Aha - genius Jelv! You could well be on to something there...
Whilst I don't graph the connection type (kinda difficult!) I note from some seperate logging that the drop-off to ~2kbps on Saturday does indeed coincide with a change to an xDSL connection type of ADSL2 compared with ADSL2+ both before and after... A quick Google search suggests that a 3dB differential is to be expected given the doubling of the frequency bands with the latter.
I guess that may explain the attentuation change, but presumably it is only an effect rather than a cause of my issues?
