I'm not sure why this is.

Is this not because as it's the lower tones that travel furthest, on ADSL it's the upstream that uses the lowest tones. Hence even on long lines, upstream is still at full kahuna, and just the downstream dies off.

On VDSL the 'blocks' of up and downstream tones are spread through the frequency range. Hence upstream dies back in a far more pronounced fashion. Have seen a fair few 40/10 lines where the downstream is at 40 but the upstream is already off the pace, say 7 meg for instance.

This is of course all conjecture on my part, so will happily sit and wait to be shot down in technical flames.

It really depends on what happens with band US0 - the creation of which is one of the improvements made between VDSL and VDSL2 to allow for longer range upstream.

US0 is supposed to sit at the bottom of the frequency bands, just like ADSL and ADSL2+, to make sure some bandwidth is available at maximum ranges.

Unfortunately, the modem reports that DS1 spans tones 32-859, and US0 spans 0-95. Which leads to a question... just who is using tones 32-95? Up or down?

US0 is usually labelled as optional, as it isn't really needed for nearby modems. Perhaps it is really only used for distant modems.

But the overlap of bands is strange - because it shouldn't happen. DSL doesn't work if some lines in the cabinet use the US0 frequencies upstream, while other use it downstream. Too much NEXT.

The tone allocation values show that bits/tone start being allocated at tone 7 and ramp up to tone 15. Unfortunately, you don't see any gaps after that (in my stats) where you'd expect guard bands between US and DS bands. Strange... why no gaps?

The SNR values reported by the modem do fit to the DS bands precisely, with all US bands and guard bands empty. QLN and Hlog too.

Things are too inconsistent to work out what is going on.

Things are too inconsistent to work out what is going on.

There you go, my kind of answer !