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

[#1]

It would be interesting to see the stats now after these 5 days and see how they compare to the previous 3 days.

[#2]

Here are the router stats for now.

Noise Margin:     5.1   dB
Connection Rate:  5381  Kbps
Line Attenuation: 50.5  dB
Power:            0.0   dBm
Max Rate:         6052  Kbps
 
SuperFrames:      29017701
SF (CRC) Errors:  2801
Reed Solomon:     1886150554
RS Corrected:     1333823
RS Un-Corrected:  36901
HEC:              2421
Errored Seconds:  1377
Severe ES:        16
Interleave Depth: 32
Bitswaps:         84416

Noise Margin:     4.9   dB
Connection Rate:  964   Kbps
Line Attenuation: 27.8  dB
Power:            12.9  dBm
Max Rate:         940   Kbps
 
SuperFrames:      860916
SF (CRC) Errors:  1574
Reed Solomon:     2096876
RS Corrected:     13037
RS Un-Corrected:  0
HEC:              1184
Errored Seconds:  1163
Severe ES:        0
Interleave Depth: 4
Bitswaps:         8657

Total Uptimes (From SF counts):
 WAN:   5 days, 17:01:40

According to my router the connection has been up for "5 Day(s) 10:09:10" which is much more accurate than the one "calculated" in RouterStats.
Below are graphs for Noise Margin over the last 24 hours and the current Bits/Tone

[#3]

Well that evening bits/tone shows the expected difference that we have been seeing on other after dark plots compared to the daytime one.
Looking at the stats, whilst the DS CRC errors are a big improvement over the previous ones, it's the continuing presence of those SES errors that is of concern. Looking at the Upstream, in my book that is a high level of CRC errors.
Overall, to me it suggests there is still some noise present and this is likely to be down to the quality of the D-side cable (based on the other tests) and this is supported by the continuing absence of most of the tones above 185.

[#4]

Previously the CRC errors were constantly rising with a slight change in gradient between day and night. Now that the background errors are being handled differently and appear to be showing up in RS uncorrected errors, it is easier to see the day/night variations of CRC errors. I made up an animation showing the last 12 days of graphs for Rx CRC (per 15 minutes). I think it is relatively easy to identify the exact times when SES have occurred on these CRC graphs. Those 22 SES will have happened in the last 7 days (starting on the first Monday in the animation). I will try to attach the animated png below that should display OK in Firefox (the forum won't allow my gif version, with a delay between days  ).
There is a garage business nearby and they are connected via the same pole as me. They start at 8am Monday to Friday, 10am on Saturday and not open on Sundays. That pattern of hours appears to have some correlation with the SES events. So, that may be an avenue for me to explore next.
You will be correct, as usual, about the suspect d-side cable and that may be what's making my line sensitive to the interference causing the CRC errors.
I've attached a Bits/Tone from a few moments ago, and it seems similar to the evening one. I have kept looking at these graphs more than once a day and the overall pattern that I feel is that the number of tones being used reduced visibly over the first few hours then the rate of change slowed until it doesn't seem to vary much.

[#5]

Yes, the Bits/Tone do seem to have been marginally worse since the daytime one on the 22nd.
The stats had implied the possibility of the noise/errors coming in bursts and your animated png shows that very well.
No doubt a good "D" side and nice clean connections to the "D" side in the Cab and Pole to Drop-wire would undoubtedly reduce the susceptibility to this as well as any MW broadcast pickup.

[#6]

Here's a still pdf version of the animated CRC graphs.
The first day (18th January) was when the Open reach was doing stuff, so RouterStats was sometimes collecting and graphing frozen data from the router. This did not reflect reality, for example, the horizontal line in the middle of the first image.

[#7]

All noted.

[#8]

I have been looking over the stats since the last reconnection to gain an overall impression of the condition of my connection.
Visually comparing several days of SNRM graphs, there has been a gradual reduction in the length of time that the SNRM is at or above 6dB, and yesterday SNRM did not return up to 6dB at all.
So, I've plotted the stats in a single graph since the last reconnection on 19/01/13 and added a 24 hour moving average trendline, as attached.
This appears to show on downward trend in my connection.

[#9]

Hmm, yes, I don't especially like that. It indicates a general trend in increased noise pickup. Perhaps slightly more of a concern (in my book) is the fact that it's dipping down to 4dB and just under during the night and that there is a greater swing between the peaks and troughs.
The sooner your D-side is sorted and eliminated from any issues the better!
If this trend continues I'd expect to start seeing increases in errors, and at worst a dropped connection 

[#10]

Current Stats:

DOWNSTREAM (Rx)
Noise Margin:     4.7   dB
Connection Rate:  5381  Kbps
Line Attenuation: 50.5  dB
Power:            0.0   dBm
Max Rate:         5936  Kbps
 
SuperFrames:      56853649
SF (CRC) Errors:  5992
Reed Solomon:     3695487190
RS Corrected:     3400107
RS Un-Corrected:  90374
HEC:              5274
Errored Seconds:  2845
Severe ES:        32
Interleave Depth: 32
Bitswaps:         168975
UPSTREAM (Tx)
Noise Margin:     5.8   dB
Connection Rate:  964   Kbps
Line Attenuation: 27.8  dB
Power:            12.9  dBm
Max Rate:         968   Kbps
 
SuperFrames:      649186
SF (CRC) Errors:  3431
Reed Solomon:     4223448
RS Corrected:     34937
RS Un-Corrected:  0
HEC:              2450
Errored Seconds:  2560
Severe ES:        0
Interleave Depth: 4
Bitswaps:         19381

TOTALS
Total Uptimes (From SF counts):
 WAN:   11 days, 04:29:35
 LAN:   0 days, 00:00:00
 
 
CRC:            5992            3431
ES:             2845            2560
SES:            32              0
UAS:            0               0
LOS:            0               0
LOF:            0               0

[#11]

Current Stats

DOWNSTREAM (Rx)
Noise Margin:     5.3   dB
Connection Rate:  5381  Kbps
Line Attenuation: 50.5  dB
Power:            0.0   dBm
Max Rate:         6080  Kbps
 
SuperFrames:      65595846
SF (CRC) Errors:  6402
Reed Solomon:     4263730014
RS Corrected:     3676508
RS Un-Corrected:  94432
HEC:              5621
Errored Seconds:  3137
Severe ES:        32
Interleave Depth: 32
Bitswaps:         195397
UPSTREAM (Tx)
Noise Margin:     5.6   dB
Connection Rate:  964   Kbps
Line Attenuation: 27.8  dB
Power:            12.9  dBm
Max Rate:         972   Kbps
 
SuperFrames:      91807
SF (CRC) Errors:  3728
Reed Solomon:     529843
RS Corrected:     36884
RS Un-Corrected:  0
HEC:              2662
Errored Seconds:  2791
Severe ES:        0
Interleave Depth: 4
Bitswaps:         22639

TOTALS
Total Uptimes (From SF counts):
 WAN:   12 days, 21:51:59
 LAN:   0 days, 00:00:00
 
 
CRC:            6403            3728
ES:             3138            2791
SES:            32              0
UAS:            0               0
LOS:            0               0
LOF:            0               0

[#12]

That afternoon SNRM dipping, seems to fit the recent pattern. But getting the D-side sorted should see you OK, as the error levels are certainly an improvement on what they were.

[#13]

The maximum SNRM has been below 6dB for a week now, so I planned a disconnect/reconnect this morning.
I saved the router stats before disconnecting (downstream sync speed was 5381kbps), left it disconnected for 10 minutes and reconnected. The exchange/router automatically selected ADSL2 and connected at 5377kbps (only dropped 4kbps).
After 20 minutes the SNRM was 6.3dB so I thought I would try another disconnect/reconnect, but I didn't wait 10 minutes between before reconnecting. I've found the quickest method for doing this, and possibly clean as well, is to open one of the router configuration pages and click Apply, so the router negotiates the disconnect/reconnect. This time the downstream sync speed was 5373kbps (only dropped another 4kbps).
Below are the stats before starting, 20 minutes after the second reconnect and 9 hours after the second reconnect.
                                                                                                                                                                                                                                                                                                                                                                         

	ADSL2+ at 09:24 Connected 15 d 9 h 13m	ADSL2 at 10:21 Connected 0d 0h 20m	ADSL2 at 18:38 Connected 0d 9h 0m
DOWNSTREAM (Rx)
Noise Margin:	5.7 dB	6.3 dB	4.4 dB
Connection Rate:	5381 Kbps	5373 Kbps	5373 Kbps
Line Attenuation:	50.5 dB	47.5 dB	47.5 dB
Power:	0.0 dBm	0.0 dBm	0.0 dBm
Max Rate:	6116 Kbps	6492 Kbps	5992 Kbps
SuperFrames:	80305789	70692	1905983
SF (CRC) Errors:	8211	5	189
Reed Solomon:	924908966	4595022	123888944
RS Corrected:	5125543	4290	82468
RS Un-Corrected:	112063	59	2300
HEC:	7130	3	160
Errored Seconds:	3860	13	119
Severe ES:	50	0	0
Interleave Depth:	32	32	32
Bitswaps:	240262	154	5675
UPSTREAM (Tx)
Noise Margin:	5.6 dB	6.0 dB	5.8 dB
Connection Rate:	964 Kbps	1008 Kbps	1008 Kbps
Line Attenuation:	27.8 dB	27.7 dB	27.7 dB
Power:	12.9 dBm	12.9 dBm	12.9 dBm
Max Rate:	968 Kbps	1008 Kbps	992 Kbps
SuperFrames:	232985	64236	789209
SF (CRC) Errors:	4584	50	355
Reed Solomon:	1617823	770475	3610876
RS Corrected:	43614	281	2187
RS Un-Corrected:	0	0	0
HEC:	3212	46	272
Errored Seconds:	3462	35	264
Severe ES:	0	0	0
Interleave Depth:	4	4	4
Bitswaps:	28489	15	367

[#14]

A little early to tell, so not very scientific, but DS looks marginally better, but US marginally worse

[#15]

Here are the stats at 14:26 today (1d 5h 54m uptime)

DOWNSTREAM (Rx)
Noise Margin:     5.4   dB
Connection Rate:  5373  Kbps
Line Attenuation: 47.5  dB
Power:            0.0   dBm
Max Rate:         6092  Kbps
 
SuperFrames:      6332591
SF (CRC) Errors:  607
Reed Solomon:     411618462
RS Corrected:     233104
RS Un-Corrected:  10131
HEC:              535
Errored Seconds:  312
Severe ES:        3
Interleave Depth: 32
Bitswaps:         19307
UPSTREAM (Tx)
Noise Margin:     5.9   dB
Connection Rate:  1008  Kbps
Line Attenuation: 27.7  dB
Power:            12.9  dBm
Max Rate:         1008  Kbps
 
SuperFrames:      94202
SF (CRC) Errors:  665
Reed Solomon:     358153
RS Corrected:     4480
RS Un-Corrected:  0
HEC:              514
Errored Seconds:  520
Severe ES:        0
Interleave Depth: 4
Bitswaps:         1373
TOTALS
Total Uptimes (From SF counts):
 WAN:   1 days, 05:54:14
 LAN:   0 days, 00:00:00

And here are the stats from the Plusnet end at 14:15:

Sync Status:	Circuit In Sync
General Information
NTE Status:	 	NTE Power Status:	PowerOn	Bypass Status:	 
 
 			Upstream DSL Link Information	Downstream DSL Link Information
Loop Loss:		27.6				47.5
SNR Margin:		6.0				5.3
Errored Seconds:4					1
HEC Errors:		5	 
Cell Count:		60102				463856
Speed:			1007				5373
 
Maximum Stable Rate (KBPS):		2272	Fault Threshold Rate (KBPS):			1817
Mean Time Between Retrains (Seconds):	86400	Mean Time Between Errors Upstream (Seconds):	232
Indicative Line Quality:		A	Mean Time Between Errors Downstream (Seconds):	232
WBC 160K - 24M Medium delay (INP 1) 6dB Downstream, UC Low delay (INP 0) 6dB Upstream (ADSL2+)