Where's the DLM when you need it?

[#1]

The interference on the AM radio was still present on Friday evening with the router unplugged and switched off (I remembered to check while moving the router to try it in another socket).

[#2]

the best option would be to see when its on and walk to the electric input point in your house (master fuse box switch) then flick it to off (yes I know all the pain this will cause in resetting everything afterwards)
if its something in your house the signal will die, if it remains for 70 seconds after you throw the switch its external or if its internal its motion operated and battery powered.
of course if you have a monitored alarm system then you best not do this method.........

if the signal dies when the power goes off chuck the power back on within 90 seconds (I sugest you time it) and *most* items will hopefully retain there programing cheaper items will be drained and require setting backup though
if the signal doesn't die when you pull the power to your house ask PlusNet to arrange for REIN monitoring to be conducted over a 7 day period 
if it does stop/start with the power flick then investigate in the following fashion....

if its internal then once you turn everything back on does it come back ?
if yes run round and unplug everything, is the signal still present?
if yes then individually isolate each thing and check if the signal is still present (simple for any modern items like water/heat/stairlift etc as they all have separate isolator switches).
if nothing is found to be "causeing the signal" in its default state at test time, use the override functions (where available) and see if you can activate the fault on any item or set piece of equipment, for power supply units this will involve creating a maximum power draw through them one by one(plug in the thing they power and for mobile chargers have the phone switched on with all Bluetooth/wifi/music turned on and similar for other devices like multifunction alrarms steroes)
if the nefarious signal is triggered by a power fluctuation then in theory turn off/on the main while under max draw will trigger the suspect item to cause its failing broadcast

I realise its a lot of work but if its internal then finding and replacing the item will be a lot cheaper than paying for repeat engineer visits where nothing is found or where they do find the problem definitively not being in there equipment (I would personally always look at telecoms equipment first as its often mass produced and poorly QA'd)

[#3]

You will always find some interference when listening with a MW & LW radio. This trick is to learn about the Interference when you Broadband problem is not there, and what is different when the problem is there. Tune up and down the band in both situations (not just around 612kHz).

[#4]

There was no interference on the AM radio early this morning coming from the 2 previous places where it was strongest: both houses' satellite dish and cable, nor anything from in the hall on the inside of the external wall that has the Post Office Telecom plastic box that the incoming telephone cable goes to. 3 star wired internal sockets probably doesn't help, each one appears to be independently wired to the external box.
When the interference was present, I think I could detect it on the AM radio just above the nearest BT covered hole in the pavement, which was located in the direction of the neighbour that also had the problem, but that is the opposite direction of the shortest route for my telephone line to go to my cabinet.
I'm surprised 612kHz is quiet considering there's a radio station on 603, and digital AM tuning goes in steps of 9. Of course I won't be able to hunt down the source until it returns. The timing would be consistent with something switched on during weekday evenings, and all 09:30 Sat to 01:10 Sun, but not the rest of today (yet) - maybe the user of it is somewhere else so far today.

[#5]

As I feared, the interference resumed at 19:10 today. It seems to be present in every even switched off electrical socket, light fitting, even metal pipes, pretty much everywhere really. 

[#6]

You shouldn't be surprised by the fact that 612kHz is quiet, even if it is when your problem is present - it very much depends on the source/cause of the interference in the first place. The digital tuning shouldn't make too much of a problem in locating interference - generally when a frequency is affected it's bandwidth is wide enough that you'll find it either just above or just below.
It's soundly like some kit someone uses when they are home and not when they are at work Mon-Fri or out for the day. Perhaps try and spot who's in and out when it does and doesn't happen.
So who came home sometime before 1910 - perhaps they'd been to Silverstone or some other event.

[#7]

612kHz is quiet when the interference isn't present, and loud when it is. I was surprised it was quiet (without the REIN) considering there's a radio station so close. I recall reading some time ago that certain MW bbc local radio broadcasts have even been shut down, or replaced with just a message telling you who to contact if you're listening to it. I'm guessing from the url that the BBC World Service on 648kHz was shut down around Feb 2011, although the "Why can't I listen on 648 kHz medium wave?" on this page just redirects to another general help page which doesn't answer the question.

[#8]

As you're allowed to take your own thread off-topic   I believe it was 26th March 2011

[#9]

well it will be some kind of clock pulse working in the hertz range (like an analogue car clock) that's parked right on top of a BT manhole fairly close to your house or next to a green street cabinet right near your house,  and the signal is being picked up by the phone lines and causing you problems, ask your neighbours up and down a few doors if there having internet problems at the same time as you are, it will narrow down the scope of the problem (if its a vehicle clock it will likely only be a few doors up/down the road from where its parked)
a common problem with older vehicles where the restorer didn't wire up the clock properly with the required diode to supress the pulse signals ......or a bodge it mechanic who doesn't know about it....either way given the timings and longevity its unlikely to be static equipment (a mains powered piece of kit would be in its death throws and failing to work at all over a period in excess of months)
as AO says everything will produce noise, the real test is the distance from the device, if the noise is present it should be undetectable at a distance of 100 cm from source, basically when your sat on your sofa holding your portable set tuned to the frequency you have noticed is most often noticeably problematic (so 612khz in this instance) providing the antenna is 100cm from your electrical devices you should get nothing on the radio normally and during the fault time you will have a signal, when this happens take a walk outside down the path to the street with your portable radio, first walk down the road and see if the signal gets weaker or stronger, if it starts to decay keep walking until its almost imperceptible then turn around and walk the other way, you will be able to find the source if its local as being the point at which the signal is strongest on the street walk

[#10]

Scratch head - was this the intended thread for your post? If so, don't agree that the source will be "in the hertz range", it'll more likely be 10's or even a 100 or so of kHz with lots of nasty odd harmonics, most likely emanating from a Switch Mode Power Supply (SMPS) that's developed a fault in a TV set or Sky box or something similar. It'll be being induced into any nearby wiring, phone, power, whatever.
Ring main and other wiring in a house can produce some odd electrical fields, so there is no guarantee that you will not pick up other interference when the problem source is off.

[#11]

yeah the problem he's picking up clearest is the harmonic wave propagation, that's most likely going to be caused by an extremely low frequency device like a car clock with a large em signature causing the 50v dc phone line to act as a transformer amplifier and be producing a lot more lower+higher frequency noise as you get close to the source point .......
if it was a problem that only ever occurred as the natural absorption layers in the atmosphere changed (decayed) then it would be a higher power item that was always there but not always causing a problem but the time frame would suggest something more along the lines of someone on benefits who's partner pops round for the evening then goes home and stays over on the weekend (only allowed to sleep over so many nights a week and all that) or some kind of illicit booty call that takes a few hours here or there in the evenings......
if it was a household power supply you would expect the pattern of usage to change quit a lot over time rather than be such a concise pattern, if it was totally random but longer lasting bouts then a dodgy hardwired bit of electrical kit would also be indicated .....

[#12]

It stopped at about 01:50 this morning. There was nothing parked near the nearest BT manhole, it's only detectable very close to the BT cover. I checked this morning and there was no interference coming from the gas meter and pipes in the garage, and far less coming from the electricity meter and fuses / breakers. I will have to attempt to fully check whether it could be coming from some part of the central heating system if/when it returns this evening.
My BT green cabinet is quite far away, up to the top of my road, up another road, then along to near the end of another road, about 600m by road.

[#13]

Obviously your hypothesis is completely different from mine. It seems to me to be very simple, something that is switched on most/all of the time when whoever is home (and awake) and switched off when they are not.
I don't agree that anything operating in the hertz range, ie. having a fundamental frequency in the hertz range, is going to have any harmonics of any strength at frequencies high enough to affect ADSL.
@ejs
A couple of thoughts - v3's (although it might be firmware dependant) can be badly affected by cordless &/or mobile phones and their chargers, so switching them off should be in the mix as well.
Also, I presume your bell-wire is disconnected at the master at least.
if you can get preferably the full version of RouterStats running or possibly DSLstats, you should get an idea of what frequencies are affected from a Bits/Tone plot.

[#14]

I think the times the interference starts and stops are the most useful evidence I have, but that would imply it isn't anything in my house nor in the nearest neighbour who also had the problem (at least nothing they could think of on Saturday morning anyway). The interference might extend a bit further in that direction, so it could maybe be in the next house along in that direction. The interference doesn't seem very strong on the other side of the road.
There are no cordless phones in our house, there aren't any mobile phones or their chargers anywhere near the router. The bell wire was disconnected inside each of the internal sockets some time ago, I believe they are all independently wired to the external box. There is no difference in in sync speed, nor with the affects of the interference, between either of the downstairs internal sockets. The router is usually plugged into the socket that isn't the "master", but again, it doesn't make any difference.
The behaviour of the DG834Gv3 stats are a little strange (but consistent) when the interference is present: the SNRM remains largely unchanged, the Severely Err Secs count ticks up at about 1 per second, while the Error Secs count remains unchanged. But then it doesn't keep count of either of those error seconds counts in ADSL1 mode anyway, or if it does it didn't print them in the stats. People seem to say ADSL2+ might be unstable on long lines, sometimes giving a reason that sounds good but doesn't tally up with how things actually work in reality, and yet on ADSL1 it would be losing sync every couple of seconds. Although the behaviour under ADSL2+ is barely more useful. Although I suppose if it were dropping sync 100 times a minute, then the DLM might notice the problem. I suppose when the 582n arrives I can give it a try, it will probably be mercilessly tortured if it performs worse than the dg834gv3.
I can obtain a bits/tone graph, I didn't think it was particularly insightful, but I might post one this evening if the problem re-occurs, assuming I can manage to load this website of course.

[#15]

If that's an external NTE, make sure the bell wire (Orange/white) is disconnected there.