Out-Of-Band Rejection
Signal filtration performance describes the degree to which a frequency-selective network reduces the amplitude of signals outside its designated passband. In digital and analog filters, stopband attenuation represents the minimum rejection level applied to unwanted high-frequency components to prevent them from corrupting the measurement. This suppression is expressed in decibels and is a major factor in determining the overall dynamic range of the sensor system.
A high rejection level is necessary to isolate weak sensor signals from strong external transmitters and environmental noise.
Calibration Limit
Metrological standards specify the minimum allowable rejection level for instruments used in custody transfer and billing applications. The measurement of stopband attenuation is performed during factory calibration using a sweep of high-amplitude sine waves. This test verifies that the filter meets the designated rejection limit across the entire specified stopband.
The calibration certificate documents this value to guarantee the instrument’s performance under dense signal conditions.
Filter Boundary
Hardware limitations in the analog stages, such as board-level crosstalk and power supply coupling, can limit the achievable stopband attenuation in the field. Even if the digital filter has a theoretical rejection of one hundred decibels, analog leakage can reduce this to eighty decibels. Engineers design shielding enclosures and dedicated ground planes to maintain the required separation between the input and output stages.
The performance is verified by testing the fully assembled instrument in an anechoic chamber. It is checked across several power supply voltage levels to confirm that the analog circuitry remains immune to system-level electrical noise.
Distortion Analysis
Digital filter coefficients must be calculated with sufficient precision to maintain the designed stopband attenuation. In fixed-point processors, round-off errors and coefficient quantization can degrade the stopband performance by creating a noise floor that limits the rejection. Software tools simulate the filter behavior with different word lengths to find the optimal balance between calculation speed and rejection depth.
This simulation ensures the digital implementation matches the theoretical model.