Getting worse after support ticket raised

[#1]

Keep running the stats - you need to watch that SNRM to see if there is any time consistency on tge events.

[#2]

Yes I am monitoring it...
Unfortunately it is not time consistency but randomly the connection drops
I have got about 6-20 times reset / resync after a short period of time from restarting the router, looks like it is communicating for the optimum speed....but it did not happen before the engineer changed a couple of cables.
I am a little bit concerned of the quality of my current cable which just been switched...
I will keep on monitoring and hope the engineer visit on Friday this week would solve it.
Thanks.
Regards,
Chong

[#3]

Keep looking!
You certainly have REIN, a different circuit might reduce the susceptibility to REIN but it will not eliminate it.  Whilst the SNRM does crazy things as per the plots you posted, you will continue to have problems.  REIN is a real pig to find... Wandering around with a portable AM radio tuned to 615kHz might locate it.

[#4]

Thanks Kevin.
I'm not quite sure how the REIN would work with PN to be honest...
Is it possible if I use the radio to find the location of REIN then pass to PN and they would send someone to fix it?
Otherwise I don't think it is worth to detect and had nothing done to improve....
Regards,
Chong

[#5]

Chong,
Have a read about REIN here - http://www.kitz.co.uk/adsl/rein.htm
It is not something to be ignored if you want the best from your broadband and you are suffering from it.  That SNRM plot suggests you might be.  REIN can only be 'fixed' by the owner of the equipment causing the issue - that might well be YOU! 
Kevin

[#6]

Kevin,
thanks for your info
I do monitor it, and found it happened again this morning at around 4-5 am.....
The CRC errors and FEC did vary a lot, and SNR went up and down for 1-2 hours before it actually became stable at only 224kb/s
However I don't think it is because of REIN, as from the link you gave me it looks like REIN is repeatable, and most of the times it should be caused by domestic appliances or obvious street electricals, of which would normally between 6am to 12pm. In my case the fault was happened in early morning so I am guessing it is not because of REIN?
Anyway, my fault is totally unpredictable in terms of times.....thus makes it more difficult to locate and fix, I understand the engineer but still hope they could fix it....
Anyway, it seems that after the engineer changed the cables it is getting worse, this kind of very slow speed happened twice within this week... 
Thanks anyway for all your help.
Regards,
Chong

[#7]

Hi Chong,
The characteristic of REIN totally depends on the cause.  Some causes are very much what you suggest, some are not.  Anything electrical could be the cause.  I previously had a problem in which the SNRM plot could be used to monitor railway traffic timeliness!
The problem sounds so bad that it knocks the speed down to a crawl to try to get some level of stability, after which the SNRM will rise to quite a high figure when the REIN has gone.  Do all of the plots have a similar shape?  Posting a few here would be helpful.
It rather sounds as though the 'new' line is more susceptible to interference.

@CRT, What is the line's balance characteristics please?
Kevin

[#8]

Thanks Kevin
Finally another OR engineer came, and no surprise he could not detect anything wrong with my line....apart from SNRM of  which he thought it is too high, but I'll be happy with it if the line is stable then.

[#9]

A high SNRM will give more stability, but at the expense of speed.  Whilst the interference persists you are not likely to get what you want - higher speeds with stability.
Note that if the target SNRM is high then there is a good reason for that - the DLM is trying to deliver stability.  If it's just the current SNRM which is high, that implies the the interference present at the time of the last re synch has now gone.

[#10]

Thanks Kevin.
I am still monitoring my line and it seems good in terms of noise.
However I found a lot of minor drops of the connection, which is even not captured by PlusNet when they are monitoring as well.
So wondering if you, or anyone knows why this happened, please? There is no sign of high errors when the connection dropped but just dropped....
Sometimes it just seems that the line is quite strict with the 6db SNR margin, as it went to about 7db the connection resets and the speed increased a little bit and SNR margin went back to 6db again.
I'll post some pics here later today when I got home.
Thanks a lot.

[#11]

Hi Chong,
What do you mean by "minor drops of connection" - exactly what did you see?  There are occasions when programs (for example IE) report connectivity issues, when there is not a synch / PPP session connectivity issue.  In such circumstances there will be no indication of an issue in PlusNet's PPP session logs.
Conversely, PlusNet's PPP session logs can be used to incorrectly infer a loss of synch.
For me when looking at such issues, the router log indicating a loss of xDSL or PPP are the only true indication that the link has been lost.  Other symptoms might point to transit issues as opposed to connectivity issues.  Some such symptoms are local monitoring programs not receiving a response from the router appearing to indicate that the link is down.  Routerstats is quite prone to that with the TG582n router.
There is some suspicion that a lot of noise on the US can result in a loss of synch WITHOUT a loss of the PPP session.  Firm evidence of this though has not been found.
The line is not exactly strict with the 6dB SNRM, it is the way the system works.  SNRM is a MARGIN metric; each time the link resynchs the router and exchange will establish a connection at the synch speed which at that time delivers a TARGET margin of 6dB above the then present noise.
Thereafter the synch speed will remain consistent until some event causes a resynch.  During that time, the reported margin will vary depending upon the changing noise conditions.  Relative to the noise present at the time of the resynch, if it decreases the SNRM will go up, if noise increases the SNRM will fall.  If it falls too far, the link will drop.
Continental radio interference can cause a 0.5dB to 2dB lowering of the SNRM during the hours of darkness.  Thus if you resynch during the evening, you will see a slightly slower speed, then during the day the SNRM will rise as you have observed.

Hope that helps answer your questions?

[#12]

Thanks very much for the reply Kevin
What I mean by the "minor drops" is that, I can see drops from the DSLStats's results, but the result from PlusNet's monitoring suggests I am online consistently..... I'll post some pics later today when I got home to assist my explanation.
Also, for the SNR margin, what I mean is, because I cannot see some obvious cause for the connection drop (e.g. high CRC errors), the line seems to adjust itself to match the target 6db. Other than that I cannot find a reason why the connection just reset itself TBH.....
I have also tested this on a different router (current one is TG582N) which is ZyXEL SBG3300, and for this router it is even worse that, during the re-connection period of my line, the router is struggling to get back to a DSL connection....
Again, I will post some pics later today to assist my explanation.
And thanks for your time.
Regards,
Chong

[#13]

Please see the attachments for what I mean. I have highlighted the time that speed varies thus disconnections, and CRC erroe low while SNR margin seemed to maintain to a certain level (6db).
I have also attatched a screenshot of PlusNrt's monitor result which suggested I am online at that time

[#14]

Just to be clear here, are you plotting DS max rate or synch rate?
There is clearly a small variation of the speed plot, but no clear evidence of a resynch.
What is the sample interval?  It is possible that the sample interval is greater than the time it takes to establish a resynch and therefore is not seen clearly on the graphs.
Is there anything in the router log which suggests a resynch took place?

[#15]

Kevin,
I think so.... please see telenet data from DSLstats, as well as screenshot directly from router's management page, suggested 29 resets.
The connection is really lost because when this happened I cannot log on to any of the website while downloading stopped as failed....
It is the sync rate as max rate is not available for TG582n...
The sample interval is 20s