Hi again UncleUB,
Kitz will probably be much better to explain this in detail but I think it's the noise spikes which really affect the different modems' performances.
If you look at some of the graphs in
http://forum.kitz.co.uk/index.php?topic=4399.45you will see they are all over the place. Whereas the ones now are absolutely flat lines.
Note also that we now have a constant noise margin of 6 dB whereas yours is less at 4.8 dB and, as the dB is a logarithmic measurement, that's quite a difference.
It's also important to realise we're dealing with (Lies, damned lies and) statistics. The routerstats data that comes from a 2wire is only to the nearest whole dB whereas the data I obtained from the DG834GT had a resolution of 0.1 dB and the sample time is only 15 seconds, so there's a very good chance that some spikes could be missed.
However in my latest examples, the line is dead flat for many many hours so it might be reasonable to say we would have seen at least one spike of over 1 dB during the period if any had happened.
It took me some time to realise that just because Routerstats joins the 15 second plots in a straight line, it DOES NOT mean that the actual noise margin is an absolutely constant figure.
Bearing in mind the router is (I think) averaging the noise margins of all the tone frequencies in use, it would be quite astonishing if the value did remain constant.
The decision you have to make is whether your line is stable enough to take advantage of the 2wire, or more accurately the AR7 chipset and 2wire's firmware.
My inclination would be to take a conservative (small c) aproach and
not to risk the 2wire.
Now lets see what others suggest !
Kind regards,
Walter