Filter Component
Digital signal processing component that reduces the sampling rate of a data stream while preventing aliasing by removing high frequency content. A decimation filter sits after an oversampling analog to digital converter to produce a lower rate signal with higher bit depth. It defines the usable bandwidth of the sensor output by suppressing noise outside the frequency of interest.
Signal Preservation
Low pass characteristics are essential to ensure that signals above the Nyquist frequency of the new rate do not fold back into the baseband. The decimation filter achieves this through a series of mathematical operations that weight and sum consecutive samples. This process improves the signal to noise ratio by spreading the quantization noise across a wider initial bandwidth.
Computational Load
Finite impulse response structures are commonly used because they provide linear phase response and guaranteed stability. Implementing a decimation filter requires significant arithmetic resources when the decimation factor is large. Polyphase architectures allow the processor to skip unnecessary calculations by only computing the samples that will be retained in the final output.
This efficiency is critical for low power battery operated sensors.
Latency Tradeoff
Sharp cutoff requirements often necessitate a high number of filter coefficients. These additional stages in a decimation filter introduce a delay between the physical event and the digital representation. Real time control loops must account for this lag to maintain system stability.
