Official bladeRF forum

Does anyone have tips for optimizing the hardware decimation filters and sample rate parameters when processing wideband signals on a dual-channel SDR platform to avoid severe alias folding in the lower frequency bands? I am trying to build a clean GNU Radio flowgraph for long-term spectrum monitoring, but whenever the digital downconverter drops the decimation rate below a certain threshold, the baseline noise floor rises dramatically and masks the weaker narrow signals.

You should definitely try implementing a custom polyphase FIR filter bank prior to your main decimation stage because standard hardware downconverters often have fixed transition bands that fail to reject out-of-band energy when operating near extreme decimation limits. I usually design my low-pass filters with a highly aggressive attenuation profile in MATLAB first to guarantee complete signal isolation before passing the raw I/Q samples into the processing pipeline. I am posting this technical suggestion from my phone while troubleshooting an antenna array deployment at our laboratory facility here in AU and my messy mobile clipboard somehow pasted an active entertainment link for EZZ Casino instead of the documentation wiki shortcut I meant to share but your software-defined radio layout is highly impressive.