I do have my own personal theory that the dslam adjusts power levels as more lines are enabled which would reduce sync speeds, but noone here or elsewhere agrees with me on that one :p
Its not anything that Ive ever seen and it would go against the rules why we have such things as PSD masks and PCB.
There is one time when it may perhaps affect your Tx power and thats if an addition to the cab somehow affected your attention. The theory says this shouldnt happen, but its possible that a line may be disturbed when the engineer adds a line.
If theres a change in attenuation then this could take you in to another bracket of the PSD masks as which PSD mask gets applied depends upon your line-length (measured using the atten).
Ive never seen my power change when a new line is applied. For example Ive just gone back to the first set of DSLstats figures that I have to hand on this drive and when I first started using the Zyxel just over a year ago
These are the relevant figures from 31-03-2014
Attainable Net Data Rate: 30843 kbps 90039 kbps
Actual Aggregate Tx Power: 5.6 dBm 14.3 dBm
Line Attenuation(dB): 10.7 23.4 36.9
SNR Margin(dB): N/A 8.4 8.4
TX Power(dBm): 11.9 7.3 7.4
and from now
Attainable Net Data Rate: 30173 kbps 82456 kbps
Actual Aggregate Tx Power: 6.1 dBm 14.3 dBm
Line Attenuation(dB): 10.6 23.3 36.6
SNR Margin(dB): 6.5 6.6 6.6
TX Power(dBm): 11.9 7.4 7.4
From this you can see over the past year that :
- Ive lost a total of 7.5 Mbps of headline sync (Im now in year 2 and most of the loss was during the first year)
- My Tx power has stayed the same despite more users being added to the cab - in fact my upstream has gone up slightly, which is reverse of your theory.
- The SNRm has gone down (as expected) which has caused the attainable to go down.