Print

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

[DeadMan]

Task takes no more than 1-2 seconds according to error log.

Modem is away from PC and monitor about 5 feet from HP LP2475w IPS LCD and 6-7 feet from  PC. Can test on a laptop from another room with PC etc completely off if required. Interstitial plate is off and plugged directly into master test socket. Specs of PC Intel E7600 core2duo, 8GB DDR2, Asus PQ5 Deluxe mobo, Enermax 530w PSU, various hard disks and peripherals including a USB DJ controller and two Logitech webcams., BT Converse 1300 phone, 3COM US Robotics 56K Professional Modem (for software answerphone), D-Link DIR655 Router with wireless on.  I will try a complete stripdown to wireless laptop but wil be near impossible to turn off wireless to test. Will need to be during a period when US SNRM is in the low noisy phase.

Modem PSU is plugged into extension yes. Not surge protected. Was on a surge strip but moved it. Can try both but this would mean disconnecting modem which risks more DLM.

Lights are ruled out. So is fridge. Microwave. Boiler is off but fan is always running. Would be a pain to switch it fully off though as it's a bugger to get started again.

Oh and I have been around the house with a AM radio at 80 x10Khz (I think this is 800?). HP LCD makes a lot of noise if held close but when I powered it off it did not seem to have any effect on stats. Panasonic Plasma puts out MASSES of noise when on. But has no effect on line because it's off most of the day otherwise I would see a big spike when it is on.

EDIT: I hardwired the laptop to the modem port 2 and monitored the US SNRM (at it's low ebb) then turned everything off in the house I could think of excepting the modem itself. All lights were off. Turned my mobile phone off. Short of turning the mains switch off in the cupboard (no idea which one is connected to the loop the modem is attached to). I could not turn anything else off (well the boiler of course but as I said I don't think that is an issue as I have been with the AM radio all around that and no noise coming from it that I can tell)!

US SNRM did not budge. It has to be external.

Attached is graphs during the period I had nearly everything in the house turned off including the PC, router and all peripherals.

[burakkucat]

There is something weird going on with the data-harvesting. Every fifth minute a record is missing . . . 

[Bald_Eagle1]

There is something weird going on with the data-harvesting. Every fifth minute a record is missing . . . 

That's just the way the tics, grid and labels along the x axis are split up for shorter periods.

Plotting only 30 minutes worth of data from my own connection looks similar along the axis, but the data itself is still plotted.

I may get round to tweaking the graphing program at some stage, but it's not actually a priority at this time.

[DeadMan]

Anyone want to explain some of what I should be looking at to tell if there is a problem? I can see the US SNRM margin going up and down but everything else I am in the dark. The basics would be nice. Thanks!

[Bald_Eagle1]

There is an explanation of what the stats mean here:-

http://www.kitz.co.uk/adsl/linestats_explanation.htm

That page does contain links to further explanations.
The details are for ADSL connections but many of them are still relevant for VDSL2 (FTTC) connections.

ADSL connections use Super Frames (SF) The VDSL2 equivalent is OverHead Frames (OHF).

WWWombat over on the TBB forum has provided quite a detailed explanation for some of the other stats, particularly related to Interleaving.

Regarding SNRM, the target at sync time is 6 dB for VDSL2 connections, to allow for sufficient downward fluctuation during 'noisy' periods, usually in the evenings & lasting until daylight hours.
Some connections fluctuate quite a lot & some only fluctuate by very small amounts.

Resyncing whenever SNRM is at its peak usually delivers higher sync speeds, at the expense of seeing more errors that can actually slow down throughput etc.

Some connections have very high SNRM levels.
This is either because DLM has taken action following a burst or bursts of errors & has capped/banded sync speeds at a lower level.
When this is the case, attainable rates are usually much higher than sync speed.
Another scenario for high SNRM levels could be a connection with a short D-side but the user has opted for a 40 Mb service.
Attainable rates could be much higher (say up to 125 Mb, but as the connection is capped at 40Mb, SNRM could be as high as 30 dB or so, showing plenty of margin for speed increase if the user chose to move to an 80 Mb service.

Basically, an ideal connection would have spare SNRM, be on fastpath (D: value as shown in the raw data stats of 1) & with very few error counts.

FEC/RSCorr errors kick in when Interleaving is applied, improving stability & INP (Impulse Noise Protection) at the expense of some sync speed.
DLM fully controls this aspect, based upon continual connection monitoring & making adjustments whenever it deems it to be necessary.

Unlike ADSL connections, at the moment, ISPs have no control whatsover over Interleaving levels & cannot adjust SNRM levels.

As my connection could happily run with SNRM as low as say 3 dB (it hardly fluctuates at all - even overnight), I could achieve higher sync speeds if the target was lowered.

Bursts of massively increased errors (whether random or with a timed pattern) tend to suggest a REIN issue, but could just as well be caused by an intermittent line fault such as a corroded cable joint etc.

My connection experienced that issue from a poor underground cable to pole monted DP (Distribution Point) joint.
As the weather became hotter & dryer, the error counts would shoot through the roof & the connection would randomly & frequently resync.
DLM didn't like that & would regularly cap/band my connection at very low sync speeds in order to attempt to provide some stability (as low as 8 Mb for a while).

Permanent 'noise' increase (as portrayed via the QLN graph (Quiet Line Noise), tends to suggest increased croosstalk from other users' connections, possibly due to naturally induced interference from other cables within a D-side bundle or more rarely, due to cable damage/water ingress.

I hope that helps somewhat, but please feel free to ask specific questions - we may be able to answer them between us.

[Chrysalis]

most isps use the standard DLM profile (the middle one).  I dont know what the exact thresholds are but my gut guess is you have too much FEC for the DLM to consdier it worthy of reinstating fast path.

My line under normal conditions is considered stable by DLM.  I had a few occasions where I was getting weird bursts of noise during the night between 1 and 3am.  This would trigger a DLM profike change as it generated 10s of thousands of errors in very quick time.

Since these were one off events not 24/7 I had quite fast DLM recovery. The first time this happened the line recovered to fast path within a few days, I was getting maybe a few hundred FEC or so a day if I remember right and no crc errors at all, I think I got 1 crc error the 3rd day or something.

The most recent time I was on interleaving before my fault got fixed, DLM was slower at recovering, I plugged in a fritzbox 3370, and the error count went to extreme low levels to the point I was also getting almost 0 FEC.  Needless to say that triggered a very fast recovery to fast path.  Your FEC counts are way higher than what I had by all means not extreme but not in the same league as what I had, and you still getting 10k crc errors a day whilst interleaved, so sorry to say my gut guess is you on interleaved for the forseeable future.  Unless you can resolve the cause of those errors.  To give you an idea, me on fast path now I am averaging under 1/20th of the CRC errors you are getting whilst interleaved.

My error rate has risen tho which is some concern for if my line will be fast path forever, hopefully vectoring comes before that happens.  I was only 2 months or so ago averaging under 300 a day.  Especially my US error stats which now seem about 1/3 of my DS. 

Previous 1 day time = 24 hours 0 sec
FEC:            0               0
CRC:            665             0
ES:             276             182
SES:            1               0
UAS:            0               0
LOS:            0               0
LOF:            0               0

Since Link time = 49 days 17 hours 2 min 46 sec
FEC:            0               0
CRC:            64902           21904
ES:             26942           19501
SES:            38              141
UAS:            0               0
LOS:            0               0
LOF:            0               0

Also if errors come in bursts thats a bad thing for DLM, if they slow an steady stream throughout the day thats better.  So eg. my 15 min stats.

Latest 15 minutes time = 5 min 0 sec
FEC:            0               0
CRC:            1               0
ES:             1               0
SES:            0               0
UAS:            0               0
LOS:            0               0
LOF:            0               0
Previous 15 minutes time = 15 min 0 sec
FEC:            0               0
CRC:            2               0
ES:             1               1
SES:            0               0
UAS:            0               0
LOS:            0               0
LOF:            0               0

I currently have a margin on both directions.

Max:    Upstream rate = 27437 Kbps, Downstream rate = 72536 Kbps
Path:   0, Upstream rate = 20000 Kbps, Downstream rate = 66387 Kbps

I didnt resync when the downstream attainable increased instead deciding to keep the extra snrm as a buffer, plus knowing my luck if @i did take advantage of it then it would revert although when I start tests next month using new CPE issued to me my sync will be reset then anyway.

[NewtronStar]

Well the last time the DLM reduced my profile in June it took exactly 2 weeks to get back to normal I think the Noise was a one of occurrence, but if the noise is consistent then I can't see the DLM increasing the profile until the engineers find out what is causing the noise/errors.

The first thing I do is start looking at ongoing stats and then fire up the Multi Band Receiver and tune it to 586KHZ, My old Sony FreeView Box (2006) is causing a hell of a mess of interference up to 4 meters from receiver and Modem & Router sit inch's away from freeview box, so I turn off the freeview box now at the plug (time for a new FreeView set top box) 

[DeadMan]

I've been investigating a bit. The faceplate makes no difference so I put that back on and also my phone. They are not the cause that I can see. Secondly I put a high quality ADSL Nation 1 meter RJ11 cable between the modem and the faceplate. A marginal improvement in overall SNRM but not enough to brag about. The SNRM still dips. It's Sunday and the 1.5dB SNRM dips still occur.

I went around with a cheapo old AM handset to see what I could hear. Unfortunately it's not digital so it was mainly guesswork on the frequency. Also a local Asian radio station broadcasts around that frequency.

Following the line out the window from the master socket going to a bunch of cable tied cables. Some are coax (probably satellite etc). The phoneline is on a balcony that goes onto the accountants/car repair side (the car place is under the accountants) into a big black box between the accountants and the next premises. We are all terraced. I have been that side before so know about the black box but did not go over there today with the radio because the accountants windows were not blinded so I did not want to risk being accused of trespassing on their property.

What I did find out though is by going outside with the radio everything became crystal clear as soon as I stepped away from my premises. If I went back the Asian radio station when slightly detuned closer to the frequency mentioned would get completely blotted out by noise.

I don't think the SNRM issue is being caused by the garage as they were closed (unless they have something that goes on and off at various times of the day). It could be the hairdressers to the other side of me but seems unlikely as the line goes the opposite direction to them. My biggest suspect is the premises next to the accountants. Reason being the times of the SNRM drops are usually in the afternoon to very late at night. The premises to the accountants/car place? A Chinese take away.

Of course it could be much further down the line. I don't know where my line goes. I think it might got to a small green cab down the road but from where that goes to the FTCC cab I have no idea. I will have to walk that stretch and have a look around.

One other thing of note is I do have an alarm system and an intercom system for the communal door (I live in rented accomodation). They put out noise but it is constant and I can't see any reason why they would affect SNRM in the pattern I am seeing. Constant noise perhaps. But not the SNRM dips.

Basically my premises and those around me put out loads of noise so it could be anything.

[Bald_Eagle1]

I don't think the SNRM issue is being caused by the garage as they were closed (unless they have something that goes on and off at various times of the day). It could be the hairdressers to the other side of me but seems unlikely as the line goes the opposite direction to them. My biggest suspect is the premises next to the accountants. Reason being the times of the SNRM drops are usually in the afternoon to very late at night. The premises to the accountants/car place? A Chinese restaurant.

Ah. It seems like you have found the source.
It'll be the industrial sized cat/dog microwave oven. 

The longer periods of interference will be due to the cats - remember they have 9 lives, so it takes a bit longer before they become edible.

[DeadMan]

I called PN to see if they would remove me from banding and to update me on when BT is due to visit. No good on either at present. They won't remove me from banding until the engineer visit and they don't know when that will be yet. One thing I remember from a previous visit was the engineer mentioning I was on the small green cab down the road which goes onto where I don't know. I went for a walk that way and could not see any large fibre cabs. I have a fibre cab literally around the corner on the other side. But I reckon I am not attached to that one. I wonder if they can get me a new pair that goes to that cab. I will ask the engineer about it when they visit.

[Bald_Eagle1]

From what I have gleaned from similar queries from myself & others, ISPs have no direct control over banding, whether it is banding to a service profile such as 40 Mb / 80 Mb or other banding at lower levels.

ISPs have to request a service profile band via an order to BT & any banding below the service speeds is automatically initiated by DLM.
On an engineer's request (during or following a home visit) DLM can be reset /re-calculated by BT to a wide open profile within a given service, but this cannot be done directly by an ISP.

It is then again left for DLM to monitor connections & make adjustments accordingly to suit conditions.
When conditions worsen, DLM appears to take action quickly, but it does appear to take a very long timer for DLM to acknowledge when coditions have improved.

If you enter your phone No. here, it should confirm the number of the cabinet that you are currently connected to:-
https://www.dslchecker.bt.com/

Most cabinets have their number stenciled on, so you may be able to see which is which.

I have also been told that it is very, very unlikely that users can have their connections rerouted to an alternative cabinet, but it wouldn't do any harm to ask the engineer.

[Black Sheep]

It's a dead certainty that the EU will NOT be re-routed to another PCP. We have been specifically told, not to engage in the process. Just for info guys.

[burakkucat]

It's a dead certainty that the EU will NOT be re-routed to another PCP. We have been specifically told, not to engage in the process. Just for info guys.

That make absolute sense. (b*cat thinks about the utter pandemonium and chaos that could occur if such a process was allowed . . . )

[Black Sheep]

Not just the pandemonium of 'Routing & Records', but the very real issue of 'Where do we draw the line ??'. As some of us are aware, Mr BE resides on the very outskirts of his serving Exchange, and his immediate neighbours are fed via another closer Exchange. You can imagine the flood of requests that would follow, should we 'move' BE to the other Exchange ??

[burakkucat]

Indeed. That is exactly the sort of situation I was thinking about.