Internet connection keeps dropping

[#1]

@pnf,
I suspect that there are two sources of REIN here - one causes the big SNRM shifts another causes the smaller one.

@DCT,
Any suggestion on what might be done to look at this from a line quality / susceptibility to REIN perspective?

Kevin

[#2]

Events today are similar in principle to the previous 2 days, just the timing of some of the events are different.  This indicates to me that whatever is causing this is not on an automatic timer but is either controlled manually or by some other, non time related event. 
For the last 3 days:
- a 6dB drop has occurred between 07:56 and 08:16. 
- a 6dB / 7dB rise has occurred at 10:55, 10:53 and 15:38.
- a 2dB drop has occurred between 17:05 and 17:14
- a 6dB / 7dB drop has occurred between 19:45 and 19:52
- a 5dB rise has occurred at 01:27, 20:12 and 19:52
So far then, the 'on' events, causing the drops in SNRM, are more predictable than the 'off' events ('on' and 'off' as in some device being switched on or off).  As you say Townman, it does appear that there might be two things going on here.  I suppose time will tell.
Since I changed my router and since the PlusNet team put my line on a 6dB profile and activated interleaving, the line has been stable, for which I am grateful.  When the line kept on dropping it was very disruptive but now it is usable.  Obviously this has come at a cost, namely the speed of the connection, which has dropped from around 13,000 kbps to 8,800 kbps.  So it would be nice to get that speed back at some point in time.
Today's graphs are attached.

1st May 2014 21:15
---------------------
Link Information

Uptime: 1 day, 12:51:38
DSL Type: ITU-T G.992.5
Bandwidth (Up/Down) [kbps/kbps]: 444 / 8.855
Data Transferred (Sent/Received) [MB/GB]: 39,17 / 1,35
Output Power (Up/Down) [dBm]: 12,2 / 20,6
Line Attenuation (Up/Down) [dB]: 18,4 / 34,0
SN Margin (Up/Down) [dB]: 20,2 / 10,4
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): 12 / 1
FEC Errors (Up/Down): 2.844 / 4.006.755
CRC Errors (Up/Down): 0 / 14
HEC Errors (Up/Down): 0 / 145

[#3]

Here's a graph of your connection:
<img src="http://community.plus.net/visualradius/generated/image1399023934833.png"/>
The latest drop would have been me testing your line - apologies! I can't see any evidence of REIN as there doesn't appear to be a particular pattern of drops. The line is now stable following Interleaving being applied and the SNRt set at 6db - this is sometimes all it takes as line conditions from time to time can change.

[#4]

Chris,
If this is not REIN, what other characteristic might give rise to such marked step shifts in SNRM, albeit not enough to cause a connection drop, is sufficient to impact data through put speeds?
Kevin

[#5]

Well, it could be many things, that's a very open question to try and answer.
It could be a fault on the Local Network, so, for example, if there's a joint that needs to be re-made it could be that that's causing the issue. It could be the port itself that needs resetting or even re-seating. There are quite a lot of different things that could cause it - it's difficult to try and identify what's causing such a large amount of drops so sometimes it can be a balancing act of getting DLM or Interleaving to control it or if it's a fault that's likely to deteriorate the service negatively and DLM can't do anything to resolve it, then that's when an engineer visit would be looked at.

[#6]

Chris,
The disconnects seem to have settled down on this line, which could be down to dryer weather and the possible joint issues your referred to being less troublesome presently.  A higher target SNRM can mitigate such issues albeit at the expense of raw synch rate.
There is now no more help I can provide to the OP.  The OP had referred to an open fault report; is this likely to lead to a BTOR visit to eliminate the possibility of those problem sources mentioned above... or is this now on a wait and see status, because if a fault is not manifest, it cannot be found and rectified?
Kevin

[#7]

If the line has stabilised and continues to do so, no further action will be required.
If the line does start to drop out again or anything of a similar nature, then we can take another look at things and potentially look at getting an engineer out, we deal with such issues on a case by case basis.

[#8]

Thanks for a view of that chart of my line Chris, it looks pretty horrific
I'm beginning to think that something may have gone on at my local exchange. Prior to all of this happening, on a 3dB profile I would get, at best, a sync speed of  13,400 kbps.  But I've noticed that after the problems started, on a 3dB profile the line would sync at up to 14,700 dB and on a 6dB profile I was getting close to 13,000 kbps. Today the router resynced at 10:44 and at 6dB the sync speed was 14,700 kbps.  Then at 11:16 the router resynced on a 3dB profile at a speed of 15,919 kbps.  I've never had that kind of speed on my line before.  With all of the BT Network Maintenance that's been going on throughout April could BT have done something to cause this?
Everything is currently going pear shaped again.  CRC errors are currently accumulating.  See graphs.

[#9]

Yes, we're seeing more disconnections too and the errors have crept up. For some reason, even though your SNRt is set at 6db, your connection is having a go at connecting at 3db again.
This is back with the Faults Team anyhow and further testing is required.

[#10]

Thanks for your help Chris.  The faults team have arranged for an engineer to visit on 7th May.

[#11]

Hi,
What is the sampling interval on router stats?
There is something not right about the CRC plots.  I'm guessing these are change since last sample, in which case negative plots mean that "this" sample reported zero (or got no response) thus the change is negative. If the next sample gets a correct response then you get a positive spike.  Be warned test tools can mislead if there are errors in the sampling method.  This tool does have a couple of gremlins.
However the story told by the SNRM plot is quite clear, repeated resynchs (reason not clear) at lower SNRM / higher synch speed until it coughs with a noise event (SNRM line hump).  Something is not quite normal here.
Keep monitoring and have the interesting graphs printed ready for the BTOR visit, especially the high to low SNRM transitions happening at predictable times.
Kevin

[#12]

Apologies for my late response, the Bank Holiday intervened.
The sampling interval on RouterStats is 10 seconds.
With respect to the CRC plots, I think you are correct, the 'spikes' that report those large positive and negative values occur when the router resyncs and when a lot is going on in a very short period of time and RouterStats is attempting to make sense of it all.  I rescaled the plot so that the values could be seen, but whenever those 'spikes' occur it just indicates that the router has resynced.
I'm about to print a selection of interesting graphs for the BT engineer!
I've spent the last hour analysing the key 'events' from 14th April to date.  The attached PDF shows the results.  Although there is no definitive pattern of events, there are clusters that occur regularly between 10:00 and 12:00 and again between 17:00 and 18:00, with random events in between.  It's definitely not normal. 
Thanks again for your help Kevin.

[#13]

If you can get access to Tones, Qlog and Hlog what do they look like ?
To me it seems like it could be a high resistance fault personally as this is sounding like the issue i had to begin with as well.
Do you have any noise on the line or dial tone not present when this happens ?

[#14]

Hi pnf, I've been having the same problems for the last 3 months! 5 engineers visits 2 " no shows" and still the connection drops for differing periods of time. Can't get PLUSNET to get their act together.
Still waiting for yet another OR engineer to come. I was also told that it might be a REIN problem - but frankly, I think they are clutching at straws.

[#15]

A very helpful BTOR engineer came along this morning and ran tests on the line.  Most of the tests were OK with green ticks showing up on the test device but there were thousands of FEC errors with only one CRC error showing.  The engineer replaced the BT faceplate on the master socket with a BTOR faceplate and then went off to the exchange to move my connection onto another port.  The engineer called back later to say that the equipment at the exchange had tested OK but that there were still errors on the line which he had seen and logged.  He was therefore going to pass the problem over to the REIN team.
@kenneth50,
I've attached a couple of 'Bits/Tone' charts taken earlier today (pdf file).  I don't know what Qlog or Hlog are.  There is no noise on the line at all.  I have carried out quiet line tests both when the problems occur and at other times and the line is always quiet.  The dial tone is always present too.
@diamonddave51,
I hope the REIN team can locate the problem, I don't like to think of having this going on for 3 months!  When the line keeps dropping it is really disruptive.  I have to say that I have found the PlusNet staff very helpful and I think the help I've been given on this forum has been invaluable. Thanks everyone!  I'm sorry to hear you are having so much trouble and I hope you get it sorted soon.