The thing is, regarding line 4, in a very recent SNRM 24-hour graph, I saw a 4dB sharp decrease in downstream noise level (upwards-going in the graph), from 4dB to 8dB vertical cliff. So I am getting far more or less noise, depending on your point of view, at times. This has not been seen before on downstream, yet this kind of step transition is very noticeable with the _upstream_on many of the lines, with either a single on/off cycle with a 24 hour period or a higher frequency cycle of several cycles per day.
So it could be that there is some equipment being switched on/off somewhere near NSBFD, in civilisation, at the exchange-end, as there is no human habitation or electrical activity anywhere near here, not even a wall or a tree for three miles upstream from this house along the cable run, and only in the last mile does one find a lot of houses.
Theory: I’m wondering if I have enjoyed my historic 3Mbps downstream sync rates while the noise is low and recently line 4 has started experiencing periods of high noise each day. If this noise extends to higher frequencies it kills the downstream performance. If it is confined to lower tones, it only affects upstream which also is on a higher 6dB margin so more resilient. And I presume there’s no DLM on upstream? But I don’t know. Anyway, if I resynched during the low noise times, then at 4dB less noise, I can get my usual high sync rates, but the 4dB noise increased level is in effect at that point in the day, then I get knocked out. Presumably this confuses DLM because it can’t understand what’s going on, what with the on-off noise pattern.
That’s the best I can come up with so far.
I think it was an excellent idea from AA about turning DLM off. In my ignorance I didn’t even know that that could be done with some BT lines.
I don’t have any similar data for line 1, as I haven’t yet installed a stats monitoring server.
I don’t suppose BT will do anything about an external noise source. But given that the other lines are not getting it, surely it has to be an issue of increased _susceptibility_ in that line, no? Lack of balance perhaps? Would that make sense? Insufficient noise rejection, because of inadequate symmetry between the two sides of/conductors in that pair?