Massive amount of noise on my FTTC line - any thoughts on the cause?

[#1]

Hi

I'm at the end of a fairly long FTTC line (about 1.9km) so I should only get between 11 and 19Mbps because of attenuation.  Sometimes I'm only getting 3Mbps because of noise.  See the attached QLN and Bit loading charts - which show that the noise is drowning out signal in most of the frequencies I need.  OpenReach have done their best with the line engineers who have swapped pairs, cards and polished the copper till it shines - but it's had little effect.  Strangely most of the noise peaks seem to be on MW radio frequencies (like Radio 5 Live).  Normally a twisted pair shouldn't notice this as they work on the difference in the signal between the two wires and the MW wavelength is over 300m so much more than the 1 mm between the wire cores. If it is MW radio then how could it be getting in?  Maybe a split pair in the wiring?  Maybe a problem with earthing/wiring in the cabinet? Maybe another wiring problem somewhere? The BT 'REIN' engineer couldn't find anything but he only seemed to wander about with a radio listening for noise on 612kHz - which isn't a problem on this line!

Any ideas would be most welcome!

[#2]

Hi @PaulK

I have tested your circuit, and whilst I can't find any faults which would be causing the noise, I've noticed that your profile is banded to 6.4Mbps. 

As you are syncing at the full banded high cap, I've requested that this be lifted so your profile can work on the full 40/10 service. 

Once this happens, we will be able to investigate the noise issue more thoroughly as tests wont throw up a banding fault. 

Kind Regards, 

MoR

[#3]

Hi MoR

Many thanks for taking the time to look into this.  OpenReach have had various attempts at playing with the profile and speed limits but I think they have failed because of the bit errors generated in tones with high noise levels.  Higher tones (say 320 upwards) don't seem to have much noise and are behaving as normal - with decreasing bit loading as signal attenuation increases with frequency.  The trouble would appear to caused by the noise (which needs to be dealt with) and the way the DLM handles it.  Basically, I think, the DLM starts off with a high nominal speed so it tries to use tones all the way up to around 600 - normally successfully but with big gaps in the lower tones due to poor SNR.  It then spots lots of page faults, FEC errors etc etc and says to itself - must cut the line speed, and hence the highest tone it tries to use - which it does repeatedly with obviously no effect as the noise is at lower frequencies.

The QLN test result was attached to the original post and shows the main peaks quite clearly - these are mainly at MW radio station frequencies which does suggest an aerial type problem.

OpenReach were out again yesterday and have played with the profile and said they are going to monitor the line for a couple of weeks.  I hope I'm proved wrong but I suspect that this will do nothing as the noise is still there.  They say they have checked for split pairs on the line - which could cause 'aerial' type noise.  Similarly any relatively lengthy untwisted wire sections could reduce the rejection of common mode aerial effects.  Interestingly, one of the previous engineers disconnected my line completely and connected his analyser direct to the card in the (ECI) cabinet - he saw the same QLN and bit loading picture as he was seeing on my line.  So - it could be that the noise is getting into the cabinet from other connections or even from a cabinet wiring/screening problem (I gather that ECI cabinets are known to have this problem sometimes).

So - this appears to be a horribly complex issue that will only really be fixed by OpenReach sending in a 'proper' telecoms Engineer with some useful test kit (spectrum analysers etc) rather than their hard working and highly trained technicians.  Or they may find it's cheaper to just run some fibre down the road!  In the meantime I am stuck as upload speeds of around 300kbps make any serious attempts at working from home a bit flakey, reverting to ADSL will be much worse as the exchange is a few more km away and the noise is at prime ADSL frequencies.

I should have thought that from Plusnet's perspective, you are, on my behalf, renting a card in a cabinet and line from OpenReach and it is really clear that OpenReach are not providing what we're paying for.  It is surely up to them to fix this problem as the monopoly supplier in this area?

[#4]

Hi @PaulK, thanks for getting back to us. I can see my colleague has identified the banding of your speeds and got this raised to be removed overnight. However, upon checking the previous engineer notes it looks like the set speed profile was put in place by the previous engineer in a attempt to see if the line becomes more stable as the 6.4Mbps looks to be the estimate speed for the length of your line. It looks like he's done this to try bring stability to the connection however we can see the connection continued to drop there after.

That brings me on then to the length of your line - we can see when testing that the length of your line looks to be 2274m meaning that the download speed and upload you've experiencing during this fault has been open is around the expected speed for fibre connection of such length.

I've included a graph below which should explain the attenuation and how as line length reach over 1500m with FTTC the speeds really start to plummet.

It's not often that we see FTTC is available for customers pushing almost 2300m and I feel we are investigating and keep passing back this fault to our suppliers that won't realistically be resolved. Upon checking over your fault and our suppliers notes I can confirm we look to have completed all quality gates in investigating the issue. Moving forwards then, I've updated your ongoing fault with my verdict of the issue which can be seen here.
Please let us know if you reply over this support channel and we'll happily get this picked up for you.
 

[#5]

Hi @BD - many thanks for your swift response.

Without going into great length - I really understand how signal losses increase with frequency and cable length which in turn sets the highest frequency that can usefully be used for data transmission.  Our line is physically about 1.96km long with maybe an extra 50m or so for crossing the road a few times.  The length estimates from OpenReach vary widely according to the quality of connections on the line when they measure it!  And, as your graph shows, our current performance is equivalent to someone around 3500m from the cabinet - so there is clearly something wrong.

The problem with our line is not actually the signal attenuating into noise at high frequencies but over 30dBm of extra noise at relatively low frequencies.  The noise appears to be predominately MW radio station stations (or at least on precisely their frequencies) - and it's fairly new.  We used to get 10-15Mb/s so the problem isn't our distance from the cabinet - it's something that's changed to increase the noise level.

So I am hoping against hope that - 

* OpenReach finally send out a specialist who can find out what it is in their cabinet/wiring/cable run that has increased susceptibility - and maybe trace the source of the noise.

* or maybe a Plusnet community member has seen this before and has a solution.

My hunch (and it's no more than that) is that the problem is with screening/earthing of cables or perhaps the ECI cabinet.

All further positive help welcome!

Paul

[#6]

Hi PaulK, I wish there was a solution to this we could recommend however the performance does seem to be congruent with the length of the line which means there's no appropriate action we'd be able to take. 

I've just added an update to your account, here's a direct link to the ticket: https://www.plus.net/wizard/?p=view_question&id=185414886

Adam 

[#7]

@PaulK Without being too precise, what is your location? Is the line over/underground? What times does it occur? Have you considered it may be a REIN issue?, see https://kitz.co.uk/adsl/rein.htm

[#8]

Many thanks @dws1900  - yes I've looked into REIN.  Searched the house with a MW radio, rechecked with everything other than the modem and router off.  Tried a different modem and router.  etc etc.  It is clearly radio emitter caused - but it is not very intermittent.  This is why a cable earthing, screening or split pair issue is my favourite theory at the moment as this would be consistent with the excellent radio 5 live reception on the line.  It was interesting that the Broadband engineer who came out picked up the same noise in the ECI FTTC cabinet when connected direct to the 'card' and with my line completely disconnected - so the cause could be from another line (either to/from another property or one of their many dead lines) or within the cabinet itself.  This is one of the reasons that I am really really frustrated with the Plusnet team who keep saying 'its the line length' - I bet that nearly every other user of our cabinet has the same problem but most of them have 30Mb/s so loss of maybe 5Mb/s is just an irritation to them.  For obvious reasons, the OpenReach guys are not keen to disconnect other users for a few minutes to check the noise levels on their lines - and for some reason they can't just dial up the relevant cards and get the QLN/bitloading data off them.

for info - I'm about 2km from cabinet 7 of the bosham exchange - near West Ashling in Sussex.  Our line is a mix of over/underground and seems to keep a vast number of Open Reach engineers employed!

[#9]

Hi Adam @adamwalker 

There is a solution - investigate the noise source with the right kit and the right expertise.  The BB engineer a week or so ago who came out observed the same noise and bitloading when connected direct to the card in the cabinet - ie without the losses etc associate with the 1.96km of twisted pair to my house.  So - this would suggest that the issue is not (only) due to my line but something to do with the noise that is getting into the FTTC cabinet.  As a professional physicist and CEng with more than a passing understanding of things to do with signal processing and RF it is really frustrating that the combined might of Plusnet and Openreach are not able to do a simple 'chase the noise' exercise.  For example, can you connect to my card/slot in the cabinet (which I assume you rent from BT) and pull off the QLN and bit loading data from your end?  If you can, do you have other customers in the same cabinet and could you pull the data off their cards too to see if they are seeing the same noise levels.  

On a more positive note - it is clear that OpenReach specialists are trying to improve things by adjusting the details of the line profile - experimenting with G.INP and interleaving.  These should improve resilience to noise (which is welcome - and we have maintained connection speeds around 5 to 6 Mb/s over the last day or so) but will not deal with the fundamental problem - where is the noise coming from?

Paul

[#10]

Hi @PaulK

I am afraid what my colleagues have said are correct.

We have exhausted all possible quality gates to try and fix the fault. Sending engineers for this issue will not make a difference as the last engineer has mentioned that due to the length of your line, you would not be able to receive the estimated speeds. Higher speeds than what it is currently on at the moment would only make your line stability suffer which is what is prioritised over speeds. As shown below, the connection has been stable since the last engineer visit and with the current line speeds you are getting.

The engineer also checked other customers fed from the same distribution point and they were all running similar speeds which is expected for the length of you your line.

We're sorry for any inconvenience this might cause.

[#11]

@PaulK 

See http://downloads.bbc.co.uk/reception/pdfs/AMLWtrans.pdf

for a location list of MW transmitters, include Radio Five Live.

Can you determine the exact frequency the noise is on?

Other possibilities are

Any internal extensions not using CW1308 cable.

Only one wire connected to the master box (does your phone function ok, my neighbour had this but didn't realise there was an issue as he never used the landline)

Extra wires teed off the incoming line, like a defunct extension for an old Sky satellite receiver?

Bell wire connected.

All these should be  determined by the OR engineer.

Having the latest Master socket (5C) helps as the filtering has improved according to OR. Have OR supplied one, its thier responsibility as they own the Master socket.

Did you get a GEA test report from plusnet, to see if that shows interference?

Re REIN, if you search on here you will find my experiences with REIN when I was on adsl with my previous supplier.

[#12]

An old thread, but it's relevant to my own circumstances as I live a long way from the exchange, but there's an AIO FTTC cabinet nearer to me than it is to the exchange. Unfortunately, from what I read elsewhere, the VDSL signals have to be transmitted at a level no higher than the ADSL signals that are passing through the cabinet. If what I read is correct, the distance of a PCP & FTTC cabinet from the exchange will influence the VDSL transmit signal levels.

[#13]

Hi Cancunia - that's an interesting observation which could affect the high frequency end of the signal.  I think that normally, VDSL signals are actually lower than ADSL - but as you say - ADSL has to go further.  My problem is actually in prime ADSL bands (frequency channels 160-300 - so approx  600kHz - 1.2Mhz) with huge noise peaks at common MW radio stations such as radio 5 live.  Normally, a balanced twisted pair should not notice these frequencies as they have wavelengths >>> wire spacing. But in this case we have noise levels of over 40dB above the "background" which is much more than the total signal loss from the FTTC.  As the higher frequencies (it syncs up to channel 750) are OK I think the wire balance is OK - so the noise must be getting in somewhere else.  Curiously one of the OR technicians was able to run a QLN test on my "signal" direct from the card in the cabinet with my line to the house disconnected - he saw the same noise spikes.  So, our conclusion was that somehow MW RF is getting into the cabinet.  We know that there are a fair number of "dead" and broken lines in the cable bundles so my theory is that one of these could be acting as an aerial and might be shorted onto a poorly earthed frame or something in the cabinet.  Another OR engineer has commented that wiring and earthing in ECI cabinets can be poor.  Unfortunately noise from aerial/radio sources is not the same as "REIN" or "SHINE" which are both intermittent and we have not yet managed to raise an OR engineer who has the slightest understanding or radio aerials - they are all much better at looking for dodgy switched mode power supplies etc (which are also a problem for many).

Plusnet have not been much use as they don't seem to have many deep technical people these days and seem to have their hands tied by the contracts/processes they have agreed with OR.  No matter how much I try they seem to be unable to escalate this problem to a level at which OR will engage someone who can investigate and solve this problem.  What's doubly frustrating is that we had much better internet a year ago - so we know it's not simply range from the cabinet!

[#14]

From what you've written, it does sound like some noise is leaking into the cabinet, maybe a stray pair acting as an aerial etc. I'd have thought that others using the cab would have similar problems. 

In fairness to Plusnet, I know from recent work with BT GS that Openreach work strictly to their own rules and act at arms length to any supplier, it's very difficult to get anything escalated up into the organisation. It's a twist in the open market competition rules so that no supplier is shown favouritism, OR are often suspected of hiding behind these rules in order to save costs. 

Perhaps there's a chance that OFCOM can help?

https://www.ofcom.org.uk/complain-to-ofcom 

or 

https://www.ofcom.org.uk/phones-telecoms-and-internet/how-to-report-a-complaint/dispute-resolution

[#15]

@PaulK

Are you near a MW radio transmitter?

This is my ADSL2 QLN, apparently the MW transmitter about 2 miles away from me broadcasts TalkSport on 1053 kHz.

Telephone grade cabling has fairly few twists per inch (or whatever unit of distance) and its noise immunity is really not that great. And my line is all underground. If you are fairly near a MW transmitter, and especially if your cabling is overhead, I'm not convinced there is anything abnormal about your QLN.