Looking at your posts on the PN forum, the main issue you seem to be facing is that the upstream bitloading is pretty bad - especially in the U2 band starting around tone 1980. This is resulting in a low sync speed.
It doesn't look like DLM has artificially lowered the upstream speed, as this low sync speed is achieved with the standard 6dB noise margin. "Banding" by DLM, or artificially setting a low sync speed, will result in an SNRM level notably higher than 6dB.
In fact, from the "--stats" output that is visible in the "pastebin" link, we can see that DLM hasn't intervened at all yet. The error rate (600 ES's in 10 hours) isn't enough to trigger DLM - at least not on a PN line.
You appear, at first glance, to be getting decent downstream speeds. Ordinarily, that would suggest that you are either suffering from noise and interference that is particular to the upstream frequencies, or you are transmitting at too low a power level - though there isn't necessarily anything you can do about either.
However, the bitloading graph shows your downstream loading has been done very differently from the norm too (see link later for a comparison example). Ordinarily, I'd expect a lot more of the D3 band to be used - instead, you seem to have a small "blip" of D3, and a lot of extra bits in D2 and D1.
Overall, it looks strange, suspicious even, and suggests your line isn't working as normal at all ... so perhaps your latest discovery of a tree rubbing the distribution cable is indeed the cause of the issues.
It might be useful to post a few more things from the HG612:
- The Hlog graph, which might give an idea of the state of the copper
- The QLN graph, which indicates how noisy your line is
- The SNR graph
- The data from the "--pbParams" command
All of these are things you can get from tabs on DSLstats; the latter two are within the "telnet data" tab.
It is probably easier to post images over on PN's forum, with a post here to say you've done it.
- Where you see reduced bit loading below tone 180 ... is normal. The FTTC cabinet is configured to reduce power on lower tones, so it doesn't interfere with ADSL signals from the exchange. The setting varies for each cabinet, so for some you see this notch at different tones.
- Your attenuation is 18.8dB, which is consistent with the speed estimates that BT provide. 80Mbps is certainly feasible downstream, just about, but it could easily drop somewhat as it gets affected by crosstalk - which will happen with increasing subscriber numbers.
Other posters are getting hung up on how long your line is, but they are forgetting that the gauge (thickness) of the copper is important, as well as the length.
An old line of mine had a 16.7dB attenuation, and was about 370m from the cabinet. That would suggest your line is behaving
like a line that is only 410m long. However, it is entirely possible that your line uses thicker copper than normal - perhaps 0.6mm instead of 0.5mm.
That line started with an attainable speed of about 90Mbps, but that dropped as crosstalk took effect over 3 years, dropping eventually to 78Mbps.
The upstream attainable was well over 20Mbps throughout. Ordinarily, you ought to be in that region.
For comparison, here is what the graphs look like for that line:
- I've had an even older line, that was 650m of copper. This was back in the days when only the 40/10 package existed, but it achieved a 10Mbps upstream easily, and reported an attainable of 16Mbps.
- The agent who told you "favouring download over upload" is talking nonsense. The distribution of upstream frequencies vs downstream ones is fixed nationwide, and everyone gets favoured the same.