Slightly slow ADSL2+ speeds � possible PSTN issue?

Hi all,

First-time poster here although hello to any long-time Plusnet customers who might remember me from their forums.

I left Plusnet last year when moving home, and am now with Andrews & Arnold. When doing up this new place before moving in, a modern phone line and NTE5 had to be installed for us because all the house had was an old GPO junction box � the new line was wired in to that for us by the Openreach guy.

The thing is though, the line's never performed quite as well as I was expecting. I'm on cabinet 1 and only half a mile from the exchange by road (shorter as the crow flies) so was expecting 20+ Mb/s sync speeds. Instead however my line gets between 16�18, with throughput lower than that obviously. Occasionally the ADSL drops too, perhaps for a few minutes, perhaps for an hour, which isn't ideal. Annoyingly my modem is showing 0dB for both up and down attenuation, but it is kind enough to tell me that my down SNR is 3dB whilst I'm currently connected at 18Mb/s, so I'm guessing when I get 16Mb/s the SNR has been upped to 6dB. It also reports 58 UAS errors in six minutes' connection time.

If I do a quiet line test with a cordless phone (via the base station directly plugged in to the test socket), I hear occasional quiet bursts of rustling static over the general hum of the handset, and I'm fairly sure these aren't cordless interference because if I pull the line out of the socket the rustling seems to disappear. If I then test with an old corded phone I hear music!!

Last time my line dropped for a significant period of time I reported it to A&A as a PSTN fault; an Openreach engineer came round but couldn't find anything wrong (though acknowledged the existence of music through the corded phone) so nothing changed. A&A's suggestion was that it was the corded phone picking up interference somehow, but otherwise they've been relatively unhelpful, not really advising whether they think there is a problem that needs further investigation or not.

So I put my quandary to you, the esteemed TBB community: in your opinion is there something iffy going on with my line, and if so how might I further test/prove it?

Need a modem that is known to report line stats correctly for a start.

If the phone is a corded and un powered one, then unplug and turn off all the ADSL kit and connect phone to master socket, is the line quiet, if not e.g. the music present then report as a voice fault to who you pay voice line rental, as a second step also remove the lower part of the faceplate and test at the test socket.

Borrowing another corded phone that is known good will help too.

Thanks for your reply.

As mentioned above I've already done the testing through the test socket etc., and indeed reported it to A&A as a PSTN fault as I have my phone line through them too.

Unfortunately I don't know anyone else with a corded phone. I could buy a cheap one I guess, but I'm still not sure it'd help much: if I can't hear music then that probably suggests I don't have a PSTN fault, but if I do hear music it still doesn't prove the existence of a fault because presumably A&A could just say that new phone is picking up intereference too.

Two different phones picking up noise that is audible music would be so rare that it points more towards an issue with the phone line/wiring rather than the actual handset.

http://www.argos.co.uk/search/corded-phone/ start at £5.49 and while not stellar you don't need anything fancy to indicate the noise is not the phone hardware itself

I'd also suggest if you can hear the music enough to try and match it to a radio station playing same thing at the same time.

I wonder if the OP�s line has a reasonable length of aerial cable on its route from exchange to property ?

I have known this scenario pick up radio transmission, maybe also with a minor earth contact in the mix too.

Possible

In reply to a post by Be3G:

Thanks for your reply.

As mentioned above I've already done the testing through the test socket etc., and indeed reported it to A&A as a PSTN fault as I have my phone line through them too.

Unfortunately I don't know anyone else with a corded phone. I could buy a cheap one I guess, but I'm still not sure it'd help much: if I can't hear music then that probably suggests I don't have a PSTN fault, but if I do hear music it still doesn't prove the existence of a fault because presumably A&A could just say that new phone is picking up intereference too.

But AA don't provide a phone service, so their support and the Openreach engineer may be confused. Looking for what they may be calling a phone fault when there can't be.

Yes it's the Public Service Telephone Network, but no telephone service. PSTN may be the acronym in parts of the world for Public Service Telecommunications Network, but I don't think it is here.

As for hearing music when the broadband equipment is disconnected at your end, very odd indeed. Almost as though it is parked on a "dummy" service to keep it alive when the broadband is not. That could explain the engineer finding no technical fault.

I didn't see your post before writing mine Zarjaz. That could explain the music, but shouldn't a fault have been detected?

They don't provide a voice as a in make phone calls, but a bare bones WLR is still available.

Yes, for the avoidance of any doubt that is what I have, albeit with an added incoming calls facility that A&A were offering as a free but optional extra when I signed up with them.

RobertoS, if you're interested: http://aa.net.uk/broadband-phoneline.html

I went with this specifically because of the ease of getting faults fixed when both ADSL and PSTN are via the same supplier. (Well, that and their much lower-than-average monthly charge for a phone line compared with the inflated costs most others charge.)

Interesting idea. I can't at first glance see anything obvious though: my line seems to be just a normal telegraph line that travels to the pole across the road, and it looks like it goes underground after that (I can't see a line leading to another nearby pole).

There is one other little titbit I forgot to mention earlier, in case that helps prove (or disprove) your theory: I only hear music when the phone is plugged directly in to the master/test socket, not through the micro filter.

Also, re. MrSaffron's suggestion earlier, unfortunately I can't identify the the radio station (assuming it is one) at all. It's Indian music and so far all I hear is music, no talk that might give me a hint.

I did explain that in an earlier draft, and that the OP and many readers may not know or if they did may have forgotten, but it began to read like a treatise .

I cut it to the essential. My point was that the CS staff and the engineer, who may or may not have been from Openreach and in any case could be a recent recruit, could have been getting confused. For instance, following "Occasionally the ADSL drops too, perhaps for a few minutes, perhaps for an hour, which isn't ideal", we have:-

In reply to a post by Be3G:

Last time my line dropped for a significant period of time I reported it to A&A as a PSTN fault; an Openreach engineer came round but couldn't find anything wrong (though acknowledged the existence of music through the corded phone) so nothing changed.

That almost certainly means a phone engineer was sent, not a broadband engineer. When the problem was the broadband had dropped for a significant time.

Then we have:-

In reply to a post by Be3G:

I've already done the testing through the test socket etc., and indeed reported it to A&A as a PSTN fault as I have my phone line through them too.

Uh uh!

We don't really know what the OP said to CS, but the odds (in my opinion) are that a phone/voice fault has been reported to Openreach. The wrong sort of engineer was sent. Would they be bothered about music on a line that didn't have a phone service, given that the engineer could well know nothing at all about the effect that would have on broadband?

Zarjaz's post was relevant, as it tends to annul my theory any such lines are terminated on a service that plays music, as at the moment Openreach and particularly BT Wholesale that I think the OP is on (I think their ADSLx service is only on BTW not TalkTalk but I could be wrong) do not generally do a line-only connection for broadband.

In reply to a post by Be3G:

There is one other little titbit I forgot to mention earlier, in case that helps prove (or disprove) your theory: I only hear music when the phone is plugged directly in to the master/test socket, not through the micro filter.

That is important information. The micro-filter passes everything on the line through to the modem. It removes all signals in the broadband frequency range from the phone port.

That is to make it possible to converse on the phone. The modem just looks for the expected broadband frequency range and ignores the rest.

However, although it is important, it is confusing in itself. As music is in the audible range so should get through the filter anyway.

In reply to a post by Be3G:

Yes, for the avoidance of any doubt that is what I have, albeit with an added incoming calls facility that A&A were offering as a free but optional extra when I signed up with them.

You've got an intriguing problem, I wonder if the music is possibly an incoming calls test facility that is triggered by your particular corded phone when doing a quiet line test (?) via the test socket. The music is not heard when the phone is connected via a filter and perhaps due to the filtering of the frequencies a test is not triggered.

Just a wild guess

It was indeed a voice fault that was reported, and the engineer who visited was tackling the problem as such. That is generally the best approach to take though, is it not? I've been using ADSL now for fifteen years with all manner of issues at my previous abode and it was always the case that if there was a broadband fault which coincided with a phone/voice problem too, it was best to raise a PSTN fault first as that would in all likelihood fix the ADSL problem too.

Bear in mind that I do have a line with phone service. I can receive calls and make outgoing non-chargeable calls � it's really a normal phone line just with outgoing chargeable calls barred.

(And yep, I'm on BTw.)

Also, whilst not entirely relevant since you have an incoming calls and outgoing non-chargeable calls service, A&A say this:

"No telephone service is provided on this copper pair, it is purely for broadband use.

Whilst there is no telephone service as such, the pair does have some audio when connected to a telephone - this is purely to stop any engineers thinking the pair is unused and taking it for another installation or fault repair. It also allows a quiet line test so one can hear crackles or interference that may be the cause of a fault. As such we still expect customers to have a telephone handset available for fault repair and diagnostics." - https://www.aaisp.net.uk/broadband-phoneline.html

They don't say what the audio is though...maybe normally just a dial tone?

Edited by 4M2 (Sun 15-Oct-17 18:29:54)

So wild that I'll discount it

Yep OCB dial tone I expect.

I feel the post by Zarjaz and the fact that the phone does not produce music with an ADSL filter in place are very relevant here.
The ADSL signals are in the same frequency bands as LW and MW radio signals, which are a significant source of interference with ADSL.
If, as suggested, the overhead line were picking up a radio station it is possible that OP's phone could be rectifying that RF signal and producing the audio that's being heard. The micro-filter is there to remove all but audio frequencies from the phone, it doesn't effect the ADSL at all.
A different phone might not behave in the same way but it seems clear that the offending signal is on the line at radio frequencies rather than audio. That must be why there is an ADSL issue.
The OP needs to ensure the fault is reported as an ADSL one rather than a phone fault so that he gets the right engineer who knows how to deal with the issue

Thanks for your input MatHal, sorry it's taken me a few days to reply.

Inspired by that post I've done some more testing. I have a couple of spare microfilters but also a simple telephone socket doubler, so I thought I'd see what happens with either of those, in case the situation was that inserting anything between the corded phone and the NTE5 killed the music, not specifically a filter. But I can confirm that your and Zarjaz's theories seem to hold up, because I don't hear music through any of the filters but I do hear it when the doubler is in place instead.

With a bit of luck, I also happened to �tune in� when there was a news broadcast rather than music, so I listened very very carefully and after a few minutes I heard the radio station being announced: Punjab Radio, which is on 1584kHz in London. This fits neatly within the ~2MHz maximum frequency of ADSL2+.

So, with that in mind, does it look like there's now no point to me purchasing another corded phone to test with, because it might or might not let me hear the radio station so won't be conclusive either way?

An RF3 filter might cure/block it out ... but you�ll need an SFI visit from Openreach to source one.

Where in the country are you?

London.

North East - Ponders End or Waltham Abbey maybe? Or Southall?

Yep, North East indeed. What's your theory? Is there a nearby transmitter for the station do you think? (I've tried Googling but nothing obvious came up, presumably because these MW stations don't need any massive infrastructure�)

The transmitter is in Sewardstone - just off the A112.

How very interesting! Yes, that's not a million miles from me. With that discovery I've been able to find a planning application for the transmitter which I note is only a couple of years old, so it's quite possible it wasn't operational when I had my line installed at the beginning of last year. Thanks for the info/sleuthing.

It is probably at the south-east corner of Northfields Nursery.

If it is very close and your line just happens to be of the right length it will act as a tuned circuit.

In reply to a post by Zarjaz:

An RF3 filter might cure/block it out ... but you�ll need an SFI visit from Openreach to source one.

Would I be right in thinking that wouldn't be a chargeable visit?

Edit: But then wouldn't the noise still be hitting the ADSL, just no longer audible?

Edited by RobertoS (Sun 22-Oct-17 09:26:55)

So an RF3 filter really would help - that�s what they were originally made for, before their use as �REIN suppressors� came to pass.

May do ... or may not.

Do you no the frequency characteristics for an RF3? If they work to stop AM transmissions then they will also kill and ADSL signals - the problem the OP is seeing is 1584kHz being picked up, and that is right in the ADSL2+ spectrum.

Screened and earthed cables might help but how far back would they need to go?

SFI visits are paid for by the CP, whether they then pass them on to their punter is up to them.

RF3 filters do improve this kind of issue with broadband, technically how they do this I don�t know ...

{You�ve seen me Bob, clearly employed for my looks innit }

As I just mentioned to Bob, not sure of the technicalities of it ... I�ll see if I can discover details.

Problem here is that the OP is extremely close to the transmitter so the field strength will be very high and probably swamping the ADSL signals.


Have a look at your VDSL plots in DSL Stats and you will see notches with zero bit loading where the major AM transmitters broadcast ... so even miles away, it can have an effect. Close up, it can cause wideband issues.

So your suggestion would be to move to Virgin then

In reply to a post by MHC:

Have a look at your VDSL plots in DSL Stats and you will see notches with zero bit loading where the major AM transmitters broadcast ... so even miles away, it can have an effect. Close up, it can cause wideband issues.

Whilst not directed at me, that suggestion has spurred me in to seeing if I can get some more detailed line stats from my modem � I don't generally have access to that kind of thing 'cause it's a router being used as a PPPoE bridge. Anyway, I found it has a CLI so reset the modem/router to gain access to it via telnet, and have been able to obtain bit loadings and per-tone SNR info. It doesn't look like I can upload attachments here but suffice it to say there're quite a lot of blips and troughs in the data, more than I'd expect for a line that's mostly underground and no more than 1km from the exchange. The largest of the troughs lines up perfectly with the 1584kHz radio station: it completely knocks out three frequency bins and reduces the throughput of about 30 more. There're a couple of other noticeable troughs (affecting approximately half as many bins as the one I can hear) and they both tally with MW radio stations, with several other narrow blips too.

I've also been able to obtain a downstream attenuation figure, which is reportedly about 20dB, which does seem a bit high for the expected line length. Do people think that's going to be caused by the radio interference issues or will be that be down to something else entirely (like the old GPO box)?

Anyway, thanks everyone for your input. I feel like I might actually have a fault diagnosis now, which will hopefully make it slightly easier to get fixed!

Interference won't be altering attenuation. Conductor type, gauge and length do that.

For me the next steps would be pushing your ISP for an SFI visit from Openreach ..... but then you are going to have to give them good cause to send one out, and I'm sorry to say, there's a few noobies out there who wouldn't have a clue if you started to explain your issues.

Good luck.

Keep us posted.

In reply to a post by Zarjaz:

SFI visits are paid for by the CP, whether they then pass them on to their punter is up to them.

RF3 filters do improve this kind of issue with broadband, technically how they do this I don�t know ...

{You�ve seen me Bob, clearly employed for my looks innit }

I asked the REIN helpdesk recently about this, apparently the innards of the RF3 filters are included in SSFPs now so rheees no need to fit RF3 filters if a mk4 SSFP exists.

I�d like to get this clarified by high up though as no one seems to know for sure.

Well I am by no means claiming to be �high up� but yes, there has been RFI filters built into SSFP�s since the mk3 version.

For me though, I�d still rather see it fitted where the line enters the property, seemed to reduce the errors more that way.

A colleague removed the active components from an RF3 connected wires to crimp on to then encased it all in a blob of Dexbond ..... these he called �REIN-drops� too convenient a name to pass up I suspect. If having no joy in gaining access to a property thought to be the source he could put them across the pair on an OH DP or at the 66, or in the UG radial ..... it helped.

How close are you to the Panjab transmitter? If you are very close even underground cables will be susceptible and the overheads if any will be wonderful receive antennas.

If the transmitter is the cause - you will have a problem getting it resolved. Filters may not work due to power levels and the radio station are fully licensed.

In reply to a post by Zarjaz:

A colleague removed the active components from an RF3 connected wires to crimp on to then encased it all in a blob of Dexbond ..... these he called �REIN-drops� too convenient a name to pass up I suspect. If having no joy in gaining access to a property thought to be the source he could put them across the pair on an OH DP or at the 66, or in the UG radial ..... it helped.

Don't put them on over heads though ... they might fall on your head!

I'm about 3km from the transmitter, so not right on top of it though not a huge distance either. The fact that my line seems to be picking up a fair bit of interference from other radio stations too means I'd be keen to have it looked at, though I take your point that there may not actually be much that can be done to improve it.

I am probably over 30km away from the closest main MW transmitters - I have dips in my bits/tone graph that coincide with frequencies used for transmissions - 909kHz which takes out 6 tones, 972kHz (about 8 km away) taking out 5 tones, 1089, 1152, 1215, 1413, ... and more, all causing a loss of 10 bits/tone over 2, 3 or 4 tones and in all cases degraded tones on either side.

In reply to a post by Zarjaz:

Well I am by no means claiming to be �high up� but yes, there has been RFI filters built into SSFP�s since the mk3 version.

For me though, I�d still rather see it fitted where the line enters the property, seemed to reduce the errors more that way.

A colleague removed the active components from an RF3 connected wires to crimp on to then encased it all in a blob of Dexbond ..... these he called �REIN-drops� too convenient a name to pass up I suspect. If having no joy in gaining access to a property thought to be the source he could put them across the pair on an OH DP or at the 66, or in the UG radial ..... it helped.

Yes, more flexibility. Like you say they often were fitted where the cable enters the house. In the Hall maybe, whereas the NTE5 could be somewhere else. I like your coleague�s idea and have contemplated similar!

In reply to a post by Zarjaz:

A colleague removed the active components from an RF3 connected wires to crimp on to then encased it all in a blob of Dexbond ..... these he called �REIN-drops� too convenient a name to pass up I suspect. If having no joy in gaining access to a property thought to be the source he could put them across the pair on an OH DP or at the 66, or in the UG radial ..... it helped.

Are you saying that one method of transmission from source property to victim property is that the interference is induced onto the source's phone line, and then onward onto the victim's phone line?

And that an RF3 on the source's line helped improve the victim's line?

Yes, especially useful when the commonest cause was faulty equipment which was connected to the line in the �source property� thus giving it a readymade aerial to transmit RFI.

Right!

I typed up a long e-mail to Andrews & Arnold yesterday detailing my concerns/findings, and they've replied this morning to say that some tests they've carried out indicate I have an unbalanced cable, and they reckon this means there's a corroded join somewhere. As such it's being raised as a PSTN fault rather than ADSL, with a resolution date of Tuesday.

So� good news, I think. The one thing that's bothering me though is why this wasn't picked up by A&A when they raised a fault earlier this year for the same reasons, albeit without me having given them such a detailed analysis. One of the reasons I signed up with A&A for the line at this property was for their elite diagnosis and BT-badgering skills, as I forecast problems even before moving in judging from the age of some of the wiring.

As such it's being raised as a PSTN fault rather than ADSL, with a resolution date of Tuesday.

Tsk, tsk, tsk ........

Unless they have a clear diagnosis of CIDT fail, and have raised a CIDT fault, then what you'll get is some new boy come, PQ test at the NTE, and then wander off whilst you get a bill for right when tested.

An SFI is what they need to raise.

To be fair, A&A's response did say the fault is one Openreach will be fixing without needing an appointment to enter the premises, so it sounds like they're on the right track in that respect.

I�ve seen a video A&A published a while back, I think on YouTube. I get the impression that they seriously dislike the SFI product!

You know, I'm getting quite fed up with Andrews & Arnold. I thought they were supposed to be good�?

So, BT engineer came on Tuesday and determined that there was a problem underground which he wasn't trained to fix, so had to leave and send someone else out yesterday to work on this unbalanced cable problem. My line went down for a couple of hours, and when it returned my sync speed had improved by about 0.5Mb/s � so somewhat less than the 5Mb/s or thereabouts that I believe I'm missing, but a start nonetheless.

I was at work yesterday so today I tested my phone line again and� I can still hear the same radio station. So back I go to Andrews & Arnold to explain that there still seems to be a problem, and for the second time this year, they blame my handset!! Despite all of the evidence I supplied them with last week thanks to all the help here.

Additionally, they didn't e-mail me at all between last Friday and today � nothing to keep my updated after the fix was postponed for a day, nothing to respond to my question re. why this unbalanced line problem wasn't picked up when I raised a fault earlier this year.

Hmph.

This isn't the first time their e-mail support has been less than exemplary either (although sometimes it's good).

Getting rid of the unbalanced cable fault will help with the noise cancellation properties of the pair. The magical properties of twisted pair is being undone by the imbalance.

Agreed, raise the pairs A/C balance.

The pass mark for a PQ test is 55db, but I think 60 or above is better for this kind of thing, problem is the value of the earth the tester uses is an unknown.

Ah yes, thanks, that makes perfect sense.

Latest update for those who're interested: Andrews & Arnold are saying there is no service-affecting fault so are currently unwilling to do anything.

HI Be3G, I believe this fault is with our escalations team, I'll see what's going on with it and will get an update to you. I'd offer to send a second corded phone yo test with, but postage will cost as much as buying one

Arrgghh my hideout in the General Broadband Chatter forum here has been rumbled, you've found me!

Just kidding. Thanks Andrew.

In reply to a post by andrewhearn:

HI Be3G, I believe this fault is with our escalations team, I'll see what's going on with it and will get an update to you. I'd offer to send a second corded phone yo test with, but postage will cost as much as buying one

So you will credit his account with the purchase price and a goodwill gesture for the inconvenience? .

(Sorry Andrew, but it was an irresistible open goal ).

Because I know you're aaaalll on tenterhooks�

My line has been fixed! Sort of.

Slightly begrudgingly it seemed, A&A pushed through a new PSTN fault to Openreach last week after finding my thread here. I wasn't expecting much to come from this, as it seemed safe to assume I'd have the same experience as mentioned in my OP after the first visit a few months ago. I was even starting to doubt myself for getting on A&A's nerves but a couple of days ago my broadband inexplicably dropped for six minutes, which proved that after the recent work to fix the unbalanced cable the line still wasn't operating perfectly, with or without the geriatric corded phone.

So the Openreach guy arrived today and this time, unlike the previous visit, did believe me when I said there was a problem. He was probably about twice the age of the previous engineer, and sounded like he had a fair bit of experience under his belt. He immediately suggested one of the filters that've been talked about in this thread before; I was a bit apprehensive of negative impacts on my ADSL signal, but I figured it was best to let Openreach do what they want to do. So he put it in in place of my old GPO box and voila: no more audio on the line!

After reconnecting the router I find, much to my interest, that my speed is� the same. So it's definitely not had an adverse effect (e.g. by filtering out useful ADSL signals in the same range as unuseful radio ones), but there's also no indication that it's fixed the ADSL side of the equation. Sure, I can live with my speeds not improving, but the fact they haven't makes me suspicious that the ADSL signal stiff suffers from interference and may therefore still be liable to drop.

So I guess I'm just going to have to keep monitoring the line, wait for any disconnections, and then report an ADSL fault in future if I do get any. I'm very pleased though that the PSTN side has been fixed now, so thanks everyone for your help and for giving me the confidence to keep pushing my case with A&A.

Whilst your speed hasn't improved are you able to see if the SNR has increased? It could be the SNR has gone up but that DLM is now holding your line back rather than a fault.

Sadly the SNR and sync speed are almost exactly the same as they were before the engineer's visit today, so I don't think there's much to look forward to in that respect!