PlusNet Slower than TalkTalk for me :(

[#1]

It's still going to be star-wired unless something like @jelv's suggestion is implemented, so that the ADSL signal is restricted to just the wiring to the router.
That sounds like it may well eventually be the optimum way forwards.  
Testing from the test socket would indeed confirm the level of the problem caused by the internal wiring picking up interference (from wherever), by looking at the router stats as mentioned earlier. No need for that to be a permanent change at this stage.
Note that he (@jelv) said:

Quote from: jelv

If it is better at the master

IF - it does need checking. 

[#2]

I thought I'd made it quite clear, testing from the test socket at this time is likely to prove NOTHING because of the way banding seems to currently work, and we know the wiring is not upto standard, yet you suggest he disrupt his household for no purpose.
Also lets be clear, star wiring is irrelevant if the wiring is up to standard or filtered - as I was suggesting, using at vDSL plate (Mk.3 btw). If the wiring to the socket used by the modem/router is CW1308 ie has the Blue+white trace & White+blue trace then there will be no problem. Why do you make it sound unnecessarily complicated.

[#3]

So if he uses the master socket he sees a higher noise margin and it is still synced at the maximum for the current band it proves nothing? And if he sees errors being logged at a lower rate it still proves nothing?
On the other hand if things are still as bad when using the test socket in terms of noise and errors (and maybe disconnections) it proves that while the internal wiring isn't helping, it isn't the only issue and an engineer needs to be booked.
I forgot to say that when the face plate is removed to use the test socket it needs to be confirmed that all the other extension sockets are dead.

[#4]

Exactly. 

[#5]

jelv & HP, the way Banding seems to work on a lot of connections these days, the SNRM does NOT change. I won't bother repeating what I've already said about unnecessary disruption, nor what I've already said about drops in sync and the aggressiveness of DLM. It would be far better to have no drops for a day or two. When the wiring issue is resolved if errors etc don't look better, then is the time to test at the master socket.
And perhaps you'd care to explain what error reporting on the v2 is likely to be of some use? I don't have a v2 to comment.

[#6]

Dave, does your Master Socket look like the one on the left with the raised Openreach logo, or is it one that has the grey print Openreach logo? Not that it should make any difference as far as fitting a Mk.3 vDSL plate is concerned (I've PM'd you about that).
As mentioned previously (by jelv) there are A&B terminals on the vDSL plate which can be used to connect the extension that will be used for the modem/router. However the wiring to this extension must be CW1308 standard - ie twisted pair - normally using the Blue+white trace on 2 from B & White+blue trace on 5 from A.
All the other extensions can be wired from the front plate terminals 2, 3 & 5 - the bell wire can be included because it's all filtered here and you don't need a filter at the extension socket end and because its filtered it doesn't have to be twisted pair. You may need the bell wire for some phones to ring, most modern ones only use the 2 wires but not all do. Where you don't have central filtering, the plug-in filter generates a local filtered bell-wire on decent filters. Which leads me to the next point.
All plug in filters where used should be Microfilters that look like this. The type shown in your photo is generally of inferior quality with sub-standard filtering and some have an unfiltered bell-wire straight through - apart from the bell-wire interference problem they will drop your sync speed by a couple of meg or so.
On the socket that you are going to use for the modem/router providing the wiring to it is CW1308 as mentioned, you can use a decent filter at the socket for the phone & modem/router. You don't have to have separate wiring or sockets.
If you need to replace any cable for the modem/router socket (or elsewhere if not going for the vDSL plate - but I strongly recommend you do) then the following may be of help CW1308 cable & Cable clips and an IDC tool good for a small number of jobs. A reputable supplier and P&P the lot should only be £1.
HTH.

[#7]

Oh, forgot to comment on your SNRM graph on your reply #25 (distracted by the other stuff) as you can see that after 0854 the line becomes virtually flat. As you get to an hour or more after sunrise, MW/AM propagation generally drops right off, and the fact the line becomes flat also suggests that most of the problem is coming from that.
But because of your wiring standard, you may at other times also have picked up the other odd switching spike or other noise from electrical/electronic appliances which isn't there consistently, ie some SHINE as opposed to REIN type interference.

[#8]

There are acceptable alternatives to CW1308 - the crucial thing is it is twisted pair. Ethernet cable (especially CAT6 is better than CW1308) as it has more twists per meter.

[#9]

To what point? - Cat 6 (or 5) is a bulkier uglier cable and will not add any improvement to ADSL2+ performance.

[#10]

If someone already has a suitable length of Ethernet cable it's cheaper than buying CW1308 (and in the case of bad noise may be better than CW1308 because of the higher twist rate).

[#11]

The photo showed CW1308 and reasonably long tails, so hopefully, IF the OP decides to follow Jelv's suggestion and then AO's advice and use a Mk3 faceplate the existing wiring is all CW1308 and can be reused easily.
Haven't heard from him for a fair few posts now, lots of comments to digest first.......................
Just to add that despite various "debates" here, I would do exactly that as well, and "regularise" all the sockets of course.  

[#12]

Hardly seems like "bad" noise when you look at the last SNRM graph. It all looks fairly typical of what I'd expect to see with the way the wiring is currently, and as I mentioned it will probably catch an odd biggish spike as well.

[#13]

Well I have been working on this all day and the line has understandably been up and down all day as I disconnected things.
Sorry I have not been reading the posts as they came in though, I have had a quick look through them and will return to the points once you guys have had a look at what I found below. Anotherone seems to have a good summary for me to follow. Twisted pairs from my experience elsewhere seems a good way to go, I need to check whether it is one already, I suspect not.
I have done quite a bit of bell wire disconnecting as suggested, I came to the same conclusion as Anotherone that the resistor that HPsauce (thank you) pointed out needs to go, along with the bell wire. As I did this routerstats indicated a 1db increase in DL Noise Margin at the time. 
I then inspected ALL extensions at the ends and removed any bell wires I found and the "other" connection, leaving just 2 & 5 wires in place. All extensions had their ADSL filters replaced, I seem to have a lot of filters in the 2 off TalkTalk & PlusNet router boxes (i had not used) most of which were in the loft. Checked all phones worked OK, with some of the rings being slightly delayed, not a problem.
Rebooted the router SNR  seem to have improved slightly may be 1db, I think was mainly due to the initial resistor removal.
Checked the "Bits & SNR / Tone" almost identical to the test sample I took before I started. Speed obtained as expected and confirmed in Anotherone post remains the same.
So I upset the one person left in the house by disconnecting ALL extensions including pulling the wiring out the Master socket so I was left with just the master socket to router and nothing else.
I rigged a power lead to the area and brought the router down, I plugged it in. Please take a look at the before 2014Dec29 and after graphs of the "Bits & SNR / Tone" the only real feedback that I have of an improved line other than SNRM 2014Dec30. I was disappointed.
I then changed the routers power block for an old transformer version (original), I had swapped a day ago and retested no real difference in the power block graph. Tried some other filters, no difference.
At least I have cleared up the bell wire issue. For info I have included some pics of the Master Socket on and off.
All now back to as it was, Router in the Bedroom 4, with all lines having new filters and no bell wires. I need the network cables plugged back in before all hell lets loose.
I understand that the internal wiring may pick up SHINE which may be knocking the line down, which twisted a pair to the router and a new master socket filter may improve.
In the mean time I would like a retune the line / profile reset after all this change and hopefully good work, what do I need to do to get that to happen?

[#14]

Can you just clarify what that device is in the first photo that you have opened up?
It looks like it may be some sort of plug-in extension lead (which rarely has twisted-pair cabling), if so where does it go and what for?
When you tested at the Test socket the interesting data would be the error rates which Routerstats could record/display, obviously the longer the better, and information such as the graphs of noise margin over time that you recorded previously. The actual connection speed wouldn't change much, or at all, because of the restrictions now imposed by DLM.
Remember it's the errors that DLM responds to. If they reduce, over time it will reduce its aggressive stranglehold on your line.
I don't know what equipment you have available, but if the router is situated near the master socket you can run a standard RJ45 LAN cable to the original location and use a cheap switch to then supply the LAN network. Might not be elegant to start with but will work.
As a final comment for now, if you are in the mood for a bit of a temporary lash-up remember that the router only needs 12V DC, which can of course be supplied by a thin 2-core extension cable from a PSU brick in a power socket some distance away. Be careful to get the polarity right though.

[#15]

Quote from: HPsauce

Can you just clarify what that device is in the first photo that you have opened up? It looks like it may be some sort of plug-in extension lead (which rarely has twisted-pair cabling), if so where does it go and what for? ....... As a final comment for now, if you are in the mood for a bit of a temporary lash-up remember that the router only needs 12V DC, which can of course be supplied by a thin 2-core extension cable from a PSU brick in a power socket some distance away. Be careful to get the polarity right though.

first photo - the header for the extension lead to the router, I was checking that the bell wire had been removed, I had done that on this extension in 2009 but forgot. The cable does look twisted but I would need to take it apart and check which wires were paired me thinks. Or just get a new one.
I had thought about a network line up to a network switch, if I got a massive improvement in line performance with the router at the Master Socket, it appears not to make much difference even with no other line connected. I was looking for a jump up in noise margin that didn't happen.
I can't leave the router down by the Master Socket for days, but I do know that is the best place for it.
I hadn't thought of the remote 12V DC. An idea to add in my brain. Polarity point noted.
Thanks
Dave