Joe, K1JT, recently made WSPR 2.1 available which now employs I/Q-SDR. Primarily it seems to be intended for the SoftRock RXTX Ensemble.
I figure, with the correct tweek, it could be made to operate on the 600m-WSPR range as well. Not sure about the precise working of the I/Q-SDR option of Joe's software, so 2 scenarios are possible:
- An ideal case would be that the software employs its own SDR-LO, meaning that it would be sufficient to have the WSPR-band within the coverage of the SDR-receiver.
- Could be that the software does not provide a "virtual local oscillator". In this case, the SDR-center frequency would need to coincide with the "dial frequency" for the respective band.
In case 1, a SDR center frequency of 500kHz would be just fine, hence, the transceiver design could be based on a simple 2.000MHz CMOS oscillator.
In case 2, the SDR center frequency would need to be 502.4kHz. This would chance the design in so far, that a XO would have to be pulled to 2.0096Mhz. A pull by 9.6kHz could be ambitious on 2MHz, even in a Pierce oscillator, but certainly worth a try.
Keep it simple, I would very likely opt for 4066 switches. Other stuff imaginable....
We are digital, using carriers, linearity is not the primary goal at this stage. So, simple/cheap possibilities would be IRF510, IRF820 etc.
Frequencies are kinda low, so, this all could be done in pure CMOS, which would have the advantage of a) being low power and b) easily creating 8V output to drive the MOSFET.
I hope I can realize this before the expiry of the experimental license.