Print

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

[ColinS]

Just when everything is going fine  , the fickle finger of fate  reaches down to show you who's really in charge!

So, this morning my line's cruising along quite happily with attainable at ~91372/31076 (and capped at 80/20).
09:15 - it must have had an extra weatabix for breakfast this morning and decides to up the ante to ~95704/31796!  wooeee!
by ~09:50 it's got bored with that and reverted to ~90688/30901.  It hangs about at those rates (give or take a little) until ~11:17 when it suddenly feels all faint and drops to ~82476/29858. oh  missus!
11:20 Aargh! Resynch & comes back at ~36167/19582, dropping to ~25324/19322 a minute later  Crickey!
11:22 Resynchs again and comes back at ~87820/38732 from which it slowly climbs back to ~89194/39595, as I write this!

Please see the attached, and note the significant increase in REIN intensity at the 11:22 resynch, and the 'dent' that fate has chosen to put in my HLog!

I'm thinking of powering it off for half an hour to let the DSLAM have a bit of a rest from all this excitement, and see if we can't get a proper Yorkshire resynch this time.   

[burakkucat]

I am a firm believer in powering off over night (and disconnecting the powered off modem from the line), if one thinks that a period of DSLAM relaxation is desirable. 

[ColinS]

I am a firm believer in powering off over night (and disconnecting the powered off modem from the line), if one thinks that a period of DSLAM relaxation is desirable. 

Good advice, thanks.  8)

[ColinS]

 

[Bald_Eagle1]

I wonder, could this have been caused by an engineer at the cabinet, possibly setting up a new service for another user?

Possibly even a pesky cat clawing at the newly installed dongle, thinking its kind owner had provided him with a new toy.

I have recently read somewhere of some house pet causing disconnections by leaning against the cable connection into the master socket.

I did see a 4 day total loss of service on my connection along with 'disturbances' whenever engineers climbed to pole where the DP is mounted (to work on other users' connections - possibly telephone rather than broadband related).

The DP issue turned out to be a very poor joint from the underground cable.
The engineer told me that it just fell off in his hand when he went to check it (some 3.5 months after I had reported the issue).


I never really found out what caused the 4 day loss of service, but my suspicion was/is that the connections had been disturbed at the PCP.

[edward]

I am a firm believer in powering off over night (and disconnecting the powered off modem from the line), if one thinks that a period of DSLAM relaxation is desirable. 

What would be the effect of doing this when away on holiday, would it have much influence on the profile if you disconnected for a couple of weeks, or longer?

[ColinS]

I wonder, could this have been caused by an engineer at the cabinet, possibly setting up a new service for another user?

Possibly even a pesky cat clawing at the newly installed dongle, thinking its kind owner had provided him with a new toy.

I have recently read somewhere of some house pet causing disconnections by leaning against the cable connection into the master socket.

I did see a 4 day total loss of service on my connection along with 'disturbances' whenever engineers climbed to pole where the DP is mounted (to work on other users' connections - possibly telephone rather than broadband related).

The DP issue turned out to be a very poor joint from the underground cable.
The engineer told me that it just fell off in his hand when he went to check it (some 3.5 months after I had reported the issue).


I never really found out what caused the 4 day loss of service, but my suspicion was/is that the connections had been disturbed at the PCP.

Sorry Kitizens - no cats, kids or dogs around here However, (no disrespect to BTOR but), I do have a sneaking suspicion like you that "somebody's been sleeping tinkering in my cabinet/joint box/DP/etc." 

There is low-level audible noise on the voice QLN test this morning, and a wetting current from using the phone gave some relief on SNR levels.  It's not bad enough (yet!) to go through the necessary hassle to get a voice engineer out, especially when the line is currently 'in-between' transfer from BE to Plusnet.  I'm hoping it will just settle down again after a bit. 

But the ever-watchful eye of ongoing stats will see all! 

[burakkucat]

No metallic pathway should suddenly 'decide' to change its attenuation. Of all the parameters for a pair, that should be essentially constant! 

I suspect: (1) a defect in the metallic pathway coupled with (2) Openreach 'fiddling' with another pair (up the pole, down the hole or in the cabinet). 

[burakkucat]

I am a firm believer in powering off over night (and disconnecting the powered off modem from the line), if one thinks that a period of DSLAM relaxation is desirable. 

What would be the effect of doing this when away on holiday, would it have much influence on the profile if you disconnected for a couple of weeks, or longer?

If you think a disconnection will be beneficial, then the longer it is left disconnected, with the DSLAM port relaxing, the better. 

It will have no effect on the VDSL2 profile . . . it is 17a, no 'ifs or buts'. However the DLM would have had a chance to record a constant 'no signal present' event for the duration.

[ColinS]

OK, there's definately been a voice fault on my line since the first signs on Saturday reported above. Problem is it's intermittent. 
I powered the modem off on Saturday night, and didn't power it back on until late Sunday night as I had been attending a wedding (no,not mine!)
This morning a voice call I made had very noticeable interference, to the extent that I had to apologise to the other party.  And, in current stats there have been no less than 13 resyncs in the 2hr period between 11:29 and 13:28, after which I switched the modem off in the hope of preventing DLM from taking action (probably already too late by the time I discovered it), and looked out the 'wired testing phone' that we have all purchased and keep specifically just for such testing purposes (don't we  ?)
Plug the phone directly into the NTE5 test socket and of course it's gone all quiet again!  Repeat the test several times while restoring the setup back to 'normal' and of course it remain quiet. Which it will do probably until tomorrow morning, I guess, when I shall be hot-foot back to the test socket with the phone ready to phone ... um who, exactly, I'm not quite sure at this point ... and hopefully will get the fault reported while it's still noisy.
Why I am unsure is that Plusnet sent the takover request for the line to BTOR on 1 Apr, but the line won't actually be transfered until 16 Apr (why for a database update is beyond me).  BTOR notified BE, who believe that Plusnet now 'control' the line, even though they will continue to bill me until then for it, both rental and usage; but I suspect Plusnet might say it's not theirs yet.  So hopefully the noise will hang around long enough for me to try to report it first to BE, and then Plusnet if necessary, and possibly BE again, I don't know! 

[Bald_Eagle1]

Presumably the broadband is being provided by Plusnet, regardless of who provides the telephone part of the service at the moment.

Do all these resyncs also show in the ROUTER's log?
If not, you may have the devil of a job to prove they actually occurred, despite having your own evidence.

If they do show in the ROUTER logs, they should also show in Plusnet's Radius log, a copy of which is usually provided in any fault tickets, or on request.

If you do receive a copy in a fault ticket, I would advise that you save it for future reference as Plusnet's graphs do expire, as I have noticed when looking up old fault tickets.

I'm not sure how switchover from one supplier to another works, but maybe some cabling needs swapping from one piece of equipment to another?

Maybe BS could explain in more detail if he if he happens to wander through this thread.

If that is the case, I suppose it could be that the cabling has already been swapped & a now slightly less than 'clean' joint is causing the current issue(s), particularly if it is being slightly disturbed by 'other' work going on in the immediate vicinity - hence it being intermittent.

My first port of call regards reporting this would be Plusnet.
At least they MAY be able to confirm the broadband disconnections etc.

[ColinS]

The DLM has now sat on me (at 05:56am), as a result of all these (~24) disconnects yesterday associated with voice-band noise. 

Attainable 91.5/34.5, Actual 35/20    SNRm now up from ~9dBto 22.5dB DS, and down from ~15dB to 13dB US.  Wouldn't mind so much but I rarely use the US! 

Sigh!  Until the fault worsens and/or remains on long enough to allow a BE CS rep to hear it, I guess I'm just going to have to tough out the next 7 days until the landline transfers to Plusnet, and then since my actual DS is now 50% of my estimated (nevermind recent actuals), I can get Plusnet to send the SFI out.

Thanks to all for their interest and help.   

[Black Sheep]

Presumably the broadband is being provided by Plusnet, regardless of who provides the telephone part of the service at the moment.

Do all these resyncs also show in the ROUTER's log?
If not, you may have the devil of a job to prove they actually occurred, despite having your own evidence.

If they do show in the ROUTER logs, they should also show in Plusnet's Radius log, a copy of which is usually provided in any fault tickets, or on request.

If you do receive a copy in a fault ticket, I would advise that you save it for future reference as Plusnet's graphs do expire, as I have noticed when looking up old fault tickets.

I'm not sure how switchover from one supplier to another works, but maybe some cabling needs swapping from one piece of equipment to another?

Maybe BS could explain in more detail if he if he happens to wander through this thread.

If that is the case, I suppose it could be that the cabling has already been swapped & a now slightly less than 'clean' joint is causing the current issue(s), particularly if it is being slightly disturbed by 'other' work going on in the immediate vicinity - hence it being intermittent.

My first port of call regards reporting this would be Plusnet.
At least they MAY be able to confirm the broadband disconnections etc.

Yes, when switching providers new 'Frames Mapping' is given to the Exchange guy, and he will run new 'Jumper wires' to the new equipment on the MDF.
They (we) get a 10-point defect if the connection is not soldered correctly (assuming it is a soldered tag. It may be an IDC connection block). That then means we will be shot at dawn, so is not a favoured practice. Other faults on the MDF can happen through 'Jumper wire burn', whereby older circuits lying lower down in the mass of jumper wires, can get their insulation stripped when redundant circuits are being recovered. There is no easy way to avoid 'Jumper burn', even pulling the redundant wiring out very sloooooowly can cause friction burns, and subsequent faults.

Just info on how it works as requested, have not got involved in the OP's issue.

[ColinS]

[Just info on how it works as requested, have not got involved in the OP's issue.

Thanks BS.  Succinct explanation.  That gives everyone time to avoid those things as well (if possible) when it actually happens ... which is not for another 7 days yet! 

However, perhaps, you, or B*Cat may care to venture _any_ kind of explanation as to what kind of line fault can present itself such that we get a perfectly FLAT SNR line across tones ~260-500?   

B-E himself assures me that it is not a graphing artifact but rather an SNR at this point( of last sync) of 26dB but no bitloadings at all on those tones.  This and the kink in Hlog (see B*Cat's exploded view above) leaves me wondering if something in the cabinet has gone  ?

[burakkucat]

Sorry Colin, I don't have any idea as to the cause of those graphed anomalies. 'Pandemonium' over a small section of the Hlog graph is something that I have only recently seen once before, as is the 'flat lining' in an otherwise healthy SNR graph. 

What I would be interested in seeing is a TDR trace, looking from your NTE5/A towards the DSLAM (and, thus, exchange battery) whilst the 'ringing' voltage is applied to the line. From the intermittent noise during telephony calls and the 'skitter-ish' effects shown in some of your graphs, above, I am fairly convinced that you have a HR or semi-conducting joint in the pair.

I don't suppose you have a 'Mole' or a 'Hawk' tucked away or could borrow a either a JDSU or an Exfo and thus perform the test, above?