Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[sheddyian]

An elderly friend of my Father's had commented that her grandchildren said her computer was very slow, so I volunteered to have a look at it.

The computer was indeed slow, taking some time to boot, and never felt very responsive.  Windows 7, quad core 1.8Ghz CPU, 4Gb RAM.. shouldn't have been that slow.  Anyway, that bit is for another day.

The internet was also rather sluggish, and I noticed it'd resynched 3 days previously.

Looked at the wiring, router plugged into master NTE5, 2 extensions star wired from that.

To cut a long story short, synch speed was 5,200.  In test socket I got 7,600.  Extension wiring included ring wire AND it's other half.  Disconnecting both of these wires and putting it all back together gave me the same speed as using the test socket alone - result!

So, by disconnecting the ring wire we got an extra 2,400 synch speed.  Was very pleased.

Curious : What's the best results you've seen for simply disconnecting the ring wire?

One oddity I noticed, on one of the extensions the orange/white wire was correctly marked, but it's other conductor was entirely white.  A fair bit of outer sheath had been stripped back, but I couldn't see any orange stripes on it to indicate which pair it was.  This had clearly given the original installer some headaches, as the end had been stripped back a bit - probably to buzz through with a meter, as no other wires had been stripped back in that way.

Ian

[ColinS]

What's the best results you've seen for simply disconnecting the ring wire?

Congrats Ian on a good job, well done. 
It's quite a good few years ago, but I recall an uplift of approximately the same order from disconnecting the bell-wire in ~10m of otherwise good extension wiring.  Master+3 extensions daisy-chained, not star.  As I seem to recall it lifted the ADSL2 DS synch from ~8Mb/s to ~10.5Mb/s at the time.  Subsequently on ADSL2+ with a  3dB SNRm it was easily capable of 12.5Mb/s and up to ~14.5Mb/s on a good day. It is about 2.5Km from the Exchange
The same wiring introduced ~the same 2Mb/s drop when my VDSL2 was reconfigured from centralised (SSFP) filtering to distributed (using Z350 filter in dongles or integrated faceplates on a per extension basis). It has an attainable DS of ~91.5Mb/s, obviously capped (when there's no concurrent voice fault  ) at 80Mb/s. It is about 150m from the PCP
Hope that this is of some interest. 

[burakkucat]

  for another excellent Sheddy-achievement!

[guest]

I think I linked something up for snadge in "another place" about how a close noise source (bellwire/ringwire) can have a surprisingly bad effect on even good quality TP but I'm buggered if I can find it. It had a good graph showing various problems like that and how to diagnose them.

[neilius]

Hey folks,

My first post here but I've been lurking for quite some time... I'm a sound engineer (music) by trade but also have an unhealthy interest in telecoms.

Thought I'd chime in with my experiences  - my parents moved into a new bungalow recently and their phone wiring was set up with 3 extensions star wired to an NTE5A in the loft, with bell wires galore. Outside on the wall was a BT junction box that took the drop pole cable into the loft to feed the NTE. The line had an attenuation of around 37dB and they were achieving a 6MB downstream sync speed. I could also hear a not so nice 50Hz hum on the phone as well as other background hiss. Not good...

I relocated their NTE to where their living room extension was, ripped out their other extensions as they use DECT phones and stuck a BT85B in the loft with gel crimps joining the incoming cable to the newly located NTE. To top it off, we put in a Pressac ADSL filtered faceplate. Voice quality is fantastic, with no noise, very clear speech and ADSL downstream speeds are now hitting 11MB with 6dB SNRM. Quite a massive improvement.

Suffice to say, everyone is very happy indeed with the result and I do find it disheartening to know that there are lots of people still living with this kind of poor setup, blissfully unaware.

I notice that the OP mentions that his setup was also star wired - this can have a bad effect on ADSL on top of the bell wire issue. If extensions are really needed, it's much better to daisy chain/bus wire them.

[Black Sheep]

Hello, welcome, and congrats on gaining great results.

As you say yourself, some folk are blissfully unaware of the negative effects of star-wiring, bell wires etc. A large proportion of my job is to resolve these issues, and when the EU sees the gains, they readily exclaim they will be telling their friends and relatives. So, as with the monkeys and typewriters, eventually there'll be no 'pre-DSL' type wiring left.

Hopefully, I'll get the very last EU on the day I retire. 

[neilius]

Hi Black Sheep and thanks for the welcome!

I guess it feels a bit like whack-a-mole or painting the Forth bridge at times!

[Black Sheep]

Ha ha, I've got to be honest, I actually love 'rehashing' the wiring config to see the speed gain that will usually follow. Especially when the EU has 'put up with it' for ages, and it has taken months before their ISP will get an Openreach engineer out to them.

[sheddyian]

Ha ha, I've got to be honest, I actually love 'rehashing' the wiring config to see the speed gain that will usually follow. Especially when the EU has 'put up with it' for ages, and it has taken months before their ISP will get an Openreach engineer out to them.

100% agree, It's very satisfying !

Ian

[JGO]

Neilius comments  " I notice that the OP mentions that his setup was also star wired - this can have a bad effect on ADSL on top of the bell wire issue. If extensions are really needed, it's much better to daisy chain/bus wire them. that a ring/daisy chain "

Would be interested why ?
If a frequency can arrive by multipaths then when components are near equal and out of phase you get a null, wired, wireless or w/g.    Does a ring in telephone cable shows more attenuation round the long path so the null isn't as deep as with a stub ?

Also, before I had interrupted the ring wire BT had set the S/N margin up to 15 dB. Reducing that had as much effect on speed as the immediate effect of the interruption.

[burakkucat]

Two words to consider: Bridging tap. 

[sheddyian]

In this particular case, the star wiring

[1]======[M]======[2]

1 2 : extensions
M   : NTE5 master socket (BT line entering here)

didn't seem to have an impact, as I got the same synch speed, noise margins and attenuation if I plugged the router into the NTE5 test socket as I did with the wiring reconnected without the bell wire.

But there were only 2 spurs and neither was particularly long (maybe 5 metres).

Ian

[JGO]

Neilius comments  " I notice that the OP mentions that his setup was also star wired - this can have a bad effect on ADSL on top of the bell wire issue. If extensions are really needed, it's much better to daisy chain/bus wire them. that a ring/daisy chain "

YES  ! - just  realised although the mechanism is much the same, with  a stub the same wire length is used twice so the effects cuts in at a lower frequency.  Calculating for 2 MHz, L/4 for stub is 37.5 metres but L/2 for ring is 75 which lets out most houses.  (For VDSL or with a big house Oh brother !   Blenheim Palace must be a problem !!!!!)