Download speed up during 'open questions', down again when closed

[#1]

Hi. This is my first posting on this forum. A little more than a month ago I began a 'Broadband Trouble Shooter' because my connection frequently dropped while I was using it. I use a lap top using wifi and have smart TV with wireless connection to the Technicolor router. I HAD the habit of switching off the router when it wasn't being used at night and when I was out at work. When I got feedback from an engineer, I was told the frequent drop outs were because I was switching the router off and I must leave it on while tests were carried out, and from then on.
I mentioned the average download speed I was getting, which is 2.4mbs. My first broadband provider was O2 and I used to get 4mbs. I was told that because I kept switching off my router, my speed was being 'banded' because the system thought I'd got a bad line. What I needed to do was connect the router to the test socket while line tests were carried out, leave it on all the time, and only test download speeds with the computer connected to the router via cable. The engineer suggested my line was capable of handling about 6mbs in his opinion.
Lets just say, I left the router on. After about a week, while the line was still being tested, and I kept going back to online speed checkers, I noticed I was getting a download speed of 5.5mbs. Wirelessly. This was born out by the fact that I put my smart TV on YouTube and was able to watch HD videos with no freezing or stalling, which I'd not been able to do before. I got a phone call from the engineer while I was doing this. We discussed the line had been stable under test, the broadband speed was up, and I was happy to close the question.
Over the weekend, using the Technicolor Gateway, I noted I'd had two disconnections that were nothing to do with me switching the router off (there had been none during the previous test period). After the first, I noticed the download speed had dropped to 4.8mbs. After the second on the Sunday, it dropped to 2.4mbs again.
After a week of the speed remaining at 2.4mbs, with a smattering of disconnections all while I was asleep at night or at work during the day, I started another trouble shooter. I didn't hear from an engineer (a different one this time) for nearly a week. I kept monitoring download speeds and the Gateway for disconnections. During this period the Gateway log also recorded my router had been 'powered up' for only a few hours - despite me not having turned it off for over a fortnight! If I'd have had a power outage in the house, I have an alarm clock with no back up that goes back to 00:00 hours and flashes - that hadn't happened, so I don't understand the router log.
The engineer who called me this time said broadband disconnections can occur for many reasons (he couldn't explain the router 'time since power on' record to my satisfaction and told me they don't use the Gateway much!), that this time round my line was NOT banded, that banding occurs after 'x' number of drop outs in 'y' number of days or hours (I think it was about six drop outs in 72 hours, which my line wasn't doing) and in HIS opinion, my line should be able to handle 3.5mbs (not 6mbs like the earlier guy). He said my line was in good order. The following day, using Ookla speed test, my line was delivering 3.5mbs and I could still watch HD TV wirelessly on my LG TV at that. The next day, I was back to 2.4mbs and there it has remained since. The same speed as when I was turning off my router (but now with no disconnections during use, which was my original complaint). We left that last 'question' open for 14 days in case of further problems and about four days remain.
I am afraid that apart from knowing what download and upload speeds are, a lot of the jargon associated with internet connections is foreign to me - attenuation, db etc. I know I keep being told that broadband speed tests should ONLY be conducted with a cable connection to the computer and probably with the router connected to the test socket, to get a TRUE reading. As far as I'm concerned this is smoke and mirrors. If my broadband speed is 2.4mbs wirelessly, it is 2.4mbs when I've plugged the laptop into the router and pulled my BT faceplate off and plugged the router into the test socket. Which I have done during the past month after the initial problem. When my line has temporarily been delivering higher speeds while my line was being monitored by Plusnet engineers, the laptop shows those higher speeds wirelessly using Ookla and my LG TV can play iPlayer and YouTube HD videos wirelessly without freezing. As soon as the speeds drop back to 2.4mbs, it can't and wouldn't be able to even if I connected the TV to the router via cable (which is logistically impractical).
These are the details. My question is, why is my download speed dropping AFTER Plusnet engineers have conducted line tests and declared my line okay? And why did one engineer say I should be getting 6mbs (where I got 5.5mbs for a time) and another say I should be getting 3.5mbs (which I did for a time) before it all drops back to 2.4mbs and stays there? (Which was the speed it was getting when I was 'banded' and now I'm supposed to be not 'banded')
Regards,
Mick

[#2]

Hi Mick,
Welcome to the forums.  I can understand your frustrations, but you need to understand that this technology is actually quite complicated, especially when it is not working ideally.  In such situations, the problem has to be addressed onion like - a layer at a time.
What you describe being asked to do by the PlusNET support people is a professional approach to skinning the onion.
Doing a wired speed test eliminates wifi in your local environment (conflicts with other wifi stations) out of the equation.
Doing tests into the master socket test point eliminates your internal phone wiring from the equation.
If your speeds are the same for each of the above, then that rather points to a line issue, as does your speed deteriorating after PlusNET intervention.  This is where you have to face learning the ins and outs (the jargon) if you want to get this fixed.  I suspect that PlusNET have been performing SNRM resets. Mathis puts your line into a 10 day training mode during which BT equipment works out the best STABLE speed for your line.  It explains why at the time you see higher speeds and after drop outs you see lower speeds.
The question which needs to be explored is why is the BT system lowering your speed.  There are many reasons from a straight forward line fault to rein and shine (go listen to Monday's radio 4 you and yours programme item on this subject - see BBC radio 4 listen again).
Please also go read the speed issues thread at the top of the adsl forum and gather and post here the requested information.  From that by looking at the line's attenuation we can determine what speed your line ought to be capable of (never mind the very conservative estimates) and start looking at what might be wrong.
Is your phone line quiet?  Dial 17070 option 2 using a corded phone plugged into the master socket.  If you do not have a corded phone, please borrow one.
Once the obvious has been eliminated, we can consider rein and shine, but that will need to real effort from yourself.  As the BBC programme explains, rein is hard to address and typically comes from something in your home.
Kevin

[#3]

Thanks for the reply Kevin. I do understand the theory behind eliminating things such as the wifi and cable extensions in order to pin down a fault. I am an ex-mechanical engineer so understand fault finding and the process of elimination. Unfortunately electronics and computing is a whole different world to me and that is partly why I changed my trade (lots of reasons though). The point I was making there though is that based on my own methodology in trying to pinpoint faults, if several wifi devices getting an 'excellent' signal in my home all show broadband speed test results identical to those that I get if I cable up, and they all move up and down as Plusnet have intervened and then ended their tests simultaneously, it would be reasonable to assume the wifi isn't the issue nor (probably) line extensions in the home that make the siting of equipment more convenient. Of course I don't understand what equipment Plusnet and BT have got for testing lines and delivering the service, so I guess that was the basis of my question as to why things are up when under test, and down again when everyone says things are optimal and no faults found.
I'm going to have to read your response more carefully and take any further steps you've advised at a later time and get back to you - I'm about to start my working day!
Thanks,
Mick.

[#4]

Mick,
Your analysis is quite sound, this does look like a line problem or an interference problem.  However, on each new PlusNET contact, they will go through their basic process of elimination, especially when the previous fault condition has been closed off.  You need to ensure that your next fault ticket does not fall closed until at least 3 weeks after the next reset.  Running the problem on here is likely to facilitate that.
This stuff is not actually that hard, just at first pass there are a lot of numbers to understand; with your back ground and a helping hand you will soon grasp the basics.  As a starter...
Attenuation is a measure of the line's quality.  It is largely related to the length on the line and the gauge of the wire.  Based on some dependable assumptions one can estimate the expected synch speed at a SNRM of 6dB
SNRM is (in simple terms) the level of the of the signal over the background noise.  It is normally 6dB.  If it has to be higher to maintain a readable signal, the synch speed will need to be reduced.  If the line is (electrically) very quiet and it can be lowered, you will see a higher synch.
From these an awful lot can be assessed about a problem.
Kevin

[#5]

Sorry to hear your still having issues.
If you can follow through the steps Townman has advised you on we can investigate further.

[#6]

Okay, here goes. first the quiet line test. I tried this before, after finding out about it from reading another post, but I just did it from upstairs (where the corded phone is). So, I moved it downstairs to the main BT socket, pulled the splitter and put the phone straight in (but NOT the TEST socket, just into the normal input. This of course disconnected my router!).
What am I meant to hear? My tinnitus doesn't help! Faint background hiss, steady, unfluctuating as far as I can tell. This is a Geemarc Rio and to be fair has always been a bit hissy. To be honest, my analogue cordless BT Freestyle 70 is totally silent doing the same test, me stood near its base station, connected to an extension and through a filter, and with my router on! Does this conclude anything? I recently changed a filter, for a spare I already had, at the point where my router and Freestyle 70 and BT R70 answerphone all are (all the latest stuff, no less!) because I'd got frying bacon on the line, which was terrible. The replacement filter appeared to cure it. This was before I began my FIRST support question over disconnections because I thought the dodgy filter had been the reason, but in this case wasn't.
So in conclusion my Geemarc corded phone is not silent but never has been, even plugged direct to the BT main, but my BT analogue cordless phone IS silent running through extensions and filters. I dare not plug the BT phone right into the BT socket right now for fear of recording another disconnection in quick succession and getting banded again!
Next - following Oldjim's direct 'link' to get DSL connection info for my 582n I get the information below. But I'm a bit confused (and this is why I prefer mechanics to electronics!) because he goes on to say "For the new 582n, take the following steps..." which provides a link telling you how to log into your Gateway, which I know how to do, then initially as far as I can see I can get LESS information than by following his original link!
Anyway this is what I got, after I'd already disconnected my router not long before, remember.

DSL Connection
Link Information
Uptime: 0 days, 0:23:58
DSL Type: ITU-T G.992.5
Bandwidth (Up/Down) [kbps/kbps]: 447 / 4.861
Data Transferred (Sent/Received) [MB/MB]: 199,60 / 337,36
Output Power (Up/Down) [dBm]: 12,0 / 19,5
Line Attenuation (Up/Down) [dB]: 23,7 / 44,5
SN Margin (Up/Down) [dB]: 19,5 / 12,2
System Vendor ID (Local/Remote): TMMB / ----
Chipset Vendor ID (Local/Remote): BDCM / TSTC
Loss of Framing (Local/Remote): 18 / 0
Loss of Signal (Local/Remote): 2 / 0
Loss of Power (Local/Remote): 0 / 0
Loss of Link (Remote): -
Error Seconds (Local/Remote): 904 / 1
FEC Errors (Up/Down): 0 / 606
CRC Errors (Up/Down): 0 / 6
HEC Errors (Up/Down): 0 / 28
Doing it the other way, I get this - the same stuff but just a few minutes later?

DSL Connection
Link Information
Uptime: 0 days, 0:44:59
DSL Type: ITU-T G.992.5
Bandwidth (Up/Down) [kbps/kbps]: 447 / 4.861
Data Transferred (Sent/Received) [MB/MB]: 199,91 / 338,81
Output Power (Up/Down) [dBm]: 12,0 / 19,5
Line Attenuation (Up/Down) [dB]: 23,7 / 44,5
SN Margin (Up/Down) [dB]: 19,5 / 12,1
System Vendor ID (Local/Remote): TMMB / ----
Chipset Vendor ID (Local/Remote): BDCM / TSTC
Loss of Framing (Local/Remote): 18 / 0
Loss of Signal (Local/Remote): 2 / 0
Loss of Power (Local/Remote): 0 / 0
Loss of Link (Remote): -
Error Seconds (Local/Remote): 909 / 1
FEC Errors (Up/Down): 0 / 1.650
CRC Errors (Up/Down): 0 / 12
HEC Errors (Up/Down): 0 / 41

The results of the BT performance test using Oldjim's first link are as follows -

FAQ
Results Image not loaded

1. Best Effort Test:  -provides background information.
Download  Speed
2.01 Mbps

0 Mbps 4 Mbps
Max Achievable Speed
Download speedachieved during the test was - 2.01 Mbps
For your connection, the acceptable range of speeds is 1.2 Mbps-4 Mbps.
IP Profile for your line is - 4.29 Mbps
2. Upstream Test:  -provides background information.
Upload Speed
0.25 Mbps

0 Mbps 0.45 Mbps
Max Achievable Speed
Upload speed achieved during the test was - 0.25Mbps
Additional Information:
Upstream Rate IP profile on your line is - 0.45 Mbps

We were unable to identify any performance problem with your service at this time.
It is possible that any problem you are currently, or had previously experienced may have been caused by traffic congestion on the Internet or by the server you were accessing responding slowly.
If you continue to encounter a problem with a specific server, please contact the administrator of that server in the first instance.
I was connected by wire to the router, not wifi, but the router was still in its usual home on the end of an extension cable. To leave it PERNAMENTLY plugged into the main BT socket  (I'd then need a 3-way splitter to the phones) would mean a mains cable going round the lounge door jamb, which isn't desirable or practical. Of course I've had 5.5mbs and 3.5mbs recently with this set up for a time, and line tests show no faults (?!)
HANG ON, HANG ON........
Next I got to Oldjim's "Check your line speed from 'here'... " link, and it says check using your test socket. I don't know what that link 'does', because it just leads me to a page saying my line speed is 'such and such', which is 2.4mbs, which I already know. So, despite my previous protestations, I've turned my house upside down, pulled the phones out, disconnected the router (again), taken off the BT face plate and dangled a mains cable round the door jamb, stuck the router direct to the test socket, plonked my laptop on battery power in the middle of the hallway on a coffee table and am starting again! Yep, I can live with that for good broadband...NOT! 
So, I'll take a break for ten minutes and start all over again, leaving the previous results for comparison. See if there's any difference. Coffee time. See you shortly....
.....and, I'm back. Been away longer than for ten minutes actually. It was lunch time. Had some toast while watching a few minutes of Carry On Doctor on More 4. Ooh, Matron...oh well, back to the world of digital jargon. At least its given the line more time to settle, such as it can.
First,  the BT Freestyle 70 cordless went into the actual TEST socket before I put the router in. Yes, I know its not corded, but I know the Geemarc corded phone hisses all by itself anyway. There was a faint unobtrusive background hiss, but nothing that I would otherwise regard as abnormal or a barrier to holding telephone conversations. Nearly all electrical audio equipment produces a slight hiss, doesn't it? (Except DAB, presumably).
Next, DSL info using Oldjim's link...
DSL Connection
Link Information
Uptime: 0 days, 0:50:57
DSL Type: ITU-T G.992.5
Bandwidth (Up/Down) [kbps/kbps]: 444 / 7.108
Data Transferred (Sent/Received) [MB/MB]: 1,18 / 5,20
Output Power (Up/Down) [dBm]: 12,0 / 20,3
Line Attenuation (Up/Down) [dB]: 23,4 / 44,0
SN Margin (Up/Down) [dB]: 18,1 / 12,2
System Vendor ID (Local/Remote): TMMB / ----
Chipset Vendor ID (Local/Remote): BDCM / TSTC
Loss of Framing (Local/Remote): 0 / 0
Loss of Signal (Local/Remote): 0 / 0
Loss of Power (Local/Remote): 0 / 0
Loss of Link (Remote): -
Error Seconds (Local/Remote): 0 / 0
FEC Errors (Up/Down): 16 / 347
CRC Errors (Up/Down): 0 / 0
HEC Errors (Up/Down): 0 / 0

BT Wholesale Speed Test - I'm now being told I've got to change my broadband login user name within my router for the test to finish! Why? It didn't do that before! This is beyond me at the moment.  I don't see why that has changed.
Re-run the test, seems to have gone straight to the results without performing the test this time! Results below...

FAQ
Results Image not loaded

1. Best Effort Test:  -provides background information.
Download  Speed
2.14 Mbps

0 Mbps 7.15 Mbps
Max Achievable Speed
Download speedachieved during the test was - 2.14 Mbps
For your connection, the acceptable range of speeds is 2 Mbps-7.15 Mbps.
IP Profile for your line is - 6.27 Mbps
2. Upstream Test:  -provides background information.
Upload Speed
0.15 Mbps

0 Mbps 0.45 Mbps
Max Achievable Speed
Upload speed achieved during the test was - 0.15Mbps
Additional Information:
Upstream Rate IP profile on your line is - 0.45 Mbps

We were unable to identify any performance problem with your service at this time.
It is possible that any problem you are currently, or had previously experienced may have been caused by traffic congestion on the Internet or by the server you were accessing responding slowly.
If you continue to encounter a problem with a specific server, please contact the administrator of that server in the first instance.
Please visit the FAQ if you are unable to understand the test results.
The link to my current line speed still says this...
Estimated line speed:
There's no speed estimate currently held on your account.
Current line speed:
2.4 Mb
Well, I hope someone can make more sense of this than I can. Yes, I can see that certain results are different plugged into my test socket than they were when I began mucking about with everything in situ where it normally lives. But the basic level of instruction that I was working on before indicated that if I've got issues with the wiring inside the house, my broadband speed will go up if I plug straight into the test socket. Before I began the process of reporting issues, this didn't happen. Then when I reported issues, I was told it was because I was banded. Now I'm not supposed to be banded, but plugging into the test socket hasn't directly influenced my usable broadband speeds. I'm going to see if I can temporarily tidy up my 'direct into test socket' installation of the router for a few days, plugging my phones back in, but it isn't pretty and can't stay like that forever.
Why, if turning the router on and off causes problems, does things like the BT line test tell you to reboot your router before performing a test?
Thanks,
Mick.

[#7]

Got to dash, but your story appears to confirm that there is a problem with your internal wiring given that the sync speed went up considerably when in the test socket. When PlusNet's current line speed catches up you should see an improvement in speed tests.
I'm sure others with more time (and knowledge  ;)) will be along soon with their observations...

[#8]

Hi Mick,
You have certainly been working hard on this!
First the phone test - I think we can safely say that at present, there is no discernible noise.  You should check this out from time to time - something has caused your target SNRM to be elevated to 12dB.
The connection of the router direct into the test socket, gives some idea of how much impact your internal wiring has on the synch speed...

Quote from: mickthefitter

Bandwidth (Up/Down) [kbps/kbps]: 447 / 4.861 IP Profile for your line is - 4.29 Mbps Bandwidth (Up/Down) [kbps/kbps]: 444 / 7.108 IP Profile for your line is - 6.27 Mbps The link to my current line speed still says this... Estimated line speed: There's no speed estimate currently held on your account. Current line speed: 2.4 Mb

That is a 45% improvement.  This synch speed is "right on the nail" for your line's attenuation at 6dB SNRM.  If you add the benefit you might see from having the SNRM reset to 6dB, you might see another 1.2Mbps on top.
You have proved that the internal wiring is taking a sizable chunk out of your line's potential performance, who you choose to address that is a decision which only you can make.
From here, a request to have a SNRM reset will deliver additional benefit, but cannot address the challenge from your internal cabling and might regress if the cause of the elevation returns.
As for the data speed tests, these will not improve until the PN profile catches up with the latest change to the BT Profile.

Hope this helps?
Kevin

[#9]

Hi Kevin. Okay, I followed most of that. After tidying up the cabling (not that a mains extension lead around a door to a router near the test socket on the other side of the wall looks good) I've come back online with a thought to telling PN what I'd just done on my still 'open question' from last time round. So, to clarify, I have to ask for an SNRM reset (I've forgotten what that represents if I've already been told) but are you saying that the broadband speeds might go up by themselves based on the BT profile?
When I did the first 'Troubleshooter' in February they turned on something called 'interleaving'. I didn't know what it was, so looking it up it appeared that, from what I could gather, the default setting was 'on' and you can request to have it turned 'off' if you are a gamer and want quicker ping times. I'm not, and hadn't. So I don't know why it was off. Apparently 'on' gives you a more stable line. I assume that is best left as it is.
With regard to my cables, I concede that if I get some better, stable results with my temporary lash-up I'll feel compelled to look at running some new cables. I own a carpet gripper tool for putting carpets back, though it's bad on the knees! I certainly do not like visible cables going round the edges of skirtings and round doors!  Now though, in a matter of weeks I've gone from someone who turned off the router when not in use, to being terrified of interrupting its connection or power supply for fear that as soon as I do, my broadband speed will be down and contrary to guidelines, stop there!
Thanks,
Mick

[#10]

Hi Mick,
If you want any of the previous clarifying, please ask - given the wide spectrum of skill and understanding (or lack of it around here) it is difficult to get the "pitch" right. 
SRNM reset - this puts the line back into training mode and releases the DLM "lock" or control of the line for poor stability issues.  Given that you line is quiet, it should be capable of doing better at 6dB rather than 12dB.
Interleaving - in simple terms, it is a non-retransmission error recovery.  Additional bits are inserted into the sych bit stream which enables the router to recover bad data blocks without them needing to be retransmitted.  It adds a little processing overhead and takes some DATA speed out of the transmission capacity.  If the line has a low error rate, not having interleaving delivers slightly better data throughput rates.
The DATA speed will improve when the PlusNet profile (see as "current speed" in the user portal) catches up with the BT profile reported in the further diagnostics in the BTw speed test.
Hope that clarifies better.
Kevin

[#11]

Sorry to clarify/correct something, but:
The only way not having interleaving will improve download speeds on a line with low error rates is if it connects at a higher "synch" speed without interleaving. Otherwise, a Plusnet "current line speed" of 6.2 achieved with interleaving will see the same download speeds as a current line speed of 6.2 achieved without interleaving.

[#12]

Hi ejs,
Can you expand on that please?
As I understand it, interleaving inserts additional (non-data) bits into the packets to enable FEC recovery.  If interleaving is switched off, then are there not more bits available for carrying data?  We've seen / had the discussion about reduced US rates when interleaving is on.
Whatever, it is marginal!
Kevin

[#13]

Yep, er, semi clarified....I've learned about the training mode and that there can be a lock on speed on bad lines - so SRNM reset starting the training mode again is fine.
Now you are having your own discussions on the effects of interleaving! 
I'm having a spot of bother identifying where you've got the 6db and 12db from in the data I've provided. I can see Output Power, Line Attenuation and SN Margin with db references, but I can't spot a 6 so I'm unsure where you've read that from in determining how much more capable my line is when I've gone straight to the test socket.
Cheers,
Mick

[#14]

Hi Mick,
Sorry - 6dB is knowledge!

Quote from: Townman

Attenuation is a measure of the line's quality.  It is largely related to the length on the line and the gauge of the wire.  Based on some dependable assumptions one can estimate the expected synch speed at a SNRM of 6dB SNRM is (in simple terms) the level of the of the signal over the background noise.  It is normally 6dB.  If it has to be higher to maintain a readable signal, the synch speed will need to be reduced.  If the line is (electrically) very quiet and it can be lowered, you will see a higher synch. From these an awful lot can be assessed about a problem.

Quote from: mickthefitter

DSL Type: ITU-T G.992.5 => ADSL2+ Bandwidth (Up/Down) [kbps/kbps]: 447 / 4.861 => how fast the line is going (synch speed) Data Transferred (Sent/Received) [MB/MB]: 199,60 / 337,36 Output Power (Up/Down) [dBm]: 12,0 / 19,5 Line Attenuation (Up/Down) [dB]: 23,7 / 44,5 => this is where we start working out expectation SN Margin (Up/Down) [dB]: 19,5 / 12,2 => This we expect in ideal conditions to be 6dB System Vendor ID (Local/Remote): TMMB / ---- Chipset Vendor ID (Local/Remote): BDCM / TSTC Loss of Framing (Local/Remote): 18 / 0 Loss of Signal (Local/Remote): 2 / 0 Loss of Power (Local/Remote): 0 / 0 Loss of Link (Remote): - Error Seconds (Local/Remote): 904 / 1 FEC Errors (Up/Down): 0 / 606 => non-zero indicates interleaving is active CRC Errors (Up/Down): 0 / 6 HEC Errors (Up/Down): 0 / 28

A change of 3dB SNRM is "worth" 400-1200kbps depending on the line - increase in SNRM leads to a decrease in synch speed - decrease in SNRM leads to an increase in synch speed.
So given you might have 6dB to spare, if your line is good, you might hope to see +1Mbps synch speed.
Your initial synch speed (as above) plus what 6dB shift in SNRM might deliver leaves you well short of the expectations for you line.  That therefore pointed to losses elsewhere - for example internal wiring, which you have kindly proved.  If you really want to know more about speed estimation, look here - http://www.kitz.co.uk/adsl/max_speed_calc.php - and have a browse around that site.  The line checker is very useful too. 

Sadly CRT have not picked up this thread today to do the SNRM reset.  You could phone up the CSC, but I would not do it at this time of evening... but looking at the call stats graph here - http://portal.plus.net/supportpages.html?a=212&helpheader=rhmcallstats - current call levels are looking very low.
Kevin

[#15]

Townman:
The speeds reported by the modem do not include any interleaving overheads. If you measure the download speed, and add the TCP/IP headers, you can get a result very close to the profile value, even with interleaving on.
mickthefitter:
The 12db figure was from your SN Margin. 6db is the default SN margin. To get an estimate for what speed your line might be capable of, you can subtract 2 from your downstream line attenuation figure (because your stats show G.992.5 - ADSL2+), and put the result into http://www.kitz.co.uk/adsl/max_speed_calc.php - this gives about 1Mb higher than the 7,108 you got with a 12db margin, and estimates what you might get with the noise margin reset to 6.

Quote from: mickthefitter

but the router was still in its usual home on the end of an extension cable.

What kind of extension cable? Flat telephone extension cable is bad for broadband as it probably won't be twisted pair, so will pick up far more interference than twisted pair cable of the proper standard (CW1308) or better.