Markus wrote:

> that looks not bad at all,

Looks promising doesn't it?  It can probably be improved: the
blanker and filter settings are just a reasonable guess, not
likely to be optimum.

This apparent 14:00 to 19:00 'window' might just be because
the sferic noise at that time suits the chosen sferic blanker
threshold.  A more aggressive blanking might open that out.

I've loaded up the workstation with jobs for the night:
copying in the raw file for the 21st from the outdoor PC,
then re-running the whole 11 days with a couple of different
blanker thresholds.  Then plot a whole bunch of spectra
spanning a range of times.  Should be done by the time I
get up in the morning.

One of those 128 core EC2 boxes would make short work of homing
in on the optimum settings - they have enough memory to load
the entire 11 days of raw signal into RAM. The problem is just
how to upload 11 by 24Gbyte files over an ADSL.

--
Paul Nicholson
--