Line errors

Rik · Sep 29, 2011, 09:16:02

Voice of experience, Lance?

Lance · Sep 29, 2011, 23:40:02

I'm mainly thanks and little complaints, but I'm working on it!

Rik · Sep 30, 2011, 10:16:57

davej99 · Oct 01, 2011, 11:25:38

Griff, I think you need a big vote of thanks for the work you have done here!

I much admire your persistence. As a result we are all forewarned of the susceptibility of FTTC to DC noise. Moreover we have been given an object lesson in how to deal with BT and, dare I say it, with ISPs too. (Your mobile number is given at post 75 BTW)

Not much point in ranting about Openreach/BT, except to say that it is sad they can't manage to provide a line with basic DC integrity. We are not talking about the finer points of RF perturbation of phase modulation in frequency division transmission, but tightening up the screws so the joints don't crackle. A great deal of science has gone into transmitting quite remarkable data rates over a copper pair. The abject and miserable failure of Openreach/BT is that they cannot organise a basic DC connection.

Worse than that, BT cynically hides the tools needed for fault diagnosis from the end user and the ISP. They obfuscate, dissemble and write quality of service and support rules that would make a snake oil seller blush. I like the one where blinding incompetence cannot find a dry joint, so lets charge the customer.

We all must take Griff's lead here; stick at it, stay cool, gather that data, maintain the chronologies, escalate to the CEO. In this case I would go back BT with the full history and ask for an explanation as to why a DC connection cannot be guaranteed. You have to ask why our ISPs cannot do this. It seems the end user is on his own; ISPs are ineffective and BT are both cynical and hapless. Perhaps this sorry tale should go to Ofcom and a few MP's Ministers or Shadow Ministers. But then why report the incompetent to the ineffective?

For sure we should follows Griff's lead and stick at it.

Thanks and well done Griff.

Rik · Oct 01, 2011, 11:34:45

Quote from: davej99 on Oct 01, 2011, 11:25:38
(Your mobile number is given at post 75 BTW)

Thanks for pointing that out, it isn't anymore.

.Griff. · Oct 27, 2011, 12:02:12

Time for an update!

I'm sorry I didn't reply to peoples comment after my last post but that was intentional on my behalf. Within minutes of me confirming everything was perfect and people congratulating me on my persistence things took a turn for the worse.

Good news -

Since a HR fault was confirmed on the line and it was subsequently swapped for a spare pair I have seen a definite improvement in sync speed and stability.

Bad news -

The sync speed still isn't a high as it should be and interleaving is still in place on the line. I'm only 320 metres from the DSLAM cabinet (confirmed by OR) and from August 2010 to January 2011 I had the full 40/10 sync with fastpath.

The errors -

You'll remember I was concerned at the amount of errors taking place on the connection and how relieved I was when the line was swapped over that no errors were detected. That lasted about half an hour and they started again.

AS of this very moment my modem has been synced for 1627513 seconds (452 hours or 18 days) and in that period there's been 21,297,628 (yes 21 million) CRC errors and 7,430,628 HEC errors. That equates to 13 CRC and 4 HEC errors every second of every day.

I've spoken to 20+ other FTTC users and countless ADSL/ADSL2+ users and no-one has even come close to that amount of errors. The highest FTTC error count I could find was 17,000 in over a month. Just to put that into perspective I get that many in less than half an hour.

The reason I intentionally didn't post earlier is Olivia Garfield (CE of Openreach) had assured me she'd liaise directly with Idnet to try and get to the bottom of this issue. Four weeks have passed and to cut a long story short she did nothing of the sort but simply forwarded a six week old email, relating to another subject, to Simon at idnet.

So tomorrow will see my fifth or is it sixth, I forget, OR engineer visiting me in just over a month. Let's hope he doesn't run a five minute test and run out the door like the last guy.

.Griff. · Oct 29, 2011, 14:23:53

Another update!

Richard (OR) attended yesterday and was probably one of the most friendly and professional OR engineers I've ever had the pleasure to meet and we were both a bit shocked to find out he was the same engineer who installed my best friends FTTC/VDSL2 on Tuesday. Small world!

I showed him the modem CP stats/errors and he was shocked at the amount that had occurred on the connection. He had an EXFO so tested the line for errors but this device only has a maximum test period of five minutes before it closes down and displays the results. As my errors typically appear within 15-30 minutes of the modem syncing a five minute test was always going to be somewhat flawed.

The modem, now my third, was exchanged and he left apologising for not being able to do more.

As if that was disappointing enough my IP Profile has just dropped from 38,718Kbps to 22,626Kbp, my level of interleaving has doubled and the errors are still occurring.

Wonderful service!

Rik · Oct 29, 2011, 14:37:20

esh · Oct 29, 2011, 16:49:58

For comparison, I am getting 0.0081 CRC errors per second. I think that says enough really. What a mess

.Griff. · Nov 01, 2011, 10:44:38

As of 4.30PM yesterday IP Profile reset back to normal Phew!!

As for the errors Simon is going to have a think about what Idnet/OR can try next.

Glenn · Nov 01, 2011, 10:49:40

Wish my profile would reset, it's been stuck at 25mb since 2nd June.

Rik · Nov 01, 2011, 11:25:37

Can support give it a kick for you, Glenn?

Bill · Nov 01, 2011, 11:27:45

Quote from: .Griff. on Nov 01, 2011, 10:44:38
As of 4.30PM yesterday IP Profile reset back to normal Phew!!

That seems to have happened to quite a few recently (including me), and I for one don't believe in coincidence... a possible explanation?

It's probably relates to the BT change from the original 8c profile to the new 17a profile, in readiness for next year's speed doubling.

Glenn · Nov 01, 2011, 11:29:07

Quote from: Rik on Nov 01, 2011, 11:25:37
Can support give it a kick for you, Glenn?

I have already contacted Brian, they can't get it reset without calling out an engineer.

Rik · Nov 01, 2011, 11:31:24

Bugger.

.Griff. · Nov 01, 2011, 11:45:29

Quote from: Bill on Nov 01, 2011, 11:27:45
That seems to have happened to quite a few recently (including me), and I for one don't believe in coincidence... a possible explanation?

It's probably relates to the BT change from the original 8c profile to the new 17a profile, in readiness for next year's speed doubling.

Some people have already been moved to the 17a profile albeit the 40Mbps cap is still in place. I thought this might explain the sudden profile resets but I'm still on 8c -

# xdslcmd info --show
xdslcmd: ADSL driver and PHY status
Status: Showtime
Retrain Reason: 2
Max: Upstream rate = 17383 Kbps, Downstream rate = 47004 Kbps
Path: 0, Upstream rate = 9999 Kbps, Downstream rate = 39999 Kbps

Link Power State: L0
Mode: VDSL2 Annex B
VDSL2 Profile: Profile 8c
TPS-TC: PTM Mode
Trellis: U:ON /D:ON
Line Status: No Defect
Training Status: Showtime

.Griff. · Nov 10, 2011, 15:20:59

Final update (I hope)

After my last update I told Idnet I'd live with all of the errors because we'd collectively exhausted all avenues of finding the cause and after six engineer visits in two months I'd pretty much had enough.

So things continued as before with a million errors accumulating each day and nothing I could do about them. That is until early this morning when my DSLAM was updated from 8c to 17a VDSL2 profile. What this "update" consists of I'm not sure but I suspect it may be a firmware update.

As a result my errors have all but disappeared completely. I've gone from 12 CRC errors every second to 0.02 CRC errors every second. In addition my TBB quality monitor always showed a slight amount of packetloss all day long. That has virtually vanished as well.

If it stays like this then I'm a very happy man indeed.

Lance · Nov 10, 2011, 15:24:59

Lets hope!

Rik · Nov 10, 2011, 16:07:03

Or pray.

Lance · Nov 10, 2011, 23:09:20

Lets be safe and do both!!!

Rik · Nov 11, 2011, 02:46:43

Tacitus · Nov 11, 2011, 08:07:37

Quote from: .Griff. on Nov 10, 2011, 15:20:59
That is until early this morning when my DSLAM was updated from 8c to 17a VDSL2 profile. What this "update" consists of I'm not sure but I suspect it may be a firmware update.

How were you able to tell that the DSLAM had been updated? Or is this information only available to 21CN/FTTC customers?

Lance · Nov 11, 2011, 08:31:21

It's FTTC only

esh · Nov 11, 2011, 19:34:44

Does this imply that Griff is/was getting very strong, somewhat intermittent interference at the 8MHz band which is gone now it has shifted to 17MHz? Either that, or some bugs fixed by the firmware update... Remarkable change either way!

.Griff. · Nov 11, 2011, 20:30:38

Quote from: esh on Nov 11, 2011, 19:34:44
Does this imply that Griff is/was getting very strong, somewhat intermittent interference at the 8MHz band which is gone now it has shifted to 17MHz? Either that, or some bugs fixed by the firmware update... Remarkable change either way!

Simon_Idnet is as equally baffled as me.

I keep checking my stats half expecting to see 1million+ errors have suddenly appear but so far so good. It's been over 36 hours and they're not even past 3000.

My personal guess, which could be completely wrong, is some sort of firmware update was applied to the DSLAM and that's fixed whatever niggle was wrong before.