[hpsdr] Proposal for a 'low-end' RX/TX board for HPSDR

[CT1IZU]

> Hello Ray!
>
> > This is still in the conceptual stage. Many details to be worked out
> > yet. Comments and feature wishlists will certainly be considered.

This is a great idea and will bridge a serious technolgy gap for those who
want to wet their feet on HF SDR.


>
> Please consider making the board large enough to:
>
> 1) include a Tx PA that can provide 3 to 8 watts PEP into a 50-ohm load;
>
> 2) include the Rx "LNA" for impedance matching and with sufficient
> dynamic range to do the rest of the system justice;
>
> 3) include the BPF and LPF needed for both Tx and Rx on the band(s) of
> operation, including switching if more than one band (latching relays
> are OK, or FSTxxx switches);
>
> 4) include suitable T/R switching (not using relays) so QSK CW is a
> possibility.
>
> This need not be the bleeding edge, but it would be enough to provide a
> practical radio that can be used without having to add a bunch of other
> "stuff".

I wonder about the advisability of rf section on the back plane. Digital and
rf can be a pain to seperate.  In any case the PA/LPF and BPF are not small
footprints. Typicaly a 7ele LPF with minature selection realys runs around
70x130mm and a 3 stage 8w PA about 60x90.

Would  make sense to put the DDS and control on the Atlas and then just the
coaxs and IC2 out to die cast box with the rf section. I'd really like to
see use of the AD9951 as I find the spurs on the 10 bit parts no longer
acceptable. If the front end is 0-30MHz  and used the divide by 2
quadrature generator that the QRP2001 used for their Tayoe then at a max of
60MHz we should get fairly acceptable phase noise with minimal
complications.

>
> In a similar vein, I've also thought about a "Sasquatch Lite" based on a
> DSP that is available and cheap, and supported by free or nearly free
> (as in liquid refreshment) development tools.

Option A on a DSP-X ? hi.  The STAR transceiver certainly shows how much
better this option can be. I like the idea of reducing the CPU overhead
anyway

Shel CT1IZU.



 1153519966.0