Strange wiring and syncing problem

I was trying to help a friend out today. He has Freeserve/Wanado/Orange or
whatever it is called today :-)

Some time in the past he had trouble with his line as the drop wire was fouled
by a tree. When Mr BT turned up he said there was not much he could do as the
tree was someone else's. My friend pointed out that there was a clear line of
sight from the pole to his barn, now converted to workshop and office, and from
the barn to the original ingress point on the house.

Mr BT thought that was a good idea. So, my friend added, how about a socket in
the barn/office whilst you are at it? So Mr BT fitted one and everyone was
happy. Even the ADSL worked when it arrived.

Recently though the modem, Speedtouch 330, has been just blinking the ADSL
light with no solid lock. The *only* way he can get a connection is to use the
POTS modem of his laptop to dial out - whereupon the ADSL light will go solid
and you can hang up the POTS call and the ADSL will stay on for some time, but
eventually it will drop out.

What I really went round for was to swap the Speedtouch for a WAG300N that he
had bought so he could work in the kitchen etc. Turns out the WAG300N does the
*same* thing. It tries, ADSL LED blinking, and then gives up. Sometimes it will
lock and always the POTS dialing trick will bring it on solid.

I checked out all the house wiring and that is when it started to get strange.
It seems that Mr BT ran the drop wire from the pole to a junction box on the
outside of the barn and then ran one cable into the office and one to the
house. Both have a NTE5 fitted and only two wires are connected. There is
nothing else on the office wire and the house one has a single DECT base daisy
chained off the NTE5 faceplate. Pulling that faceplate made no difference to
the lock/no-lock. Everything is correctly filtered.

ISTR something similar recently, but it may not have been this group, and of
course I can't find it now.

My friend thought it started after he installed *something* on the XP laptop.
Now I can see something possibly buggering a USB connection but not an ethernet
one. In any case that is a complete red herring because it would not connect
with only my laptop wired to the WAG300N - and my laptop does not run Win.

--

Regards

Dave Saville

NB Remove -nospam for good email address

On Thu, 01 Feb 2007 20:58:08 +0000 (GMT), "Dave Saville"
<(E-Mail Removed)> wrote:

>Some time in the past he had trouble with his line as the drop wire was fouled
>by a tree. When Mr BT turned up he said there was not much he could do as the
>tree was someone else's. My friend pointed out that there was a clear line of
>sight from the pole to his barn, now converted to workshop and office, and from
>the barn to the original ingress point on the house.
>
>Mr BT thought that was a good idea. So, my friend added, how about a socket in
>the barn/office whilst you are at it? So Mr BT fitted one and everyone was
>happy. Even the ADSL worked when it arrived.
>
>Recently though the modem, Speedtouch 330, has been just blinking the ADSL
>light with no solid lock. The *only* way he can get a connection is to use the
>POTS modem of his laptop to dial out - whereupon the ADSL light will go solid
>and you can hang up the POTS call and the ADSL will stay on for some time, but
>eventually it will drop out.
>
>What I really went round for was to swap the Speedtouch for a WAG300N that he
>had bought so he could work in the kitchen etc. Turns out the WAG300N does the
>*same* thing. It tries, ADSL LED blinking, and then gives up. Sometimes it will
>lock and always the POTS dialing trick will bring it on solid.
>
>I checked out all the house wiring and that is when it started to get strange.
>It seems that Mr BT ran the drop wire from the pole to a junction box on the
>outside of the barn and then ran one cable into the office and one to the
>house. Both have a NTE5 fitted and only two wires are connected. There is
>nothing else on the office wire and the house one has a single DECT base daisy
>chained off the NTE5 faceplate. Pulling that faceplate made no difference to
>the lock/no-lock. Everything is correctly filtered.

What are the cable run lengths like? Words like "barn" suggest quite
big distances are involved. If the feeds to the two buildings
effectively amount to a Y-shaped arrangement, with either branch
comparable to a quarter wavelength at ADSL carrier frequencies, which
I understand go up to about 1MHz, then it might act as a quarter wave
stub and cancel at some frequencies causing the unpredictable results
you describe. For 1MHz, a quarter wavelength would be 75 metres, so if
the cable runs are comparable with that, this could be the cause.

If so, the cure would be to get one of the Y branches removed so there
is a single cable run to one termination socket, and use a cordless
phone or a filtered extension cable after that.

Rod.

Dave Saville wrote:

Connect a router and look at the status page.

Anything else is guesswork.

Graham

On Fri, 02 Feb 2007 04:10:36 +0000, Eeyore wrote:

>
>
>Dave Saville wrote:
>
>Connect a router and look at the status page.
>
>Anything else is guesswork.

It *is* a router, well a modem/router - And if it does not connect the status
will tell me exactly what? Not connected or link down I suspect.

--

Regards

Dave Saville

NB Remove -nospam for good email address

On Thu, 01 Feb 2007 21:31:37 +0000, Roderick Stewart wrote:

>What are the cable run lengths like? Words like "barn" suggest quite
>big distances are involved. If the feeds to the two buildings
>effectively amount to a Y-shaped arrangement, with either branch
>comparable to a quarter wavelength at ADSL carrier frequencies, which
>I understand go up to about 1MHz, then it might act as a quarter wave
>stub and cancel at some frequencies causing the unpredictable results
>you describe. For 1MHz, a quarter wavelength would be 75 metres, so if
>the cable runs are comparable with that, this could be the cause.

No cigar I am afraid - "It's 40 m from the pole to barn and 20 m to the house."
Junction box on barn to NTE5 is at most 3m and is drop cable. On the house it
is drop cable for 4 m or 5 m into the house, a really old junction box, and
"normal" inside type BT cable for maybe another metre to the second NTE5. As I
said pulling the faceplate from this one makes no difference. I can't (legally)
isolate it any more than that. :-)


--

Regards

Dave Saville

NB Remove -nospam for good email address

On Sat, 03 Feb 2007 11:06:38 +0000 (GMT), "Dave Saville"
<(E-Mail Removed)> wrote:

>On Thu, 01 Feb 2007 21:31:37 +0000, Roderick Stewart wrote:
>
>>What are the cable run lengths like? Words like "barn" suggest quite
>>big distances are involved. If the feeds to the two buildings
>>effectively amount to a Y-shaped arrangement, with either branch
>>comparable to a quarter wavelength at ADSL carrier frequencies, which
>>I understand go up to about 1MHz, then it might act as a quarter wave
>>stub and cancel at some frequencies causing the unpredictable results
>>you describe. For 1MHz, a quarter wavelength would be 75 metres, so if
>>the cable runs are comparable with that, this could be the cause.
>
>No cigar I am afraid - "It's 40 m from the pole to barn and 20 m to the house."
>Junction box on barn to NTE5 is at most 3m and is drop cable. On the house it
>is drop cable for 4 m or 5 m into the house, a really old junction box, and
>"normal" inside type BT cable for maybe another metre to the second NTE5. As I
>said pulling the faceplate from this one makes no difference. I can't (legally)
>isolate it any more than that. :-)

If I understand correctly from your original description, the ADSL
modem is in the barn/office, which is 30m from the end of the line,
which is in the house, or in other words, the ADSL modem is not at the
end of the line. Is this right?

Admittedly the run lengths you give suggest that problems would only
occur at higher frequencies, but I still wonder if there could be
enough cancellation to intervere with the waveform of the ADSL signal?
If I'm right about your cable geography, it might be worth trying the
modem at the end of the line instead of 30m from the end, and with
nothing connected to the intermediate point. Otherwise I'd have
thought there was a fault on the modem, but you say you've tried two
different ones with the same result, which would suggest it is
something to do with the cabling.

Rod.

On Sat, 03 Feb 2007 15:49:37 +0000, Roderick Stewart wrote:

>Admittedly the run lengths you give suggest that problems would only
>occur at higher frequencies, but I still wonder if there could be
>enough cancellation to intervere with the waveform of the ADSL signal?
>If I'm right about your cable geography, it might be worth trying the
>modem at the end of the line instead of 30m from the end, and with
>nothing connected to the intermediate point. Otherwise I'd have
>thought there was a fault on the modem, but you say you've tried two
>different ones with the same result, which would suggest it is
>something to do with the cabling.

Yup - Two modems and three filters.

I may suggest trying the far end but 1) He wants the ADSL in his office and 2)
It *was* working for a long time with just this setup. Could it be *anything*
at the exchange end? I have seen reports of ADSL being mucked up by using POTS
but this is the exact opposite.

--

Regards

Dave Saville

NB Remove -nospam for good email address

On Sun, 04 Feb 2007 12:54:28 +0000 (GMT), "Dave Saville"
<(E-Mail Removed)> wrote:

>On Sat, 03 Feb 2007 15:49:37 +0000, Roderick Stewart wrote:
>
>>Admittedly the run lengths you give suggest that problems would only
>>occur at higher frequencies, but I still wonder if there could be
>>enough cancellation to intervere with the waveform of the ADSL signal?
>>If I'm right about your cable geography, it might be worth trying the
>>modem at the end of the line instead of 30m from the end, and with
>>nothing connected to the intermediate point. Otherwise I'd have
>>thought there was a fault on the modem, but you say you've tried two
>>different ones with the same result, which would suggest it is
>>something to do with the cabling.
>
>Yup - Two modems and three filters.
>
>I may suggest trying the far end but 1) He wants the ADSL in his office

Well, if a test showed that the ADSL modem worked on the *end* of a
cable run but not an intermediate point, the practical thing to do
would be to make the office the end of the cable by removing the run
to the house, and putting a cordless phone base station in the office.

> and 2)
>It *was* working for a long time with just this setup.

I must have missed that point, as I thought it was the installation of
the socket in the office that caused the signal to go bad. If it
wasn't, then was anything else about the installation (knowingly)
changed at the relevant time?

> Could it be *anything*
>at the exchange end? I have seen reports of ADSL being mucked up by using POTS
>but this is the exact opposite.

I could be wrong, but still think it's most likely something to do
with your local cabling. Another possibility that occurs to me is that
maybe one leg of the cable has become broken or disconnected
somewhere. ADSL, being an RF signal, could perhaps still get through
intermitently even with only one wire present, and connecting a phone
might provide a path to earth which could alter its behaviour.

I'd want to check all the connections in the termination boxes first,
and then continuity of all the cables. Strictly you shouldn't touch
any of this of course because it belongs to BT, but I'd be tempted to
have a look....

Rod.