ADSL rate just capped?

[#1]

Hi

I reset my connection last night and this morning I found that the connection had been reset again at 6am this morning and it appeared to have capped my connection rate at around 6.6Mb/s. Seems strange - although there have been power outages recently I'm pretty sure I have stayed below 5 resets in any 24 hours (until this morning AFTER I noticed the cap at 6am).

I have reset the connection a couple of times since then with different target SNRM and it always connects at 6.6Mb/s. By the time I apply AQM the actual achievable throughput is around 5Mb/s if I'm lucky.

Additional

I think my line is/was configured at the BT Wholesale Normal stability profile (the middle one, not aggressive and not stable).

It has been this way for some time now (9-12 months sort of time), but I'm finding it to be quite inconvenient; when I stabalise the line where I want it the DLM still decides to step in and "improve" things for me (ie it reduces the target SNRM)!

FYI - I usually try and settle my line with a target SNRM of around 9-10dB giving a connection rate of around 8-9Mb/s (connected at night when the line is at its worst performance). This works quite happily with very few errors (which is a problem!) until the power drops (usually at a time when the line is at peak performance ie in the morning) and the DLM decides the connection is too good for its current target/profile so it drops the target by 3dB before reconnecting - which then leads to trouble in the evening as the SNR/M is now way too low and line performance is considerably worse than at synch time.

I think it is probably easier for me to manage the connection if it is returned to the default aggressive profile; I can add a few dB to the target SNRM and the DLM will just be happy (believing it is connected with the minimum target SNRM) and leave me alone!

I'm not too concerned about latency right now so interleaving can be left at auto/default/whatever-you-want (it seems to have very little affect on the achievable throughput or the stability tbh).

Could somebody check if I'm capped and if so remove the cap and change the stability profile please? And maybe offer up any explanation about why the cap was applied at this time - when my line has been as stable as its ever been?

Thanks

PS I attached the most recent weekly stats that I could find (I've stopped collecting them lately) - from 13 May 2022. This is just to demonstrate that I know my line is far more capable than 6.6 Mb/s... the stats show 8.6Mb/s down avg at 10.9dB down avg SNR with an average of around 100 errored seconds per day.

I usually run with a little more bandwidth and a little less SNR (and a few more errors) - but when it synchs this close I just leave it and wait for the next reset.

[#2 · Accepted solution]

Hi there, I've just tested your line and we're not seeing signs of any issues or that DLM has lowered the speed profile at all. Have you tested since last week and are things looking any better now?

[#3]

Hi

I'm sure you noticed... I started another less wordy post as this one seemed to get overlooked (I know your busy, so no problems - I appreciate being able to get anything done this way rather than having to queue on a telephone all day!!). The new post was just asking for a DLM reset... which you have now done, thanks!

Up until the remote reset yesterday around 7pm (by remote I mean I didnt do it!) I could do nothing to raise my connection above around 6.6 Mb - no matter what SNR target I chose (or even left at default)... on every one of my resets the max speed always showed higher (as high as 12Mb, but mostly 10Mb) - so it must have been capped/limited/banded whatever the DLM does.

It doesn't matter too much, and you sorted it out now - I'm just a bit puzzled what prompted the timing of the banding... the line had been pretty well behaved... had a couple of power outtages when the tumble dryer burst into flames (!!)... but never reached 5 resets in 24 hours in my stats and Errored Seconds rarely exceed 200 in any 24 hours (this may have exceeded 288, I wasn't watching closely because it has been so consistent). It's very bursty, so many 15 minute periods can exceed 3 errored seconds I expect... its not something I measure though. Even so, I would have expected this to just increase the target SNR Margin (I think DLM was targeting 12dB during the period immediately before the cap/banding). If it had targetted 15dB, accounting for the approximate -2dB setting that would have remained in my modem... I would have ended with SNR margin in excess of 13dB and my line rarely has an error at that level.

I did run the tumble dryer a few times without electrical noise suppression after the capacitor spectacularly exploded (gotta get the clothes dry!)... but the errors didnt seem to increase?? I had replaced the capacitor/suppressor by the time the banding was applied so I ruled that out as the cause.

As I say... its not that important, just a curiosity!

Thanks again for looking and resetting.

[#4]

Well I must say, this all sounds very familiar to me! It seems to be the equivalent of my own experience quite recently on ADSL 2+.

https://community.plus.net/t5/Broadband/Line-Banding/m-p/1867169#M346521

After attempts to get my speed back up to what I know my line was capable of, by asking for a DLM reset, I also found my line was disconnected early in the morning hours and put back to a banded speed range, even though there was no obvious signs of the traditional reasons given for banding and nobody could explain why my line was banded. So eventually I gave up fighting 'the system'.

I concluded - advisors/people on here seemed to disagree - that this was a result of new Openreach/BT 'policy', given the current day reality of the environment in which ADSL/Broadband has to operate. But that this 'policy' - stability over speed - has not been made explicit or is necessarily understood by people dealing with broadband issues. (The old texts on banding etc. on here still in place state that it comes about from disconnections/faults etc.) The estimated ADSL speed range from Openworld for my address has steadily reduced over the years and I usually find mine is actually above their estimated maximum.

I could of course be wrong. (Surely not?)

Anyway! Having given up fighting 'the system' I learned to enjoy the new arrangement. Stable it is - every day, every time, router on/off, whatever. Unless there is an actual fault (noises on line...) my speed is ALWAYS exactly the same up and down. It is at the very top range of the top banded range for ADSL 2+. Really this is fine for me as I don't download films, play games etc.

Anyway it gave me an entirely stable and predictable ADSL speed, come what may. It was a little bit slower than the top speed I knew my phone line could cope with, but in the end I accepted it as a worthwhile compromise. This is my take on answering your curiosity - other opinions are doubtless available.

PS.
Hah! As I was preparing this post my router disconnected. Ironically, my changeover to FTTC, so it'll be different from now on.

[#5]

...Curiously, my new FTTC download/upload connection sync speeds are both (at present!) slammed up against the top end of the quoted range. Much like my previously banded ADSL 2+.

Can't be that much wrong with 'the wires'. 🤔

[#6]

Yeah... weird! And probably quite annoying! I mean, I'd sell my grandma for 14Mbps... but I suppose when it has dropped to that level for no apparent reason then it is still annoying.

Thanks for the heads-up - I'll be keeping an eye on it anyway! I'm sure I just missed something in the stats (they are held in memory and I don't bother to back them up currently - so they are lost on power-down/reset).

I often think that a good way for ISP's to convince people to migrate to Fibre would be to "nobble" their ADSL!

Sadly, my cabinet is never getting FTTC... I have FTTP available, but I could not currently imagine life without a land line - too many power outtages and no mobile signal (and I'm too old!!)... so I just have to keep making the most of what I can get out of these rusty old (long) wires!!

Knowing that all that speed (and probably stability) is just on the other side of my garden gate is excruciating!!

Your decoking routine sounds interesting though... does it work? Care to share!? Is it necessary on all sockets - I got one of them filtered/split faceplate thingamies.

Good luck with your FTTC... what are you going to do with all of that extra bandwidth!!?

[#7]

---

@Anywho wrote:  I often think that a good way for ISP's to convince people to migrate to Fibre would be to "nobble" their ADSL!

While this has crossed my mind - more in relation to BT/Openreach - I am not into conspiracies and believe it is really a case of a deteriating environment for ADSL. Most likely due to crosstalk in copper phone lines due to the increasing use of high speed fibre BB.

@Anywho wrote: 

"Your decoking routine sounds interesting though... does it work? Care to share!? Is it necessary on all sockets - I got one of them filtered/split faceplate thingamies."

Electrolysis!

Rightly or wrongly I have - after ten years or so of ADSL experience - eventualy concluded this is/was the issue behind the majority of my Broadband faults. (So maybe other BB users too?)

It was the issue I eventually discovered behind longstanding, highly intermittent, peculiar faults I had with my original BT 5A(?) master socket. There being a design vulnerability inside the consumer plate - this would have become known to BT/Openreach too, as they must sample or look at removed faulty equipment.

The newer Openreach 5C master socket was specified, I imagine, to cut costs and call outs to customer premises (IMV it also adressed the above 5A design vulnerability). However, it isn't perfect! It works until it doesn't - crackles on the line show it's time. Out comes the consumer plate and the Brasso and after a cleanup of the blackened surface, on the same wire side every time, we are back in business for another few months (hopefully).

But what causes the corrosion and why is it always on the same side of the pair? Well I recently noticed it's not just black, it's copper coloured under the black. But that copper gets cleaned off too with the Brasso leaving the original surface of the silvery metal. What gives?

I think this... the surface copper plating is what eventually corrodes. But where does the copper come from? The metal in the consumer plate is a silvery colour. Perhaps it comes from the springy (bronze?) contacts in the main/test socket in the wall unit it plugs in to? Why is it always on the same wire side? Because it is an electolytic cell? 50 Vdc in the existing telephone system. This is a variety of what happened with my original 5A master socket.

If so, then when BT close down the conventional POTS and there is no 50 Vdc on our phone lines we might find Broadband has a lot fewer intermittent faults!

@Anywho wrote:

"Good luck with your FTTC... what are you going to do with all of that extra bandwidth!!?"

Probably nothing. I think I may have made a mistake here, I was out of contract and missed an email I got, apparently offering me another contract for ADSL at a lower price than I was paying. So I looked at what was on offer on the Plusnet website and as their cheaper fibre offer seemed to be less than the advertised ADSL, so I signed up. I should have signed up for ADSL for another 18 months and gone to fibre after that as I don't really need the speed. Oh well, it's done now! 🙄

Obviously the sync speed is much higher at 40MBps, though I *think* the data throughput is only about twice as fast as ADSL. Subjectively I don't really notice any difference.

PS. I hate this Lithium BB software.