Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[jeffbb]

Hi
Remember that the SNR margin will vary with Noise on the line .So a reduction of your margin at a particular time is NOT a good sign .When analysing SNR margin you need to take into account Your attenuation ,your synch Rate ,The changes to the SNR margin over time .
That's why A graph over time is so important ,as shown by your last one . Its obvious that there is something Drastically wrong .Is this the replacement router ?.
BTW how far do you think your exchange is ?according to your attenuation it should be around 1.3 KM

Regards Jeff

[burakkucat]

I'm thinking of plugging old router back in, in a few hours, to see if the SNR jumps back up to the 38/39 mark again - what do you think ?

Please don't keep "chopping and changing" with the modem/router. All you will eventually prove is that the TT DLM has decided that your line is unstable and so it has taken the only action available to it . . . which results in a degraded sync speed.

[sparklyman]

Hi Jeff,

I would have guestimated about 2K, so your 1.3K figure (where'd you get that from ?) may be correct.

Yes - This is the new router/filter combo that arrived today. 10/10 for the SOuth African office that agreed to send a repalcement to try....but I don't think it's worked.

Burakkucat - The TT Forum post has been closed ?   SO I've opened a new thread in the "Broadband connection & stability" section

http://www.talktalkmembers.com/forums/showthread.php?p=722100#post722100

TT have placed the 256K cap on my line - they did so on 7th Oct in order to stabilise the connection - is the DLM a person or equipment ?

[burakkucat]

Unfortunately connection is back to it's old tricks - here is the latest capture

That is, to be blunt, appalling.  

At certain times, on that graph, the noise is at a greater intensity than the ADSL signal that you are trying to receive.

[burakkucat]

Burakkucat - The TT Forum post has been closed ?   SO I've opened a new thread in the "Broadband connection & stability" section

http://www.talktalkmembers.com/forums/showthread.php?p=722100#post722100

TT have placed the 256K cap on my line - they did so on 7th Oct in order to stabilise the connection - is the DLM a person or equipment ?

Yes, that's right. Perhaps you missed my comment in an earlier post to this thread -- Barney, another TT forum member advised me to lock that thread to stop any further nonsensical, argumentative and unhelpful posts being to it. Apparently an OCE, forum moderator or myself are the only people able to post to (or unlock) that thread.

DLM -- Digital Line Management; a "thing"; monitoring and controlling software.

EDIT: Now that you have been able to register as a TT forum user, once you have provided all the background information to the OCEs they will do their best to resolve this problem.

[jeffbb]

Hi
quote (where'd you get that from ?)
got the distance from here  using your attenuation figures .

I am amazed that any ISP would quite happily slap a 256K profile on a line that is capable of synch at ~20000 Kbps on ADSL2+ .It should be obvious that there is some serious problem .
Just to recap
Do you get this problem whilst connected to the test socket ?
Have you tried quiet line test during a bad episode as seen on the  RS graph .
Can you superimpose the Synch graph on the SNR margin Graph ? to confirm the actual loss of synch. Because  you are capped it will re-synch at same speed .
quote from earlier post
** You may have an option to superimpose the synch rate date onto the SNR margin graph .Right  click on the SNR margin graph and see if you have the option.
Continue monitoring both the snr Margin and the synch rate the graphs times may reveal a pattern.

Regards Jeff

[sparklyman]

burakkucat - thanks for the info - I didn't see that earlier (probably speed reading)

Jeff - The cap was made on 07/10/11 by TT - last engineer I spoke to (yesterday) advised they had done so becuase they had tried every single "profile" without stabilising the line, the 256K was a last resort to acheive stability (which it has done, although not completely....still dropping out)

I've been in the test socket for the last 7 weeks...

I haven't tried the Quiet Line Test during a drop out - I will next chance I get...

Routerstat LITE won't allow me to superimpose anything. I have been capturing the sync graph at the same time as SNR - I haven't been posting it though, as it's a straight line @ 256K....except for when I get the drop-outs - these coincide  exactly with the drops on the SNR graph. I'm now sampling every 60 second in order to get a 4hr capture per page. I'll try to keep my PC running tonight and get a lobger capture.......QUBE have re-schedule visit for tomorrow 3-6pm.

[burakkucat]

As you have a fully LLU'd line, you won't be able to do a quiet line test via 17070 option 2 -- the best you can do will be to key in one digit and listen to the line before the exchange equipment (waiting to receive a full subscriber's number) times out.

[sevenlayermuddle]

Hi again Sparklyman, good to see you're getting some interest on the TT forums.

But I'm a bit confused, as they have homed in on your quoted upstream attenuation of 73.5dB.  That figure is, indeed, very odd and worthy of investigation.   But I couldn't understand how I'd overlooked it in my own comments, so I checked your earlier post and you quoted...

Line Attenuation      
up    10.0
      
down    19.0

So, please can you clarify... is the upstream attenuation perhaps flailing around between two extremes?

As for today's reported SNRM of 14/15 dB, it would only be 'normal' if it co-existed with a 'normal' connection speed.  That's why I said yesterday that the ridiculously high SNRM was 'consistently ridiculous' with your ridiculously poor capped speed.    But it all lends credence to the theory that there's something badly wrong with either your line, the exchange equipment, or your home wiring/router.  You seem to have done all you can to eliminate home wiring and router, so that leaves the exchange or the line.  Hopefully that much will be blindingly obvious as soon as anybody with engineering knowledge gets involved.

[Oranged]

Those latest stats posted on TT forum may indicate false readings from the router......RouterStats influence perhaps if not then a duff router :

No downstream Output Power and a totally false 75.7dB upstream attenuation 

[sparklyman]

...this is very interesting....

The previously posted 10 & 19 Up/down attenuation figures were yesterday, with the previous router.

New router arrived today, installed - I've been monitoring since about 3pm this afternoon and have consistently had 75 & 19 up/down. SNR has averaged about 33

Now, I've had 5 or 6 drop outs so far, the last one was about 10/15 minutes ago when I was picking up sevenlayermuddles post. When I got a connection back, I've cheked the router stats page & routertstats lite application - both are giving the same readings. However, after this last drop/re-connect, the attenuation has reverted to 10 & 19 up/down, and the SNR Margin has now jumped upto 42, now stabilised to about 38 ? Heres a capture;

Oranged - since posting the capture below - connection is now totally haywire.Been monitoring  downstream output, with constant page refreshes, and it remains at 0.0.  ?? See DSL-2689 screen capture

Jeff- plugged phone into test socket during last drop out - dialled 1 - VERY faint background "hiss" - constant, no crackles etc. Only able to listen for 3 secs, then voice advising "sorry, there is a fault". Tried this 5 times, same result each time.

[attachment deleted by admin]

[sevenlayermuddle]

Oh dear life is never simple   

I think Oranged is probably right, the new router is reporting false readings.  That's not going to affect your problems (capped line, and frequent disconnects), but it is going to mess things up a treat if it leads a visiting engineer to jump to entirely the wrong conclusions.

You could just re-install the old router which may make the  line problems easier to diagnose, but that may confuse the TT thread, and could be seen as 'non-cooperation' by the TT support folks you've been talking to.  I think I'd be inclined to add a 'PS' to your TT posts, explaining that the new router may be reporting false stats, and offering to reconnect the old if and when they'd like you to do so.

When new & old routers report entirely different figures it would normally be possible, of course, that the new router was right and it was the old router which was reporting false stats, but I don't think that is the case here.  The stats from the old router seemed entirely consistent, the only mystery being why the speed was so low - and that was explained by the fact it had been capped.  And of course there remains the mystery as to why your line keeps dropping, neither router is giving (or likely to give) any clues as to why that happens.

[sparklyman]

...had 7 or 8  drop outs in the last few hours.

QUBE visit tomorrow.......now approaching 4 months after fault first notified - I'm hoping the TT post will bring a swift conclusion now. Surely this can't go on for too much longer ?

[attachment deleted by admin]

[Oranged]

It might be a good idea to stop RouterStats running as this may well influence the performance of the router while there is an obvious connection problem which is being masked by capping.

We know that it's capped and the capping method may well be manipulation of the Target SNR and RouterStats is not likely to add any more meaningful information.

[jeffbb]

Hi
quote It might be a good idea to stop RouterStats running as this may well influence the performance of the route
As far as I know RS does not affect the router as it is only reading the counters/registers  .
quote :When new & old routers report entirely different figures it would normally be possible, of course, that the new router was right and it was the old router which was reporting false stats
seeing as how the new router now seems to agree with the old one then is it not  also possible that these unstable  readings are another symptom of the problem ?
Regards Jeff