Print

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

[raislander]

Hi guys, hope you can help me understand what's going on with my line.

I recently was upgraded to BT uncapped speed for Infinity 1, I know my line can't go further than 63-65M sync, because of distance and other factors, it was synching ok at 65M, however I had a resynch coming from the DLM and since then I've been checking the ES/CRC and it's weird. I have no errors whatsoever from 7AM till 4PM, but then ES/CRC spike again from 4PM until 7AM, and it's been following that pattern for the last few days.

I have checked some interference from electronic devices, but heating is off at 11PM, and there is no electric elements running at home during those times. I thought it could be crosstalk, but it's weird that the errors keep running even after 12AM-1AM, so do you know what could be causing this? And more importantly how I could avoid this from happening?

Thanks a lot for the help.

[burakkucat]

From the timings that you have mentioned, it appears that your circuit's problems appear to correlate with the hours of darkness.

Do the street lights remain on for the full duration of the hours of darkness or do they go out around midnight? Is one of those street lights faulty? Flickers? Fails to come on?

Do you know the route that the cable takes from your home to the pair of cabinets?

It might be worthwhile walking the route during the hours of darkness and looking for anything electrical that could be responsible.

[kitz]

Havent done a full stats analysis, but the Err/Secs are within normal limits..  and about on par to what I see on my own line on a 'good' day!
In fact you're quite lucky not to get any at all for some periods of the day.  Traffic lights are all green.  Night time errors could even be atmospheric RFI

Ive had a look at SNRm which seems a wee bit noisy.  I was thinking from the 7 day graph that it could be more like SHINE rather than REIN..   but when I zoom in on what I thought looked like what may be a particularly noisy period it doesnt really show that much that causes me concern.   
Its not crosstalk...  it could just be that since the upgrade speed you are pushing you line to its maximum capabilities and therefore it is usual to see more errors.

Best monitoring your SNR Margin for a period of time as I'd expect that to give more of an indicator that something was wrong.
 

[kitz]

From the timings that you have mentioned, it appears that your circuit's problems appear to correlate with the hours of darkness.
Do the street lights remain on for the full duration of the hours of darkness or do they go out around midnight? Is one of those street lights faulty? Flickers? Fails to come on?
Do you know the route that the cable takes from your home to the pair of cabinets?
It might be worthwhile walking the route during the hours of darkness and looking for anything electrical that could be responsible.

I don't hold out much hope of spotting something causing 20 Err/Secs ph. 
Cant find any odd patterns in FEC or LEFTRS.    There was a CRC spike several days ago, but again 600 isn't particularly high.

One thing that does look strange is QLN.    Theres some x-talk in D3, but its D1 that isnt very pretty.  Hlog is fine.
The fact it ceases after ~ 1650 kHz makes me wonder about RFI.     What do you think?


@  raislander

Do you happen to live near any Radio Transmitters?
http://www.kitz.co.uk/adsl/rein.htm#MWradio
 

[kitzuser87430]

Wonder if it the high pressure that affected freeview over the last few days?

[tbailey2]

That's what it says on the BBC web site:

Freeview blames air pressure for TV disruption

Freeview has said that high air pressure is the cause of disruption being experienced by some of its users in England and Wales.

[tubaman]

I suspect the weather may be doing something as I've had a few days in the last week when my US ES are way higher than normal.
Today I've so far got over 1600 US ES where I would normally get about 150 - 200.
DS is keeping well in check, but it's probably G.INP doing that.

[kitz]

From the MDWS stats I looked at, I couldn't see anything particularly wrong with the line other than what could be possible RFI/EMI.

Admittedly I at first thought there may be some SHINE going on, but when I zoomed in there wasn't anything and I think its just the way MDWS compresses average stats  (Thats not a complaint @ Tony, we dont have a long term period of stats to look at and I'm quite happy at the way things work).
Looking at the SNRm, it could be EMI/RFI and with the errors being worse at night time, why at first I thought of atmospheric EMI.... because from what stats we did have it didn't show me anything REIN worthy.


THEN when I later saw the QLN affected only up to ~ 1650 kHz it rang alarm bells for RFI.   I haven't analysed each tone, but noise in that particular area is usually from MW transmitters or Radio Hams.   
B*cat is the resident expert as tracking down RFI to location by analysing the individual tones & whilst it is something I could do... Im afraid its not something I have patience for as it can be very time consuming. 

OTOH b*cat has a lot of the many MW frequencies stored in memory; So much so, that he is practically able to pinpoint location of EU from which tones are affected...  and its why I threw it back at him to ask his thoughts on the probability of it being RFI.

Need the OP to also confirm if he lives near to any transmitters.

I wasn't aware of the current atmospheric issues affecting transmissions, so thanks to KU87430 for bringing that up. That obviously wont be helping a line showing some sort of noise at <1700kHz    

[WWWombat]

I saw this kind of pattern with errors on my Mum's ADSL2+ line. It started up around 4pm, sometimes triggering resyncs.

That was on an extension (with a blanked master), with a ring wire connected, and a poor filter. Sorting out the wiring helped a lot, suggesting a good chunk was RFI induced into the circuit. Setting the modem to ADSL2 (not 2+) also improved things.

[konrado5]

raislander: could you attach your Hlog and QLN?

[burakkucat]

B*cat is the resident expert as tracking down RFI to location by analysing the individual tones & whilst it is something I could do... Im afraid its not something I have patience for as it can be very time consuming.

One really needs to have sight of the raw QLN data and yes, it is very time consuming . . . though quite satisfying to successfully geo-locate a member from the data.     

OTOH b*cat has a lot of the many MW frequencies stored in memory; So much so, that he is practically able to pinpoint location of EU from which tones are affected...  and its why I threw it back at him to ask his thoughts on the probability of it being RFI.

Need the OP to also confirm if he lives near to any transmitters.

I've only just taken a look (via MDWS) and the FEC plot shows that the error detection & correction process is working hard!

As for the QLN plot, I have two observations --

• the broad bases to some of the higher frequency peaks hints of a quite strong signal strength, possibly overloading to analogue front end at the respective frequencies
• the peak resulting from the Droitwich Transmitting Station (198 kHz) is minuscule in comparison to that of the highest frequency (1583 kHz) present

Both of those observations, when taken in conjunction, hint at a non-U.K. located member! However there are areas in the U.K. which could give rise to such a QLN plot.

[burakkucat]

raislander: could you attach your Hlog and QLN?

K5, please make use of MDWS to view those plots.

I appreciate that you have had trouble in trying to use MDWS but I'm sure if you persevere -- and ask Tony for some assistance -- it will ultimately work for you.

[tbailey2]

@kitz and others

Although you can't zoom in enough to see much better than about every ten tones (I'll see if can improve that), all the data is actually there that you need but it's impossible to fit up to 4000 bits of data in a plot around 900 pixels wide!

EDIT:
It will now zoom out to show every tone on a dragged selection - you'll need to zoom then zoom again  until you get what you want i.e. you can zoom a zoom 

The data below (just a subset around 1.5Mhz shown) is what the graph is created from and it is possible to look at it if you know how (you may already know how of course but there will be others who do not):

Code: [Select]

[326, null, null, null, null, null, -117.1, 'red', 'Tone: 326 SNR is -117.1 dB @ 1405.875 Khz'],
[327, null, null, null, null, null, -115.6, 'red', 'Tone: 327 SNR is -115.6 dB @ 1410.1875 Khz'],
[328, null, null, null, null, null, -115.6, 'red', 'Tone: 328 SNR is -115.6 dB @ 1414.5 Khz'],
[329, null, null, null, null, null, -117.3, 'red', 'Tone: 329 SNR is -117.3 dB @ 1418.8125 Khz'],
[330, null, null, null, null, null, -115.3, 'red', 'Tone: 330 SNR is -115.3 dB @ 1423.125 Khz'],
[331, null, null, null, null, null, -117.0, 'red', 'Tone: 331 SNR is -117.0 dB @ 1427.4375 Khz'],
[332, null, null, null, null, null, -115.3, 'red', 'Tone: 332 SNR is -115.3 dB @ 1431.75 Khz'],
[333, null, null, null, null, null, -114.9, 'red', 'Tone: 333 SNR is -114.9 dB @ 1436.0625 Khz'],
[334, null, null, null, null, null, -115.3, 'red', 'Tone: 334 SNR is -115.3 dB @ 1440.375 Khz'],
[335, null, null, null, null, null, -115.3, 'red', 'Tone: 335 SNR is -115.3 dB @ 1444.6875 Khz'],
[336, null, null, null, null, null, -112.9, 'red', 'Tone: 336 SNR is -112.9 dB @ 1449 Khz'],
[337, null, null, null, null, null, -108.3, 'red', 'Tone: 337 SNR is -108.3 dB @ 1453.3125 Khz'],
[338, null, null, null, null, null, -88.8, 'red', 'Tone: 338 SNR is -88.8 dB @ 1457.625 Khz'],
[339, null, null, null, null, null, -107.0, 'red', 'Tone: 339 SNR is -107.0 dB @ 1461.9375 Khz'],
[340, null, null, null, null, null, -112.4, 'red', 'Tone: 340 SNR is -112.4 dB @ 1466.25 Khz'],
[341, null, null, null, null, null, -113.9, 'red', 'Tone: 341 SNR is -113.9 dB @ 1470.5625 Khz'],
[342, null, null, null, null, null, -115.4, 'red', 'Tone: 342 SNR is -115.4 dB @ 1474.875 Khz'],
[343, null, null, null, null, null, -114.4, 'red', 'Tone: 343 SNR is -114.4 dB @ 1479.1875 Khz'],
[344, null, null, null, null, null, -115.1, 'red', 'Tone: 344 SNR is -115.1 dB @ 1483.5 Khz'],
[345, null, null, null, null, null, -116.1, 'red', 'Tone: 345 SNR is -116.1 dB @ 1487.8125 Khz'],
[346, null, null, null, null, null, -117.0, 'red', 'Tone: 346 SNR is -117.0 dB @ 1492.125 Khz'],
[347, null, null, null, null, null, -116.1, 'red', 'Tone: 347 SNR is -116.1 dB @ 1496.4375 Khz'],
[348, null, null, null, null, null, -116.0, 'red', 'Tone: 348 SNR is -116.0 dB @ 1500.75 Khz'],
[349, null, null, null, null, null, -115.1, 'red', 'Tone: 349 SNR is -115.1 dB @ 1505.0625 Khz'],
[350, null, null, null, null, null, -116.1, 'red', 'Tone: 350 SNR is -116.1 dB @ 1509.375 Khz'],
[351, null, null, null, null, null, -115.4, 'red', 'Tone: 351 SNR is -115.4 dB @ 1513.6875 Khz'],
[352, null, null, null, null, null, -115.1, 'red', 'Tone: 352 SNR is -115.1 dB @ 1518 Khz'],
[353, null, null, null, null, null, -114.8, 'red', 'Tone: 353 SNR is -114.8 dB @ 1522.3125 Khz'],
[354, null, null, null, null, null, -114.7, 'red', 'Tone: 354 SNR is -114.7 dB @ 1526.625 Khz'],
[355, null, null, null, null, null, -114.1, 'red', 'Tone: 355 SNR is -114.1 dB @ 1530.9375 Khz'],
[356, null, null, null, null, null, -113.8, 'red', 'Tone: 356 SNR is -113.8 dB @ 1535.25 Khz'],
[357, null, null, null, null, null, -112.8, 'red', 'Tone: 357 SNR is -112.8 dB @ 1539.5625 Khz'],
[358, null, null, null, null, null, -111.3, 'red', 'Tone: 358 SNR is -111.3 dB @ 1543.875 Khz'],
[359, null, null, null, null, null, -101.7, 'red', 'Tone: 359 SNR is -101.7 dB @ 1548.1875 Khz'],
[360, null, null, null, null, null, -111.4, 'red', 'Tone: 360 SNR is -111.4 dB @ 1552.5 Khz'],
[361, null, null, null, null, null, -109.8, 'red', 'Tone: 361 SNR is -109.8 dB @ 1556.8125 Khz'],
[362, null, null, null, null, null, -109.1, 'red', 'Tone: 362 SNR is -109.1 dB @ 1561.125 Khz'],
[363, null, null, null, null, null, -107.5, 'red', 'Tone: 363 SNR is -107.5 dB @ 1565.4375 Khz'],
[364, null, null, null, null, null, -105.0, 'red', 'Tone: 364 SNR is -105.0 dB @ 1569.75 Khz'],
[365, null, null, null, null, null, -102.2, 'red', 'Tone: 365 SNR is -102.2 dB @ 1574.0625 Khz'],
[366, null, null, null, null, null, -97.3, 'red', 'Tone: 366 SNR is -97.3 dB @ 1578.375 Khz'],
[367, null, null, null, null, null, -85.0, 'red', 'Tone: 367 SNR is -85.0 dB @ 1582.6875 Khz'],
[368, null, null, null, null, null, -91.8, 'red', 'Tone: 368 SNR is -91.8 dB @ 1587 Khz'],
[369, null, null, null, null, null, -99.5, 'red', 'Tone: 369 SNR is -99.5 dB @ 1591.3125 Khz'],
[370, null, null, null, null, null, -103.4, 'red', 'Tone: 370 SNR is -103.4 dB @ 1595.625 Khz'],
[371, null, null, null, null, null, -106.1, 'red', 'Tone: 371 SNR is -106.1 dB @ 1599.9375 Khz'],
[372, null, null, null, null, null, -107.6, 'red', 'Tone: 372 SNR is -107.6 dB @ 1604.25 Khz'],
[373, null, null, null, null, null, -109.4, 'red', 'Tone: 373 SNR is -109.4 dB @ 1608.5625 Khz'],
[374, null, null, null, null, null, -110.6, 'red', 'Tone: 374 SNR is -110.6 dB @ 1612.875 Khz'],
[375, null, null, null, null, null, -111.3, 'red', 'Tone: 375 SNR is -111.3 dB @ 1617.1875 Khz'],
[376, null, null, null, null, null, -112.5, 'red', 'Tone: 376 SNR is -112.5 dB @ 1621.5 Khz'],
[377, null, null, null, null, null, -113.0, 'red', 'Tone: 377 SNR is -113.0 dB @ 1625.8125 Khz'],
[378, null, null, null, null, null, -113.5, 'red', 'Tone: 378 SNR is -113.5 dB @ 1630.125 Khz'],
[379, null, null, null, null, null, -113.5, 'red', 'Tone: 379 SNR is -113.5 dB @ 1634.4375 Khz'],
[380, null, null, null, null, null, -113.8, 'red', 'Tone: 380 SNR is -113.8 dB @ 1638.75 Khz'],
[381, null, null, null, null, null, -114.2, 'red', 'Tone: 381 SNR is -114.2 dB @ 1643.0625 Khz'],
[382, null, null, null, null, null, -115.1, 'red', 'Tone: 382 SNR is -115.1 dB @ 1647.375 Khz'],
[383, null, null, null, null, null, -114.6, 'red', 'Tone: 383 SNR is -114.6 dB @ 1651.6875 Khz'],
[384, null, null, null, null, null, -114.8, 'red', 'Tone: 384 SNR is -114.8 dB @ 1656 Khz'],
[385, null, null, null, null, null, -115.3, 'red', 'Tone: 385 SNR is -115.3 dB @ 1660.3125 Khz'],
[386, null, null, null, null, null, -114.8, 'red', 'Tone: 386 SNR is -114.8 dB @ 1664.625 Khz'],
[387, null, null, null, null, null, -115.5, 'red', 'Tone: 387 SNR is -115.5 dB @ 1668.9375 Khz'],

You need to look at the code contained in the frame you are looking at to get at it by viewing the frame code. The menu needed varies between browsers but if you (using a desktop PC)  right-click just about a millimeter to the left of the Refresh button at bottom left the pop-up menu (in FF) will show This frame > View frame source and that will display the source code in a new window. Scroll down and you'll find the data above. In Chrome it's just View frame source in the pop-up  menu.

[kitz]

I've only just taken a look (via MDWS) and the FEC plot shows that the error detection & correction process is working hard!

I couldn't detect noticeable pattern though to give any clues.   It could tie in with the small oscillations on the SNRM.    However SNRM doesnt appear to dip low enough to give much of any specific indication pointing towards REIN or SHINE. - Just 'noise'.  Whilst FEC and LEFTRS are working hard, it wouldn't mask any symptoms that should show up in SNRM which is why I suggested monitoring for a longer period in case anything does show.

As for the QLN plot, I have two observations --

• the broad bases to some of the higher frequency peaks hints of a quite strong signal strength, possibly overloading to analogue front end at the respective frequencies
• the peak resulting from the Droitwich Transmitting Station (198 kHz) is minuscule in comparison to that of the highest frequency (1583 kHz) present

Both of those observations, when taken in conjunction, hint at a non-U.K. located member! However there are areas in the U.K. which could give rise to such a QLN plot.

From what I can see OP is in UK.  I think we need to know from him if he is near a transmitter and rough geo location.
The RFI symptoms may be a red herring, but atm from QLN its looking like the most possible reason.

Whatever it is causing the problem I doubt Openreach would do anything.   DLM appears to have things under control and whilst it looks like he has High Re-TX and INP  (would need full stats from modem), the SNRM variance whilst not perfect is well within normal parameters and better than a lot of lines we've seen.
Hlog looks good meaning physical line is ok.   
Whilst G.INP will be masking Err/Secs... its just doing what its supposed to and I doubt any engineer would class 20 Err/secs per hour as a fault.  As long as the line is performing within Openreach estimates then I should imagine the ISP would say no fault found and refuse to take it further.   20 Err/secs ph isn't service affecting and the line is remaining in sync. 

[kitz]

@kitz and others

No I wasnt aware of that - thanks Tony, I'll bear that in mind for future.


..  and wow quick of the mark too re zooming a zoom.    Good work that man