Unresolved support ticket open for more than six months - is this a record?

[#1]

Very strange. Something throbbing on your line   That's certainly a bit regular. Don't suppose there's a discharge lamp going off and restriking anywhere nearby? (Street light, fluorescent etc).
Can you hear any noises on the phone line?
I was about to suggest get the MW portable out and have a wander.

[#2]

Weird, I'd say! I've just tried listening to the BT quiet line test through the "throb" and it's perfectly quiet. Similarly, the AM radio test didn't reveal any additional noise at the time of the "throb" - or is it a heartbeat? 
There's no street lighting in my road - the nearest is in the main road (just over 100m away), and I didn't notice anything amiss when I was taking the PCP pics (just after 7pm). This regular change in noise margin started just after the blip at 18:29. I guess this could be related to the tie-pair mod as I've never seen this sort of behaviour before - I only started Routerstats when I returned from work (just after 6pm), so this could have been going on all afternoon.
Will have a wander around the house and down the road with the AM radio in a moment.

[#3]

No sign of life out here, alien or otherwise! AM radio didn't pick up anything strange and I couldn't see any lights on the blink either.
Will leave pc on overnight with Routerstats running to see if anything else crops up.

[#4]

That graph shows a choked signal.
Is there an in-line RF filter fitted to the line anywhere that you're aware of?  
Basically, your line itself is capable of supporting a data rate right up to the highest 'buckets', pretty much.  However, the 'bits per tone' is being restricted (even to the point that the upstream looks choked, although there is plenty of spare capacity in the upstream, usually.
What is most concerning is that the maximum bits per tone that you can see looks to be around 5-6.  Down in the lower frequencies, that should carry around 10  bpt.  The graph should gently slope off to the right, and should not be completely flat  (see here for an example of a decent line)
The pulsing looks  very unusual too, but it's definitely indicative of a periodic problem.  It looks from the graph to have a period of approximately 7.25 minutes, although it would be interesting to increase the sampling frequency to try to time it more precisely.  I'm sure someone will correct me, but aren't the peaks on the graph showing periods of 'good quality' signal, and the troughs showing a slightly worse signal?
For normal periodic noise, I would expect to actually see  that graph upside-down, with a sharp decline when noise is apparent, and a more gradual return to usual levels - or perhaps I'm just reading it upside-down
Whatever it is, it's still apparent - the key might be in the 7.25 minute cycle.  Anything close by - a local airfield for example, (although a >7 minute pulse isn't likely to be radar, as it's too slow)?
B.

[#5]

There is an RF3 filter on the property Barry, but that was put in the other day and made no difference to before and after.
@catweazle
I did wonder (as opposed to wander) whether you'd come across ET. I expect any locals that saw you wandering will have been wondering "what's that looney upto?"
I also wondered did that most recent "engineer that came use any thing like -
http://www.trendtest.com/trendweb/resource.nsf/vlFileURLLookup/MC2/$FILE/al.mc2plus.html
Anyway, I can't understand why if they were in the exchange messing with tie -pairs, why they didn't do the lift and shift. Maybe they don't have any spare ports - well if so, they've now had plenty of time to get a spare fully testing and working line card to put in
Chris.

[#6]

Quote from: Barry

Is there an in-line RF filter fitted to the line anywhere that you're aware of?

Yes, BT filted an RF filter on 3rd Feb but that didn't seem to make much difference, as @anotherone said.

Quote from: Barry

Whatever it is, it's still apparent - the key might be in the 7.25 minute cycle.  Anything close by - a local airfield for example, (although a >7 minute pulse isn't likely to be radar, as it's too slow)?

I'm about 3 miles from Gatwick Airport, as the plane flies. Also about half a mile north of the normal flightpath. But this regular pulsing phenomenon is new - I've not seen anything like it in the last few weeks. I've set Routerstats to monitor the line every 20 seconds, but I may try reducing the period tonight when I have more time.

Quote from: Anotherone

I did wonder (as opposed to wander) whether you'd come across ET. I expect any locals that saw you wandering will have been wondering "what's that looney upto?"

Even more so if they saw me taking flash photos of crusty old ex-GPO boxes - I'm expecting Special Branch to visit very shortly...

Quote from: Anotherone

http://www.trendtest.com/trendweb/resource.nsf/vlFileURLLookup/MC2/$FILE/al.mc2plus.html

Will have a look at the trendtest site tonight.
Routerstats is still showing the periodic variation in noise margin, but I've also noticed that the Telnet Rx Noise graph picked up some pronounced upwards and downwards spikes during the night. This isn't shown on the standard noise graph nor is there any evidence in the log (which is also updated every 20 seconds). I've attached the Telnet graph plus another DMT snapshot.

[#7]

Airfield have a number of radio devises, not just the radar - but even that will beam will sweep your are about 3 or 4 times per minute - not once every 7.25 min.
Instrument landing systems, VHF AM at about 120MHz  - they wouldn't do anything like that.
Maybe a Non Directional Beacon that are in the medium wave band around 170 and 1500 kHz - so right in the same frequencies of ADSL, but they wouldn't do anything every 7 minutes IIRC - they send out their |D in morse code.
Aircrafts radars is another possibility - but I can't remember if airliners have them on when on the ground - I think maybe not ?.
But all this would affect your neighbours too .................

[#8]

7.25 min is too long for the gap between successive planes - also Gatwick operates batches of landings then batches of take-offs so you would see that variation.

[#9]

Idea: Do you leave your mobile phone laying anywhere near your router?

[#10]

No, not normally - it's usually left in another room (probably about 3-4 metres away, at least). But I will try switching the mobile off when I get home just to see if that makes a difference - thanks for the idea.

[#11]

I trust you've updated your ticket with the latest info. It might be worth asking the question, which tie-pairs did they alter, I assume there were no external ones involved?
The strange thing about this is that the frequencies affected seem smack bang in the range of a LW frequency band pass. An iffy bit of cable could do this but it would be really really unusual as one would expect the higher frequencies to be affected more. At the moment I'm stumped, but as the lift and shift has not been done, the linecard has not been eliminated as the cause.

[#12]

I don't really suspect the GSM phone - and although I agree it's best not to leave it close to the router, I have done just that many times I'm afraid - and it hasn't affected my ADSL noticeably.
I forget how often the cellphone transmits when in the standby mode - but if it is getting a good solid signal from the base station it will not be too frequent ( I've done a quick search but find it ) - it will transmit in respomse to a polling call from the base station which does it to keep track of all phones registered on that site, and of course the cellfhone will transmit in response to a call from the base station.
So could could monitor your ADSL with routerstats etc, leave the cellphone by the router and then call the cellphone from your landline to see if that effects the noise level on the landline.

[#13]

Quote from: Anotherone

I trust you've updated your ticket with the latest info. It might be worth asking the question, which tie-pairs did they alter, I assume there were no external ones involved?

Yes, I have updated the support ticket and referred back to this discussion (with a suggestion that the info, graphs etc were passed on to BT), but the advice from PN support (Logged Faults team) was just to continue monitoring the connection i.e. as the line had stayed connected for 24 hours. But, after I challenged this, they've now passed the fault back to the escalation team.
I've had no info on what was changed for the tie-pairs, but my impression was that BT could do this mod remotely, whereas the lift and shift would require an engineer to go to the exchange.
Would any of the PN guys like to comment on this (or any of the posts from yesterday)?

Quote from: mal0z

So could could monitor your ADSL with routerstats etc, leave the cellphone by the router and then call the cellphone from your landline to see if that effects the noise level on the landline.

Yes, thanks for that - I'll give this a try tonight when I get home.

[#14]

Quote from: mal0z

I don't really suspect the GSM phone

Clutching at straws was the expression you were looking for.

[#15]

Quote from: mal0z

I forget how often the cellphone transmits when in the standby mode

The cells are defined to be part of a larger group known as a "location area", and mobiles are meant to do a location update whenever they see that they've changed area (which is a piece of the data broadcast by each cell). They then do a periodic update if they haven't changed area recently - but this period is configurable. The trick, from a cell provider's perspective, is to set the time short enough to keep track of phones that are still turned on, but not so short that the network is inundated with updates. I think you'd normally be talking times like 20-30 minutes.
Having said that, some cells do get configured in odd ways to cope with non-normal traffic - and I can imagine that Gatwick might qualify for that, as a place where people are turning phones on & off more than usual.
If your house is very close to the border of two location areas, then the vagiaries of RF might make the phone re-register on a more frequent basis.  Again, I could understand if cell planning around Gatwick meant a larger number of smaller cells (to cope with high traffic levels), combined with small location areas. That would increase the likelihood that you're in such a border region, I guess. But this would be more irregular.
You normally hear the intereference noise of location updates when your phone is left next to the radio in the car. A 5 second burst that sounds almost like morse.

Quote from: mal0z

So could could monitor your ADSL with routerstats etc, leave the cellphone by the router and then call the cellphone from your landline to see if that effects the noise level on the landline.

Sending a text message to it would be fairly similar to the location update work it performs.
But... wasn't it pointed out the the SNRM graphs are showing regular bursts of clear, interspersed with longer periods of noise, rather than vice-versa?
Mike