Do BT know how to run an ADSL2+ service properly?

[#1]

I have received a response from one of Plusnet's senior support agents who will be looking into the fault.
Something I should have mentioned earlier:
When Plusnet's fault team started looking into the fault, a diagram was uploaded to my support ticket showing Session Start and Session Close events on my line.
After 1 November 2009 at 19:00 the diagram showed "No connection for 10 days", i.e. as if my router had been completely disconnected or switched off. Then from 14 November 2009 at 16:30 the diagram showed "No connection for 32 days". The diagram showed no Session Close event where these long gaps appeared.
Both of these "No connection" events are impossible, because I had been using the Internet every day and I am at home all day long. The more recent Session Start and Session Close events correspond exactly with the times when I rebooted my router during the initial training period after my upgrade to the Plusnet Pro 20 Mbit/s service. So something screwy is going on.

[#2]

Today, Plusnet recommended that I try using a BT I-Plate (now known as a 'BT Broadband Accelerator'): http://www.productsandservices.bt.com/consumerProducts/displayTopic.do?topicId=25075&s_cid=btb_FURL_...
That was an interesting suggestion, bearing in mind that the I-Plate only works on a BT NTE5 type master telephone socket with the split faceplate and "BT" logo. I don't have that type of socket, as had already been explained to Plusnet under the heading "BT master socket testing" in the document I sent them (see my original post).

[#3]

Hi R1chard,
Did any of the other advice in Adam's reply to your ticket help at all?

[#4]

Quote from: R1chard

When the faceplate is off there are two cables coming out of the wall and into the Line Jack Unit (for two phone lines). So in effect, yes, it's just one cable.

so that suggests to me there is a connection coming into that box and a connection going out of that box.
As itsme says, there might be a junction box further up the line .  Is there anyway you can trace the cables either to or from that box ? Do you have overhead lines as this might allow you to trace the run of the cable ?

[#5]

I'm assuming that only one of the sockets is live? I'm concerned with the wiring as from the photo it look like the right-hand socket has 2 blue pairs going to it and the left-hand has 2 orange pairs going to it. Hope this is not the case. I would expect seeing one cable feeding each of the sockets.

[#6]

Matt Taylor,
Yes, some of the information was useful and I am thankful someone is now trying to help me.

[#7]

Quote from: itsme

I'm assuming that only one of the sockets is live? I'm concerned with the wiring as from the photo it look like the right-hand socket has 2 blue pairs going to it and the left-hand has 2 orange pairs going to it. Hope this is not the case. I would expect seeing one cable feeding each of the sockets.

Stallan & Itsme,
Yes, the inactive socket has two blue pairs going to it and the active socket has two orange pairs going to it.
The upper cable inside the unit seems to go horizontally straight back through the wall; then I think it goes outside to a grey box on the wall outside the three flats. The upper cable has two green/white banded wires that are loose (not connected to any circuit).
The lower cable appears to go down behind my wall plaster towards the skirting board.
Circuit 1 - live, used for the telephone:
1 -
2 - ORANGE/white from upper cable & ORANGE/white from lower cable
3 - WHITE/green from lower cable
4 -
5 - WHITE/orange from upper cable & WHITE/orange from lower cable
6 -
Circuit 2 - inactive when I plug the phone into it:
1 -
2 - BLUE/white from upper cable & BLUE/white from lower cable
3 - GREEN/white from lower cable
4 -
5 - WHITE/blue from upper cable & WHITE/blue from lower cable
6 -
Colours in upper case are the larger bands of colour on each wire. Connections 1, 4 and 6 are unused on both circuits.
The lower cable also has what appears to be two strands of cotton thread coming out of it - one white, one brown. Bizarre.

[#8]

The cotton thread is there to help strip the outer sleeving. It's standard on telephone cables not so important on 6 core cables but at the exchange it's not unusual to strip several feet of outer sleeving and the cotton thread is used to do this .
I now understand how the socket is wired. The cable going outside is bringing the lines in and the other is going to the extensions, maybe via a junction box along the way. I would still recommend removing the wires going to terminal 3. 
Edit: The brown thread is probably a batch maker, this will be a different colour for different batches.

[#9]

itsme, thank you very much for your advice, as well as educating me about the cotton thread.

[#10]

Thanks itsme, I have also learnt what the cotton thread is for 
If circuit 2 is inactive, and I presume is feeding the other 3 unused extension sockets in the house, could you just pull the wires off the back of the master socket? This will allow you to ensure there is NO interference of any kind on that MASTER SOCKET.. As you have a record of the wiring you could always put them back at some point should you wish.
You mentioned a grey box on the outside. do you have access to this? I was think of you checking the wiring and trying to match up the wire colouring.

[#11]

Stallan, that's a very good point regarding circuit 2. However, I've been testing with the router connected to the master socket today and it is impossible for me to get any wireless connection to the router from my PC, so after using a borrowed laptop to test with I've had to connect the router back to the extension socket I normally use.
Connecting to the master socket made no difference anyway, other than a small rise in SNR Margin. The sync rate was still only around 6653 kbit/s, despite the router stats reporting an Attainable Rate of around 12000 kbit/s.
I probably can't reach the grey box on the outside without a ladder (which I don't have) and I wouldn't want to mess with BT's property anyway. I presume it is something to do with routing the incoming lines into each of the three flats in the building.
The ring wire has been disconnected now. The router reported a permanent rise in SNR Margin each time more ring wire was disconnected. SNR Margin with the router back on its normal extension socket is now at 25dB (it was between 17 and 19dB previously) and Loop Attenuation has dropped slightly from 15dB to 14dB.

[#12]

I'm assuming that your sync rate stayed the same with the SNR margin increasing to 25dB. If so the good news is that the ring wire was introducing noise of the A/B pair. The bad news if the sync rate remained the same is that your are on a banded profile and this can take several weeks for the DLM to remove it. I have been on a banded profile for the last 3 weeks at 2268kbps  You could request PN to try and remove it.

[#13]

Out of interest, what time of the day are you performing this work ? It's a known fact that 09:00-15:00 is the optimal time to achieve a higher sync rate, while later in the day/evening seems to incurr a higher Noise Margin and lower sync rate.
To me this all points to the wiring of the master socket and socket extensions. I understand the point of demarcation on the master socket, but to prove a point I'd but an NTE5 socket (ebay are awash with them) and replace (on a temp basis) the master socket (omitting the extension sockets) and then see what the connection data says. You can always put everything back once the exercise is over. Or pay BT (approx £150 odd) to come and install an NTE5. At least that way you know the wiring coming into the house is correct.

[#14]

Quote

while later in the day/evening seems to incurr a higher Noise Margin and lower sync rate.

Think you are confusing sync speed and SNR margin with a fixed speed and adaptive speed services.
With fixed speed the sync will remain the same and the SNR margin will change throughout the day depending on conditions. Normally the margin is higher in the day than night.
With adaptive speed ( up to 8m and 24m)  the sync speed is the highest that can be attained with the target SNR margin.
So you are correct that a lower sync rate will happen in the evening/night but the SNR margin will still be the same straight after a resync. If no resync is performed the next day the SNR margin will increase in daylight hours returning to the target value at the same time as the last resync
As r1chard SNR margin increased after the removal of the ring wire and I assumed that the sync speed stayed the same this is an indication that the line is on a banded profile and the ring wire was introducing noise on to the A/B pair.

[#15]

Quote from: itsme

I'm assuming that your sync rate stayed the same with the SNR margin increasing to 25dB. If so the good news is that the ring wire was introducing noise of the A/B pair. The bad news if the sync rate remained the same is that your are on a banded profile and this can take several weeks for the DLM to remove it.

The sync rate did stay the same.
The router had an unintended reboot at around 7 AM the next morning when my flatmate pulled out the plug; he thought it was his phone charger despite the prominent "DO NOT UNPLUG" label attached to it. After the re-sync at 7 AM on the master socket, the sync rate was still just 6652 kbit/s.
Stallan, I don't want to even temporarily replace the master socket because Plusnet want to book a visit by a BT SFI (specialist faults investigator). I know that my line is capable of achieving up to 19 Mbit/s because I witnessed that happening during file downloads in August *2008*, despite the fact I was on an 'up to' 8 Mbit/s service. Speedtest.net confirmed the high speeds being achieved, but being a highly sceptical person I emailed Speedtest.net to try and make sense of it.
Speedtest.net responded: "I'm not sure why the results would be higher than being advertised by your ISP - it's possible they simply allow it to burst to a higher speed if they have the capacity available, as we have seen in some instances."
After some investigating, I came to the conclusion that at the time I witnessed the 19 Mbit/s speeds BT were probably testing the exchange for their ADSL2+ service. The timing seems right because Plusnet wrote: "our first staff member was migrated [to ADSL2+] on 1st October 2008".
I do not trust BT's DLM (Dynamic Line Management) system works properly, hence the title of this thread. I've been fiddling with computers since 1982 and working in the IT profession since 1997, but I've lost faith in it because there are so many fools involved in providing the technology that could be giving the UK the lift it needs.