response in the CapSense system. The pattern in the data is triggered by an event, such as a handheld product placed
into a pocket, or the power supply voltage (V
DD
)
dropping suddenly in a camera phone when the camera flash circuit is
being charged. Following are the common responses used with an event-based filter:
To block the CapSense data transmission until the pattern returns to normal
To reset the level of the baseline reference defined in SNR
To drop or ignore the sample of data when the event occurred
An example for an event-based filter is dropping the sample or resampling of the data when the interrupt occurred. I
2
C
is one of the common communication protocols used in CapSense applications. Because I
2
C interrupts are
asynchronous in nature, they may occur when the sensors are being scanned. Interrupts that occur during the scan
may increase the noise and, therefore, decrease the SNR. In such cases, you may implement an event-based filter,
such as ignoring the raw count sample corresponding to that scan when interrupts occurred and rescanning.
3.4.6 Rule-Based Filters
Rule-based filters are another special filtering method, where a pattern observed in the sensor data causes a rule-
based response in the CapSense system. Unlike the event-based filter, the rule-based filter acts on patterns in the
sensor data that are encountered during normal operation of the touch surface. The rule-based filter takes into account
special scenarios on how sensors are used.
For example, with a set of radio channel selection buttons, two buttons can be touched accidentally, but only one should
be selected. The rule-based filter sorts out these kind of situations in a predefined way. Another example is having a
virtual sensor in CapSense applications. The virtual sensor is not expected to be triggered during normal operation of
the sensors, but it may occur in unexpected scenarios such as presence of water (for example, the guard sensor) or
when the system is subjected to RF noise. Therefore, when the virtual sensor is triggered, all the actual sensors are
turned off so that there is no unintended triggering of the actual sensors.
3.5 Power Consumption
Minimizing power consumption is an important design goal. For many CapSense systems, extending battery life is
critical to the success of the product. In systems that do not use batteries, power consumption still plays a role in
optimizing power supply designs to reduce costs and PCB area.
3.5.1 Active and Sleep Current
Active current is the current consumed by the device when all the selected analog and digital blocks are enabled and
the CPU is running. In typical applications, the CapSense controller does not need to be in the Active state all the time.
The device can be put into the Sleep state to stop the CPU and the major blocks of the device. Current consumed by
the device in Sleep state is called sleep current. Sleep current is much lower than the active current.
3.5.2 Average Current
In typical applications, sleep state can be invoked periodically to reduce power consumption. This means that during a
preset time period, the CapSense controller wakes up from sleep state, performs all necessary operations in the active
state (scan all sensors, update all baselines, check if any sensor is in the TOUCH state, and so on), and then returns
to sleep state. The resulting instantaneous current graph is shown in
