Five months.....

[#1]

Hi Steve,

I'm back at the PC and have done a detailed look through the graphs you've provided.  Here's my thoughts on th e(thus far) limited information available.

Sat 14:25

• SNRM transition 6dB -> 7dB suggests that some source of RFI stopped
• SNRM transition 7dB -> 4.8dB suggests some source of RFI started, but worse than before
• Note the corresponding changes in the max (theoretical) rates, HOWEVER there is no change in synch rate implying these events did not drop line synch

Sat 15:45

• SNRM crashes loosing DSL synch
• SNRM 'returns to 6dB (line resynchs at the TARGET SNRM) at a slower speed than before
• Note how flat the SNRM is after this

Sat 17:20

• SNRM crashes again
• Resynch is very slightly faster

Sat 17:45

• SNRM crashes quite badly (-2dB or less)
• DSL resynch is slightly slower

Though there are few data points from which to draw firm conclusions, I am struck by the close time correlations - 20 ish and 45 minutes after the hour - this is worth watching for.

In response to other comments you've made...

• A mains filter is not going to stop RFI getting into the phone line
• In respect of "nothing in the house has changed" that is only nothing that you KNOW about - PSU power bricks used for powering so many things in our households can slowly decay ... they still function to all intents and purposes, but in doing so they spew out a lot of RFI.  I once spent months looking for intermittent issues (which seemed dusk related) and eventually by accident I discovered it was the PSU brick used to power the CCTV cameras - that had been there for years!
• I the DSL synch drops everything above it will drop
• A problem with the line (including the phone socket) can make a line more susceptible REIN issues

I suggest that you keep RS running - 10 second sample rate - 720 points to a graph - timed capture - will give good granularity to events, keep the graphs covering a 2 hour period and save them all in the RS directory.  I think it is worth doing a few days of this monitoring before invoking further assistance from PN / BT.  If the line is still unstable, you'll need good evidence and these RS graphs will help to provide that.  One problem that repeatedly crops up in circumstances such as this is that the cause of the intermittent disconnects is never manifest when the engineers attend!

[#2]

• SNRM transition 6dB -> 7dB suggests that some source of RFI stopped

But oddly enough I have NEVER seen 7dB before

Line was moderately stable until yesterday and has been stable since (apart from when I removed the Master socket front plate after finding it was loose ) - the gap on Friday was the engineer

And to be honest this burst of instability is just another annoyance in the packet loss and multicast issues I've been having for 6 months...

[#3]

As I said, the 6dB to 7dB transition is indicative of the RFI which was present when the router last synched disappearing for a brief moment.

When a line resynchs it establishes a synch rate which will deliver a 6dB SNRM TARGET over the then present back ground noise.  If after that the background noise subsides, the ACTUAL SNRM will rise ... which is why the max rate also rose.

Whilst you have had the packet loss and multicast issues over recent months have you also been monitoring the SNRM at this level of detail?  Generally people only turn to this detailed level of monitoring when all else has failed ... at which point they find stuff which unknowingly has been on-going for months.

OMG just had a thought - part of my previous problem turned out to be caused by passing WCML electric trains - but only the 11 coach Pendelinos . Cheltenham - you are not by any chance near the GSWR main line are you?

[#4]

Been monitoring it on and off since the start - and its just wobble along between 5.5 and about 6.1 :

Back in FebOddly enough since removing the master socket front plate:

So using the filter provided with the Hub 1 gives a much less wobbly Down SNR

I'm on the other side of town from the railway - only big "Industrial" thing round here is the Sainsburys

[#5]

Considering this is FTTC, I am thinking the SNRM transition from 6dB to 7dB and then to 4.8dB could be someone else's FTTC line re-syncing, and resulting in worse crosstalk after than before. Probably unusual for the amount of crosstalk to change so much, but I suppose various things could have happened to the other line, maybe it was re-synced to restore its usual speed after previous operating at a lower speed, resulting in more signal and more crosstalk.

[#6]

Considering this is FTTC, I am thinking the SNRM transition from 6dB to 7dB and then to 4.8dB could be someone else's FTTC line re-syncing,

Could be - the line back to the cab is somewhere between 700 and 900 metres with the last 50 metres or so overhead from a pole. From the pole to the Master socket the wiring is less than 14 years old but I suspect the pole to the cab dates back to when the houses were built in about 1971.

The line is syncing about as fast as BT say it can:

So I'm quite happy with all of that.

Will be interesting to see if the packet loss has gone away now I've dumped the BT master socket ..

Edited to add at 18:56 : the Noise Margin just spiked to 8.5 down (with a corresponding spike to 46Mbit on the Max Data Rate), and then dropped back down to where it was before.

Edited again to add : in the early hours of the morning the Noise Margin down spiked to 8 (Max data rate of 47) and has now settled at 7.2 (Max data rate 44.4) with the connection not dropping and staying solid at 41.7

[#7]

@Gandalf / @Chris,

Does the SNRM variability evidence point to anything which can be investigated further by BTOR?

[#8]

Down Noise margin slowly decreased with the sync holding at 44000 Kbps. Got down to 3.8 and stayed solid for most of a day then spiked up past 6, then disconnected.

Yesterday the Noise marging was 4.8 then it spiked up to 6, dropped to 3 and disconnected. When it reconnected it connected with a noise margin of 5dB and a speed of 42.6.  At 19:40 last night it spiked to 7 and then back down to 5.8  (with no disconnect) and is now sitting at 5.9

I'm getting more disconnects than I used to do.

[#9]

This really needs to be plotted to provide documentary evidence of some source of interference. Note that the link will drop on a fall of SNRM, not a rise. As suggested by ejs this could be cross Talk leaving the only answer to be a higher SNRM at a slower speed.

[#10]

Noise margin has dropped with no disconnects - I there are spikes to values like 7 or 8 which sometimes have a line drop after them and sometimes dont.  As the line can apparently hold a 44Kbps connection (which is right at the limit of the line) and give me download speeds just about on 40 at 3.8Db.  

The sudden drops and spikes have no pattern to them

I'm seeing drops when there is no sudden change in margin.

I've got the graphs

[#11]

---

@SteveA wrote:
I'm getting more disconnects than I used to do.

---

It looks like there are a few disconnections in recent days.

Tests aren't identifying any issues...

Test Outcome	Pass
Test Outcome Code	GTC_FTTC_SERVICE_0003
Description	No problem found, OAM test is not currently supported on this line.
Main Fault Location	OK
Sync Status	In Sync
Downstream Speed	42.6 Mbps
Upstream Speed	7.0 Mbps
Appointment Required	N
Fault Report Advised	N
NTE Power Status	PowerOn
Voice Line Test Result	Pass
Bridge Tap	Not Detected
Radio Frequency Ingress	Not Detected
Repetitive Electrical Impulse Noise	Not Detected
Cross Talk	Not Detected
Estimated Line Length In Metres	943.5
Upstream Rate Assessment	Good
Downstream Rate Assessment	Very Good
Interference Pattern	Regular Interference Observed Daily
Service Impact	Retrains Observed
Interference Duration Longest Occurrence	From 14:45 to 15:00
Interference Location	Unknown
Home Wiring Problem	Not Detected
Downstream Policing Discard Rate	0.0
Customer Traffic Level	Upstream and Downstream Traffic Detected
Technology	VDSL
Profile Name	0.128M-80M Downstream, Interleaving Low - 0.128M-20M Upstream, Error Protection Off
Time Stamp	2017-06-30T11:45:00

Parameters	MIN	MAX	AVG
Down Stream Line Rate	34.3 Mbps	47.1 Mbps	39.8 Mbps
Up Stream Line Rate	6.6 Mbps	7.1 Mbps	6.9 Mbps
Up Time	24.0 Sec	900.0 Sec	898.0 Sec
Retrains	0.0	4.0	0.0

Current and Last 15 Minute Bin Performance
Parameters	Last Traffic Count(Upto 15 mins)	Current Traffic Count(Upto 15 mins)
Start Time Stamp	2017-07-13T10:08:39Z	2017-07-13T10:23:39Z
Ingress Code Violation	0	0
Egress Code Violation	0	1
Errored Seconds	0	1
Severely Errored Seconds	0	0
Unavailable Seconds	0	0

@Townman wrote:
Does the SNRM variability evidence point to anything which can be investigated further by BTOR?

 We'd just have an option to book an SFI at this point because no faults have been identified.

[#12]

What is to be made of "Regular Interference Observed Daily"?

Does this infer cross talk or REIN or something else?

[#13]

Spike up - no disconnectSpike up then down then dropMargin down and then slow rise no drop

But Ive also had this:

low noise margin, good connnect speed and stable connection

But the sudden increase in random drops (not always associated with Noise Margin changes) is annoying

[#14]

---

@Townman wrote:
What is to be made of "Regular Interference Observed Daily"?

Does this infer cross talk or REIN or something else?

---

It could be anything.

There don't really seem to be enough disconnections to look into arranging another engineer.

[#15]

Been away for 4 days.

According to wife there have been multiple disconnects.

SyFy is currently just about unwatchable.

I have had enough...  I will be cancelling my TV subscription tomorrow and expecting some compensation from Plusnet as I've been paying for this service and for the past 6 months I've had problems. I've had to take half days from work to be in for engineers who have criticised previous engineers, complained about the external wiring, have said paperwork has to be raised and then haven't done it ( causing yet another engineer visit) or just haven't turned up. The problem isn't directly with the TV service but I might as well save myself some money whilst I look around for a new ISP.

Noise Margin Graph: