Anyone throw any light on this strange DNS problem?

This query is addressed to those of you who are experts on ADSL line performance (ADSL1) and the terminating kit at the local exchange.

Just this last week a problem's occurred in which my access to the Internet, both browser and POP e-mail, disappears randomly. When it disappears, it stays disappeared for quite a long time. Some days I've had no Internet access at all. And yet if I log into my router, the line to the exchange is most certainly sync'd and running at a reasonable speed, even though access to the Internet is impossible. THe TCP/IP settings in Windows are all fine. It's a fixed WAN IP address and the router gets the ISP's DNS servers automatically.

If I run the utility DMT, I'm able to see graphically that, at the physical level, there's a connection right back to the kit in the exchange. The only significant clue I have is that my POP3 e-mail client reports when it tries to send mail "The host <my ISP's DNS name> cannot be found". When browsing fails, Windows just reports it, saying it can't find the website.

Initially, the non-access was preceded by my landline phone in the house ringing fast and uncontrollably, including at 2am in the morning on one occasion. Now, I do know that the phone doesn't use the ring wire. Most landline tone-dialing phones don't. So this can't be anything to do with the ring circuit in the back half of the master socket. The phone recreates its own ring circuit purely from the A-B pair, inside the handset.

At the master socket I use an ADSL Nation XTE-2005 master faceplate filter. The one extension line is solely for my wired router and is terminated in an RJ45 socket into which the router's RJ11 lead is plugged. This upgraded wiring arrangement has worked fine for some two years so far. Cabling between the XTE-2005 and the RJ45 is Cat5e and implemented to a high standard..

Testing of the line the other day, conducted initially by me with the BT Faultline (the 151 service), indicated that in the absence of my extension and any of my kit, there was some sort of problem with the voice line. Accordingly, the next day a BT engineer turned up. But he knew nothing about the ADSL Broadband side of things and so he merely conducted a ringing test on his test phone and a line quality
check, which of course showed little or nothing wrong on this 2.5km line. He refused to do any checks at the exchange, though, saying he wouldn't know what to look for. Sod's Law, the phone problem wasn't there when he visited, neither was the ADSL problem. But within an hour of him departing, I'd lost Internet access again.

In my own experiments I've found that if I log in to my router and I nudge the kit at the exchange into responding by me re-entering and applying my Internet login details so as to momentarily take down the 'Internet' indicator on my router, it will restore full operation at all levels. But I lose the Internet if I run DMT and try to re-sync the line to a slower speed, something with which until recently there was no overall problem. The line resyncs and it's all there but there's no actual data-packet communication going on (and so any resolving of IP addresses).

I've double-checked all my physical connections at my end and even re-wired one part, but accessibility is still iffy. I've even restored the router (a wired Netgear DG834) back to its factory defaults and then reconfigured it but it's made no difference.

Personally, I find it difficult to tell whether this might be the router on its way out or whether instead it's down to the terminal kit (DSLAM, etc) at the exchange malfunctioning. The DSLAM has often refused to respond in the past to my router renegotiating the line and it's sometimes taken me several tries with DMT to regain sync. On the other hand, I've begun to wonder if the filter in the XTE-2005 has become damaged by a surge on the line, as I can't see how the phone could have otherwise gone berserk without there being a hard fault on the line.

For a day, I lashed up a bypass arrangement on my XTE-2005 and on my extension wiring, and used a completely independent plug-in type filter for the phone. Everything ran okay. However, I didn't as before remember to force the DSLAM to renegotiate the line, so maybe I need to repeat that particular test.

The real question, though, is: What does that e-mail client error message suggest? Is it saying that the problem lies somewhere within the router, or is it indicating that the problem's somewhere on the far side of the DSLAM? Browsing fails because clearly there's no translation going on (website name to IP).

I've of course been in touch with my ISP about this. My connection at the local exchange is, as far as my understanding goes, an LLU arrangement, as my ISP doesn't have a major presence in that exchange.They say that at present there's no problem further up into the Web in their domain (although a couple of weeks ago, I know that they did have a major e-mail server failure that took a couple of days to fix).

The line to the exchange is wholly copper.

I've primed my ISP that they might need to get, via their own mechanisms, a BT engineer out to the exchange to specifically investigate there and maybe also conduct some meaningful ADSL testing from my end with a BT laptop and router. I had to do that once before, several years ago, and although it was a totally different problem then, it turned out to be a fault at the exchange. Unfortunately, I myself have no alternative router to try. But I'll only press it with my ISP if I can first eliminate my end as being the probable source of the problem.

Your IP traffic is tunnelled to the ISP, so it is not a DNS problem at the exchange.

When you have this problem can you visit a website via its IP address

Haven't tried that and, right now, it's not worth doing, as everything's working fine.

Remind me about the form. Is it:

http://xxx.xxx.xxx.xxx ?

Know of any suitable IP addresses I can use? TBB's, maybe? Then I'll give it a try tomorrow. Am busy on the Web with something else at the moment.

http:// 80.249.99.130 which s us

I'm a bit confused. When you say:

In reply to a post by meditator:

my access to the Internet, both browser and POP e-mail, disappears randomly. When it disappears, it stays disappeared for quite a long time. Some days I've had no Internet access at all.

do you mean you can't access the Net at all or just you can't get addresses resolved?

If the former then, as you are synced to exchange, then it is an issue with authenticating with your ISP. If just the latter then it's an issue with the DNS you have set. In which case raw IPs, as previously suggested, should work.

In reply to a post by meditator:

"The host <my ISP's DNS name> cannot be found".

Don't understand this! DNS are normally specified by raw IP not name.

an LLU arrangement, as my ISP doesn't have a major presence in that exchange.

Don't understand this either! LLU is quite the reverse; LLU ISPs do have major kit presence at exchange.

If it is a DNS issue as you seem to suspect, why not change to a public DNS and nail it?

Open DNS : 208.67.222.222 & 208.67.220.220
Google DNS: 8.8.8.8 & 8.8.4.4
Norton DNS: 198.153.192.50 & 198.153.194.50

In any case, as you are synced OK to exchange, it has nothing to do with exchange or your local wiring.

I'll take your responses one by one, XRay:

1) When it happens, I can't get the Net at all. My outgoing POP e-mail message gets as far as my Outbox but then the client says that it can't find the dns e-mail server belonging to my ISP. When Internet access is lost I've no idea if the addresses for either the e-mailing or browsing are getting resolved because no other error reports are available, eg. if the browser (IE) can't access the Net, you just get a standard Windows failure browser page. I'm sure you know the one I mean.

2) The POP e-mail error message names the server (generically) it looks for but can't find. It's one of the ISP's e-mail servers, that's all.

3) All I know is that it's 'some sort of LLU arrangement'. I think my ISP piggybacks my account on to one of the LLU providers. You and others know more about that side of things than I do, I'm sure.

4) Regarding your last comment about it being nothing to do with the exchange or my wiring, you may well be right. But don't forget that this all started by my phone ringing strangely. It was as if a call was coming in, the phone rang but then there were no pauses between the rings. That's to say, it continually rang at a fast rate and even continued for a few minutes like that after I picked up the receiver. It did the same when I was rung back by the BT Faultline people. Knowing that the ringing tone is generated entirely in the phone off the A-B wire pair, it's as if the A-B pair from the exchange was being rapidly modulated, if you see what I mean. So, this may well be a DNS problem but undoubtedly it's also linked with a line problem of some sort that's affecting the phone. I've an identical phone, BTW, which I also tried, and that too became subjected to the rapid ringing. The BT engineer who called mentioned in passing that, some years ago, they had a facility at the exchanges for forcing rapid-ring calls on to customers' lines (goodness knows for what purpose), but he thought it was never used nowadays. To try to re-create this phenomenon, I've used the old BT ringback no. (17070) but thus far the ringback has worked and the phone has rung in the correct fashion, ie. this particular problem's not recurred. The other day, when this was at its worst, a phone contact of mine found that when he rang me he heard the start of the ringing but then the line went dead.

Edited by deleted (Sun 22-Apr-12 00:34:39)

All I can reply is that I don't know what you mean by a "dns e-mail server". You can either have a Domain Name Server (DNS) or an email server but not a hybrid.

Anyway, I've already advised you to try a public DNS. Also next time problem occurs try raw IPs; having collected some in advance by doing e.g. 'ping bbc.co.uk'

Yes, the strange phone ringing does seem strange, but we can't say if it is connected to your Internet issue until you've eliminated the DNS. It could be just a scam call

In Re LLU: Which ISP? Which exchange?

Right, I've had an opportunity to try a raw IP address or two when the failure occurs. None of the IPs are successful. There's no connection to the Internet from the browser.

The ADSL tones (bits per tones plot), as seen in the DMT utility, look pretty ropey at present. They tail off quite severely in amplitude as you move up the frequency range and there are several big gaps. Mind you, I am on a long line, so that spectrum has never looked very uniform. Synchronisation is still there, and if DMT's figures and those in my router are anything to go by, at a speed, SNR, etc that I normally get.

As I mentioned before, I find I can retrieve access with the Internet if I log into my router and re-apply my Internet login details. When I do that, the 'Internet' indicator (only that indicator) on the Netgear router extinguishes and then returns. This surely means, doesn't it, that the connection through to the Internet has been reset in some way at some position beyond the DSLAM? The 'Internet', as far as that indicator goes, is surely a level beyond my line?

On the more hardware aspect of this problem, I'm mindful that the XTE-2005 contains transistors amongst its filter components. Thus, I've been wondering whether maybe a surge on the line has partially or fully damaged one or more of those semiconductors and that the filter is possibly now loading the line in an abnormal way. Perhaps the line is okay enough to support the tones but if a line resync occurs, resolving of server sites fails?

XRay, you've misunderstood my description of the e-mail server. Its name, in this case, just happens to include the word 'dns'. So, unfortunately I described it as a dns server.

When the e-mailing fails, the full error message that the client gives me is:

The host 'pop.dnsmaster.net' could not be found. Please verify that you have entered the server name correctly. Account 'Skymarket', Server: pop.dnsmaster.net, Protocol POP3, Post 110, Secure SSL: No, Socket error: 11004, Error no. 0x800CCC0D.

Addendum: Do you happen to know whether the equivalent faceplate filter marketed by Clarity uses any transistors? I'm referring to the one here:

http://www.clarity.it/telecoms/adsl_faceplate.htm

Edited by deleted (Sun 22-Apr-12 11:46:02)

How does a faceplate failure cause a loss of IP throughtput, but the ADSL layer remain? That is assuming the noise margins remain sensible at above 3dB 99.999% of the time

The issue you appear to have is an authenticated session going stall, and a drop of the PPP session appears to resolve this. I would advise stop worrying about the ADSL hardware.

The problem may be a factor of the modems software behaviour so worth trying a different modem.

Well, that simple test has eliminated any DNS issue.

'dnsmaster.net' is an unfortunate, and stupid, domain name for an ISP's domain. No wonder you are confused.

Your ISP, Skymarket, is not LLU going by the speeds it sells. It resells BTw ADSL Max BB. Or is it just your email host?

Everything points to an issue with your ISP's Authentication Server (LCP/PPP), remote from the exchange.

I agree with MrSaffron to forget the ADSL hardware, other than trying another router, and cut down on the wordy expositions. It is difficult for us to see the wood for the trees.

Mr Saffron, the question you asked there (kinda thinking to yourself) does indeed throw up some weird, and probably impossible, interactions. But myself, I was only summising. I'm not an expert in the upper-level workings of ADSL and so that's precisely why I've posted this and have been hoping that you'll all coax me along without yourselves using too much jargon or assuming that I have the same level of expertise as yourselves.

I might point out that the Netgear router (router-modem) is one that I've been using for several years and which, thus far, has given no problems, though I have noticed in recent months that when I go to a website from my browser's address field it sometimes fails and I have to tell it to do it again. Invariably, it gets the website on the second go. I don't, in any event, have another router to try in its place. It'd be a lengthy and costly exercise to buy a replacement, just in the hope that the router was the source of the problem.

So, XRay, you reckon it's due to an issue with my ISP's authentication server (whatever that is)? Unless you've since had a change of mind about that, or Mr Saffron vehemently disagrees, I'll get on to my ISP and put that to them and get them to investigate. There was a major outage of their e-mail server just over a week ago. We users were unable to use any e-mailing for varying numbers of days, though I myself was affected for only a few hours. Their backup e-mail server kicked in apparently but soon got overloaded and that too then crashed. When things were finally returned to normal after 2 or 3 days, the MD of Skymarket sent out an official letter of apology to everyone by e-mail. I had all of that in mind when my connection first started playing up this last Monday and I duly asked my ISP support people if that server and the service generally was definitely now okay. I was told yes. I lose browsing as well, of course, and to me (with my limited knowledge of such things) it didn't seem logical that a failed e-mail server could also affect that.

Afterthought: By 'authentication server', perhaps you're referring to the entity that checks my Internet logon details before allowing me access into the Internet? Hmm, why would my authentication details fail by me just resyncing the line? Until just this last week it's never done so before.

Edited by deleted (Sun 22-Apr-12 14:22:33)

In reply to a post by meditator:

Afterthought: By 'authentication server', perhaps you're referring to the entity that checks my Internet logon details before allowing me access into the Internet?

Exactly! They sometimes play up for all ISPs and by resyncing/rebooting you give them a kick up the backside

EDIT: MrSaffron won't disagree. He came to exactly same conclusion.

Edited by XRaySpeX (Sun 22-Apr-12 14:45:16)

The only remaining oddity to this saga is the malfunctioning of the voice circuit on my line, ie. the uncontrollable ringing of my phone on Sunday evening and during the Monday. That coincided with my initial full loss of the Internet. The strange ringing appears to have now gone away, though, as over the last two or three days I've run a ringback test from time to time and it's rang fine every time.

In reply to a post by meditator:

... The only remaining oddity to this saga...

Glad to hear you are nearly all sorted - was it the 'transistors' in the end?







Sorry TBB

b4dger, Saffron and XRay,

I'm afraid the problem is far from resolved. I've been communicating with my ISP on this in a far more pressing way these last few days and they tell me that they can find nothing amiss in their domain (I use the term loosely).

Over the last few days they've been monitoring the authentication server and have been telling me that they've seen no evidence of any authentication failures with my account. This is odd because I know it's happened many times in that period. In fact, I find I can engineer the failure myself quite easily - by getting, via DMT, my router to negotiate a resync with the DSLAM. All access to the Internet disappears. After that, I'm able to restore access to the Internet by logging into my DG834 router and re-applying my Internet login. That re-application causes a momentary reset of some sort much further up the hierarchy, somewhere between the exchange and the wider Internet. During all of the non-access I can see, via DMT and other things, that the line is synchronised and working normally in that regard.

Skymarket's latest reply to me this morning was seemingly addressing the point I made to them about my raw IP addresses failing as well as the normal URL. They say: It is quite normal for the hosting IP address of a website to fail to display anything, as the hosting server is rarely set up to display a page when someone browses to its IP address. The majority of web servers host multiple sites and they therefore leave the "raw IP address" without a binding. This doesn't make much sense to me - but it might to you!

They also say: To date, we have still not seen any authentication issues for your user account on our Radius server.

Note that they had a major e-mail server failure a few days before my problems began, and in the end they had to completely replace that server. As far as my problem goes, was that just coincidence or what?

Incidentally, using my backup arrangement, I restored my PC back to the status it had on 15th April, just before these problems first began. I've recalled also that when the symptoms first started appearing on that day outward-going e-mails from me were getting badly corrupted - in some instances, attachments were getting stripped off before they reached the recipient, in others the body text arrived at the recipient completely garbled - and then every such corrupted e-mail was later returned to me - but up to nine copies of it! Meanwhile, my Sent box showed the original message and the attachment quite perfectly. Since that day there have been no further occurrences of e-mail corruptions.

In reply to a post by meditator:

It is quite normal for the hosting IP address of a website to fail to display anything, as the hosting server is rarely set up to display a page when someone browses to its IP address. The majority of web servers host multiple sites and they therefore leave the "raw IP address" without a binding.

I don't know. That may be so in some cases but these 2 particular URLs lead to the same webpage:

• http://www.bbc.co.uk
• http://212.58.246.91 (found by pinging www.bbc.co.uk)

The former needs assistance of the DNS the latter doesn't.

You can re-create the problem using DMT, but does the problem show up if you reboot the router, e.g. turn if off and on.

If not, then whats happening is you've found a way to rebuild the ADSL layer connection, without triggering the PPP to rebuild. Resulting in a stall session, rather than an authentication failure.

The multiple sites on one IP can be the case, hence why you test this when the connection is working fine, to know which hosts to use.

Saffron,

Right, I've just run that quick test. I powered off the router, gave it a minute or so, and then powered it back on again. 'PPP' came back, then some seconds later 'Internet'.

When I then went into DMT, DMT wasn't 'connected' and so to sync the line and get a plot I had to specifically use DMT's Connect button. This resync'd the line but (as to be expected) at a speed about twice that of which it's actually capable, in stability terms. I then used the Target SNR facility to bring the speed down. The DSLAM appeared to negotiate okay on this occasion and, once more, the line's now running okay.

I'm still getting strange things happening, though, especially with the e-mailing side of things. A few minutes ago I sent yet another test e-mail to myself, one including a small attachment. When a few minutes later I read it back off the server, the name of the attachment had become appended to the subject line. Also, the subject line now said '1/7', indicating that this was likely to be the first of seven identical copies coming back to me in due course. This was happening about a week ago, when indeed I did subsequently receive multiple copies back.

I can't see that anything I can be doing or any possible fault in my router could cause such corruptions of my e-mails other than something much further up the ADSL chain, most likely the Skymarket e-mail server. And whatever it is also affects access by browser. Do you agree?

DMT does things at a lower level, so no surprise it behaves oddly sometimes.

On the email - no idea on how good or bad SkyMarkets email service is.

In reply to a post by meditator:

Skymarket

That's the first time I think you've been prepared to name your ISP! Not a company I've heard of - just looked at their website - WOW not sure I like their pricing/allowances!

In case you missed it they posted details about their email outage here: http://dmtrk.net/964-RNJB-32/sv.aspx

No, I didn't miss it. If you read my posting you'll see that I mention it more than once and that the MD of Skymarket actually later sent out a message of apology to all Skymarket ADSL users (which I of course got). And actually about a week before that there was also a significant interruption to the service when a crucial BT node somewhere in the Midlands, I seem to recall, broke down. It affected lots of ISPs across the country, including Skymarket, and was reported on TBB's News section. So, not Skymarket's fault on that occasion.

I might add that, over the years that I've been with Skymarket, their service up-times have been excellent. This recent blip is not, I hope, the start of a downward trend.

Skymarket's pricings stem from their concentration on providing services to commerce. They've never aimed to be up there with the big boys like TalkTalk, Virgin, Sky and BT, as their target customers are not really joe public. Whilst their fees for ADSL Broadband are comparatively high, I've always found their tech support to be responsive and helpful, and that counts for a lot in my book. They don't employ people who are at the end of a phone in India or who simply read off scripts as solutions. All that inhouse service, however, has to be paid for in some way or other. Over the last year or so I've tried to persuade Skymarket to increase its monthly allowances (now fairly mean, by most people's standards) and to also move toward FTTC, but they've shifted very little on that. I may well switch to a different ISP in due course.

Anyway, this is a digression and is not getting the baby's nappies washed. I'm still no nearer finding a cause for the e-mail havoc and non-access I've been experiencing.

Edited by deleted (Wed 25-Apr-12 00:01:19)

b4dger, Saffron and XRay,

I think I may have finally found the reason - well, certainly ONE reason - why my e-mailing and my Internet access generally has been malfunctioning. I think it is probably down to a configuration mistake made entirely by me and concerns a server setting in the client account. Errh, doh! I've contacted my ISP to double-check with them.

Will be back here later to explain.

I look forward to your explanation.

The only "server setting in the client account" I can imagine is a email server setting in your email client, which would have no effect on your ability to browse or connect to the Net in general, but would certainly affect any emailing.

Yup, you guessed right, XRay. The setting was "My server requires authentication". Despite me having looked over several times this week all the settings in my client account, I'd not noticed that that one was not checked.

However, this has turned out to be a false positive, as my ISP has responded that, for smtp e-mail, the checking (ticking) of that setting is only optional. It's not actually required to be checked. Also, of course, Web access via browsing wouldn't also be affected by that.

So, as you were. I'm still no nearer to finding the cause.

So null effect! In fact you never need "My outgoing server requires authentication". when using the default port-25 SMTP Server of your own ISP. It knows who you are by the mere fact that you connected to them for the Internet.

The earlier email error you reported concerned failure to connect to their POP3 server, not SMTP, had nowt to do with this and was certainly due to it being down as covered by their admittance of being blighted by email issues .

Don't get confused with my namings. Mostly, I've been referring to my e-mail connection generically as a "POP connection", and doing so merely to distinguish it from webmail. I do realise that outward messages from me go via an smtp server, whereas inward ones come in via a pop server.

The latest explanation coming from my ISP is that two different things are happening with me - corruption of e-mails and loss of access to the Internet. Well, I agree with that. However, they've come up with a theory, which they say they've tried out and confirmed, concerning the corrupted, multiple e-mails. This is what they contend:

They say that if the 'break apart messages longer than x KB' setting in OE is checked, a single e-mail message will get split up into a collection of separate e-mails and will therefore get delivered as separate e-mails. The setting in OE is found in the Properties of the POP3 e-mail account, in the Advanced tab. They suggest that this is why any attachment in the message will be corrupted!

Quite apart from me not using that setting, my understanding of that setting is that it merely segments packets of a message during initial launch into the Internet but that the packets are reassembled at the receiving end. This is a hangover from the days of dial-up, when speeds were far slower and there were more bottlenecks on the Internet requiring messages to be split up and sent by different routes. No multiple e-mails were produced from the sender;s single e-mail, as far as I'm aware. The implication of their explanation is that the recipient would need to open several contsituent e-mails in order to read a single, full message. Really? How would you ever send someone a sizeable attachment, using that setting?!

As you don't have that setting ticked, this is not a solution..

If you had used it the msgs would have been auto reassembled at recipient. It's mainly used for mail servers that place limits on emails, but the mail client on PC or Web should reassemble them.

Still leaves unexplained loss of access to the Internet.

Yes, you agree with me on all points, XRay.

In fact, I have this morning tested out that "break apart messages ... " setting and sent myself some long e-mail test-messages, first without an attachment included and then with an attachment included. In both cases, I only ever got back one complete message, and in both there was also no corruption. It'd simply be daft for that setting to work in any other way, wouldn't it?

Am still no nearer determining why I've from time to time received multiple copies of corrupted e-mails and why randomly I lose all access to the Internet, even though my router never drops the line.

I'd have thought there was a clue in the e-mail client's error message, in that it's telling me it's a 0x800CCC0D error and, as you know, it's saying that it can't find the e-mail server. I've googled for that error code but all I've turned up are vague explanations. Surely, Microsoft should have at some stage published a precise definition of that particular error code? Clearly, what I need to know is whether the inaccessibility is happening somewhere at my end of things or instead at some server much further beyond the exchange.

In fact, my ISP has turned out to be correct about that "break apart ..." issue, as when they prompted me to look in my Deleted Folder in my e-mail client, I found appended multiple copies of my test messages, albeit that none of them contained any body text. All they each contained was a .dat attachment. Thus, whilst the recipient does get a full message, with or without attachment, other copies of the message are also received but bypass the Inbox and are dumped into the Deleted Items folder.

Incidentally, can any of you advise me as to whether BT are likely to bill me for their call-out a couple of days ago? I'd contacted their 151 Faultlline, reporting that not only was my landline phone ringing strangely, with no real incoming call there, but also I was randomly losing access to the Web. There was no evidence to me that any of my internal wiring was the cause. Indeed, the phone still rang wildly (and a spare one too) when plugged into the internal test socket on the NTE5 (via a filter).

A BT engineer accordingly turned up the following day but said he knew nothing about ADSL and he merely proceeded to test the line for voice operation. The fault, though, was affecting both the telephone and my ADSL. He tested in the usual way by unplugging my faceplate filter and using the test socket. He of course found nothing amiss (isn't it always the case?!). He tested none of my extension wiring (not that I needed him to). Thus, he departed having found apparently nothing wrong with my line to the exchange. He advised me to get the problem further investigated via my ISP, which I've now done. As you'll be aware, the non-access problems are continuing, being random. The phone peculiarities haven't recurred.

So, where do I stand as regards any sort of call-out bill from BT? They were called out because of a genuine problem but the engineer failed to find anything amiss with the line back to the exchange. So, is BT likely to bill me for the visit, and if so for how much?

In reply to a post by meditator:

In fact, my ISP has turned out to be correct about that "break apart ..." issue

That doesn't prove anything as you had not used that setting before your tests and those deleted items were as a result of your tests.

The way it works is:

You send a large message under "Break-apart". Your client breaks it up as instructed but only stores the complete msg. in Sent Folder.

It arrives on recipient's server broken into parts, like:

From: Subject: Date: Size:
Me Test Mail Break RadioDatabase4.db3 {10/1... Apr 27 2012, 01:23 AM 35.3 kb
Me Test Mail Break RadioDatabase4.db3 {08/1... Apr 27 2012, 01:23 AM 60.9 kb
Me Test Mail Break RadioDatabase4.db3 {09/1... Apr 27 2012, 01:23 AM 60.9 kb
Me Test Mail Break RadioDatabase4.db3 {07/1... Apr 27 2012, 01:23 AM 60.9 kb
Me Test Mail Break RadioDatabase4.db3 {06/1... Apr 27 2012, 01:23 AM 60.9 kb
Me Test Mail Break RadioDatabase4.db3 {04/1... Apr 27 2012, 01:23 AM 60.9 kb
Me Test Mail Break RadioDatabase4.db3 {05/1... Apr 27 2012, 01:23 AM 60.9 kb
Me Test Mail Break RadioDatabase4.db3 {03/1... Apr 27 2012, 01:23 AM 60.9 kb
Me Test Mail Break RadioDatabase4.db3 {02/1... Apr 27 2012, 01:23 AM 60.9 kb
Me Test Mail Break RadioDatabase4.db3 {01/1... Apr 27 2012, 01:23 AM 60.8 kb

The recipient (you) POPs his mail and these parts 1st arrive in his client Inbox where they are recombined to the complete msg and the parts discarded to the Deleted folder.

BT has no involvement/responsibility in your BB. Their responsibility ends at the test socket. It would have been wiser to not mention BB to BT and to concentrate on the phantom ringing.

I would think they will charge you about £120 as no voice fault was found.

Edited by XRaySpeX (Fri 27-Apr-12 01:56:02)

I was under the impression that BT Wholesale owned the phone terminating kit and the ADSL DSLAM in their exchanges. Therefore, if either or both of those are suspected of having faults, isn't it BT's responsibility to investigate them? I guess you're right, though.

In the past when the aforementioned kit has been put under suspicion, my ISP has instructed an engineer to investigate both at the exchange and at my end. From what you're saying, though, it sounds as though that engineer was NOT a BT or Openreach engineer.

Perhaps this is why my ISP has not yet suggested that an ADSL engineer go to the exchange? (cost!).

This morning I've replaced the faceplate filter with the Clarity version. It made no difference. I then deleted the e-mail account in the client and then re-created it, very carefully reconfiguring everything. I'll now have to wait and see how access to that ubiquitous server goes over the next few hours. If there's still no change, then the only thing left to swap out will be my router. I may or may not be able to borrow one. And it certainly won't be the same brand/model.

Am of course gutted at the prospect of having to pay BT £120. If I'd known that they weren't going to investigate the line ADSL-wise and the kit at the exchange end of it, I'd have never called up the BT Faultline. But then, you see, the problem all started with the phone going crazy and remaining so for the rest of that day. So, the problem's been a mix of phone and ADSL. I wasn't to know that the next day, when the BT fella turned up, that the phone side of it had gone away. An expensive lesson for me. What would happen, though, if it turns out that the problem HAS been at the exchange end, all along? Could BT/Openreach still charge me the £120?

Different bits of BT own different bits of the hardware, and they are setup with firewalls between them, so that they cannot just go talk to an engineer in another division, they have to report it and use the same systems as TalkTalk and Sky would

In reply to a post by meditator:

I was under the impression that BT Wholesale owned the phone terminating kit and the ADSL DSLAM in their exchanges.

They do, but they have no contract with you for BB. It's up to your ISP to call them (it will be an OR eng) out and it's up to you to call only your ISP for all BB issues.

Even if it eventually turns out to be a fault at exchange, you still called your voice provider (BT Retail not BT Wholesale) out for a BB issue.

The piper only plays the tunes for which he is paid.

Point taken. And it's a hard lesson learnt by me. But anyone around me at the time would have found the peculiar phone-ringing bizarre. And the fact that my access to the Internet then started to become difficult, indeed impossible, meant that I needed to call Openreach in to investigate the line. Once the phone rang, it was hard to stop it, even with the receiver picked up. (A duplicate phone did the same). So there was most certainly a fault on the line, as I can't see that my ADSL filter or any of my ultra-simple extension for my ADSL connection could have caused that. It's just unfortunate that, by the next day when the OR guy turned up, the phone problem had mysteriously gone away, although I still had the Internet access problem.

You never call Openreach.

Well, you know what I mean. I called BT Faultline.

In reply to a post by meditator:

Well, you know what I mean. I called BT Faultline.

And they'd probably be only too eager to earn a quick buck!

Yeh, I was kinda bounced into calling in Openreach. In my old age, I must be losing my touch.

Incidentally, since the last changes I made, e-mail access has remained up. I'm keeping my fingers crossed. If it stays okay, and browser access to the Internet similarly, then it would appear that my mail account had somehow become corrupted. My view now is that the phone phenomenon must have been something separate and was caused by some sort of change at the exchange being made. I've noticed that with both of my high-grade faceplate filters - the one in use at the start and then the new one I fitted yesterday - there was an extra 3dB of SNR (and hence more speed) suddenly available after all the troubles began. Remember, I'm on ADSL1, not ADSL2 or FTTC. That's still the case. According to the OR guy, some exchanges do have the ability to generate continuous phone rings on customers' phones, so perhaps during some sort of upgrade to the ADSL kit at the exchange, that phone function was unintentionally triggered, and then maybe later that day it was found and turned off. Otherwise, I can't see how anything at my end could have mysteriously simulated an incoming phonecall. In the early hours (2am) of the preceding night of the beginning of all of this, the phone had rung in the normal way for about 3 mins. Being in bed at the time and semi-conscious, I didn't answer it. As far as my limited knowledge goes, it's invariably during the 'early hours' that engineers at exchanges perform switchovers and other changes.

Now that I've the Clarity faceplate filter in place - the so-called Modified NTE5 BT Faceplate Filter - I'm leaving it like that. The ADSL Nation one employs one or two transistors and it seems to me that, although that one's probably still good, it's perhaps better to use a non-active one, as at times the phone line can be a hazardous environment for such semiconductors (high levels of noise, voltage surges caused by lightning, etc). It's my understanding that the filter in the Clarity one is non-active. Had I had the time to do it, I would have removed the back cover of the Clarity one and had a look at its pcb, to check.

In reply to a post by meditator:

I'd have thought there was a clue in the e-mail client's error message, in that it's telling me it's a 0x800CCC0D error and, as you know, it's saying that it can't find the e-mail server. I've googled for that error code but all I've turned up are vague explanations.

This is the Error you get when you have no connection to the Net and so is part of your bigger issue and nowt to do with the other email issues you had.

The latter I'd put down to the problems your ISP had with email which were so severe that the MD had to issue an apology letter which, in itself, is so unusual. As you are no longer having these email issues I'd take it your ISP has fixed them and you should now forget them and not connect them with your lack of Net connectivity.

I'm wondering if the ringing phone was a complete red herring and the ISP's attempts to fix the email problems created authentication problems. Tiscali and a few others were absolute stars at creating such conundrums out of simple problems.

Correct me if I'm wrong but I think that Error 0x800CC0D, and possibly also 0x800CCC0E, is non-specific as to cause. They both indicate a condition of non-access to the Internet but it's not possible to say whether the blocking of the access is occurring at the output of the e-mail client, within the router or modem, or instead somewhere further up the chain at, say, the ISP's server. In my case, the overall error message was saying that the e-mail client simply couldn't find the Skymarket server - but that could have been for any number of reasons.

Edited by deleted (Sat 28-Apr-12 16:14:37)

True, but the 1st 2 causes are very unlikely when you know everything your end is connected (and works sometime later without being touched) and the 3rd cause is the most likely when you know you are syncing with exchange OK.

In reply to a post by meditator:

..... They both indicate a condition of non-access to the Internet......

What the error code indicates is that the client cannot connect to the server. There are many reasons for this including but not dependent upon not being connected to the internet.

Well, thanks for all your indulgences in this, but the saga of my problems with my phone and with my Internet access has just taken an extraordinary turn.

I live in Hampton, a SW suburb of London, and the local free newspaper dropped belatedly on to my doormat half an hour ago. What should I find in it but:

Underground cables stolen.

Hundreds of households and businesses have been left without phone or internet connection after thieves stole two underground cables.

BT said its engineers installed new wires overnight on April 23rd, but some traders and residents in the Hampton area will suffer disruption until the weekend.

A BT spokeswoman said: "There has been a lot of damage and this is complex work and we expect it to take us into the weekend."

Well, it was the 21st April when I first posted here about my problems. I'd already been wrestling with the problems since 15th April and, as you know, had had an Openreach engineer attend. The newspaper report doesn't state the date when the cables were thought to have been removed but it does sound to me as though the uncontrolled ringing of my phone and my subsequent difficulties with accessing the Internet could be explained by either that removal process or by BT's revamping of everything in the days that followed. There must have been considerable chaos caused. It seems that the repairs, and presumably reconfigs of the associated kit at the exchange, have continued throughout this last week. It's only today (28th) that my e-mailing and my access by browser have been stable. Earlier today, I'd concluded that that was because of reinstalling my e-mail account, but this news report certainly now throws a different light on the matter.

It's perhaps possible that the bouts of phone instability that I had on the eveniing of the 15th and then on the 16th were due to repairs being made to my and others' lines. Thus, when the OR guy turned up, the line had - at least for phone usage - been fixed by then. But maybe on the ADSL side of things it's taken much longer to correct things and get them fully working again?

Hmm, maybe I won't have to pay that £120 + VAT charge after all? Perhaps if I can get in touch with the billing people at BT and tell them about this outage, they'll waiver that charge? After all, it's happened well outside of my control.

In reply to a post by meditator:

...... Hmm, maybe I won't have to pay that £120 + VAT charge after all? Perhaps if I can get in touch with the billing people at BT and tell them about this outage, they'll waiver that charge? After all, it's happened well outside of my control.

PM me if you need to contact BT's CEO. I find this "short circuit" very effective in favourably resolving disputes.

My technical problems aren't yet fully solved, as this evening (Sat) I've started to lose access again in a very intermittent fashion. I'm streaming radio while occasionally catching up on e-mails but the Internet's been disappearing and coming back again a minute or two later.

Well, that fits with the stolen cable being between your exchange and the ISP.

But it doesn't fit with being able to regain Net connectivity within secs by resyncing the router reauthenticating to ISP.

Unless there is more than one route from exchange to ISP.

EDIT: Which still looks like a ISP authentication/PPP session issue rather than a missing cable.

Edited by XRaySpeX (Sat 28-Apr-12 22:26:29)

I've now pleaded with my ISP to send an engineer along to the exchange to find out what the devil's going on. I doubt whether I'll now get a response until Monday.

BTW, you've got it slightly wrong. My resync'ing the line didn't bring back access, it took it away. I found that doing that kinda simulated the loss, which was otherwise occurring randomly. But with access impossible I could still see that my line was still sync'd and up and running. To then restore access I went into my router and re-applied my logon details. That then restored access almost immediately, by effecting some sort of reset at the 'Internet' (as seen on my router's frontpanel indicators).

Edited by deleted (Sat 28-Apr-12 21:57:19)

That's what I was referring to ... corrected. Same principle!

Edited by XRaySpeX (Sat 28-Apr-12 22:48:43)

Is this the same company???
"our Support is second to none when it comes to fixing whatever problems you might encounter"

However, having looked around, it seems that any problems are pretty much limited to you. I think the cable theft thing is another red herring and it's time to substitute your router, dump the faceplates and run from the NTE5 test socket whilst using something like RouterStats.

Access to the Internet this morning has been very much on and off. When on, streaming football interviews from the BBC website has resulted in extremely jittery video. It's more-or-less unwatchable. Normally, it's fine. I use the latest Flash, etc.

Anyone know what the 'server' setting in DMT 8.07 is all about? It's under the Miscellaneous menu in DMT. Don't recall what that's about. It appears to consist of just two configurable settings:

DMT server enabled (not enabled at present).
DMT server port (currently showing 7784).

It says 'Server is offline'.

To which server is it referring?

And anyone know much about Ethernet adaptor interface settings (found in the Properties of the adaptor in Device Manager)? The built-in PCIe Ethernet adaptor of my PC is a Realtek chip but Realtek don't publish the definitions of its somewhat cryptic settings. These are what they're set to at present, and am wondering if they all seem okay:

802.1Q/1p VLAN tagging - disabled
Auto disable PCIe - disabled
Auto disable PHY - disabled
Checksum offload - Rx and Tx enabled
Flow control - enabled
Green Ethernet - disabled
Jumbo frame - disabled
Large send offload - enabled
Network address - not present
Shutdown wake-on-LAN - disabled
Speed and duplex - auto negotiation
Wake-on-LAN capabilities - Pattern Match and Magic ...
WOL and Shutdown link speed - 10M bps first

The report of the cable theft and the mayhem it caused has been included in the local newspaper's webpages:

http://www.richmondandtwickenhamtimes.co.uk/news/966...

Edited by deleted (Sun 29-Apr-12 13:20:40)

My, you are one for grasping at straws!

Forget the cable theft. They did not nick your copper phone line between you and exchange and there is no copper beyond exchange.

Yep, I agree that I'm clutching at straws.

If the outage caused by the theft had absolutely nothing to do with the problems I experienced and which, on the ADSL side of things, I continue to experience, then how do you explain away that my phone went completely berserk for a day or more and simply became unusable? In fact, one of my contacts later told me that he tried to phone me on the day all of this began and my number was unobtainable. Nothing concerning my PC could have caused that.

I did, on the second day of it, buy an extra long extension cable and, with it connected into the test socket, connected up my phone and router, thus bypassing the faceplate filter, but it made no difference to it.

And BTW, I've not contended that the actual theft of the cable necessarily directly affected my phoneline. What I suggested was that the theft may nevertheless have had a knock-on effect.

Must go now, as I've other work to get through today.

In reply to a post by meditator:

..... I did, on the second day of it, buy an extra long extension cable.....

What sort of extension cable? CW1308 or better must be used. Cheapo extensions can cause the problems you are experiencing but two failing is unlikely unless they're routed near a noise source.

You need to run tests from the NTE5 hidden test socket via a very short RJ11 cable. The cable supplied with the router should suffice.
If that isn't possible, site the router near the test socket and use a long ethernet cable. Tests must be conducted via fixed cable only - Wireless connections can be a problem in themselves.

The fixed extension is with Cat5e cable (screened at that!). Just two wire connections end-to-end involved, nothing else. It was a year or more ago when I revamped my extension, deciding to do it with the best materials and tools and to the highest standard. I'd challenge anyone to show me a better domestic arrangement. The length terminates in an RJ45 wall socket and a screened RJ11 jump lead then connects from that socket to the router.

I refute your advice about using a very short RJ11 cable for the temporary test. The arrangement would be essentially no different to my Cat5e permanent extension and, like the Cat5e, merely represents an extra 10 metres of a connection to the exchange that totals around 2.5km.

Gotta go, cooker is calling!

In reply to a post by meditator:

..... I refute your advice about using a very short RJ11 cable for the temporary test.....

It's actually very sound advice where the word "extension" is used. I suspect that the majority of extensions are the flat cable with cotton thin parallel conductors bought from the local DIY/Pound Shop.

If we are going to be able to help, we need to discount every possible problem that may be present on your side of the NTE.

Edited by Deadbeat (Sun 29-Apr-12 16:13:35)

Back again.

I selected the temporary extension lead quite carefully. It certainly wasn't the non-twisted type. Whilst not ideal, it clearly served its purpose. The vast majority of households where extensions are in use probably do not use even anything as good as my temporary one. I do know about cabling, you know, as I worked as a professional engineer/scientist for over 30 years and structured cabling systems was one of my many specialisms. I wrote specs for it and supervised people installing it. I've plenty of practical experience.

Well, aside from completely replacing the router, I'm certain we've covered just about every single part of my kit that could be responsible for these problems. I've no replacement router to hand and I'm certainly not going to buy a new one just in the hope that that's the problem. I'm a pensioner now and can't afford to waste money. Any replacement router, in any event, is unlikely to give me the re-sync'ing control that I have with this one.

Bear in mind that I've already returned the router to its factory defaults and reconfigured everything. As far as I can see, the router is behaving itself and whenever the Internet access fails the ADSL signal is found to be still up and running. The occurrence of the stolen BT cables happened, as far as we can tell, at pretty much the same time as my phoneline went crazy and the Internet access problems began. It's just far too coincidental to not be connected in some way with my problems. As a result of the BT cables incident, I'd put money on there being an obscure fault with some of the ADSL kit at the exchange.

Local cabling and everything OP's side of NTE5 is another red herring. OP has never reported trouble with his sync.

@OP: You have never posted router stats, particularly when Net lost.

As XRaySpeX says I didn't think you had a sync issue, but just kept dropping your PPP (login) session?

Perhaps you should try running RouterStats in case you are getting confused with what is going on?

In reply to a post by b4dger:

As XRaySpeX says I didn't think you had a sync issue, but just kept dropping your PPP (login) session?

I don't think so either but there's already one possible Openreach charge on the cards and I was looking to save another.

In reply to a post by b4dger:

Perhaps you should try running RouterStats in case you are getting confused with what is going on?

I had already suggested that but the OP seems to have ignored it.

Even from the first post the problem sounds like it's the router.

Thirding the suggestion to run routerstats, and/or get a different router.

OP told a number of times to try another router, but like Routerstats has so far ignored.

I wasn't ignoring trying a different router, it's been a matter of just getting hold of one. Would you seriously buy a new one, just in the hope that that was the explanation?!

As for the PPP dropping, no it most certainly hasn't been dropping, if the PPP indicator on the Netgear router is to be believed. I have proof, anyway, from several angles that my router stays up and running during these losses of access. And were it not for the little nudge I subsequently give to the non-access condition, I'd stay locked out for ever.

As it happens, today I've managed to secure the loan of a router. It's quite an old and noticeably slow-in-responding one, though, made by Billion. It's not compatible with DMT, so I can't tweak the SNR, speed, etc with that utility. I was told that it's not guaranteed to be fully working still but I thought I'd give it a go nevertheless. I've managed to get it working but it might be days or even weeks before I can draw any conclusion, as it's running at a stupid top speed of 6.7M bps which I know that my line cannot support. So, the connection is now going to drop out continually, especially in the evenings. I normally run with an SNR of about 15dB, giving me a sync speed of around 4M bps (because my line is long and noisy), but this router has negotiated a default SNR of 4.5dB.

I've tried telnetting into the Billion from the PC via a Command Prompt and am fairly convinced that, in principle, its CLI menu of commands will enable the use of:

set wan adsl targetnoise <value>

where value, in decimal, is 512 x required SNR. The value needs to be converted to hex before inserting. I tried using hex 1800 but the Billion just never responded to the command. So, I guess it's simply a command that, in practise, is not usable with that model (5100).

The mayhem from the theft of the cable has continued. I was talking to a neighbour of mine earlier today and he said that, even after all this time, some residents around here with landline phones (like me) have been without the service for 4 or 5 days. It's had serious repercussions for those people needing to book GP appointments and the like.

Edited by deleted (Mon 30-Apr-12 17:56:39)

Here in may be the problem, you tweaked the old DG834 such that it was not following all the suggestions it would receive from the DSLAM.

Without you sniffing every command over the ADSL link, you don't know if the other modem was ignoring PPP drop requests. So that the modem sat there thinking it still had a good PPP session, when it had actually gone stale.

Once you start tweaking, you go outside the realm of what the original hardware manufacturer would have tested under.

If a line is only stable with a 15dB target, the logical thing is to get your provider to raise the margin for you.

In reply to a post by meditator:

I wasn't ignoring trying a different router, it's been a matter of just getting hold of one. Would you seriously buy a new one, just in the hope that that was the explanation?!

Yes, just get one off ebay and sell it on ebay when you've finished.

In reply to a post by meditator:

As for the PPP dropping, no it most certainly hasn't been dropping, if the PPP indicator on the Netgear router is to be believed.

What PPP light on Netgear? Netgear DG834 series only have one light for the WAN connection, doubling up on ADSL and Internet.

Are you saying that this is always solid green and that the only inkling you have of losing the Net is being unable to browse?

What does the router log say when you lose the Net?

Having to run with a NM of 15 dB seems rather high. You could well have a wiring issue that needs sorting out 1st.. Tweaking the router while trying to investigate this problem is going to be counterproductive.

What are your router stats?

As I said earlier, you do have and have had a phone service at all times and you personally are not part of the mayhem. So stop bombarding us with red herrings and concentrate on your own problem, for which you have not yet given us all the relevant info.

It's perfectly possible for a long line to have to run at 15dB Target SNR in order to run stably. After all, I've been doing so for a good many years now. Just because you yourself don't happen to have the same difficult line operating problems that I and some others have doesn't mean that I and those others are wrong in the way that we have to run our connections. Forgive me for my pique but all you are doing by making your assertions is that you're diverting attention from the real problem in hand.

I've been using DMT for several years now in order to promptly get the line to run at a slower rate and so remain stable when first commissioning the router or when very rarely needing to resync because of other reasons. It's worked wonderfully well in all that time. I brought the internal wiring, short and simple though it is, up to a very high standard a year ago. It's only been this last week or so when the phone suddenly went bananas and the ADSL connection started playing up that access to the Internet became randomly impossible. From what we now kniow, it seemed to have coincided with thieves having stolen BT cables in the neighbourhood and put the phone and Internet connections of hundreds of businesses and households on this exchange out of action. Even today, some phonelines around here haven't been restored yet.

Yes, my line is long, and it's spuriously very noisy as well. That's just the luck of the game, or rather the lack of it. Over the years it's become plain that some of the noise has exceeded 12dB. The source of the noise has been at some unknown point or points somewhere between my house and the exchange. Believe me, the electrical environment of my own and neighbouring properties is very clean, and in that I do know what I'm talking about! As I've already stated, my own internal phone/ADSL wiring is impeccably good and you'll just have to take my word for that.

At the moment, I'm doing just about the most sensible and logical thing I can do, namely swapping the router for a different one, so I can't see what you're complaining about. It takes time to organise these things. Over the last two weeks I've tried just about everything else that's possible.

I think that, although it certainly seemed at one point like this problem was down to my ISP's authentication of the connection, one or two of you who've still had reservations about the general integrity of my Netgear router could well be right. This is why I've gone to great pains to borrow a disused but thought-to-be-still-good router from a friend. But since, unlike with the Netgear, there's no means of user control over the line speed that it initially negotiates, I'll have to keep observing over many days now before I can come to any sort of conclusion. Clearly, if the temporary router in that time shows no sign of non-access and then when I swap back to using the Netgear the symptom quickly shows, then we'll have nailed it finally. So, I'll report back here in a few days time.

I've not heard from my ISP for three solid days now, so I've no idea what they've been doing, if anything. It's probably been all too much for them and they've probably quit the country. Heh, heh, heh!

No answers to my Qs then?

Still not clear on the symptoms.

In reply to a post by MrSaffron:

Here in may be the problem, you tweaked the old DG834 such that it was not following all the suggestions it would receive from the DSLAM.

Without you sniffing every command over the ADSL link, you don't know if the other modem was ignoring PPP drop requests. So that the modem sat there thinking it still had a good PPP session, when it had actually gone stale.

Indeed, my other thought was the router was simply ignoring or not timing out correctly. As it stands stale/dropped PPP sessions are not uncommon but a router should detect it within seconds and redial automatically. Every one of the symptoms can be explained by the OP's router not doing this.

Right, I'm back, y'all. Aside from needing to catch up on a host of other important non-IT jobs that have been neglected these last two weeks, I've been seeing how an alternative ADSL router has worked in my setup over the last three days or so.

The outcome is still quite difficult to fathom. On the one hand, the alternative router, an old Billion 5100, never once prevented access to the Internet. But on the other hand, the Billion never seemed to negotiate an SNR of anything higher than 6dB over those three days and the sync speed it reported was a constant 6.3M bps, which I know my line cannot support without falling over.

The Billion 5100 is really basic, ostensibly obsolete, and I highly suspect that it has some important firmware bugs, which might include an inability to collect the line statistics properly. This may be why the former owner dumped it after only about a year's use.

You can't use DMT with the Billion 5100, so I couldn't experiment with forcing different target SNRs. I found out that the 5100 has a CLI command set, usable from a Command Prompt, that includes one for setting the target SNR but when I tried it, it simply didn't work; there was no evidence that the command ever passed to the router. I of course checked that other things weren't blocking it.

Meanwhile, I tested the Netgear's little PSU. By all accounts, these are not very robust plugin power units. However, the measured off-load output voltage from the PSU looked perfectly okay.

When today I put back the Netgear router, it initially ran okay. But as soon as I re-sync'd the line with DMT to get a slower, more sensible speed, I got the non-access problem. Even with the SNR negotiated at 12dB I got the problem. I've left the SNR on 12dB and will see how things pan out for a day or two more.

Qasdfdsaq, you may have a point. In fact, more than one of you have summised of late that the DSLAM may have, for some time, been unable to cope with the way in which I've been operating the line. The only counter to that is that I've been using the line like this for several years and the DSLAM has always complied - until a week or so ago! That's to say, the DSLAM has always co-operated, whether I set the SNR to 15dB, 12dB or whatever. Mind you, during the use of DMT, the DSLAM has on occasions simply not responded to a re-sync at all, leaving the line in a sorta 'hung' state (even though the indicators on the Netgear show that the line is apparently still fully up and running). But a second attempt has then succeeded. Something else that's made me wonder is that, for as long as I can remember, the browser sometimes fails to reach a website from the URL in the browser's address-box and I have to try a second time. This kinda suggests that something has been edgey.

Looking at the bits-per-tone plot in DMT, the plot is very poor, in that the decline is rapid and there are many gaps as the tone frequencies increase. I don't recall the line looking quite that bad before. And yet the line at my end, in physical terms, is about as good as it gets.

For the present, what I'm concerned about is whether I have a damaged Netgear router. If the router's no longer physically working properly (due to, say, damaged line-drivers in the modem chip), then the router might not be handshaking the data transfers properly. Clearly, I'd want to avoid any downloads of important updates, etc. if the router were simply now unreliable as to its error-correcting capability.

It could transpire, though, that the problem all along has, as been suggested, been due to my pushing the DSLAM too far, in demanding that it negotiates at 14 or 15dB. It might be that I've just been lucky for several years. But that sort of SNR is barely sufficient to give me a stable line.

Maybe it's now time for me to seriously consider switching to FTTC, as I guess that then these tweaking requirements would disappear, a different DSLAM would be in use, and I'd have a less noisy line anyway? I began to consider it about six months ago but perhaps I should now do it in earnest. As I understand it, for that I'll need a new PPPoE router. But all queries over that are best left to a completely separate forum topic.

Edited by deleted (Sat 05-May-12 16:52:25)

In reply to a post by meditator:

the sync speed it reported was a constant 6.3M bps, which I know my line cannot support without falling over.

But has the Billion done so?

In reply to a post by meditator:

Something else that's made me wonder is that, for as long as I can remember, the browser sometimes fails to reach a website from the URL in the browser's address-box and I have to try a second time. This kinda suggests that something has been edgey.

That's not uncommon! Maybe the route or the end server is busy at that instant.

Router stats?

Yes, but I can't tell how many times it's done it, as I've not sat over it for the last 3 - 4 days, and the stats that the router gives you are absolutely minimal. You get little more than just the instantaneous sync speed and SNR. The Billion 5100 is a truly awful device by today's standards. It responsiveness is appalling, it taking about 3 secs typically to get a webpage. By comparison, my Netgear 834 is fast and slick.

I've been using the Netgear router quite a lot today, having set the target SNR to 12dB with DMT earlier this morning, and have been trying hard to make the connection block. But so far the connection to the Internet's been holding up. I need to give it a day or two more yet but I think we may finally, but finally, be on to something here.

Postscript: For some unknown reason I've not done this during these last two weeks but just now I pushed the target SNR down to around 9dB, and then as a separate exercise right down to 6dB. In both cases the Internet became inaccessible. Now surely that should not be happening?! As before, all visible aspects of the line remain up and running throughout and, again as before, the only way I can then regain access is to go into the router and re-apply my logon particulars.

So, one thing's for sure, this problem's not been happening because I've demanded too much of the DSLAM as regards negotiated SNR.

One thing I've just remembered is that following Skymarket's serious e-mail server failure a few weeks ago, I found that my line's sync had been affected (the line was down, with the Internet indicator permanently off but all others still up) and I'd lost the Interleaved status for my line. Normally, the extent of the noise on my line is such that, for the line to run at all stably I not only have to have a target SNR of between 12 and 15dB but also the line has to be run in Interleaved mode. Finding that its status was no longer Interleaved, I contacted Skymarket and asked them to arrange for it to be put back. This was duly done within 24 hrs. But I now wonder whether something else was done to change the DSLAM's mode of operation, something which perhaps might now account for all this strange non-access I've been getting.

Edited by deleted (Sat 05-May-12 22:57:00)

Buy yourself a cheap new (spare) Netgear DG834GT or DG834G v4, which can be tweaked even w/out DMT, on eBay and take the router out of the equation. Can be had for < £20.

Well, that's all academic now, as I've decided to go FTTC and to sign up with a different ISP to Skymarket (Skymarket can't offer FTTC at present, anyway). I'm simply not prepared to spend any more of my time on this problem. I've got too many other time-sapping and urgent things to do.

I remain of the view that the problem probably lies with either the DSLAM or with a Transport or Network level, further up toward Skymarket. Way back in the past the DSLAM was problematic and Skymarket had to send an engineer to the exchange to sort it out on one occasion. The engineer then came to mine to test the 'correction' and confirm that I was up and running fully again. It seems that these days that's no longer being done.

I think something was changed inadvertently when, some weeks ago after a Skymarket shutdown, they instructed an engineer to restore my Interleaved status, which had mysteriously disappeared.

Skymarket currently seem unwilling to discuss the problem any further with me. I think that says it all.

Well, we'll never know! You asked for help but never provided what was requested; instead launching into long verbose spiels about red herrings and grasping straws

Way back in the past the DSLAM was problematic and Skymarket had to send an engineer to the exchange to sort it out on one occasion. The engineer then came to mine to test the 'correction' and confirm that I was up and running fully again. It seems that these days that's no longer being done.

Some one is lying to you.
Engineers are tasked at the CP's request, to check and repair broadband problems every day.