BT have messed up my broadband connection, is there anything I can do?

[#1]

Latest snr graph attached.
Nothing was switched on at that time as far as I remember.
I can't seem to get it to automatically capture bits/tone.  Schedule Capture [Always] is ticked when I right click the graph.  It captures ok if I right click and select Capture Now.  The bits/tone graph seems a little temperamental, sometimes all the lines are green and sometimes they are green and blue for tx/rx,  sometimes it shows nothing and sometimes obvious garbage, though most of the time it is ok.  I am running in Ubuntu so that may be an issue I suppose, but the non-telnet charts are fine.  I should probably try it in Windows.
I have a filter plugged directly into the test socket behind the master plate with a single phone and the modem plugged into that.  It goes even slower if I allow it to connect into the extension wiring in the house, but I have a Panasonic wireless phone and the master is in the "office" so it is not a problem to connect directly into the test socket.
I have just unplugged the phone and the snr went up from 4 to 6.  I left it a bit to check that it stayed there, which it did apart from occasional spikes down as usual.  When I plugged it back in it re-synced and the speed has dropped to 1152 (from 1248) and the snr is at 6.

[#2]

The scheduled capture will only happen when the graph has completed every 2½hr stint with the settings mentioned. If you notice a significant change in the bits/tone graph, best to do a capture there and then and post it. The colour changes and other issues occur in Windows, so I wouldn't worry about that, minor bugs in the program.
OK now please remember that Disconnect procedure I outlined in an earlier post. If you can't get to an Interface to drop the Internet session in a particular modem/router then just power down before unplugging from the line, however with the 834v3 I'm sure you can Disconnect first.
So I'm not sure where you are plugged in at the moment, if it's the Test socket the please post the current bits/tone (a non-rubbish version) and SNRM graph. If you maximise the RouterStats window the captures should be a full screen which show better detail. The filter that you unplugged the phone from and the SNRM went back to 6, I assume was the one you plugged the phone back into and the resync occurred. Please mark this filter so it doesn't get mixed up with any other. Do you have any spare filters? Do you have a corded phone?
Apart from giving me a more detailed description of your setup as requested in my previous post and confirming the type of filters you are using, if you are not plugged in at the test socket can you do the following -
Do the Disconnect procedure. Plug the filter you've marked into the test socket and plug the modem/router into it but don't power up yet. Plug a second filter in to the phone socket of the first filter and plug your Panasonic into the second filter - this double filters the phone. BUT if you have a corded phone handy then use that instead and turn off your corded phones and base units. Then power up the modem/router, you may need to log into the interface and click Connect to establish your PPP session, then go grab a capture of a bits/tone and post.

[#3]

First let me say once again many thanks for investing so much time on helping with this problem.  I am retired and contribute where I can to the Ubuntu support community and with a couple for Open Source projects, so I hope you can feel your time is invested in a good cause
I should have been clearer on my current setup.  I am already plugged directly into the test socket and have run like that for a long time as it seemed to give the best performance.  The filter is an AGCOM AMF320U02, and only the modem and the single phone are plugged into that.
Attached are snmr and bit/tone graphs from before the experiment, snmr at 2143 and bits at 2146.
I disconnected via the router http interface, powered it down and plugged a second filter into the first.  This one is a Speedtouch DSL4132005 (NA30513).  I plugged a wired phone into the second filter and powered off the wireless phone.  I waited till ten minutes had passed and powered up the router.  Sadly this time it has connected at 1088, which is a little disappointing.  I think it initially connected at 6dB again but RouterStats seemed to have crashed and I had to kill it and restart which took a little time, so not certain of the initial value.
Attached are the snmr and bits graphs after power up at 2208 and 2207.
Sorry, I have just noticed that I forgot to maximise RouterStats after restarting before taking the snapshots so the graphs are not the same size.  I have taken them again now and attached those also, timed 2228 and 2229.
Blast, I forgot to switch off the handset, I have done that now, it did not affect the snmr graph, I have attached yet another bits/tone graph (timed 2236), but I don't think it is significantly different.

[#4]

Hi Colin,
Having just Googled that AGCOM filter, the picture I got was of one that has an appearance of a solid block splitter style. Mostly all filters of this type are a disaster as Broadband performance is significantly poorer than the types as per the link in my reply #44, and is probably responsible for the poor bits/tone allocations. In a way I'm not surprised that the sync speed was lower because noise levels will be higher after dark, and a sync will always occur at the Target SNRM currently 6dB and so result in lower speed. You will generally get a better sync speed in daylight hours.
If the Speedtouch one is the "rat's tail" type as per my link, then please use that only. With the corded phone there will be no need for double filtering to eliminate possible interference issues.
So to be clear, do the Disconnect procedure. Plug the Speedtouch filter straight into the Test socket, plug the corded phone and the Modem/router into it, power up (login and click connect if required) and immediately check the stats. Do a grab of the bits/tone plot which should be much improved as should the sync speed. If you post those as soon as they are done.
When you've got a good duration of Noise margin graph post that.
Also when you post back tell me what wires and their colours are connected to the the back of the NTE5 front plate which is presumably dangling on the floor, also whether the logo on it is OpenReach or if it's BT with the Piper logo?

[#5]

First another apology.  When you pointed to a picture of a micro filter I thought that was just in case I did not know what a micro filter was, I did not realise that the physical construction was significant.  Both of them are of the rat's tail variety.
Second yet another apology.  You asked about the wires connected to the NTE5 front plate.  In fact I realise now that there is no NTE5 front plate, when I looked for it in the draw I realised that it is actually a filter unit that plugs into the master socket, identical (including text on front) to the right hand one in http://www.plus.net/support/broadband/master-socket-guide.shtml#whatsMaster.  I remember now that when this was first fitted by OpenReach (two or three years ago) it gave lower speeds than a rat's tail one (confirming your comments I think) so since then I have plugged a rat's tail one into the test socket, with just the modem and phone connected into that.  Is that configuration ok?
As you asked, using the Disconnect Procedure, I have reverted to just the Speedtouch filter and the corded phone.  It has connected at 1312 and 6dB (confirming your comments).  Bits/tone graph from immediately after connection is attached.  The snrm graph is showing much as before, 6dB with spikes down as low as 2.  I will post it when it has built up some history. 
It does seem that the wireless phone is at least part of the problem.  I could fit an extension cable to the phone and move it across the room if that would help.  At the moment it is only a foot away from the router.
I have bought a DG834G v4 from ebay for 11.49 inc postage.  I could probably have saved a couple of pounds if I had been prepared to wait, but I decided it was not worth the hassle.
Colin

[#6]

Noise margin graph attached, showing two hours after revert to single filter, with corded phone.

[#7]

More noise margin data.  Removing the wireless phone does seem to have removed a lot of the spikes.

[#8]

Hi there, no need to apologise by the way, strange about that AGCOM filter as I got a different picture, however the Bits/tone graphs is looking better with the Speedtouch, so for the time being want to treat the other as suspect. The ADSL v1.0 plates are not usually a problem and should work fine, but I still assumed you had other fixed extension sockets because in reply #45 you said "It goes even slower if I allow it to connect into the extension wiring in the house". It was the colours of the wires in this wiring I was asking about - so how are these normally connected and are they all currently dead if you plug a phone into any of the sockets?
The SNRM graph is looking better as the frequency of spikes is not as great. Whether this is to do with the cordless phone or not, can't be certain at the moment. Are all the base units and phones switched off? With all the swapping about in the last 24hrs it might be an idea to keep things unchanged for the moment just to see how stable the situation is. How long is the cable that is connecting the modem/router by the way?
Cordless phones can sometime be an interference issue both from their wireless transmissions and noise fed up the cable hence the double filtering to deal with the latter when it's an issue, I'm not certain that it is yet in your case but to be safe, if you can do without it for the time being. But yes, base unit and the phone, along with mobiles are best kept away from Modem/routers, a foot is close, 3 or 4 feet would be better.
Although the bits/tone is looking better, I'm still concerned about the absence of bitloading on tones 33-37 and the fact that it seems to run out of steam at around tone 109. Longer lines than yours still have bitloading at tones 120-130. The other 2 gaps are bugging me as well as the lower bits/tone loading than one might expect.
So a couple of questions, do you live either quite high up or not too far from a Radio4 LW transmitter. Do you have a wireless alarm system, do you have any other wireless gadgets?

[#9]

Hi again.  On the issue of the other extensions, when I said "It goes even slower if I allow it to connect into the extension wiring in the house" that should have been in the past tense, when I tried connecting the extensions the broadband went slower.  Since we have the wireless phone we don't need the extensions so I disconnected them at the master socket.  The wires come straight in from outside into the test socket and hence into the filter, the modem and the phone.  No other wiring is connected.
The cable to the modem is about 3m.  I do have a 2m one I can fit the next time I am changing anything.
Not having the wireless phone is a bit of an issue as it means the remote handsets don't work and we can't hear the phone ringing in part of the house, plus we have no answering machine as that is in the phone.  Would it be appropriate to try connecting the wireless base station to the filter via a phone extension lead and moving it well away from the modem.  I am assuming that conducted interference is less likely if there is a long lead between the filter and the base station.
As far as external sources of RF are concerned we are out in the country with only one neighbour nearby and I am don't think they have any unusual wireless kit.  We are in a valley and there are no transmitters nearby, quite the reverse in fact, getting decent radio and tv reception can be problematic.  We have not got an alarm and don't even normally have mobiles switched on.  The only wireless stuff is the wifi.  So nothing obvious to explain the missing tones.

[#10]

Bitloading at tone 45-46 is almost definately 198khz Radio 4, it would make little difference if you were in a valley or mountain top due to the transmission being a ground wave.
Dect phones are terrible for interference due to the high pulsed power of the rf, analogue cordless phones have far lower peak power so cause less problem to broadband.
Paul.

[#11]

Hi Paul,
Thanks for the extra info, well he has the passive ADSL v1.0 plate which as I've mentioned below to consider. We'll also see what the situation is with the double filtering and phone moved. You don't have to go overboard with Cat5 cable though, it's ugly thick stuff. CW1308 is quite good enough, it is used by BT  Anyway we need to try and discover why the bitloading on the other tones is poor. & PS I will reply to your PM, this is getting embarrassing now 
Ok, Colin understood. Your internal wiring to extensions shouldn't affect the broadband if the cable is to the correct standard and connected appropriately. If it is 2 (or 3) pair CW1308 cable ie contains wires coloured Blue/white trace + White/blue trace, & Orange/white trace + White/orange trace and if a third pair Green/white trace + White/green trace then this is the latest standard. This can be connected but you would only connect Blue/white & White/blue to IDC terminal 2 & 5 on the back of an unfiltered face plate and you would use (quality) filters at every socket you plugged something in. If you have a filtered Faceplate such as the ADSL v1.0 (or XTE-2005 from ADSLNation) then you also connect the bell wire Orange/white to terminal 3, but note extensions can only be used for phones, not broadband. You could also use older standard wiring in this situation where the cable has Blue, Orange, Brown and Green (4 wires) but you must not use this old standard or connect the bell wire without Central filtering other broadband performance will be much poorer and on long lines it may not work.
So that is something you may wish to consider later on.
As for now, the Speedtouch filter seem better than the other so leave it in the Test socket. If you have a phone extension lead, plug that into the Filter phone socket and plug the other Filter into the other end and plug the base station into that and try it as far away as practical. If you have no drops in sync so far today then I think I would use the Disconnect procedure and power down to do that, and put the shorter lead in for the modem/router at the same time.
The other thing I want to check at this time is the standard of BT incoming cable (the drop-wire) to the Master socket. Is it a Black round cross-section about 5.3mm in diameter, or is it a largish flat black "twin flex" looking cable (figure 8 cross-section) or thinnish grey flat twin?

[#12]

The cable for the extension sockets appears to be 3 pair CW1308 as you describe, so I may try connecting them back up at some point once it is all working as well as can be achieved, to thanks for the info on that.
I have a second SpeedTouch filter so I have used both of those as you describe with the cordless phone on the end of a 10m extension with the second filter on the phone end of the cable, and the base station in the next room.  I also used the shorter modem cable.  Unfortunately this was not all that successful.  It connected at only 736k and 6dB!  Bits/tone attached (1638 timestamp).
Risking the wrath of the BT exchange I unplugged the phone (at the test socket filter) and the snmr went up to 11dB, it did not resync.  I then removed the 10m extension and instead plugged the phone with the second filter direct into the original filter, so it is still double filtered but without the extension lead, so as you originally suggested.  I have kept the phone as far away from the modem as possible (about 8 feet).  Plugging it in  had no effect on the snrm which is still sitting mostly at 11dB and again it did not resync.  So now I have a working phone with a connection at 736k and an snrm of 11.  I am hoping that since the snrm did not fall when I plugged in the phone that this will equate to over 1200 with 6dB.  Will it speed up automatically at some point or will I have to force a re-sync (but not today presumably)?  Bits/tone attached (1657)
The incoming BT cable appears to be the round black variety.  It is difficult to see but I think that the orange and the white cores are connected, with green and black left dangling.

[#13]

Well that is certainly strange with the extension cable. You will need to resync to get the speed back,so if the SNRM is still at 11dB then do the disconnect method, wait about 5 minutes and power up again. I wouldn't leave it too long as it will be getting dark soon and noise levels will increase, but still do it today as it will take a while for your profile to recover after this lower sync speed.
Edit: If the resync today doesn't quite achieve at least the best sync speed you had earlier, you can always do it again tomorrow well into daylight hours. Post up some SNRM plots and new bits/tone when convenient and I'll take a look tomorrow evening.

[#14]

An oddity that hasn't been explained:

Quote from: Jameseh

One of the things that we look at is the line loss on your line, which generally gives a gauge of distance - your router says 58dB, but Openreach think it might be closer to 80dB (IE VERY VERY LONG).  If it is more like the Openreach figure, I could understand why they don't think that they can do much more.

James - were you looking at the right line? He's seeing the same attenuation from two different makes of router!

[#15]

Hi clanla,
I have just checked your line and your profile is currently banded, I have requested that this be removed and this should take effect in the next 4 hours. This should help your speeds.