Flaky FTTC and frequent dropouts

Visiting my mums over xmas (don't panic, it's all officially sanctioned!), the internet has slowly been getting worse and worse. It's quite a large house so I put in a mesh wifi system and hoped that would help affairs. It has fixed the wifi but was still having serious reliability issues and slowdowns.

Changed VDSL modem to a known good one (a Zyxel and also a Draytek which are both mine, end up getting a similar line rate) but the issues remain. We've always had an issue with water ingress into the phone line so that is probably the most likely culprit but really, we need BT to replace the line from the mast as every time they 'fix' it, problems promptly return. I took the (5-10 year old) master socket and most of the pins were corroded with water damage. Replaced it with a brand new (5c) and still get about the same line rate. There are NO wired extensions connected as I removed them all some years ago and it is just the house phone and the modem that are connected to the master socket.

When I last lived here, it would sync at ~55mbit and now can't get it beyond 33. The symptoms of the internet issues are total connection drops (following 500 or so CRC errors as shown below). That's the simple one. The other issue that I cant narrow down is just poor performance. Speedtests will drop to 10mbit for a short period (5-10 mins) and occasionally, sites will just stop loading but it recovers itself sometimes without retraining the connection. Stats are below for my Zyxel VMG1312-B10A.

VDSL Training Status: Showtime
Mode: VDSL2 Annex B
VDSL Profile: Profile 17a
G.Vector: Disable
Traffic Type: PTM Mode
Link Uptime: 0 day: 0 hour: 2 minutes
====================================================================================
VDSL Port Details Upstream Downstream
Line Rate: 8.418 Mbps 29.998 Mbps
Actual Net Data Rate: 8.419 Mbps 29.999 Mbps
Trellis Coding: ON ON
SNR Margin: 6.3 dB 10.4 dB
Actual Delay: 0 ms 0 ms
Transmit Power: 4.4 dBm 12.5 dBm
Receive Power: -9.2 dBm -15.0 dBm
Actual INP: 44.0 symbols 54.0 symbols
Total Attenuation: 0.0 dB 23.0 dB
Attainable Net Data Rate: 8.419 Mbps 40.911 Mbps
====================================================================================
VDSL Band Status U0 U1 U2 U3 U4 D1 D2 D3
Line Attenuation(dB): 9.2 40.3 59.2 N/A N/A 17.7 47.7 73.9
Signal Attenuation(dB): 9.2 39.8 58.6 N/A N/A 25.7 47.3 76.8
SNR Margin(dB): 6.7 6.3 6.3 N/A N/A 10.4 10.4 10.4
TX Power(dBm): -0.1 -2.5 0.3 N/A N/A 10.6 7.7 -4.6
====================================================================================

VDSL Counters

Downstream Upstream
Since Link time = 2 min 19 sec
FEC: 13 0
CRC: 0 0
ES: 0 0
SES: 0 0
UAS: 0 0
LOS: 0 0
LOF: 0 0
LOM: 0 0
Latest 15 minutes time = 14 min 18 sec
FEC: 2275 0
CRC: 575 0 ********** Shortly before the connection drop/retrain
ES: 11 0
SES: 11 0
UAS: 38 27
LOS: 1 0
LOF: 7 0
LOM: 0 0
Previous 15 minutes time = 15 min 0 sec
FEC: 52 0
CRC: 0 0
ES: 0 0
SES: 0 0
UAS: 0 0
LOS: 0 0
LOF: 0 0
LOM: 0 0
Latest 1 day time = 22 hours 59 min 18 sec
FEC: 11241 7
CRC: 575 5
ES: 11 5
SES: 11 0
UAS: 66 55
LOS: 1 0
LOF: 7 0
LOM: 0 0
Previous 1 day time = 0 sec
FEC: 0 0
CRC: 0 0
ES: 0 0
SES: 0 0
UAS: 0 0
LOS: 0 0
LOF: 0 0
LOM: 0 0
Total time = 22 hours 59 min 18 sec
FEC: 11241 7
CRC: 575 5
ES: 11 5
SES: 11 0
UAS: 66 55
LOS: 1 0
LOF: 7 0

Wondering how to persue it, hopefully some here will come up with some good suggestions. Again, I imagine the only real solution would be to replace the line, but I imagine it will be very hard to convince BT to do that. The point where the phone line enters the house has been changed 3 times over the last 25 years due to water ingress and subsidence.

Cheers in advance!

Is the line free from audible noise ?

Have actually reported the broadband issue to the service provider, that would be my first step.

The downstream has been capped/banded by the DLM.
It won't/can't sync above 30Mb/s (29.999) unless the DLM is reset or the line remains very stable and the DLM removes the cap.

Banding is usually caused by too many resyncs in a short time period.
This could be because of OpenReach working on the cabinet, a line fault, or your mum rebooting the Hub over and over.

Going by the stats the line should be syncing around 40Mb instead of the 30Mb it is capped to.

If it synced higher than 40Mb previously then this has probably been lost to crosstalk over the years.

If the DLM is reset by the ISP then the line will go right to 40Mb.
With stability it could go to just under 50Mb in time with a reduced SNRM target.

Ignore the CRC count.
The DLM is interested in the ES count.
This plays no part in the capping of the sync and the ES rate is low.

Edited by j0hn83 (Sun 27-Dec-20 13:57:53)

My parents had a similar situation recently and after talking it through with their ISP (Zen) we identified a line fault. DLM had drooped the line steadily from a stable 40/10 to just 7/0.7. BT said there was nothing wrong but having insisted the line was the issue, Openreach duly came out and found ‘a cracked connection’ causing a high resistance issue. There was also noise on the line at times. The pair was repaired and everything went back to 40/9 with a stable 7ms latency and it’s been that way for several weeks.

I had to play on my parents being elderly and needing the line sorting out. Fortunately I found a sympathetic support body at BT who got it sorted.

They’ve had trouble with that line for years and the Openreach engineers have in the past said the overhead lines need replacing but no one at OR will sanction it because it’s too expensive.

Edited by MercuryRH2 (Sun 27-Dec-20 15:25:16)

In reply to a post by milesharrison:

Total time = 22 hours 59 min 18 sec
FEC: 11241 7
CRC: 575 5
ES: 11 5
SES: 11 0
UAS: 66 55
LOS: 1 0
LOF: 7 0

I would report an intermittent broadband fault and request a SFI engineer - something is causing ES, SES and a LOS.

In reply to a post by 4M2:

In reply to a post by milesharrison:

Total time = 22 hours 59 min 18 sec
FEC: 11241 7
CRC: 575 5
ES: 11 5
SES: 11 0
UAS: 66 55
LOS: 1 0
LOF: 7 0

I would report an intermittent broadband fault and request a SFI engineer - something is causing ES, SES and a LOS.

Every single time DLM resynced my line i received i burst of CRC and exactly 11 ES.

These occurred during the resync process.

There's not a single ES or CRC in 23 hours of sync time apart from those that happened during the resync.

Pointless reporting this. It's normal.

Edit: I'll add to that... Even if the 11 ES were not caused by a resync (they were) it would be ridiculous to report 11 ES in a 23 hour period to an ISP.

Edited by j0hn83 (Sun 27-Dec-20 19:51:46)

In reply to a post by j0hn83:

Edit: I'll add to that... Even if the 11 ES were not caused by a resync (they were) it would be ridiculous to report 11 ES in a 23 hour period to an ISP.

I only have 9 on my l8ne and that’s running normally, so nothing to see here.

In reply to a post by j0hn83:

Every single time DLM resynced my line i received i burst of CRC and exactly 11 ES.

These occurred during the resync process.

There's not a single ES or CRC in 23 hours of sync time apart from those that happened during the resync.

Pointless reporting this. It's normal.

Edit: I'll add to that... Even if the 11 ES were not caused by a resync (they were) it would be ridiculous to report 11 ES in a 23 hour period to an ISP.

Latest 15 minutes time = 14 min 18 sec
FEC: 2275 0
CRC: 575 0
ES: 11 0
SES: 11 0
UAS: 38 27
LOS: 1 0
LOF: 7 0
LOM: 0 0

If ES, SES and a LOS happens intermittently causing relatively frequent resyncs, which may be the case, then that suggests a line fault.

Have to agree, nothing out of te ordinary there.

I work with some comms systems where 50ES per hour is quite "normal".

Broadcom modems show 10/11 ES (and often hundreds of CRC) on every single DLM initiated resync.
Every single 1.
It happens on many resyncs not caused by the DLM also.

The line has 0 ES outside that 15 minute minute window.
0 ES means 0 CRC.
So all of those 575 CRC were at the exact same time as the ES (1 CRC is a minimum 1 ES. 100 CRC can be 1ES).
The 11 SES are caused by the 11 ES.
The DLM initiated resync is the 1 LOS reported.

To the untrained eye the error figures are low enough
To the knowledgeable they are zero.

Excluding the DLM initiated resync there's not a single LOS, CRC or ES.

There's a low amount of FEC but even millions an hour of them is ok.

That's not to say there's no issues with the line.
It is currently banded.

The error figures are low though, and don't justify contacting an ISP on any line.
Knowing how the DLM causes ES on Broadcom modems i know there are zero CRC, ES, SES or LOS in those 23 hours of stats.

Edited by j0hn83 (Sun 27-Dec-20 22:20:54)