Kitz Forum
Broadband Related => Broadband Technology => Topic started by: Weaver on November 08, 2015, 03:33:53 PM
-
[Apologies if this question has been answered before. I get the feeling that it might have been.]
In a purely theoretical scenario, talking just physics and practical engineering and materials, and leaving aside costs, will, politics or policy…
Let's say that someone wanted to try and make an improvement in quality to the copper pairs belonging to a typical sized substantial bundle. Let's say the metallic paths in question are several miles long, say 3 - 5 miles long. By "improvement" I mean (i) improving the attenuation, and / or (ii) anything at all to reduce external interference pick-up by way of screening.
So my question is:
* Do any such possible physical upgrades exist?
(Eg getting rid of non-copper, fewer joints, better joints, cross-sectional area
-
fewer joints, thicker cable, screening/shielding of individual pairs and/or the whole bundle, all of which involve replacing the whole lot with something better :)
-
A few questions then about thicker cable:
1. Do Openreach have a choice of cross-sectional area cables to pick from for long lengths?
2. What's the effect at very high frequencies? ( Is it simply the case that skin effect dominates so it's all about the circumference of the conductor? )
Screening and crosstalk:
3. Is there a choice of screening types? geometries?
4. Number of pairs per bundle?
I'd like to understand a bit about what physical choices may become available to Openreach in the future.
If this myth about 10 Mbps USO has any substance at all in it, I wonder if in five years Openreach might plan some simple cable replacements / upgrades.
-
I believe back in the ancient times when DSL was first being brought in, that BT felt it had to do some physical changes to metallic paths, what with sorting some joints out, removing evil kit and so on.
I don't know about what else might have been done. Thicker cable? Copper instead of other metals? (DACS boxes?)
-
The trouble with under ground cables is that the each twisted pair is very close together so its capacitance is very high, and not matched by its inductance. To get over this problem BT used to put loading coils into the pairs every 200 yards or less to match these parameters up. in this way attenuation was lowered on cables. Unfortunately this turns the pairs into a low pass filter and therefore cuts the higher frequencies off. So for broadband the use of loading coils is totally out !
I understand that when TV first started up, outside broadcasts were carried on bunched cable pairs to lower the attenuation.
The best lower loss cabling is the old fashioned open wire routes that were used between exchanges when telephones were first used. Shame about dirty insulators and radio interference though......
In my opinion, the best way to serve customers who are too far away, and few and far between is to use a wireless link. This was done by a gentleman in Scotland, if I remember correctly. See the link ' Diary of a Broadband Not Spot '
-
Goggled it and the link is here for a 'Register Article' --> http://www.theregister.co.uk/2010/12/03/not_spot_diary/?page=1 (http://www.theregister.co.uk/2010/12/03/not_spot_diary/?page=1)
Loved the story :D :D
Reminds me of the long distance WiFi stuff you find on the web using 'Pringle cans and hacked hardware AKA Cantennas. ;D
Some related stuff:
http://www.engadget.com/2005/11/15/how-to-build-a-wifi-biquad-dish-antenna/ (http://www.engadget.com/2005/11/15/how-to-build-a-wifi-biquad-dish-antenna/)
http://people.wallawalla.edu/%7ERob.Frohne/Airport/Primestar/Primestar.html (http://people.wallawalla.edu/%7ERob.Frohne/Airport/Primestar/Primestar.html)
http://www.turnpoint.net/wireless/cantennahowto.html (http://www.turnpoint.net/wireless/cantennahowto.html)
http://www.turnpoint.net/wireless/has.html (http://www.turnpoint.net/wireless/has.html)
I am sure the knowledge & skills to do all this is on kitz and to adapt it for 802.11n/ac. ;D ;)
-
There's a commercial wireless broadband system here in Kent: VFast (http://www.vfast.co.uk/). It now includes an optional free VoIP service, so the pricing looks quite reasonable. I actually applied for it some years ago when we were still stuck with no broadband at all, but it was rejected for topological reasons (it's a line of sight service). I would have thought that wireless would be the cheapest way to fill a lot of the gaps.
-
A few questions then about thicker cable:
1. Do Openreach have a choice of cross-sectional area cables to pick from for long lengths?
Yes. IIRC, the main options are 0.5mm, 0.6mm, 0.9mm.
2. What's the effect at very high frequencies? ( Is it simply the case that skin effect dominates so it's all about the circumference of the conductor? )
Good question, but no idea.
If you google this, "Cable reference models for simulating metallic access networks etsi", you'll find an ETSI document describing cable models for VDSL frequencies, and the parameters that apply to a variety of European cable types, including a few BT drop wires, and 0.5mm d-side cable.
There are similar documents describing more complex models for G.fast frequencies, but you need to be good at figuring out what to search for in ITU work groups; not everything is public - and BT docs are harder to find than some.
One of the leading g.fast researchers has a page here, with some documents from a Dutch perspective:
http://www.joepeesoft.com/Public/DSL_Corner/_Index.htmlhtml
The section on " Modeling Copper Cable" has a good set of documents.
In fact, that ETSI document is on that page too.
Screening and crosstalk:
3. Is there a choice of screening types? geometries?
4. Number of pairs per bundle?
I think there is a possibility of screened cable, but it isn't commonly used. Most likely one screen around the entire bundle.
E-side can number 1000+ pairs, up to 4,800. As a guide, look at the two PDFs at the top of this page
http://www.btcables.com/products/voice/outside-plant-primary-cable
D-side more likely to be 100 pairs. As a guide, look at the two pdf's at the top of this page:
http://www.btcables.com/products/voice/outside-plant-secondary-cable
I'd like to understand a bit about what physical choices may become available to Openreach in the future.
If this myth about 10 Mbps USO has any substance at all in it, I wonder if in five years Openreach might plan some simple cable replacements / upgrades.
I've certainly seen signs that BT will alter copper wiring within bduk, so it is certainly possible this could happen. I do wonder whether they'd prefer to solve the problem through deeper fibre instead of upgraded copper. After all, by then, BT will be 3 years into their 10-year plan.
My observation is that a 10Mbps minimum is a technology killer ... It requires all 20CN, ADSL-only exchanges to be upgraded, either to 21CN, or to fully-fibred capability. At least it means that where BT will be part of the USO.
I agree with benji that wireless is better in some places. But that might turn out to be 4G with BT/EE by 2020.
-
@WWombat - you're a star. thank you so much for all that good stuff, which will take me a good while to digest and to follow up.
One clarification, I realise it's a complete and utter pain entering these things, but I'm presuming I should read 0.5 mm2 area earlier, rather than a cable of diameter 0.5 mm etc etc
-
. . . I'm presuming I should read 0.5 mm2 area earlier, rather than a cable of diameter 0.5 mm etc etc
I have always assumed the measurement to be the diameter of the metallic conductor. :-\
-
Got it, thanks.
-
Wombat has covered everything in his great post ............. to be as pedantic as the most pedantic man on Earth ...... the actual sizes are 0.32, 0.4, 0.5, 0.63 and 9.0 mm.
0.5mm is by far and away the most commonly used.
-
Here's a section from our Network Records that shows an E-side cable (From the Exchange MDF to the Cabinet), and it's 'make-up' of different lengths and sizes of Cu cable.
The 'Pair No.s' refers to the Bar/Pair on the MDF ...... for example on Cable A leaving the Exchange, there are 400prs of wires that are routed to this particular Cab.
As an aside, the Transmission dB values shown were calculated years ago at 1.6Khz, and isn't a true representation of Broadband losses which are taken at 300Khz, and would be shown at 54.8dB for this cable.
:)
-
As an aside, the Transmission dB values shown were calculated years ago at 1.6Khz, and isn't a true representation of Broadband losses which are taken at 300Khz, and would be shown at 54.8dB for this cable.
Just to say thank you for that aside. It has explained two minor puzzles! :)
I had long wondered why the 2Wire 2700HGV (and the like) would report the attenuation at 300 kHz.
And now I know why, when I measured the frequency generated by my Oscar (http://forum.kitz.co.uk/index.php/topic,15579.msg289987.html#msg289987), it was of the order of 1.X kHz.
-
:thumbs:
-
If anybody is interested, the gentleman in Scotland who wrote the DIARY OF A NOT SPOT article for the REGISTER did a follow up article later called DIARY OF A NOT SPOT - FINAL CHAPTER. Sorry that I have not worked how to do the click on link to other sites yet...............
-
Just copy the address from the url bar and paste it in your post
http://www.theregister.co.uk/2011/02/03/not_spot_no_longer/