Kitz ADSL Broadband Information
adsl spacer  
Support this site
Home Broadband ISPs Tech Routers Wiki Forum
 
     
   Compare ISP   Rate your ISP
   Glossary   Glossary
 
Please login or register.

Login with username, password and session length
Advanced search  

News:

Author Topic: VDSL knowledge  (Read 12766 times)

Halestorm

  • Just arrived
  • *
  • Posts: 5
VDSL knowledge
« on: December 30, 2011, 05:27:08 AM »

Hello!  I stumbled across this forum completely by accident, and found myself interested in reading about BT's deployment of VDSL/FTTC services over there.  (I'm in the US).  I work as a technician for a company who uses VDSL2 for broadband deployment of triple play services over existing pairs of phone wiring.  As such, I've become extremely intimate with how VDSL works over short and longer distances, not to mention how the signal is impacted both positively and negatively from its travel from the fiber card to your modem.  So, I just thought I'd say hello and introduce myself, and I'll see if I can become useful for any questions that are generalized or troubleshooting related.  :) 

Good luck with your service!
Logged

roseway

  • Administrator
  • Senior Kitizen
  • *
  • Posts: 43924
  • Penguins CAN fly
    • DSLstats
Re: VDSL knowledge
« Reply #1 on: December 30, 2011, 07:27:23 AM »

Hi Halestorm and welcome

VDSL/FTTC is still in its infancy here, so your input will be much appreciated.
Logged
  Eric

Bald_Eagle1

  • Helpful
  • Kitizen
  • *
  • Posts: 2721
Re: VDSL knowledge
« Reply #2 on: December 30, 2011, 07:40:28 AM »

Hi Halestorm,

If you fancy reading a couple of long, long posts that discuss my issues in some depth, take a look at these 2 threads:-

http://forum.kitz.co.uk/index.php/topic,9726.0.html

http://forum.kitz.co.uk/index.php/topic,10142.0.html

My issues appear to be high attenuation & error levels.
I am supposedly around 1000 metres from the street cabinet (probably based upon an estimation from my attenuation levels).
I believe the true distance to be closer to 800 - 850 metres.

In the UK, we have been recently switched from profile 8c to profile 17a in readiness for a speed doubling to 80 Mb DS & 20 Mb US (for those who have decent enough connections).

During the first month, I achieved solid speeds of 33 Mb DS (actual download, not sync speeds) & other users who were switched to profile 17a saw quite a decent increase in DS speeds (both sync & actual download speeds).

The provider (BT) did "something" after the first month that left me with no broadband or telephone services for 4 days & on reinstatement the speeds & stability have not fully returned, despite numerous BT engineer visits.

Unfortunately I have no detailed stats from when my connection was performing much better (The Huawei HG612 VDSL2 modems supplied by BT are locked against user access to their stats & an unofficial unlocking method was not available at that time). 

These are my fairly recent stats (from an unlocked modem):-

Code: [Select]
# xdslcmd info --pbParams
xdslcmd: ADSL driver and PHY status
Status: Showtime
Retrain Reason: 2
Max: Upstream rate = 5878 Kbps, Downstream rate = 31480 Kbps
Path: 0, Upstream rate = 5926 Kbps, Downstream rate = 29804 Kbps

Discovery Phase (Initial) Band Plan
US: (0,95) (868,1207) (1972,2783)
DS: (32,859) (1216,1963) (2792,3939)
Medley Phase (Final) Band Plan
US: (0,95) (868,1207)
DS: (32,859) (1216,1963)
       VDSL Port Details       Upstream        Downstream
Attainable Net Data Rate:       5878 kbps         31480 kbps
Actual Aggregate Tx Power:        6.3 dBm          12.3 dBm
============================================================================
  VDSL Band Status        U0      U1      U2      U3      D1      D2      D3
  Line Attenuation(dB): 8.3 52.9 64.0   N/A 21.6 63.1 0.1
Signal Attenuation(dB): 14.6 52.3   N/A   N/A 21.6 63.1   N/A
        SNR Margin(dB): 5.8 6.0   N/A   N/A 3.5 3.4   N/A
         TX Power(dBm): -4.3 5.8   N/A   N/A 10.9 6.5   N/A

Code: [Select]
# xdslcmd info --stats
xdslcmd: ADSL driver and PHY status
Status: Showtime
Retrain Reason: 2
Max: Upstream rate = 5882 Kbps, Downstream rate = 31480 Kbps
Path: 0, Upstream rate = 5926 Kbps, Downstream rate = 29804 Kbps

Link Power State: L0
Mode: VDSL2 Annex B
VDSL2 Profile: Profile 17a
TPS-TC: PTM Mode
Trellis: U:ON /D:ON
Line Status: No Defect
Training Status: Showtime
Down Up
SNR (dB): 3.4 6.0
Attn(dB): 0.0 0.0
Pwr(dBm): 12.3 6.3
VDSL2 framing
Path 0
B: 63 175
M: 1 1
T: 64 40
R: 16 12
S: 0.0683 0.9424
L: 9366 1596
D: 473 1
I: 80 94
N: 80 188
Counters
Path 0
OHF: 121577347 1534025
OHFErr: 21334 440
RS: 1868062647 2485721
RSCorr: 9374564 17520
RSUnCorr: 1850217 0

Path 0
HEC: 546322 0
OCD: 0 0
LCD: 0 0
Total Cells: 1444465070 0
Data Cells: 61203491 0
Drop Cells: 0
Bit Errors: 0 0

ES: 1843 1237
SES: 348 0
UAS: 100 100
AS: 400330

Path 0
INP: 3.00 0.00
PER: 3.27 9.42
delay: 8.00 0.00
OR: 51.22 27.16

Bitswap: 164721 9064

Total time = 1 days 11 hours 58 min 33 sec
FEC: 28057268 19496
CRC: 99336 1440
ES: 1843 1237
SES: 348 0
UAS: 100 100
LOS: 21 0
LOF: 25 0
Latest 15 minutes time = 13 min 33 sec
FEC: 12342 0
CRC: 3 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: 20392 1
CRC: 16 1
ES: 5 1
SES: 0 0
UAS: 0 0
LOS: 0 0
LOF: 0 0
Latest 1 day time = 11 hours 58 min 33 sec
FEC: 1883677 9304
CRC: 9575 43
ES: 152 43
SES: 41 0
UAS: 0 0
LOS: 0 0
LOF: 0 0
Previous 1 day time = 24 hours 0 sec
FEC: 1971373 166
CRC: 5037 110
ES: 227 88
SES: 15 0
UAS: 0 0
LOS: 0 0
LOF: 0 0
Since Link time = 4 days 15 hours 12 min 9 sec
FEC: 9374564 17520
CRC: 21334 440
ES: 856 382
SES: 82 0
UAS: 0 0
LOS: 0 0
LOF: 0 0

& a couple of recent graph sets are attached

Paul.

EDIT:
For comparison purposes I have also attached a set of graphs from a user's decent connection (currently capped at the UK's 40Mb DS & 10 Mb Us).


[attachment deleted by admin]
« Last Edit: December 30, 2011, 09:43:40 AM by Bald_Eagle1 »
Logged

Halestorm

  • Just arrived
  • *
  • Posts: 5
Re: VDSL knowledge
« Reply #3 on: December 30, 2011, 01:08:21 PM »

Hi Paul!

I saw some of your posts earlier - I will have to read them more in detail (fixing to head to work) but I did take a quick look at your graphs & logs.

My experience is that the reported distance from a VDSL modem, be it the Huawei or a JDSU, is usually shorter than it really is.  So if yours is showing 1000m according to attenuation, you're probably closer to 1100.  A TDR would give an accurate cable measurement.  At that kind of loop length, the sync levels you're getting look really good speed-wise.  However, your signal looks like absolute crap.  It's noisy and you're taking errors out the rear.  When a loop gets longer than 915m modem noise begins to affect the quality of the VDSL signal, so that's part of it.  However, there is trouble, a bad connection somewhere between you and the cabinet.  I will read your posts in detail later on and see what I can derive from it.  :)
Logged

waltergmw

  • Kitizen
  • ****
  • Posts: 2776
Re: VDSL knowledge
« Reply #4 on: December 30, 2011, 02:43:56 PM »

Hi Halestorm,

It really is excellent news to have somebody with good practical knowledge available but not closely affiliated to you know who over here !
To have an external comment that Paul's line is noisy is music to our ears !
Sadly there is no universal service obligation here and all ADSL & VDSL services are sold as up to x Kbps on a best endeavours basis so the cure for a noisy line is to threaten a disconnection.

I have a JDSU question for you. I've observed many tests mainly on ADSL 1 & 2 lines in Surrey UK.
Invariably we find that the JDSU syncs at a much lower speed than say a 2Wire 2700HGV modem does on long very nasty noisy lines with DS attenuations over 80 dB in some cases.
Does that match your experience or has the UK JDSU been slugged not to sync at the maximum possible speed ?
(We know there is bespoke firmware running in the JDSU kit here as it uploads the test results to a BT Database.)

Kind regards,
Walter
Logged

Bald_Eagle1

  • Helpful
  • Kitizen
  • *
  • Posts: 2721
Re: VDSL knowledge
« Reply #5 on: December 30, 2011, 09:04:26 PM »

Cheers Halestorm,


However, your signal looks like absolute crap.  It's noisy and you're taking errors out the rear.  When a loop gets longer than 915m modem noise begins to affect the quality of the VDSL signal, so that's part of it.  However, there is trouble, a bad connection somewhere between you and the cabinet.  I will read your posts in detail later on and see what I can derive from it.  :)


These "issues" have been ongoing since the end of July.

This is what my ISP had to say 8th December:-

"Your connection has been stable for 8 days solidly at 27.9Mbps. As such this line does not warrant any further investigation. Especially because the line is operating above the speed that BT deem within limits. Please let me know if there is anything further I can do for you?"

My cumulative CRC / RSUNCorr error count reached 2.9 Million over its 6 day connection's duration, & the modem re-synced itself this evening.
As it was around 7:30pm & dark, the DS re-sync speed is now around 1 Mb lower than it was during the 6 days.
The RSUNCorr error count had reached 141429 within the first 3 minutes of the re-sync, & has gone up another 500 in the last hour or so.

One TDR test was carried out a few months ago.
The engineer spotted "something" NEAR the cabinet.
He found my connection had been "double jumpered" by a different engineer INSIDE the cabinet.
He removed the double jumpering, but actually REFUSED to run another TDR test that would have confirmed whether or not he had resolved the actual fault.
That engineer didn't confirm the distance as reported via his HAWK TDR test unit, but he did state that my connection was approximately 820m as shown in the Network records on his laptop.
It subsequently transpires that he didn't actually show me the actual connection route at all, but just some straight lines on an Ordnance Survey type map of what he assumed the route would be.

I have requested a further TDR test on a number of occasions to no avail.


Paul.

EDIT:
As mentioned by Walter, JDSU results are uploaded to BT's database by the attending engineer.
I have been "reliably" informed by my ISP that they have no records at their end from when my connection was performing well & that BT will have deleted/overwritten any previous records, especially those from the installation date itself.
« Last Edit: December 30, 2011, 09:10:07 PM by Bald_Eagle1 »
Logged

Halestorm

  • Just arrived
  • *
  • Posts: 5
Re: VDSL knowledge
« Reply #6 on: December 31, 2011, 09:50:35 PM »

Hi Halestorm,

It really is excellent news to have somebody with good practical knowledge available but not closely affiliated to you know who over here !
To have an external comment that Paul's line is noisy is music to our ears !
Sadly there is no universal service obligation here and all ADSL & VDSL services are sold as up to x Kbps on a best endeavours basis so the cure for a noisy line is to threaten a disconnection.

I have a JDSU question for you. I've observed many tests mainly on ADSL 1 & 2 lines in Surrey UK.
Invariably we find that the JDSU syncs at a much lower speed than say a 2Wire 2700HGV modem does on long very nasty noisy lines with DS attenuations over 80 dB in some cases.
Does that match your experience or has the UK JDSU been slugged not to sync at the maximum possible speed ?
(We know there is bespoke firmware running in the JDSU kit here as it uploads the test results to a BT Database.)

Kind regards,
Walter

Hi Walter!  Thanks for the kind welcome.  =)

I deal with the exact same thing over here.  It all depends on the chipsets used on either end of the line.  The chipset in the 2700HGV most likely is newer and better matches the signal generated at the exchange than the one the JDSU uses.  The firmware in the JDSU, even if it's been modified to report stuff to BT, wouldn't affect how the chipset of the modem inside syncs with BT.  So, most likely they're using an older JDSU model with an older chip in it to save money.  I run into the same thing with different models and versions of test sets and modems.  Newer chipsets are better at training around noise and adapting to line conditions to deliver a higher sync rate with fewer errors.
Logged

Halestorm

  • Just arrived
  • *
  • Posts: 5
Re: VDSL knowledge
« Reply #7 on: December 31, 2011, 10:41:28 PM »

Disclaimer: this information contained in this post is just that, INFORMATION, and I am not making any statements or recommendations that are a reflection of any company or practice.  This information is publicly available and not considered "trade secrets", it's straight applied electrical & radiotelecommunications theory & practice that I have learned from my own studies.  I also make NO recommendation to go and do things that may involve tampering with other companies' property, and will not be repsonsible if you ignore my sound advice and do so.  My opinions on topics discussed in this post are simply that, my opinions, and no one has any recourse to hold them against me; listen to them and follow (or not follow) at your own risk.

Okay Paul, I've read through everything, and it just screams that something is definitely wrong with your wiring.  Let me roll through some of the things I looked at that scream trouble:

Code: [Select]
Max: Upstream rate = 5878 Kbps, Downstream rate = 31480 Kbps
Path: 0, Upstream rate = 5926 Kbps, Downstream rate = 29804 Kbps

Your max rate versus your path are waaay too tight.  You are running at 95% capacity down and 99% up.  There is no wiggle room to allow for changing line conditions.  This virtually ensures that you will lose sync on a regular basis.  I would be more comfortable if someone was running at 80-85% capacity, no more than 90%.

Code: [Select]
  VDSL Band Status        U0      U1      U2      U3      D1      D2      D3
  Line Attenuation(dB): 8.3 52.9 64.0   N/A 21.6 63.1 0.1
Signal Attenuation(dB): 14.6 52.3   N/A   N/A 21.6 63.1   N/A
        SNR Margin(dB): 5.8 6.0   N/A   N/A 3.5 3.4   N/A
         TX Power(dBm): -4.3 5.8   N/A   N/A 10.9 6.5   N/A

In reading the properities of your signal, the attenuation isn't as important to focus on as the signal to noise ratio (SNR) margin.  If you look, the SNR upstream between the bands is almost the same, as is the downstream SNR.  To have a quality signal where it's not taking too many errors, the SNR upstream should be between >6-8dB and downstream should be >9-11dB.  Your upstream would actually be okay, but the downstream is what's killing you - 3.4/3.5 is waay too noisy.

Code: [Select]
Latest 15 minutes time = 13 min 33 sec
FEC: 12342 0
CRC: 3 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: 20392 1
CRC: 16 1
ES: 5 1
SES: 0 0
UAS: 0 0
LOS: 0 0
LOF: 0 0

These are the measurements that report the errors of your VDSL signal.  FEC (forward error correction) are corrected errors in the signal.  IMO, anything more than 2000-3000 per 15 minutes is where trouble starts, and you taking 12k-20k every 15 minutes is definitely excessive and indicates a line fault, bad connection or unshielded/untwisted/unbonded cable that is allowing outside interference or crosstalk from other ADSL & VDSL signals to affect yours.  If you were taking more upstream errors, I'd say that AC line interference at the modem would also be a concern, but you're not, so that's not it.  CRC errors are uncorrectable errors.  What you're taking isn't debilitating, but it would be annoying.  ES is errored seconds, SES is severely errored seconds.  That's the time of the 15 minute period that your line was taking uncorrectable errors or a crapton (yes, it is a technical term ;) ) of uncorrectable errors.  If one was to have measurable ES/SES on a daily basis, that is a cause of concern and indicates a definite fault.  The other metrics UAS (unavailable seconds) would be the time spent out of sync, LOS (loss of sync) and LOF (loss of frame) indicate the number of times you lost connection.

Code: [Select]
Latest 1 day time = 11 hours 58 min 33 sec
FEC: 1883677 9304
CRC: 9575 43
ES: 152 43
SES: 41 0
UAS: 0 0
LOS: 0 0
LOF: 0 0
Previous 1 day time = 24 hours 0 sec
FEC: 1971373 166
CRC: 5037 110
ES: 227 88
SES: 15 0
UAS: 0 0
LOS: 0 0
LOF: 0 0

For you to be taking this many errors in a single day, and somewhat consistently, to me narrows it down to a cable fault rather than an intermittent source of outside interference. 

Now, to elaborate on what a good cable path is for VDSL versus a bad one:

Shielding: Shielding of the cable feeding the neighborhood helps keep out interference in the bands that VDSL operates in - AC line noise, AM broadcast, amateur radio, etc.  If your neighborhood has unshielded cable, this greatly increases the noise present on your line.  Also, where splices in the cable are made, the shielding between them needs to be bonded, so that interference continues to have a place to go. And at every opportunity (say every pole if it's aerial, or every pedestal where cable may be buried) these bonds should be grounded to dump the interference off of the cable.  Aerial poles should each have a ground wire running down the pole to the ground.  There's nothing you can do about this, really, except you CAN walk down your cable and look at the poles and make sure the ground wires aren't broken.  Buried cable splices are usually grounded in the pedestals where the splices are brought to the service to be accessible to technicians.  (Of course, telco terminals are their property, and opening one when you're not authorized by the company is probably a crime of some sort.)

Twisting: Wire feeding VDSL, when it's not shielded, needs to be twisted cat-5 style.  This includes everything from the cabinet to your modem.  Common places where wires need to have a twist are in the cabinet (the cables feeding from the fiber cards, and where the cables enter the ground/air before their shields are bonded), at splice points, where the company's cable interfaces with yours at your home (demarcation point), and the wire run from that interface to your modem.  The wire in your house should be cat-5 or cat-6, and using shielded cat-5 or cat-6 would help immensely, providing that the shield is bonded to the ground in your demarc.  Using a shielded cable but not bonding the shield would not give the interference picked up inside your home anywhere to go and could cause trouble (most likely the problem you had when you used that shielded cat-6).  The cord from your service jack to your modem should also be cat-5.  Using a flat telephone cord or nontwisted cat-3 wire just lets interference walk right on in.

Termination: Poorly terminated wire just causes problems anywhere.  If your service is running out of an old school telephone jack with screw terminals inside, it could be fine, it could also cause trouble.  I personally prefer using punch down style jacks, that way the service cord can maintain its twist all the way to the point of termination, and there's zero chance of a wire being broken on a screw terminal.  Also, nicks on a cat-5 from it being stripped, old bad splices in telco terminals where the insulation on the wire is hardening and breaking apart, bad splicers themselves (i.e. wires twisted together with electrical tape versus using a gel-filled splicer) can introduce trouble.

AC Interference: This can happen in your home itself.  If you have any ungrounded appliances in your house, or if your house doesn't have a good common electrical ground; again, interference has no place to go.  Someone who was having trouble with this would have success with an electrical noise isolator plugged inline with the modem's power supply, and/or having their electrical wiring repaired to be properly grounded.

This is hardly inclusive, but these items are the biggest causes of trouble that I've encountered.  Of course, without physically being able to inspect all of these elements, I can't tell you what's causing your trouble.  :(  BUT... anything in your home, you CAN look at and see for yourself if the signal path is pure.  :)
« Last Edit: December 31, 2011, 10:44:08 PM by Halestorm »
Logged

Bald_Eagle1

  • Helpful
  • Kitizen
  • *
  • Posts: 2721
Re: VDSL knowledge
« Reply #8 on: January 01, 2012, 11:08:39 AM »

Hi Halestorm, & thank you for the detailed comments.

Okay Paul, I've read through everything, and it just screams that something is definitely wrong with your wiring.  Let me roll through some of the things I looked at that scream trouble:


I'm pleased that you have immediately spotted that something is wrong.
I have presented very similar data to my ISP who has been unable to spot any problems (as per their comments that I displayed in red in my previous message).

Quote
Code: [Select]
Max: Upstream rate = 5878 Kbps, Downstream rate = 31480 Kbps
Path: 0, Upstream rate = 5926 Kbps, Downstream rate = 29804 Kbps

Your max rate versus your path are waaay too tight.  You are running at 95% capacity down and 99% up.  There is no wiggle room to allow for changing line conditions.  This virtually ensures that you will lose sync on a regular basis.  I would be more comfortable if someone was running at 80-85% capacity, no more than 90%.

Whenever the modem re-syncs or is rebooted the difference between the max rate & the path rate is more pronounced.
e.g. for the connection in my example data, the max rate started off at 35112K, with SNRM at 6.3 dB.
That equates to a path rate of approx 85% of the max rate.

Over the duration of the connection, the max rate for my connection generally fluctuates a little but gradually tails off, as does SNRM.
Just before the modem re-synced itself, the max rate had dropped to 31552 & SNRM was 3.5 dB. (Those were the single values taken from the modem's GUI).

Quote
Code: [Select]
  VDSL Band Status        U0      U1      U2      U3      D1      D2      D3
  Line Attenuation(dB): 8.3 52.9 64.0   N/A 21.6 63.1 0.1
Signal Attenuation(dB): 14.6 52.3   N/A   N/A 21.6 63.1   N/A
        SNR Margin(dB): 5.8 6.0   N/A   N/A 3.5 3.4   N/A
         TX Power(dBm): -4.3 5.8   N/A   N/A 10.9 6.5   N/A

In reading the properities of your signal, the attenuation isn't as important to focus on as the signal to noise ratio (SNR) margin.  If you look, the SNR upstream between the bands is almost the same, as is the downstream SNR.  To have a quality signal where it's not taking too many errors, the SNR upstream should be between >6-8dB and downstream should be >9-11dB.  Your upstream would actually be okay, but the downstream is what's killing you - 3.4/3.5 is waay too noisy.


I believe that in the UK, the target SNRM is 6dB for DS for ADSL. I'm not sure what it is for VDSL2, but I have seen many examples where users' DS SNRM is between 20 dB & 30 dB.
These levels are from users who currently have max rates of 100000K or so, but are currently capped at 40000K (until BT raises the capping to 80000K over the next few months or so). 

I suppose the quality of the signal (SNRM etc.) is critical in maintaining a high speed, relatively error-free connection.
However, to achieve high speeds in the first place, surely my attenuation levels (especially in the upper frequency band(s)) is currently simply too high at 63 dB or N/A.

On the original 8c profile we had fewer frequency bands (2 No.) & fewer tones available for bit-loading (less than 2000 on 8c - now almost 4000 available from the 17a profile).

You can see the differece between my graphs & the other user's graphs, in that I get absolutely no bits loaded in the higher frequency band.

However I was able to sync at around 35 Mb & download at 32 - 33 Mb for the first few weeks.
That leads me to believe that considering the old cabling from the cabinet, I should really be quite a bit closer than the 1000m that has been quoted from a JDSU unit.
A BT engineer in the thinkbroadband forum has stated that he has seen as high as the full 40Mb over 1000m & as low as 16 Mb. I can only assume the 16 Mb connection included substantial quantities of aluminium cabling.
From what I can gather, all other users with lower speeds saw a substantial DS speed increase when switched to the 17a profile that now has more bands & tones available.

Quote
Code: [Select]
Latest 15 minutes time = 13 min 33 sec
FEC: 12342 0
CRC: 3 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: 20392 1
CRC: 16 1
ES: 5 1
SES: 0 0
UAS: 0 0
LOS: 0 0
LOF: 0 0

These are the measurements that report the errors of your VDSL signal.  FEC (forward error correction) are corrected errors in the signal.  IMO, anything more than 2000-3000 per 15 minutes is where trouble starts, and you taking 12k-20k every 15 minutes is definitely excessive and indicates a line fault, bad connection or unshielded/untwisted/unbonded cable that is allowing outside interference or crosstalk from other ADSL & VDSL signals to affect yours.  If you were taking more upstream errors, I'd say that AC line interference at the modem would also be a concern, but you're not, so that's not it.  CRC errors are uncorrectable errors.  What you're taking isn't debilitating, but it would be annoying.  ES is errored seconds, SES is severely errored seconds.  That's the time of the 15 minute period that your line was taking uncorrectable errors or a crapton (yes, it is a technical term ;) ) of uncorrectable errors.  If one was to have measurable ES/SES on a daily basis, that is a cause of concern and indicates a definite fault.  The other metrics UAS (unavailable seconds) would be the time spent out of sync, LOS (loss of sync) and LOF (loss of frame) indicate the number of times you lost connection.

Code: [Select]
Latest 1 day time = 11 hours 58 min 33 sec
FEC: 1883677 9304
CRC: 9575 43
ES: 152 43
SES: 41 0
UAS: 0 0
LOS: 0 0
LOF: 0 0
Previous 1 day time = 24 hours 0 sec
FEC: 1971373 166
CRC: 5037 110
ES: 227 88
SES: 15 0
UAS: 0 0
LOS: 0 0
LOF: 0 0

For you to be taking this many errors in a single day, and somewhat consistently, to me narrows it down to a cable fault rather than an intermittent source of outside interference. 

From a layman's perspective I also tend to agreee with that suggestion.
I do get a few spikes of really high error counts in one hit, but none of them coincide with any electrical eqipment being switched on/off in my home.
I live in a semi-detached house & my immediate neighbour is away quite a bit. I still get the errors while she is away & the spikes are not really at times when her heating could be switching on/off etc.


Quote
Now, to elaborate on what a good cable path is for VDSL versus a bad one:

Shielding: Shielding of the cable feeding the neighborhood helps keep out interference in the bands that VDSL operates in - AC line noise, AM broadcast, amateur radio, etc.  If your neighborhood has unshielded cable, this greatly increases the noise present on your line.  Also, where splices in the cable are made, the shielding between them needs to be bonded, so that interference continues to have a place to go. And at every opportunity (say every pole if it's aerial, or every pedestal where cable may be buried) these bonds should be grounded to dump the interference off of the cable.  Aerial poles should each have a ground wire running down the pole to the ground.  There's nothing you can do about this, really, except you CAN walk down your cable and look at the poles and make sure the ground wires aren't broken.  Buried cable splices are usually grounded in the pedestals where the splices are brought to the service to be accessible to technicians.  (Of course, telco terminals are their property, and opening one when you're not authorized by the company is probably a crime of some sort.)

I'm not 100% sure of the cable route as some is overhead & some is underground.
I have walked the 2 likely routes previously, but have not noticed anything untoward.

Quote
Twisting: Wire feeding VDSL, when it's not shielded, needs to be twisted cat-5 style.  This includes everything from the cabinet to your modem.  Common places where wires need to have a twist are in the cabinet (the cables feeding from the fiber cards, and where the cables enter the ground/air before their shields are bonded), at splice points, where the company's cable interfaces with yours at your home (demarcation point), and the wire run from that interface to your modem.  The wire in your house should be cat-5 or cat-6, and using shielded cat-5 or cat-6 would help immensely, providing that the shield is bonded to the ground in your demarc.  Using a shielded cable but not bonding the shield would not give the interference picked up inside your home anywhere to go and could cause trouble (most likely the problem you had when you used that shielded cat-6).  The cord from your service jack to your modem should also be cat-5.  Using a flat telephone cord or nontwisted cat-3 wire just lets interference walk right on in.

The shielded, but unearthed cable could have been causing a problem (others had also suggested it should be earth bonded), although it was all fine right up to losing my services for 4 days.
The latest engineer replaced the external drop wire from 2 poles away from my home, right into the master socket (which he also replaced), next to my desktop PC.
the modem is plugged directly into the VDSL socket at this master socket using the BT supplied patch lead, & the shielded cable is now simply used for a telephone extension, hard wired into the removable face plate at the master socket.

So, the problem could have been either the external drop wire or the shileded cable (or even both?).

I have disconnected everything apart from the modem for a week at a time now & then, but I see no differences in error counts, tailing off SNRM rates etc.

Quote
Termination: Poorly terminated wire just causes problems anywhere.  If your service is running out of an old school telephone jack with screw terminals inside, it could be fine, it could also cause trouble.  I personally prefer using punch down style jacks, that way the service cord can maintain its twist all the way to the point of termination, and there's zero chance of a wire being broken on a screw terminal.  Also, nicks on a cat-5 from it being stripped, old bad splices in telco terminals where the insulation on the wire is hardening and breaking apart, bad splicers themselves (i.e. wires twisted together with electrical tape versus using a gel-filled splicer) can introduce trouble.

None of the negatives from that section now apply.

Quote

AC Interference: This can happen in your home itself.  If you have any ungrounded appliances in your house, or if your house doesn't have a good common electrical ground; again, interference has no place to go.  Someone who was having trouble with this would have success with an electrical noise isolator plugged inline with the modem's power supply, and/or having their electrical wiring repaired to be properly grounded.

A lot of the house wiring is either new or less than 8 years old, & it has all been tested by an electrician under Building Regulations Part P.

One engineer did wave some sort of a gadget around the power points near the modem to check for AC interference, & gave it the all-clear.

Quote
This is hardly inclusive, but these items are the biggest causes of trouble that I've encountered.  Of course, without physically being able to inspect all of these elements, I can't tell you what's causing your trouble.  :(  BUT... anything in your home, you CAN look at and see for yourself if the signal path is pure.  :)

I THINK that I have eliminated everything within my home that could cause issues, & can only assume that the "fault" is somewhere externally, between the cabinet & my home.

If I had never achieved the higher speeds & apparently stable connection, I would simply accept that what I am achieving now is much better than the 1 Mb I achieved over ADSL.
The copper distance right from the exchange is reported as 5283 metres, which looks about right.

I can only assume that whatever work was done to reinstate my services back in July introduced the "fault" that is now restricting my connection's speed & still leaves me with some connection instability.
If only I had records of my connection's stats from when it was working very well, I might even be able to prove it.
All I really have is the attached speed test.

It may just be coincidence, but now the weather is much cooler than it was between July & September, my connection does appear to be a little more stable.
Can temperature have much of an effect on old external cabling (some of it being possibly 40 to 50 years old)?

I am sure I can't be the only user experiencing these issues, but in the main, any forum posts I have read suggest that users are generally very satisfied with their FTTC connections, managing to stay connected for hundreds of hours at a time.

One final matter, not really discussed previously, is power levels.
I have very recently read that BT can remotely adjust VDSL2 power levels at each cabinet, to avoid the VDSL2 signal swamping ADSL signals that are already quite weak, having travelled a few thousand metres from the exchange over old, possibly partially corroded copper cabling.

I just wonder if BT had to turn down the VDSL2 power levels at my cabinet in order to avoid cross-talk issues after my first month's connection.
The ADSL copper cabling would probably be around 4300 metres to 4500 metres in length by the time it reached the cabinet.

Thanks again for commenting on my issue(s) with a fresh pair of eyes & I fully accept & appreciate the comments in your disclaimer.


Paul.

[attachment deleted by admin]
« Last Edit: January 01, 2012, 11:18:55 AM by Bald_Eagle1 »
Logged