Copper Clad Steel "CW1308"

Tidied up the incoming phone line today. The house rewire included a couple of Cat5e's and some CW1308 from the NTE to the eventual cable/distribution point.
Strip back the CW1308 and think these cores are a bit stiff, wire everything back up. POTS works fine, good level, no audible noise, hum, crackles, etc.
ADSL2 (up to 8 Meg) on the other hand is not happy. Daytime sync rate was persistently 1 Mbps lower than previous that had bee rock steady at just over 6 Mbps (not bad to 3 miles of old ali). Looking at the plot of the number of symbols/carrier it is obvious that LF and MF broadcast stations are really clobbering things. It's now got dark and it's even worse, sync rate 2.5 Mbps lower and basically no symbols being above about 500 kHz, the MF band starts at 526.5 kHz...
All I've done is shorten the dropwire by 3 yards, and use this new bit of "CW1308". The wiring down to the current location of the ADSL modem is unchanged. Close inspection of the new CW1308 shows it to be copper clad steel, yes steel, not ali, it is magnetic.
Could this dubious bit of CCS be better at receiving the MF interference, have a funny resistance/impedance or perhaps joints to it doing "odd" things? I can't think how but the ADSL is definatly suffering a lot of MF interference.
Everyone home now and taking down the 'net connection would not be conducive to family harmony. Tommorow I'll switch to a pair in one of the (copper) Cat5e's and see what happens. Most odd and be interested in others thoughts.
--
Cheers
Dave.
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

A quick punt:
From memory, the CW1308 spec specifically specifies<?> copper so anything but copper is not to spec.[1]
Steel is too hard to give a gas tight seal on IDC so will be more likely to corrosion and poor joints in the long (and possibly short) term.
I can't see it being more susceptible to induced interference though as it will still be cancelled in the return leg.
Skin depth at ADSL freqs, pass.
[1] or at least specifies a loop resistance that steel cannot meet.
--
fred
it's a ba-na-na . . . .
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On 24/09/13 21:02, fred wrote: [snip]

Plus, I'd imagine the steel itself will rust far too easily.

Is it possible since this cable is crappy, the twist isn't to spec either, so any common-mode interference may be worse?
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On Tue, 24 Sep 2013 22:39:34 +0100, Chris Bartram wrote:

short)

The slight staining on the insulation at the cut ends was another hint along with the stifness that this wire wasn't copper. That end is under screw terminals, the other end (at the NTE) was cut back about 6" before being punched down into a many time used IDC on the back for the NTE's lower half face plate.

The twist looked good, a lot tighter than I've seen on some samples of telephone cable. Very similar to that of Cat5. Hum, all the pairs looked very much the same though and the cable sheath has a very uniform pattern. Unlike Cat5 where each pair has a different twist pitch, to minimise crosstalk and which gives that bobbly nature to Cat5 sheaths.
--
Cheers
Dave.
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On 24/09/2013 21:02, fred wrote:

Well, the skin depth in copper at 500 kHz is about 0.1 mm. CW1308 conductors are 0.5 mm dia (IIRC) which would make the RF resistance roughly 0.1 ohm per metre.
For steel the calculation isn't so easy since the key parameters (resistivity and relative permeability) vary wildly depending on the alloy, state of temper, etc. For a stab in the dark I'll use a resistivity of 10E-8 ohm-metre (Kaye & Laby, for mild steel) and 100 for the permeability (Wikipedia for unspecified carbon steel). This gives a skin depth of ~20 microns and an RF resistance of circa 3 ohms/m. (This is ignoring the copper coating which may or may not be insignificantly thin).
That's a factor of 30 on resistance of this section, but its probably insignificant compared to the overall exchange loop. Another factor is that the steel section will be quite inductive (the line's Zo will be higher) and this will introduce some more attenuation.
Dave - does your modem/router show line attenuation figures and have they increased?
The steel will have to go though. These fake cables are nothing but trouble...
--
Andy

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On Wed, 25 Sep 2013 12:16:27 +0100, Andy Wade wrote:

When talking about 20 microns I should imagine the copper plating to be significant.

No the line attenuation figures are the same at either 44 or 45 dB down stream. It's just (was...) an awful lot of added MF noise screwing up the SN ratio.
--
Cheers
Dave.
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On 25/09/2013 12:42, Dave Liquorice wrote:

Hmm, it might only be 3-5 microns - that's a common sort of plating thickness for appearance and protection. Fakers won't use any more copper than they have to. For most of the current to flow in a copper layer it needs to be several skin-depths thick - two or three, anyway - and that means hundreds of microns at the frequencies of interest, so the steel will have an adverse effect. This is becoming an academic discussion though; I don't the RF resistance of the conductors is very relevant...

Hypothesis: as some contributors have said, the steel won't work well in IDC (incl. jelly-crimp) connectors - you won't get reliable low-resistance connections. If connection resistances are different in the A & B wires it will have an unbalancing effect on the line - allowing the common-mode broadcast interference to enter the wanted signal path. (Discuss.)
You could try a common-mode filter (choke) such as the BT80A-RF3 (or an iPlate, which contains the CM choke component as the aforesaid block terminal.
I don't buy any of this stuff about resonant antenna lengths. We're talking about a 3m change where the wavelengths of the interfering signals are 100s of metres.
--
Andy

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On 25/09/2013 19:04, Andy Wade wrote:

AIUI, and I am open to correction, the choke in the latest front plate is wired to reduce RF pick up on the bell wire not the incomimg line. As for the resonant antenna length that is what we are talking about the total physical length. Altering that can make a difference. The worst mistake anyone can make in fault finding is to say that "It can't be that".
--
Peter Crosland

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On 25/09/2013 20:15, Peter Crosland wrote:

Not quite sure what you mean by "the latest front plate" but in the iPlate it's both. There's a bifilar-type common-mode choke for the thro' line pair (pins 2 & 5) and a separate 22mH inductor in series with the bell-wire (pin 3). The details are all in BT's patent, no. GB2445212.
--
Andy

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On 25/09/2013 22:53, Andy Wade wrote:

I was referring to the latest version that BT fir for customers that have FTTC service. I have not had the opportunity to look inside but I was assured by an experienced BT engineer who is broadband trained, that there was not an equivalent inductor to that fitted to the RF3 in the latest version of the NTE5. So now there are two directly conflicting accounts.
--
Peter Crosland

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On 27/09/2013 10:16, Peter Crosland wrote:

Ah, googling "vdsl faceplate" is quite revealing. There is indeed a new i-plate style filtered faceplate for VDSL (FTTC) and ADSL. This looks like the death-knell for the microfilter - and about time too. Using the interstitial concept opens the door to a (legit) self-installed filtered faceplate installation. FTTC self-install must be on the way...
According to http://www.run-it-direct.co.uk/btvdslfaceplate.html it does provide common-mode filtering, in addition to the longitudinal low-pass filtering for the phone o/p. There are three double-wound coils visible in the photo, neatly colour-coded for correct insertion on the PCB.
--
Andy

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On 27/09/2013 23:43, Andy Wade wrote:

I am surprised it was not so from the start... changing the front (or in this cases adding under the front panel) of an NTE5 for the new style filter can't be any harder than for the old ADSL only style...
Unless there is some other magic the BT bod does on his visit?

Which explains their extra depth...
--
Cheers,

John.
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On Sat, 28 Sep 2013 16:31:03 +0100, John Rumm wrote:

I should imagine putting a tone tracer on the end of the line so he can find the right pair in the rats nest cabinet to jumper across to the fibre cab and back. B-) It's not quite the same as ADSL from the exchange, they know which "port" the line is connected to in the exchange so can find that relatively easily.
--
Cheers
Dave.
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On Wed, 25 Sep 2013 20:15:20 +0100, Peter Crosland wrote:

You don't know how close you are to the orginal problem with that statement.
When I repodged one of the junctions the other night I sort of noticed that the ends of the unused pairs where "close" to one of the screw terminals. Had a fiddle, deliberatly connected one wire from one of the unused pairs to the A wire of the line. Instant speed drop and lots of MF interference. Remove it and modem doesn't notice, at least not in a few minutes time frame. Manual resync good speed, bung the wire back on instant drop... So that's the orginal problem solved. B-)
That night it still wasn't as good as it was though, left it through last night as well and again a larger than normal drop in SN ratio. The daytime symbols/carrier plot shows big dips at 693, 810 and 1053 kHz, 45 dB loss and a SN ratio varying. from 8 down to 3 dB. Stable sync at a tad over 6000. Night SN was getting down to 0 dB, that's not how it was.
This morning swapped out the CCS for a copper pair in one of the Cat5's. Still a large dip at 693 kHz, the one at 810 has more or less gone and 1053 is now much broader (there are three MF stations close together around there) and not as deep. 44 dB loss and much more stable SN at 6 dB and (daytime) sync a tad over 7000. That's more like it.
Will be interesting to see what happens tonight, I expect it will resync down to something over 6000 as that was "normal". At this stage it very much looks like CCS knocks the edge of ADSL for what ever reason. This is about 10 yards of CCS in between the NTE and about 40 yards of internal, proper, CW1308 that runs on a convoluted route through the house to filter/ADSL modem and POTS equipment. Single pair only all the way.
I noticed CCS "telephone cable" in one of CPCs offer flyers the other day, Pro-Power 100m 3 pair white £8.00. By comparison Pro-Power 100 m 3 pair white CW1308 (copper) £26.94. For the CCS they do quote a conductor resistance of <= 700 ohms/km and state it's CCS. TBH how many would realise the significance of those two bits of information? The price differential would prompt me to look closer but a jobbing builder?
--
Cheers
Dave.
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Short memories:
I had forgotten I had a recent experience of this (April 2012) with toolstation selling CCS not quite at CW1308 but at CW1308 prices:
https://groups.google.com/forum/?hl=en #!topic/uk.d-i-y/jg8HASJ_RJs
I was sure it was a mistake by a buyer at TS but instead of admitting/correcting the error they dropped copper telephone cable altogether and now only sell CCS and CCA.
My guess was that they had a warehouse full of Chinese CCS and weren't going to take it on the chin and bin it.
Here's a copy of my April 2013 followup (invisible on GG):
"They no longer sell copper telephone cable. They used to sell some stuff that was priced at copper levels and some cheaper CCA stuff. I bought some of the more expensive stuff which turned out to be CCS which I took back for a refund as poorly described. I then made a point of telling the buying team that someone had fooked up and misbought CCS instead of copper. As they had a warehouse full of CCS they decided to solve the problem dropping 'copper' and just selling CCS and CCA which are both of course shit. After a while, and with a little persuasion, they corrected the descriptions and sorted the prices to reflect those expected of copper plated fence wire.
Definitely a store where the buyer needs to keep their eyes open."
--
fred
it's a ba-na-na . . . .
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On Fri, 27 Sep 2013 21:05:27 +0100, fred wrote:

May '13 cat only has CCA and picture of the untwisted pair flat stuff...
As I suspected the modem did resync over night, twice, first time down to 6800 then again to 6200 (ish). Daytime SN is 7 or 8 dB.
Also did some maths on the resistance, that 10 yards of CCS would have a resistance of about 7 ohms per leg, same length of copper about an ohm per leg. That to me is a significant difference...
--
Cheers
Dave.
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Not sure the resistance matters much, after all you've got miles of BT wet string between you and the exchange. I'd be looking at the capacitance per m, and the number of twists per m, of the CCS vs. copper.
--
(\_/)
(='.'=)
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On Sat, 28 Sep 2013 13:58:55 +0100, Mike Tomlinson wrote:

True enough but at 97.8 ohms/km assuming 0.5 mm copper, which might be a bad assumption, longer local ends tend to be laid with heavier cable. 0.5 mm CCS is 700 ohms/km...

Won't that be governed by conductor size and spacing, which are essentially the same? The higher resistance will upset the characteristic impedance of the transmisson line and introduce an impedance discontinuity at the transition from copper to steel.

The sample I have is fairly tightly twisted, the orange pair about 80 twists/m, green about 50 and blue maybe a bit less than green. A lot higher than I've seen in several samples of copper.
--
Cheers
Dave.
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Dave Liquorice wrote:

Surely a jobbing builder would use this?
http://cpc.farnell.com/x/x/x/dp/CB14642
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

A jobbing, subbing builder will use whatever he can get his hands on for the lowest amount of bunce!.
Ask Bill Wright what builders use for aerial cables...
--
Tony Sayer



  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

HomeOwnersHub.com is a website for homeowners and building and maintenance pros. It is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.