Disconnections and slow speed

[#1]

@ejs, please don't try quoting me out of context. Different modem/routers can and do report attenuation differently. I did clearly say so, and that's because they use slightly different algorithms for the calculation.
Your presumption that just because a given modem/router connects at a higher speed on a particular occasion means higher tones are in use is not necessarily the case. The higher speed may be entirely due to more bits being usable on a selection of tones (generally due to less noise being present) at the time of sync.
You'll also note then when tones become unusable due to noise, and most modem/routers mark them as such, they don't recover them later if noise on those tones reduces, but the attenuation does not change. The attenuation curve is generally not significantly affected by typical changes in noise levels.
Some modem/routers do have sophisticated firmware that does recover tones that later become usable, but the 582n for example is not one of those.
Changes in noise levels would have to be huge, and have a significant effect on tone availability for there to be any real noticeable affect on reported attenuation typically in ADSL2+ mode, and whether you observe that would depend on the resolution of the modem/router reporting eg.some only report in 1dB steps. For plain ADSL(1), likely to be the case here based on the reported figures, attenuation was typically calculated at 300kHz. Tone 70 is rarely affected by any significant amount of noise, except perhaps on faulty lines.

[#2]

@Anotherone,
For reference, TG582n is not in use here - see  http://community.plus.net/forum/index.php/topic,128606.msg1119286.html#msg1119286
Whilst I accept that some routers might exhibit the described characteristic I'm yet to be convinced that it is the explanation for the marked variation of line attenuation in this situation.
Kevin

[#3]

Barry,
From what I can see on the line is that it's erroring very frequently. I've attached some statistics which make quite interesting reading. It's really difficult for us to tell though whether it's being caused by internal or external influences, that's the problem for me 😕

[#4]

Barry,
I know you have an aversion to wifi, but are you able to run with the router in the master socket and the laptop running on wifi for a few days?  I understand this might not be ideal but it is essential to discern the source of the problem.
Do you feel able to run some diagnostic software?  Look up RouterStats.  This will help identify if you have a REIN issue.
Kevin

[#5]

Kevin, Barry, it needs to be in the test socket to have that internal wiring disconnected.
Kevin, I realise he's using Netgears, mentioning the 582n was just giving an example.
Difficult to say exactly what may be happening as (IIRC) the previuous swapping around and disconnection was in a somewhat randomish manner, as I posted before it needs to be more methodical and consistent for a period to see whether there is still instability.
When swapping between different models of modem/router (even the same brand) I would not necessarily expect attenuation figures to be exactly the same, unless they happen to have the same firmware. The important thing is that over a period of time with any given modem/router there is no significant change.
The place to start here is filter, modem/router and phone plugged into the test socket. Also check the phone line for noise regularly.

[#6]

Barry, sorry forgot to respond to your query about buying the CW1308 cable on-line.
http://www.tandyonline.co.uk/catalogsearch/result/?q=CW1308+Twisted+Pair and P&P is only £1. They have quite a reasonable reputation, although their website cab be a bit random to go around!
Watch what other sites may charge for P&P it can make the overall cost significantly more

[#7]

Thanks everyone,
Just read through the latest posts. You're right that I have an aversion to WiFi, and don't have the equipment to test anyway. I could plug in the master socket for a few days if necessary to track down any fault, but would rather wait until after the weekend as I have auctions ending and possibly labels to print. After that I can do it but would need to keep a phone plugged in, even though it's hardly ever used, the law says that someone will want to get hold of me while its unplugged.
I'll buy some cable (thanks for the link), though it will be a pain to change due to the old cable going through the wall (breeze block) above the skirting and being  plastered over to fill said hole. The old cable is about 10 feet long, don't know if that's long enough to cause a problem or too short to cause a problem. Here are my latest statistics (not formatted, sorry:)
ADSL Link Downstream Upstream
Connection Speed 5466 kbps 950 kbps
Line Attenuation 42.5 db 21.2 db
Noise Margin 9.7 db 6.7 db
Thanks chaps.

[#8]

Other thing to be careful of us using telephone extension leads to extend master socket to plug router into. These leads use flat cable and not twisted pair and this will cause problems
This has caught out a lot of people

[#9]

Barry,
Does the laptop not have wifi capability?
The short length of untwisted pair multi core old style wire can have an impact; is it responsible for all of the difference you see between the master socket and your final extension?  I cannot say for sure.  Eliminating it makes your installation as good as it can be short of working off the master socket.
An alternative approach is either WiFi (which you don't like / have not got) or using a pair of Ethernet over power adaptors. You can locate the router at the master socket and plug into the Ethernet over power adaptor.  You can then put the other adaptor into any power socket and plug your PC into that.
HTH
Kevin

[#10]

Hello Kevin,
Tomorrow or Tuesday I'll move my laptop and router to the master socket. I'll plug a filter into the test socket (behind the cover) and plug my router and a phone into that. I'll post back once that's done and then I hope that plusnet will be able to determine where the problem is. I've ordered some cable and I'll replace the old cable while the"'master socket test thing" is going on. Thanks again for your help,
Barry.

[#11]

Hello chaps, thanks for sticking with this.
As of about 3:10PM today I'm plugged into the master socket test plug via a filter. A phone is also plugged into the filter. I'll leave it like this until you say otherwise and will leave the router on.
Stats taken before the move:
ADSL Link Downstream Upstream
Connection Speed 6266 kbps 935 kbps
Line Attenuation 42.5 db 21.2 db
Noise Margin 6.3 db 5.7 db
And after:
ADSL Link Downstream Upstream
Connection Speed 6685 kbps 1119 kbps
Line Attenuation 39.0 db 20.9 db
Noise Margin 7.1 db 6.0 db
I also had a disconnection yesterday requiring a router reboot before I could get on-line again. This is the first for a while.
Anyway,  I hope this helps Plusnet track down the source of any problems with the line.
Thanks again,
Barry.

[#12]

Barry,
I've been rather busy over recent days.  There are a number of things in regard of these figures which I'll try to comment on soon.
Kevin

[#13]

From what I can see, the sync rate is still lower than I would expect for the line length. That said your SNR figures are really encouraging and your upload sync rate is really good too.
I'd like to se4e if the line remain in place for a week or so to see if the sync rate fluctuates at all.

[#14]

Thanks a lot Kevin and Chris.

[#15]

Hi Barry,
Interesting figures!
1. Attenuation - all round these figures are markedly better  - the DS attenuation change is equivalent to having taken about 0.25km out of the length of the phone line.  Therefore the extension wiring is having a marked impact on the circuit performance.  I know you have plans to replace the first segment with more recent (twisted pair) specification cable, which should help.
2. You are seeing a higher synch at a higher SNRM - the higher speed could be down to just the reduction is the line's attenuation, however the higher SNRM implies either the line has been banded (I'm not convinced) or that there was more noise present when the line synch'd than there was at the time you took the stats - I think you need to consider monitoring this more closely with routerstats - see http://www.vwlowen.co.uk/internet/files.htm - and ask for help were needed.

Quote from: Chris

From what I can see, the sync rate is still lower than I would expect for the line length

3. Bless Chris for his powerful understatement    The expectation for this line ought to be in the region of 9Mbps...  Can you please post the full router stats here?
Your line is doing much better than it was, but not as good as may be it should do.  To resolve this is going to require continued connection at the master socket; I know this is a bit inconvenient, however I do not think that the cause of the issue is overly clear right at the moment.
@Chris - is this line banded please?  What do the line error stats look like?  What is the time resolution of the SNRM as seen by the DSLAM?  Might they be adequate to point to REIN and thereby avoid the chore of running local monitoring?

Cheers,
Kevin