Technicolor TG582 firmware upgrade

[#1]

The SNRM has been dead flat at 11dB since the event at approximately 06:20.  I'll check the extension sockets but I know the ring wires are not connected.  I checked that when I first got broadband several years ago.
The house wiring was fully checked last Tuesday when the consumer unit was changed.  Every socket in the house was megger tested and passed easily.  If I get the opportunity I'll do a graceful disconnect and use the test socket for a while. Meanwhile I'll continue to monitor the graph and quiet line test.
The router transformer is plugged into an extension lead.  I would have thought the power supply would control any spikes but do you think that might be relevant?
By the way, I live just outside Bedford.

[#2]

Graph showing intermittent data due to computer sleep mode but indicates flat 11dB SNRM all morning/early afternoon.
Is it possible the glitch this morning was anywhere between my router and the exchange or even in the exchange itself?  If so, as I'm so far from the exchange it could be an interesting problem to track down!
When two line faults were fixed a few weeks ago leaving me with hum on the line it transpired that the equipment in the exchange was causing it.  I don't know how fixing a line fault could suddenly cause the exchange equipment to produce hum (like mains hum)  but the Openreach guys changed something in the exchange and the quiet line test has been good ever since.

[#3]

You are obviously clued up on the bell/ring wire issue, which should be disconnected at all ends, especially the master socket. But the colours of the wires are also important as it indicats the standard of the wiring. When you decide to do a check at the test socket, have a look at the connections to the faceplate as well.
The daytime SNRM graph is looking very good, other than that small spike at 0836 ish. Don't want to see too many of them (something being switched on or off at a guess).
Whilst it's good to know that your electrical installation is fine, that will have no impact on this sort of issue which is usually down to poor/faulty switch mode power supply units (SMPS/U) used in all sorts of modern electronic/electrical equipment. Discharge lighting can also cause interference, sodium street lights, fluorescent lights and so forth, often when the tubes aren't up to spec. or are on the blink.
The Modem/Router PSU will have little effect on suppressing these type of problems, and the use of mains extension leads shouldn't make any difference. Sometimes if power cables are running parallel to the phone/router line within about 6", that can sometimes add to the problem.
Because your modem/router was obviously constantly on, as was you computer and monitor at 0620, they can clearly be eliminated from the equation.
Whilst whatever is generating the noise that went off at 0620 "could" be anywhere between you and the exchange, it is far more likely to between you and the Cab (how far are you from that btw?) or more probably very local and/or between you and the DP.
The first thing in these situations is to eliminate anything of your own or immediately adjacent - ie. a street light right outside for example - taking account of when this happened - I mentioned central heating - a controller might be responsible, but bearing in mind the weather, that might just be the hot-water side of things. But if you know the system wasn't on or didn't go off then, that could be eliminated.
Dusk to Dawn lighting, or IR lit security camera PSUs are the hot favourite IMHO (where the IR goes off in daylight).

[#4]

The plot thickens....  Have a look at the attached graph.  The first dip from about 6:10 to 6:15 was, I think, caused by me turning on the kitchen lights to make a cup of tea.  All the lights in the house are CFLs with the exception of the kitchen ones which are low voltage halogens driven by a transformer. I know the transformer switches at high frequency (can't remember how high but 20kHz plus).  However turning the kitchen lights on when the SNRM was up at 11dB doesn't cause as much of a dip.  The spikes after the rise to 11dB are, I think, caused by switching CFL lamps on or off.  However the main rise from 7dB to 11dB appears to be caused by the street lamp switching off.  It is later today due to the overcast/rainy conditions.  Our telephone wire goes right over the head of the street lamp (sodium).  Of course I need to prove this by making sure RSL is running before dusk and note the time the streetlight turns on and again in the morning now I know what to look for.  It's a bit too high for me to shin up and cover the photocell!  You realise you've now got me hooked on RSL and seeing what happens when I turn things on/off! 
I've checked the wiring to the first telephone extension after the master socket - blue with white trace on pin 2 and white with blue trace on pin 5.  This is modern wiring I think??  I'll have a look at the master socket later.  I will also try running on the test socket when the street light turns on/off to see if there is any difference.
As for the distance to the cabinet, I'm not sure.  I should have asked the Openreach guys when they were here.  I think it's about 1km.
I'll keep posting updates as and when I have more useful information.  Thanks again for your help.

[#5]

Catching what happens at dusk is obviously important as you've appreciated. I'll post back in more detail a bit later.

[#6]

Quote from: martinjbuckley

You realise you've now got me hooked on RSL and seeing what happens when I turn things on/off!

You won't be the first. It has been very useful for a lot of users REIN (interference) problems.
With some CFLs (also old CFL converters),  individual lamps can periodically be an issue, it is a case of going round and checking each one. I've even heard of the odd LED one giving a problem. Dimmer switches can sometimes also be a big problem. Sometimes fluorescent strip-lights/tubes cause problems - and you've discovered the discharge lamp (street light) outside being a significant issue. I was going to say the later time would be due to the overcast conditions, but you said it yourself. Do you have a neighbour with whom you can check his stats (no need to run RSL) and just see if their SNRM drops when the street light comes on?
The HF SMPS transformer for the halogen lights is probably running at circa 30/40kHz, as can CFLs etc and it's usually when the harmonics are excessive that problems occur.
The wiring you've checked so far is the correct modern CW1308 standard. That's a good idea to try in the test socket as well once you checked out everything else. It will enable you to see if anything might need "rejigging" (I'll explain should the need arise).
With regard the hum on the line after the fixing of a line fault, this might have been due to the line being unbalanced due a fault in the exchange. If the engineer put you on a new 'E-side' pair (Exchange to Cab) the problem could have been at the exchange end on the new pair.
If you are about 1km from the Cab, unless your 'D-side' pair is poor quality aluminium, when Fibre comes, you might be lucky and get about 20/25Mbps.

[#7]

I must be wrong about the distance from the Cab as the openreach guy said I'd be worse off with fibre .  Have a look at what happened yesterday evening when the street lamp switched on.  I haven't tried the test socket yet will do so soon.

[#8]

That is seriously bad and no wonder your line can't handle a lower Target SNRM. A Faulty Lamp and gear there by the looks of it, the igniter going off improves things a bit, but.........
You ought to look at your NTE5a asap, just to check if the engineer reconnected your bell-wire.

[#9]

I will check the master socket later but the engineer only used the test socket and didn't make any changes.  However attached is the graph from this morning.  The annoying thing is that I looked out of the window at 06:10 and the street light was already off so I suspect it may another street light.  At the risk of becoming a curtain twitcher I'll try to monitor the on and off times of the one just outside to crosscheck.  I'll also try using the test socket to see if there is any difference.

[#10]

Better keep an eye on any other nearby street or dusk-to-dawn security lighting. Just another thought, your computer clock is the right time compared to when you curtain twitched? 

[#11]

Yep, clocks are in sync.  The switch on last night exactly coincided with the fall which I why I was so sure that the light outside was the problem.  I guess I need to become a neighbourhood watcher and patrol the streets for a few days!

[#12]

So has your patrolling found (all) the culprits? Anything interesting at the Master socket?

[#13]

The short answer is no.  Using the test socket made no difference.  However on Tuesday evening as usual I got the 5dB drop in SNRM just before 8pm.  The street lamp outside didn't turn on until 6 minutes later so it definitely isn't that one.  When I turned on the PC at 5am the next morning I was expecting to see the same SNRM and for it recover about 6 but to my surprise it had already recovered during the night.  It was most definitely dark at 5am and the street lamps were still on.  My hypothesis was that maybe whichever street lamp was causing the problem had finally failed (wishful thinking!).  Wednesday evening was fine with no drop in SNRM and it was still good on Thursday morning.  However it came back last night and I lost my connection during the night.  I've just had to restart the router so I'll see what happens today.  I really need to establish whether or not it is a dusk to dawn issue so will have to monitor it for a few more days to see what happens.  There is only one street lamp between our house and where the telephone cable goes underground so I'm no longer sure about the street lamp theory.

[#14]

Well best of luck with the hunt. As you say it may not be dusk to dawn, but an "evening and overnight" issue.