[hpsdr] multi Mercury

[Ken]

Georg Wrote

>DttSP is the main code for PowerSDR. So far as I know it supports up to
>4 receivers and the receiver number is a parameter. Based on this
>information this would mean that no changes on the mercury board would
>be necessary. The communication is done through OZY.

The problem right now, with this approach, is

1) The Mercury talks to OZY through an I2C serial interface (at present
anyway); all Mercury boards would have to share the same wire (bus line A10)
- that would really strangle through-put

2) You could use 2 OZY's/Atlas's for this, but the USB interface on PowerSDR
does not support more than one connection (see notes on Flex Radio's site
relevant to this)

Joe: The no-OZY approach would just take the decimated data stream and feed
it into the on-board D/A converter(s) and you would get the I+Q out as
baseband "audio" that you would feed into a sound card (kind of a step
backwards, but that would allow Linrad and MAP65 would work without
modification). This would not produce the bandwidth you are looking for,
however.

To use polarity diversity (for EME, or any other use where you have two
antennas) I think you cannot just add the signals - at least not for
single-signal reception. As most folks know, Linrad does this type of
reception. Scratching my head over the Linrad source code "mix2.c" I can see
that what you need to do is estimate the phase angle of the desired signal
and then take an appropriate fraction  of the real and imaginary components
of each sample. That way, you peak up the one signal you are listening to,
and not the noise .... 

Apparently there are some issues with this approach when you use a digital
down conversion scheme, not the least of which is full/exact synchronization
of the H&V samples, as seen by the SDR program. You need to maintain the
sync as you tune, as well.   
 
Ken