Hi Scotty...
On closer inspection, the DMT plots show a noticable variation in SNR above 900KHz, which is most probably caused by some sort of electrical interferance such as a motor, neon lighting or similar. It's probably only hurting one of the lines due to different routing of that pair - either in a different cable at some point - or simply further away from another pair with interferance in the same cable.
The difference in SNR is a symptom of the poor quality of the bad line - if you can get the SNR immediate after a re-sync, that will show the "SNR Target" for the line - we would expect the bad line to have a higher SNR Target to maintain stability as the expense of speed. Make it stable and it will go faster, eventually...
Not sure if you saw the request re Routerstats SNR plots? - These can help a lot if you can leave it running for a day or two. It may well help reveal the source of the RFI.
Re ADSL2, I don't see any issues with ADSL2 over ADSL1, it's usually faster and more reliable, my backup ADSL connection is on ADSL2 with 61.5dB loss and sync at 3283Kb, with interleave turned off on a 6Km line... SNR sits around 3dB 24x7.
A E/D-side pair swap may well help - but it's hit and miss as there's no way to know if you will get a better or worse pair than at present - though the odds look in your favour.
Good luck with Routerstats plots,
Regards,
Robin
