block diagram
Receiver and transmitter use separate antennas for simplicity to start.
GnuRadio/Red Pitaya DDC (on the FPGA) yields streaming baseband data from RF coming in the antenna. The receiver is integral to the Red Pitaya board: ADC -> FPGA -> DDC -> CPU -> processing -> range, polarity, doppler, etc.
GnuRadio/Red Pitaya DUC (on the FPGA) yields RF from the generated phase-modulated waveform to the antenna. The transmitter is integral to the Red Pitaya board: digital phase modulated waveform -> CPU -> FPGA -> DUC -> DAC
For initial lab testing, we don't need this filter as long as we are not transmitting into an antenna.
§ 97.307(d) requires spurious emissions (harmonics are typically the largest spurious emission on unfiltered transmitters) to be suppressed to 43 dB below desired emission. This is typically accomplished via a low-pass filter. Besides its own emissions, an unfiltered transmitter can create intermodulation products with other higher-frequency transmitters nearby, leaking into our own receiver (or someone else's) even when the transmitter is off.
Fixed-cutoff frequency LPFs are available for less than $5. Initially we are interested in the 7 MHz band. When we move to transmitting on multiple bands (say adding 3.5 MHz) we would need to electrically switch between 3.5MHz and 7MHz filters as the 7MHz filter will pass the second harmonic of the 3.5MHz signal.
It has to be a hardware filter because the unwanted emissions are generated in hardware due to non-ideal behavior. There is not a practical way to eliminate all the spurious emissions in software.
Power supply: If a passive unswitched filter, no power supply. If a bandswitched filter, would need a modest power supply.
We assume there is 120 Vac utility mains power available. There might be occasional outages so we do want to system to come back up by itself without human intervention. Because of the deployments to northern latitudes where sunlight isn't available in winter, we aren't concerned with providing our own power for the first generation prototype. We don't want to be a power hog however, for future situations where there might be deployments using on-site renewable energy sources (wind,solar).
GPS 1PPS receivers are available for $50. The computer clock drifts due to errors in the timebase. These errors can be adjusted on an ongoing basis (compensating for diurnal temperature drift) by a 1PPS from a $50 GPS.
The transmitter frequency can be adjusted to remove bias in the following ways:
- monitor on-air frequency standard such as WWV, CHU or one of many other frequency/time reference stations.
- clever method via 1PPS perhaps
Power supply: Typically requires a clean power supply.
If less than about 20 dBm transmit power is used, it is possible to use simple Skyworks SPDT switches that cost about $1 and cover 300 kHz to 2.5 GHz.
The "pulse stretcher" if needed, is a generic device built from op amps that is used to make a too-short pulse longer. E.g. if 1PPS is 10 μs pulse width and we need a bit more energy say 10 ms, a pulse stretcher deterministically increases the pulse width.
Power supply: bias voltage is typically all that's needed, which is particular to the switch. I.e. the control voltage is the power supply.
if needed, GPSDO reference provides a far more accurate frequency reference. GPSDO are available new from $175
- Geppetto
- BG7TBL
- Jackson Labs (more expensive, pro grade)
Initially we are planning for simple transmit/receive antennas. Various straightforward designs have been used in a variety of HF ionospheric science efforts.
An antenna design simple enough for anyone to construct with basic hand tools is the multiband fan dipole, described in numerous sources including a 1969 SRI report. The length of this dipole will be adjusted to account for being closer to the ground that the original design may have anticipated.
Power Supply: This sort of antenna requires no power source. Although some types of antennas, including some used in HF ionospheric science, do require power for switching or remote tuning, in interest of robustness we would initially focus on passive antenna designs. The antenna designs that do require power come with a power injector box that simply screws inline with the existing coax cable.