Nice chart.   I'm glad to see that you're getting some
decodes down below zero Eb/N0, that's more like a proper test!

Are you running with list size 200?  There's about another dB
of coding again available if you increase to 20000.  Maybe
that will fill in some of the gaps.

Mostly these signals are too strong. If you can see EbNaut on
a spectrogram, you don't need to be using it.

--
Paul Nicholson
--