cancel
Showing results for 
Search instead for 
Did you mean: 

Issue after issue after issue...

briarley17
Grafter
Posts: 39
Thanks: 1
Registered: ‎21-01-2014

Issue after issue after issue...

See here for previous issue.
I currently have a ticket open (#8769332Cool, which is stalled at booking a third engineer visit after the second engineer visit suggested a REIN test but never ordered one. Exasperated at the idea of having to book a call out just so they can book a REIN test so I had left it for the last couple of weeks.

The last couple of nights I noticed that drops hadn't been happening, so I checked my router to see the uptime to confirm this and was pleased to be greeted by 3days, 9hours of uptime. However, I was somewhat less pleased by the colossal numbers of errors. (2.3Billion FEC, 17.9Million CRC, 14.4Million HEC and constantly rising).
This seems very similar to an issue I had in May, however the CRC and HEC errors were in 100,000s after a couple of days not 10,000,000s.
The errors are almost halving my line speed currently.
My line speed is set to 11.6Mb/s (as pictured).
By now I realise I'll probably have to try the test socket, then when that fails, raise another fault ticket.
Also, why oh why do you ask for contact details on fault tickets and then ignore them?!
I have done a quiet line test (17070 option 2) and this is fine.
24 REPLIES
Superuser
Superuser
Posts: 12,974
Thanks: 4,269
Fixes: 26
Registered: ‎22-08-2007

Re: Issue after issue after issue...

Hi,
The CRT team have been swamped today.  Your line is a little slow for its attenuation.  If REIN is suspected, you fantastic can get your own perspective on this by downloading RouterStats (full version) and running that to monitor the SNRM.
If you can get it downloaded and installed, I'll give you some configuration suggestions.
In the meantime, I think you should respond to the ticket.  Note even if BTOR identify and locate REIN, they have no powers to address it.

Edit: Correct typos.
Community Veteran
Posts: 5,472
Thanks: 292
Fixes: 4
Registered: ‎11-08-2007

Re: Issue after issue after issue...

Quote from: Townman
get your own perspective on this by downloading RouterStats (full version) and running that to monitor the SNRM.

Here is the link to the "RouterStats" download page - http://www.vwlowen.co.uk/internet/files.htm
Wink
Superuser
Superuser
Posts: 12,974
Thanks: 4,269
Fixes: 26
Registered: ‎22-08-2007

Re: Issue after issue after issue...

Thanks purleigh, I don't have those links readily available when posting from the iPad.
Kevin
briarley17
Grafter
Posts: 39
Thanks: 1
Registered: ‎21-01-2014

Re: Issue after issue after issue...

Townman: Your first post makes a little more sense after realising you were on ipad, "you fantastic get your own" fan is apparently nearer fantastic than can. I've just set routerstats running recording Rx CRC as the user defined graph.
I hesitated to respond to the ticket as the issue seems to have changed. Previously my internet was dropping frequently between late evening and early morning. The drops now appear to have stopped but my CRC errors are now in the millions which is causing my speeds to be around half of what they should be.
purleigh: Thanks for the link, but, having posted on the forums numerous times, I already had routerstats installed on my laptop Smiley
Superuser
Superuser
Posts: 12,974
Thanks: 4,269
Fixes: 26
Registered: ‎22-08-2007

Re: Issue after issue after issue...

Hi,
Had not noticed what a muck-up the iPad's auto correction had done there - sorry!  Embarrassed  Now I'm on a PC, here is a link to my recommended routerstats configuration.  http://community.plus.net/forum/index.php/topic,128606.msg1122269.html#msg1122269
This gives a high granularity of SNRM and allows you to see nasty spikes which might last as little as 30 seconds.  See attached graph showing the effects of passing electric trains.
Cheers,
Kevin
briarley17
Grafter
Posts: 39
Thanks: 1
Registered: ‎21-01-2014

Re: Issue after issue after issue...

Hi, I put in those settings and have left it running for a while.
There does not seem to be anything of particular note.
The CRC errors rise steadily (as do HEC and FEC).
There are no spikes in noise margins?
Rx Noise
Tx Noise
CRC Errors
Superuser
Superuser
Posts: 12,974
Thanks: 4,269
Fixes: 26
Registered: ‎22-08-2007

Re: Issue after issue after issue...

Hi Briarley17,
Actually the is a lot to note - you do have REIN issues - possibly from more than one source!
The step changes in SNRM up is REIN source switched off and down is switching on.  Can you please keep routerstats running watch the times that these changes happen and try to correlate them to something being used in the household.  I'll refrain from saying being switched on / off as step changes such as this can be associated with things being switched out of "standby".  One recent PN customer correlated the jumps to "waking up" his laptop - when the screen came out of screen saver, the SNRM dropped.  This was due to the extra power demands causing the (already switched on) PSU to start to generate RFI (REIN).
Also the expected night time AM/MW interference starts a bit earlier than I'd expect.  What extension wiring have you got please.  How many extensions and how wired - daisy-chain or star?
There is a 5dB variation in the SNRM across the day - one would normally expect this to be less than 2.5 dB.  Looking at the stats and the time of posting, I'm guessing that the last resynch was around 02:30 - 03:00 in the morning, that is during AM/MW night time interference.  In other words synch was established at the worse time of day / night, which explains the elevated SNRM and the consequential synch speed reduction.  There are however other issues which are moving the background noise.  See 19:25 - there is a step increase in noise and a marked decay in SNRM prior to dusk, which is when AM/MW interference starts to bite.
This issue has legs and will run for a while, hope you'll enjoy the challenge.  Though there might still be an issue with the line, until you have eliminated your extension wiring (if any) and household equipment from the equation, calling BTOR could be expensive if the find no fault on the line.
Kevin
briarley17
Grafter
Posts: 39
Thanks: 1
Registered: ‎21-01-2014

Re: Issue after issue after issue...

I meant to put a question mark at the end of the nothing of note line. I'm not knowledgable enough to actually know. Wink
I actually did the old 'tune a MW radio to 612kHz and look for interference' and found that my PC is very noisy, ie. lots of interference, this seems to be focused on the back of the PC so dodgy PSU?
Edit: It would appear the noise behind my PC is actually existent along about a metre of the wall even with PC off, unplugged and moved. I was able to get 'broadband' noises (or kind of like dial-up noise). I'm not sure what this is coming from. I may investigate if I get time tomorrow.
I also noticed a drop in the rx noise graph when I switched it on today, but it quickly reurned to how it was previously.
Clearly this isn't the only source of interference however as the graphs certainly do not match up to my PC on/off cycle. I don't believe the PC is near broadband wiring either, and quiet line test on 17070 is quiet even with PC on.
Struggling to find correlation for points unfortunately but I shall persevere.
I'm 75% sure it's 'star' wired, finding out would involve crawling through very small loft crawl space which I don't fancy right now! There are four extensions I believe, one was removed from our kitchen a couple of years back when it was redone, and one was added down the garden in an outbuilding used as an office.
The first BTOR engineer to come out months back discovered we had terrible wiring and no real master socket.
When we were getting frequent drops at night, the two engineers that came out to look into that came between 1-6pm and on both occasions the line was perfect, although these were not the times at which I was having issues.
Only in the last 5-6 days have the errors rocketed up. Previously, when the connection was dropping at night, there were some errors but not many, in the region of hundreds in a day, which still isn't great.
Also, before the errors, the SNRM was fairly solid at 6dB.
The drops occured even when no one was home, router in test socket, and nothing switched on (other than things like fridges that are always on...).
I'll update the imgur albums with recent captures later tonight.
Thanks for your help and expertise thus far!
Edit #2: Updated graphs.
Rx Noise
Tx Noise
CRC Errors
Superuser
Superuser
Posts: 12,974
Thanks: 4,269
Fixes: 26
Registered: ‎22-08-2007

Re: Issue after issue after issue...

Hi,
These graphs show a similar profile.  As you collect these graphs try to identify reoccurring patterns / times.  Similar shapes from day to day or changes at about the same time point to behavioural patterns.  Look at both Tx and Rx plots.  Note the best period is between a little after 12 until around 14:00.  This shows the highest SNRM (good) and the most flat line (also good).
The night / day variation is over 3dB which is rather on the high side - where are you North or South - continental interference is worse down south.  The star wiring will not be helpful..if you are going to go crawling to take a look, it would be worth reading up on internal phone wiring.  Many installations have 3 wires connected.  Only two are now required, as the bell wire is redundant.  Disconnecting it can deliver marked improvements. Read this thread for reference - https://community.plus.net/forum/index.php/topic,128606.0.html
As for the CRC plots, right click the graph and select 'plot changes since last sample' and change the axis as appropriate.  This will show CRCs as a spike when they occur, rather than as a slowly increasing line.
As a first step to improvement, you need to explore the internal phone wiring.  Do you now have a proper master socket?  If yes, is the router attached to it?
Kevin
Community Veteran
Posts: 19,102
Thanks: 443
Fixes: 21
Registered: ‎31-08-2007

Re: Issue after issue after issue...

Hi briarley17,
Townman sent me a link to your thread. You certainly have a problem there.
You obviously have multiple sources of interference appearing from time to time there, but the effect of these can be grossly exaggerated by the wrong standard of wiring. I wouldn't worry about "hunting for the interference" for a moment - more on that later. As Kevin has mentioned, checking out the wiring is more important. No need to go crawling in loft spaces at the moment, just looking behind sockets will do.
There's a lot of misinformation put around about star wiring. Was it usually far more important is the standard of the wiring and whether the bell wire is connected. In star wiring in one leg of the star just happens to be the "wrong" length, that can sometimes be an issue.
So lets start where the BT incoming cable is. Kevin has asked about your master socket, but is it wired correctly? In other words, when you remove the lower plate to access the test socket, are all the extensions dead?
Look at the extension wiring that's connected to the front plate. Are there wires only connected to terminals 2 & 5 or is 3 (in particular) or any of the others connected as well?
If the wiring here is to an old standard, the wires may be coloured as follows Blue, Orange, Green and Brown.
If it is to CW1308 standard as used these days, then it will be twisted pair cable with 2 or 3 pairs. The pairs will be Blue+white trace & White+blue trace, Orange+white trace & White +orange trace, and the third pair if present - Green+white trace & White+green trace?
Also whilst you have the front plate removes (and everything unplugged, carefully remove the rear plate from the box and look at the incoming BT wiring. Do you see a round black sheathed cable that has 2 pairs Orange and White, also Green and Black. You may also see some yellow strainer wires if it's an overhead cable incoming.?
Also have a look behind all your extension sockets.
Whatever standard of wire you find, in all places remove the wires from all BUT terminals 2 & 5 by gently pulling them vertically from the connector. Be very careful not to pull out the wires from 2 & 5. If you accidentally do that, don't push them back with a screwdriver as it will open the jaws of the connector and give a bad connection. Use something like a pair of tweezers either side of the terminal to push the wire down.
Even with old standard wiring this should give some improvement. Post back with what you find.
Community Veteran
Posts: 38,460
Thanks: 1,031
Fixes: 62
Registered: ‎15-06-2007

Re: Issue after issue after issue...

Quote from: Anotherone
Use something like a pair of tweezers either side of the terminal to push the wire down.
That is what I used but a word of advice or two
Don't use your wife's best tweezers as I found they get upset if you do but more seriously a pair of tweezers with blunt ends rather than the modern sharp ended ones would be better as there is less likely to be damage to the wires
briarley17
Grafter
Posts: 39
Thanks: 1
Registered: ‎21-01-2014

Re: Issue after issue after issue...

Hi all, thanks for all the info and advice and generally taking the time to help a stranger!
I've been away/busy the last few days so no stats (had to take my laptop).
Townman: Before I left I turned the router off and on again and it synced at about 15.5Mbps. This was also probably during that 12-2 period. Which may well be why I seem to now be getting disconnections again. I'll see what happens overnight. The errors are now also right down again. 11,000 CRC errors in the 8 hours the line has been up.
I am in the South, East Hampshire, so possible continental interference? I did get weak signal from a French radio station when doing my 612kHz scanning.
I had intended to make that change on the CRC graph, thanks for reminding me! Wink The gentle slope really wasn't saying much.
We now have an NTE5 socket, newest version with phone and broadband plugs above and below the recessed line. And yes, the router is attached directly.
Anotherone: I'll get the screwdriver out and start removing sockets tomorrow and post back with what I find.
You say "In star wiring if one leg of the star just happens to be the "wrong" length, that can sometimes be an issue." Are you defining "wrong" as being an 'odd' length? So, for example, if there were three 20m cables and a 40m cable, would the 40m be "wrong"?
Oldjim: I managed to read tweezers and think pliers so was curious as to who has 'best' pliers for a second! Hopefully I will avoid pulling any wrong cables at not need tweezers anyway!
Highlighted
Community Veteran
Posts: 5,472
Thanks: 292
Fixes: 4
Registered: ‎11-08-2007

Re: Issue after issue after issue...

When I read the "wrong length" comment,  I took it to mean that if a piece of wire in your phone extension circuit happens (by chance) to be the same length as the wavelength (or multiple) of a nearby interfering signal,  then in effect your wiring can become a tuned circuit wireless aerial - which amplifies the effect of the interference and therefore can have an unexpected but significant impact on your broadband signal.  If the offending wire was a slightly different length (longer or shorter) then the tuned-circuit effect would not be an issue.
Community Veteran
Posts: 19,102
Thanks: 443
Fixes: 21
Registered: ‎31-08-2007

Re: Issue after issue after issue...

Yes that's right purleigh. Hi briarley17.
The frequencies used by ADSL cover from just above audio to above the MW/AM radio band. Most of the "interference" after dark will come from the increased propagation of MW radio stations. In simple terms the "wrong" length acts as a "tuned" aerial more "easily" picking up a strong interfering signal which can make some of the ADSL frequencies unusable. This has the effect of reducing the Noise Margin and if it's reduce enough, you lose the ADSL signal. If the ADSL disconnects and reconnects when there is strong interference present, so that some frequencies can't be used (at all or as effectively) then the connection speed will be slower. Again in simple terms - more frequencies available with higher noise margins means higher speed.
Interference can also come from assorted electrical/electronic equipment as already mentioned. HTH.
I really wouldn't worry about the lengths, getting the standard of wiring correct to minimise unwanted pick-up is the most important thing to start with. Because of the way the "bell-wire" is connected to one leg of the phone line (via a capacitor) it acts as an aerial.