The SuperDARN Borealis I/Q streams potentially offer great value as a resource for identifying HF signals of opportunity, including from other SuperDARN radars and from diverse sources, e.g. timing transmitters and codar. These signals would be useful for array calibration and propagation modeling, among other applications. However, the current recordings are not suitable for this application (at least not without major surgery), as they contain gaps of variable length where the transmit pulses occur. Many signal processing applications (e.g. FFT) require data at a constant sample rate, and these gaps are not easily filled as they are a non-integer number of samples long.
The following modification would address this issue, permitting straightforward data analysis:
Record continuous I/Q streams instead of breaking them up into chunks
Samples recorded during signal transmission can be blanked (set to 0), or they can be left as-is. Given the duty cycle is typically <10%, this change should add <10% to the data budget.
If this change is implemented, the receive-only debug mode should produce long (>1 second) continuous streams from the full array.
The SuperDARN Borealis I/Q streams potentially offer great value as a resource for identifying HF signals of opportunity, including from other SuperDARN radars and from diverse sources, e.g. timing transmitters and codar. These signals would be useful for array calibration and propagation modeling, among other applications. However, the current recordings are not suitable for this application (at least not without major surgery), as they contain gaps of variable length where the transmit pulses occur. Many signal processing applications (e.g. FFT) require data at a constant sample rate, and these gaps are not easily filled as they are a non-integer number of samples long.
The following modification would address this issue, permitting straightforward data analysis:
Record continuous I/Q streams instead of breaking them up into chunks
Samples recorded during signal transmission can be blanked (set to 0), or they can be left as-is. Given the duty cycle is typically <10%, this change should add <10% to the data budget.
If this change is implemented, the receive-only debug mode should produce long (>1 second) continuous streams from the full array.