Migrated from Be/Sky to PlusNet - Lower than expected Sync

[#1]

Currently under attack by the REIN gods; here are two recordings taken at the approximate frequencies of a high SNR tone, and a dropped tone, respectively. The noise present at Tone 291 is present all throughout the spectrum, but primarily at the high end.
Tone 153 - 659kHz - http://www11.zippyshare.com/v/HTTv4Tfr/file.html
Tone 291 - 1254kHz - http://www11.zippyshare.com/v/ILAwdKGC/file.html

[#2]

Tone 153 is not one that is affected by REIN. That sounds like typical White Noise. I should have mentioned that in addition to the 7 standard gaps, there is a tone that has only a 2-bit loading and it can move around. In your last graph it's on 153, but if you look at the graph on reply #17 it's nearer 155
The sound on the recording for tone 291 is more like the sound you need to be looking for, which should cease when the interference goes away. BUT 291 is one of the standard gaps as mentioned in reply #21, AND the reason you can hear this noise is because of the proximity of the actual noise which on the graph on reply #17 was centred on 288, and the graph on reply #27, it's nearer tone 284. But as you haven't posted the graph that reflects your comments in replies #29 & #30 I can't say accurately where it was centred in this instant.
I would also say that 1200kHz is not ideal, you should be better nearer 1177kHz or 1242kHz, but again you need to refer to the graph to see how much up or down the harmonic is shifting.
As I mentioned, you will get some noise from a variety of in-house appliances, but generally speaking with the receiver about 6ft or more away if you don't get any noise from a particular one, it shouldn't be an issue. The ultimate check is to look at the SNRM graphs and switch the appliance off and watch for a shift in the SNRM level, likewise when you switch it on. Generally anything around 0.1/0.2db shouldn't be a worry. Shifts of ~0.5db resulting from cross-talk aren't uncommon either.

[#3]

Quote from: Anotherone

Tone 153 is not one that is affected by REIN. That sounds like typical White Noise. I should have mentioned that in addition to the 7 standard gaps, there is a tone that has only a 2-bit loading and it can move around. In your last graph it's on 153, but if you look at the graph on reply #17 it's nearer 155

Tone 153, or at least that region of the spectrum (the scanner I'm using doesn't show the exact frequencies, just a sliding scale marked in 200kHz increments, so I'm having to guesstimate here) was what I chose as a "reference" point (of no interference), due to its extremely high SNR. At the time of the test, it was heavily loaded; 13-15 bits if I remember correctly, so thought that would be a good benchmark to compare the interference higher up against.

Quote from: Anotherone

The sound on the recording for tone 291 is more like the sound you need to be looking for, which should cease when the interference goes away. BUT 291 is one of the standard gaps as mentioned in reply #21, AND the reason you can hear this noise is because of the proximity of the actual noise which on the graph on reply #17 was centred on 288, and the graph on reply #27, it's nearer tone 284. But as you haven't posted the graph that reflects your comments in replies #29 & #30 I can't say accurately where it was centred in this instant. I would also say that 1200kHz is not ideal, you should be better nearer 1177kHz or 1242kHz, but again you need to refer to the graph to see how much up or down the harmonic is shifting.

The noise recorded was present pretty much throughout the spectrum, seemingly coinciding with ALL of the gaps in the bitloading; it was also present elsewhere, at lower intensities, with what looked like a corresponding drop in usable bins.

Quote from: Anotherone

As I mentioned, you will get some noise from a variety of in-house appliances, but generally speaking with the receiver about 6ft or more away if you don't get any noise from a particular one, it shouldn't be an issue. The ultimate check is to look at the SNRM graphs and switch the appliance off and watch for a shift in the SNRM level, likewise when you switch it on. Generally anything around 0.1/0.2db shouldn't be a worry. Shifts of ~0.5db resulting from cross-talk aren't uncommon either.

I actually killed the entire power to the flat at one point, apart from the socket the router was plugged into and checked things then; no visible difference in the stats thankfully, so hopefully that pretty much rules out most of the stuff in here
Anyway, I sat here for about an hour and a half this evening, waiting for the interference to stop; eventually my fiancée got sick of hearing the constant buzzing and so I turned it off I checked it a couple of hours later, and lo and behold - no buzz like that, at any frequency. I just checked again - while writing this - and it's still absent. I managed to repair the battery compartment of the scanner - it uses those nasty cork-board contact plates which have a habit of bowing outwards over time, a cut-up PAYG Top-Up Card strategically inserted on the outsides of the corkboard sorted that, along with some resoldering of the contacts themselves - so now the next time I hear that godawful noise I can pop some headphones in and go out in an attempt to find the source
The one thing I'm concerned about is the potential that the interference may be at some point mid-pair. It's a direct-DP line, and I have NO idea the route it takes along the roads.
If I could have one question magically answered by a genie - or at least by someone more knowledgeable than myself - it would be what on earth could cause that wide a range of interference (seriously, it ranges from 800 to over 1600kHz - above that, it's also accompanied by some sort of faint clicking noise), with such a level of intensity that directionality, and moving around in the flat doesn't change the signal whatsoever when the power in the flat is out (any changes through moving around I was able to attribute to devices in my flat).

[#4]

Quote from: mdc

...... here are two recordings taken at the approximate frequencies of a high SNR tone, and a dropped tone, respectively.

Sorry, my confusion re tone 153, because  the plot that you were discussing wasn't present and my observation of the previous graph showed it as the 2 bit tone, that made me overlook your comment that I've highlighted.

Quote from: mdc

The one thing I'm concerned about is the potential that the interference may be at some point mid-pair. It's a direct-DP line, and I have NO idea the route it takes along the roads.

Whilst this is possible, I think it highly unlikely - there would be a considerable number of users on many DPs who would be suffering problem which would have resulted in some other action by now, so I would only start considering that if it turned out the cause was nowhere in the current building.

Quote from: mdc

If I could have one question magically answered by a genie - or at least by someone more knowledgeable than myself - it would be what on earth could cause that wide a range of interference (seriously, it ranges from 800 to over 1600kHz - above that, it's also accompanied by some sort of faint clicking noise), with such a level of intensity that directionality, and moving around in the flat doesn't change the signal whatsoever when the power in the flat is out (any changes through moving around I was able to attribute to devices in my flat).

I thought when I quickly read this, it was a more general query about interference across those frequencies as I'd already mentioned, I think at least twice, the likely cause of this particular problem, so perhaps a slightly more detailed explanation may be in order.
The general radio frequency spectrum these days is rather cluttered with not only the obvious transmissions eg. Radio & TV Stations, Beacons, Radio Amateurs, a multitude of Wireless devices from phones, security systems, wireless routers, home plug type devices, wireless microphones, taxi and  emergency services communications, as well as many other devices - anything with an oscillator designed to control the frequency that a device operates at, etc etc , (the list goes on) but it is also full of noise, a significant amount of which can be classed as interference of one sort or another.
All things that either use the RF spectrum directly or are capable of putting some "noise" into the RF Spectrum - that just about covers any type of electrical or electronic item - are covered by stringent regulations and specifications with regard to the frequencies & amount of spurious noise that they can emit. This is given the general term EMC - ElectroMagnetic Compatibility.  Equipment that is designed to "transmit" is also regulated in the amount of harmonics of the fundamental frequency that they can emit.
The cumulative effect of a large number of such pieces of equipment when used in a fairly small area can result in fairly high levels of  "noise" across parts of the spectrum, even where individual devices are within specification.
However when a piece of equipment is outside the legal limits, whether because it's gone faulty or because it's a poor cheap design that never complied - eg. cheap chinese copies of certain equipment, then the amount of interference can be severe.
One of the most common causes of such interference these days is Switched Mode Power Supplies which exist in many many pieces of electronic and electrical equipment. As the name implies, the power supply works by operating in a "switching mode" ie. the controlling part of the circuit is switching on and off. Note the similarity with digital circuits which switch signals on and off (0s & 1s). The waveform of these is what's called a squarewave. When you do a frequency analysis of a square wave (a complex mathematical procedure) you find not only is there the fundamental frequency which is dependant on how quickly it is switching on and off, but there are multiple harmonics - these are integer multiples of the fundamental frequency and can extend to very high multiples.
So in the case of your interference problem, you see repeated harmonics ranging up the spectrum. It is more common with SMPSs for the harmonics to be the odd harmonics ie x3, x5, x7 etc.

[#5]

Just to finish of my post above which I didn't really have time to complete, the spacing between the dips is around 62kHz.
If what we are seeing on the graphs is just the odd harmonics, that would put the fundamental at ~31kHz. If it's all the harmonics (quite possible with faulty gear) the fundamental would be around 62kHz.
Low energy fluorescent bulbs (CFLs) quite commonly operate around 30kHz, but they aren't the only equipment. Larger fluorescent fittings with electronic ballasts tend to run at the higher frequencies. This may also be true for other "industrial type" lighting, not an area I've taken too much interest in recently. But again the cause could equally be other equipment with a SMPS.

[#6]

Quick update; narrowed the source of the interference down to two potential sources; there's a MASSIVE amount of interference coming from our local Co-Op, downstairs. It seems to be coming from the area used as a bakery, although we weren't allowed inside to check the individual pieces of equipment. The interference stopped before the manager arrived on-site, so will have to try back later.
Also, each one of the overhead poles (fed by the same trunk as our line) are emitting a massive amount of RF around 1300kHz, even after the RFI from Co-Op stopped. I'm unsure what to make of these results tbh; it definitely seems like something in Co-Op is blasting the area with RF, but there's also an additional source somewhere which is so-far unexplained.

[#7]

Two interesting graphs here; showing an event last night at 7:22pm, lasting precisely 8 minutes before rapidly tailing off.
Second graph shows an event this morning at 7:08am, also lasting 8 minutes and then rapidly tailing off.

[#8]

Can I suggest 2 things in relation to the two graphs above, can you in future post the matching Bits/Tone graph which will then give a more meaningful picture, also I suggest the tones that you monitor are the means from several incidents around the tones where you see the dips when interference is present (as it would appear to shift slightly up & down each time - not totally unexpected).
Can you also post a Bits/Tone when everything is normal - ie no interference, and ~17Mbps sync speed, we haven't seen one of those yet, so unable to form an opinion of normal working.

[#9]

Also what is the sample interval on those graphs.?
Anotherone & I would usually recommend 10 seconds (as low as RS goes) to identify short duration events.

[#10]

Here's the graph when things are perfect, as requested.
Polling interval is currently set to 20s, anything lower gives me a lot of polling failures.
I went down to the Co-op and had a chat with the main manager; he claims that nothing they have there could be causing it, and "everything runs off Bluetooth", which struck me as completely irrelevant, and just a transparent attempt to fob me off lol.

[#11]

Be fair he is unlikely to know anything relevant.  A flakey fluorescent tube could be the culprit.  The right answer here is "In that case could I  wander around your store with my detector to confirm that the problem is not here next time I see the problem?"  ...please arrange this with your staff in case you are not here...

[#12]

Yeah, it was a no-go. He's happy for OpenReach to do it, but not me.
Just had another dropout, graph included, looks like 8 minutes again.
I noticed when I went downstairs that were was a Hutchinson Communications box downstairs, spitting out a ton of RFI. Seemed to fade away though, so I came back upstairs to find the SNR back to normal. I think a call to Three might be in order.
EDIT: One of these: http://www.prattfamily.demon.co.uk/mikep/jpegs/phot2k2.htm

[#13]

Just out of curiosity, are telegraph poles SUPPOSED to buzz like crazy throughout the MW spectrum? Or is this a symptom of the issue?
I tested 4 today, all of which buzzed at decreasing intensities, heading away from my flat down a side-road.
I'm just wondering whether this GSM/3G cabinet has a dodgy power supply, spitting out mediumband RFI which is somehow being transmitted through its antenna and into the overhead cabling from the poles outside, then coming through the trunking and affecting my line.

[#14]

A question for clarification
Are these poles
Carrying only electrical power lines
Carrying only telephone lines
Carrying both

[#15]

Frankly, I would forget what you are getting from telegraphs poles at present, this could be a red herring. Cables can pick up and re-radiate all sorts of things, as long as it isn't induced into cable pairs, it's not much of a concern. Are there overhead power lines on the poles as well?
Thanks for the graph. Broadcast station pickup is not uncommon with overhead lines. Looks like you have 936kHz & 1116kHz. Nothing to be concerned about. When you have time, and remembering to do graceful disconnects, you might like to compare the "normal" situation with the Mk3 plate fitted, I assume from your previous posts that it's not fitted at present.
Regarding the manager downstairs in the co-op, I would pop in again when he isn't busy and say to him something like "further to our chat before, I'm talking about interference from electrical equipment, it could be a light fitting. I wouldn't be asking for access to non-public areas if I think I've found something that suggests it could be a piece of your equipment, just for you to switch it on and off whilst I check  the interference levels. I would hope for your co-operation in that."
If he's still "uncooperative" then I would add, "well in that case I will be contacting your head office as well as OFCOM". Call his bluff.
Regarding sampling, 10 seconds is the minimum for the TG582n, I have used 8 seconds on some other broadcom boxes without issue, it depends very much on both the processing power of the modem/router and also the computer you are using. If either get maxed out by anything at any point data is required, polling issues arise. This also depends on the monitoring program, I have used DMT on a 5 second sampling for noise without issue on some broadcom boxes.