Repeated problem with FTTC connection.

I am about 1000 metres from the cabinet, and on a good day I got about 34Mb/sec down. The trouble is that it is a few weeks since I've had a good day.

The connection drops back to about 8Mb/sec and stays there until one of two things happens:

If I manage to get through to the ISP and they do a line check, it drops then reconnects at 18Mb/s.

Alternatively, If I call the landline number (there is nothing plugged into the phone socket) the ringing causes the connection to drop and reconnect at about 29Mb/sec.

All is well for a few hours, until it happens again.

Openreach engineers come out and find no fault on the line: in fact they say the line is excellent.

I suspect there is a bad joint somewhere, and the ISP's test or the OR tester clears the fault in much the same way as ringing the number does, but that's only my guess.

So what should I do? Engineers keep coming out. They agree there is a problem, but can't find any fault. I'm slightly tempted to give up and cancel the FTTC altogether, just use 4G, but the latency is MUCH worse that way.

Suggestions, anyone?

...or should I have posted this in Technical Issues?

In reply to a post by hoopla:

...If I call the landline number (there is nothing plugged into the phone socket) the ringing causes the connection to drop and reconnect at about 29Mb/sec... ...Openreach engineers come out and find no fault on the line: in fact they say the line is excellent.

Are those engineers aware that an incoming phone call can cause the connection to drop?

I had this problem with ADSL and demonstrated this issue to a specialist Openreach SFI engineer when he came to check my line. The problem was traced to a fault external to my property and was eventually overcome. Perhaps, in your case, with VDSL the fault may be somewhere between the NTE master socket and the cabinet.

Good luck.

Persist with raising the fault … it’s probably an HR fault, or less likely, a DSLAM fault.

This is a classic HR DIS fault. High Resistance Disconnection.
Caused by a corroded joint nib, or similar.
It may cause noise on the landline phone. But it will only cause noise if it is VARYING.
Varying may be caused by movement (wind), temperature, humidity etc.
But, but, the broadband line is more demanding of line quality.
The clue in this case, is the broadband is crashing when the customer's number is dialled.
The a.c. ringing current upsets the rectified faulty joint.
No guesswork is needed to locate the culprit.
Look out on the customer's end line with a TDR meter -- Time Domain Reflectometer.
A trace of the line can be seen -- even up to Km's distance.
Joints will be seen. Changes of gauge. Changes of conductor.
At the fault point, a clue is that there may be no up-down waveform.
Moreover, the best test is to call the customer's number, and watch the waveform.
The faulty joint will reveal itself, and the whole waveform may jump about with the
ringing cadences.
Gauge the distance to the disturbance. It usually coincides with a pole or joint-box position.
Or find a new good cable pair, and change the faulty pair.
QED.
What's difficult about that ? ?
.

In reply to a post by dave4real:

This is a classic HR DIS fault. High Resistance Disconnection.
Caused by a corroded joint nib, or similar.
It may cause noise on the landline phone. But it will only cause noise if it is VARYING.
Varying may be caused by movement (wind), temperature, humidity etc.
But, but, the broadband line is more demanding of line quality.
The clue in this case, is the broadband is crashing when the customer's number is dialled.
The a.c. ringing current upsets the rectified faulty joint.
No guesswork is needed to locate the culprit.
Look out on the customer's end line with a TDR meter -- Time Domain Reflectometer.
A trace of the line can be seen -- even up to Km's distance.
Joints will be seen. Changes of gauge. Changes of conductor.
At the fault point, a clue is that there may be no up-down waveform.
Moreover, the best test is to call the customer's number, and watch the waveform.
The faulty joint will reveal itself, and the whole waveform may jump about with the
ringing cadences.
Gauge the distance to the disturbance. It usually coincides with a pole or joint-box position.
Or find a new good cable pair, and change the faulty pair.
QED.
What's difficult about that ? ?
.

I was going to post that but you beat me to it

In reply to a post by dave4real:

This is a classic HR DIS fault. High Resistance Disconnection.
Caused by a corroded joint nib, or similar.
It may cause noise on the landline phone. But it will only cause noise if it is VARYING.
Varying may be caused by movement (wind), temperature, humidity etc.
But, but, the broadband line is more demanding of line quality.
The clue in this case, is the broadband is crashing when the customer's number is dialled.
The a.c. ringing current upsets the rectified faulty joint.
No guesswork is needed to locate the culprit.
Look out on the customer's end line with a TDR meter -- Time Domain Reflectometer.
A trace of the line can be seen -- even up to Km's distance.
Joints will be seen. Changes of gauge. Changes of conductor.
At the fault point, a clue is that there may be no up-down waveform.
Moreover, the best test is to call the customer's number, and watch the waveform.
The faulty joint will reveal itself, and the whole waveform may jump about with the
ringing cadences.
Gauge the distance to the disturbance. It usually coincides with a pole or joint-box position.
Or find a new good cable pair, and change the faulty pair.
QED.
What's difficult about that ? ?
.

Because based on my experience it takes 5 visits before an OR engineer finally listens when you tell him the ADSL drops on incoming calls and does the correct test . 4G was my escape.

Well, the OR man was told that it is a high resistance fault between the cabinet and the premises, and I told him that ringing the number drops the connection, then it comes back faster.

He checked the line with his gizmo and declared it perfect. I suspect that the testing gizmo did something that temporarily clears the fault.

He replaced the master socket faceplate* and the cable to the modem/router. That didn't help, so he decided it must be the modem at fault.

As it happens, I had a brand new, sealed-in-the-box, spare modem/router of the same make but a different (inferior) model**. That works slightly less well, but essentially similarly. Unfortunately, he didn't stay around long enough to see that.

Since then, the line has got itself a bad reputation, I think. Now after I call it, the connection drops as before, but the sync comes back at around 7Mb/sec, with an SNR of over 20.

So I guess I shall have to spend another hour on the phone to Shell Energy to persuade them to poke OpenReach again.

*I don't think he thought it was defective, but OR engineers seem to do this to make sure the customer doesn't get charged £65 for "no fault found".

**In fact, I have two identical spares. When I signed up originally, Shell Energy screwed up the account setup and I had to go through the process three times, Seems that getting a modem/router was an unavoidable part of the process: the part that didn't go wrong!

It only took 16 minutes to speak to someone at Shell Energy today: that was good.

The issue has been raised with OR again.

In the meantime,

DSL Status Up
DSL Uptime 2 hours 20 minutes 29 seconds
DSL Type VDSL2
DSL Mode Fast
Line Rate 3.77 Mbps 7.89 Mbps
Data Transferred 33.38 MBytes 86.47 MBytes
Output Power 7.1 dBm 10.1 dBm
Line Attenuation 16.9, 56.9 dB 28.7, 69.5 dB
Noise Margin 6 dB 5.8 dB

then call the number. While it is ringing (note the noise margin):

DSL Status Up
DSL Uptime 2 hours 22 minutes 11 seconds
DSL Type VDSL2
DSL Mode Fast
Line Rate 3.77 Mbps 7.89 Mbps
Data Transferred 33.68 MBytes 87.95 MBytes
Output Power 7.1 dBm 10.1 dBm
Line Attenuation 16.9, 56.9 dB 28.7, 69.5 dB
Noise Margin 6 dB -6.9 dB

then it drops and reconnects

DSL Status Up
DSL Uptime 51 seconds
DSL Type VDSL2
DSL Mode Fast
Line Rate 7.87 Mbps 7.89 Mbps
Data Transferred 0.07 MBytes 0.05 MBytes
Output Power 6.1 dBm 9.2 dBm
Line Attenuation 8.8, 48.7 dB 21.5, 61.0, 90.3 dB
Noise Margin 6.4 dB 21.1 dB

...but still it is only 7.89 Mbps

Hopefully you’ll get an OR engineer with a scooby this time.

Ask him to put his tester in sync, and be looking at the stats page whilst he rings the DN.

Point out how many repeat reports your line has had, and could they, please, speak to their patch lead, or senior engineer ? So as to avoid yet another repeat report.