I think that the sample clock triggers a conversion on all
the channels at the same instant.

Have a look at the data sheet for the CS42448 and see what
you think.   I put a copy here -

 http://abelian.org/vlf/tmp/cs42448.pdf

Some while ago I applied the same signal to all inputs
and there was only a small phase difference between the
channels.  I'll have to try that again more carefully to
make sure.

I'm in the process of making up detailed instructions on how
to set up a Pi and Octo or stereo card and GPS with vlfrx-tools.

I might need a bit of advice from Wolf or someone on how to
use the rawtcp: mode of Spectrum Lab to pull signal from it.

--
Paul Nicholson
--