Why is a "degraded" line not a faulty line?

[#1]

Thank you for your response on your ticket and on here too. I'm still currently awaiting the engineers notes so I can gain some understanding of what the next steps will be. I will be in touch once I have further information for you. Please accept my apologies for the inconvenience caused.

[#2]

Quote from: JohnJ

As he was getting a 5Mbps signal at the cabinet, he thinks that the most likely cause is between the cabinet and the pole and ...

How far away from your house is the cabinet?
What attenuation values did his equipment show between readings at the cabinet and your master socket? And at the DP?

[#3]

The distance to the cabinet is 325 metres. (The engineer accidentally wired my new D side pair to my old D side pair at the cabinet, which showed as a fault.  When he tested it, the distance to break was reported as 650 metres, in other words, twice the distance from my master socket to the cabinet.)
I know that the attenuation reading at my master socket was about 52dB, as it was the same as my router, and I was present to ask.
However, I don't know the attenuation readings anywhere else, but he may have included that in the notes to Plusnet.

[#4]

Following on from receiving the engineer notes, Plusnet have now carried out further testing on the line that resulted in a couple of disconnects at 14:06 and 14:37
I noticed that happening and took the opportunity to grab some statistics while at a few stages.

			14:15		14:35		14:37		15:07		17:11
					
DOWNSTREAM (Rx)					
Noise Margin:		9.3   dB	9.1   dB	9.3   dB	9.2   dB	8.5   dB
Connection Rate:	3833  Kbps	3833  Kbps	3769  Kbps	3769  Kbps	3769  Kbps
Line Attenuation:	52.5  dB	52.5  dB	53.0  dB	53.0  dB	53.0  dB
Power:			0.0   dBm	0.0   dBm	0.0   dBm	0.0   dBm	0.0   dBm
Max Rate:		4800  Kbps	4780  Kbps	4256  Kbps	4692  Kbps	4488  Kbps
					
SuperFrames:		29950		100539		3479		108566		573411
SF (CRC) Errors:	0		0		0		0		2326
Reed Solomon:		1946766		6535076		226174		7056844		37271748
RS Corrected:		1357		1604		32		937		138817
RS Un-Corrected:	0		0		0		0		34233
HEC:			0		0		0		0		2311
Errored Seconds:	161741		161741		161751		161751		161799
Severe ES:		153311		153311		153320		153320		153358
						
Interleave Depth:	32		32		32		32		32
Bitswaps:		62		207		5		219		1200
					
UPSTREAM (Tx)					
Noise Margin:		6.0   dB	5.6   dB	5.9   dB	6.1   dB	5.8   dB
Connection Rate:	883   Kbps	883   Kbps	888   Kbps	888   Kbps	888   Kbps
Line Attenuation:	33.3  dB	33.3  dB	33.3  dB	33.3  dB	33.3  dB
Power:			12.9  dBm	12.9  dBm	12.9  dBm	12.9  dBm	12.9  dBm
Max Rate:		888   Kbps	888   Kbps	888   Kbps	888   Kbps	864   Kbps
					
SuperFrames:		26948		90542		3056		97390		514173
SF (CRC) Errors:	0		0		0		0		0
Reed Solomon:		484524		1627919		55010		1753035		665195
RS Corrected:		0		9		0		102		271
RS Un-Corrected:	0		0		0		0		0
HEC:			0		0		0		0		0
Errored Seconds:	198		198		198		198		198
Severe ES:		0		0		0		0		0
					
Interleave Depth:	8		8		8		8		8
Bitswaps:		6		28		0		32		197
					
TOTALS					
Total Uptimes (From SF counts):
WAN:   			00:08:29	00:28:29	00:00:59	00:30:45	02:42:27

Below are graphs for  Bits/Tone graphs for the same time slots as above router stats, also the Noise and sync speed graphs for today.

[#5]

Taking into account the notes on the Ticket that you've sent me, it certainly seems that the engineer's testing appears to have narrowed down the major cause of your significantly reduced performance, to between your pole and the cabinet, whilst the copper tests (within their limited results) don't highlight an obvious cause.
Overall, taking all the recent data and information into account, your line attenuation has increased by 4dB since the 'E' side pair swap, there's clearly a major cable issue (possibly REIN) between the pole and cab, 'D' side swaps have had little effect, and you still get noticeable reductions in Noise Margin of which the damaged untwisted section of drop-wire will be a contributing factor and this will ultimately limit the maximum speed because of the reduced margin.
So what can I conclude from all this -
Well you could be put on a much better pair on the 'E' side, if your original pair weren't actually faulty, they may be the best - 4dB less attenuation!
The damaged untwisted section of drop-wire should be replaced - an easy job, can then be eliminated as any possible contributory factor.
Further investigation of the 'D' side near the electrical cable works is needed. It would seem maybe a REIN engineer may be needed if any other problem cannot be found.
There is a big difference between the bit allocation just before the 1406 test and resync and bit allocation after that. This suggests that there was more noise present when the previous ~2930 sync was achieved, but otherwise the Noise seems fairly constant, so not so much Random as Repetitive and probably not Impulse so much as Continuous Noise!

[#6]

Although REIN is still only a possibility, I just took a walk between me and the cabinet with a radio tuned to 612kHz. That included walking around the nearby electricity sub-station and along the route of the cable works 18months ago.
The cable route is mostly underground, but there are a pair of poles through the back of a small group of houses. Between those poles I can see a thick cable (possibly 20-30mm OD?), so that may include my wires.
Anyway, I didn't notice any changes to the radio around the cabinet or electricity sub-station.
However, the only spot where I noticed interference was when standing close to one of the poles that carried the thick cable. There was loud buzz that pulsed 2-3 times per second. I was initially listening at 612kHz so gradually adjusted the frequency. The noise got louder and clearer over 700kHz, but I could still hear it when I went up as far as 800kHz. The sound changed, but the existence of the pulsing was quite consistent.
There is a residential garage about 3 or 4 metres way in one direction. In the other direction, about 1.5 metres, is a lamp post and, about 1 metre further, the back wall of a newish restaurant.
The restaurant was converted from a pub and was opened about 14 months ago, so would have been in its finishing phase of construction about the time my line deteriorated. As a customer of the restaurant I know that they installed lots of new air conditioning and karaoke equipment in the area adjacent to where I was detecting the interference.

[#7]

You really need to try the lower frequencies (LW) where the bits/tone are much reduced.

[#8]

I managed to borrow a LW radio this morning, so repeated my walking route.
I tuned the radio off station at about 180kHz.
Around the same place as heard the interference on MW, the background buzz got loud. It was different from what I heard at the higher frequencies, more buzzing than pulsing.
When I was at the spot with the buzzing at it loudest, I tuned the radio through the LW range. The buzzing was by far the loudest at the top end of the range, but was audible to some extent through the complete range.

[#9]

The difficulty when 'tracking' this sort of thing is listening for the right noise. Some pulsing sounds aren't uncommon and not necessarily a problem. Have you noticed whether you can hear either of the noises when holding the radio close to the cable going up your pole, or near the drop-wire say just before it goes into the house?
The "buzzing" noise that was loudest at the top of the LW band , was it audible and loud down at 522kHz at the bottom of the MW band?
Do either of the noises stop/reduce when you see a reduction in the error rates you mentioned in a PM? The fact that those error rates reduce at the time slots you mentioned is quite significant, especially if it's a pattern that re-occurs.
The other trick is to rotate the radio, as if it were sat on a table and you twisted it round, so the speaker was facing you initially say and then the back of the radio was facing you.. You might be able to get a 'null' direction and see where the line along the length of the radio pointing.

[#10]

Engineer Number 3 has visited.
This engineer said that he got the same errors as previous engineers, at Master socket. However this time when he tested at the junction on the gable of my house where the cable came from the pole and joined the old cable, he said that there were no errors back to the cabinet/exchange. So he changed the part of the drop cable that was old non twisted pair (I had also pre-drilled a hole through my wall,in a new location, to make it easier).
He then tested the connection and got a speed of approx 4800kbps.
He then connected my router and this is the current ADSL Status:

ADSL Status
Parameters
DSP Firmware Version 		A2pB025f.d22k
DMT Status			No Defect
Operational Mode		ADSL2
Upstream			832
Downstream			5003
SNR Margin(Upstream)		6.5
SNR Margin(Downstream)		6.3
Line Attenuation(Upstream)	32.8
Line Attenuation(Downstream)	52.5

Chris Pettitt has increased my profile to match.
@ Anotherone - this may be where you say "I told you so"?

[#11]

Well I have to say the line is looking much better now. As we discussed, I'll keep an eye on things until Friday and we'll see how we go

[#12]

I wouldn't be so smug John, I'm not infallible!
I can only imagine that as engineer no:2 was full of cold as I recall you mentioning in a PM he may have been testing from the top of the pole to the exchange with the line back to you still connected.
Well I'm so glad that the significant part of your problem looks as though its been overcome, and that the cable entry is now also in a more appropriate location.
The only thing that you mentioned in your PM, not mentioned here, is the drop in SNRM and increase in CRC errors when the phone rings. this is generally symptomatic of a poor joint. Bearing in mind engineer no:2's comments about the top of the pole , this seems a likely probability.
In view of all the current information I'd suggest one more engineer visit was appropriate with 2 specific tasks -
1) To change your 'E' side pair back to your original pair if no better pair is available, the pair you have been put on has 4dB more attenuation and will limit the maximum sync speed.
2) To check and remake  the joints at the top of your pole - the only remaining ones that won't have been remade. (The exchange and Cab will be done by the 'E' side swap, your drop-wire joint at the gable and Master socket have just been done).
Anyway the proof of the current pudding is in the eating - any chance of seeing the Bits/Tone and SNRM plots?

[#13]

Here is the current Bits/Tone graph, the SNRM graph and the downstream sync speed graph.
Some observations:
05:00 - SNRM dropped to about 6dB for the first time in a long time.
10:00 - 11:00 engineer working on line
11:00 - 14:40 a few disconnects possibly due to DLM adjustments, PN testing, and interleave switching on and off on upstream and downstream.
At about 13:30 I remembered to switch on my cordless phone. It had been plugged into the ADSL filter all of the time, but the power wasn't plugged in   I saw the SNRM immediately drop by about 1dB so I added the second ADSL filter onto the phone line and it climbed back up again.

[#14]

That Bits/tone is looking vastly better, less interference from R4 as well, be interesting to see what it's like in daylight tomorrow (12 -1 ish  ).
If your 'E' side was sorted out and 4dB less attenuation, I'd expect to see a few more of the higher tones back on the plot.

[#15]

R4?