Print

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

[Weaver]

(I'm a new user of MDWS, viewing data and getting a feel for what I'm seeing. I haven't been able to upload anything yet myself.)

Referring to WWWombat's "bits per tone" graph, there's a huge dip around bin 329 (@~1418kHz) at which it goes down to 2 bits, but on either side of the dip the high ground of the graph is at 12 bits per bin.

What's this all about? Why the dip?

I've seen it with other users too.

[WWWombat]

The dip is caused because FTTC cabinets have to transmit at reduced power in this part of the spectrum, to make sure it doesn't cause excess interference with the ADSL/2+ signals that originate in the exchange.

The aim is to ensure that the signal level generated at the cabinet matches the signal level of ADSL as it arrives at the PCP.

The notch varies, depending on the distance between PCP and exchange, with something like 25 different settings:
- No notch if the PCP is outside the exchange.
- Deepest notch, at highest ADSL frequencies, if PCP is ~ 2km from the exchange.
- Smaller notch, at lower ADSL frequencies, if PCP is ~4km from the exchange.

Start looking at some of the other lines, and you'll see a variety of these patterns.

[Weaver]

Thank you very much for that, from the horse's mouth. So FTTC users could do with ADSL being scrapped? :-)

[Bald_Eagle1]

I'm not sure how far WWW's PCP is from the exchange, but the notches in our graphs are in quite different places.

My PCP is around 4.3km from the exchange & I'm around another 1.1km further away again.


I would probably benefit from ADSL being scrapped so that power wasn't cut back so severely at the lower tones/frequencies.

Due to being quite a long way from the cabinet (for VDSL2 purposes), my connection can't physically use the higher frequencies, so power being cut back at the lower frequencies has quite a negative effect on my connection speeds.

[WWWombat]

I reckon my line loses around one eighth of its capability from that notch, so it can be quite an impact.

The power notching is specified in the NICC's ANFP, ND1602 v5.1.1,
http://www.niccstandards.org.uk/publications/llu_spec.cfm

I've attached the appropriate graph from the ANFP - where the different coloured lines give 5 example power masks (ie max power limits) for 5 different cabs. It shows 5 of the 26 different settings; my very rough rule of thumb is that a CAL of 10 maps to 1km of exchange-cabinet distance for standard 0.5mm copper (the CAL is actually attenuation at 300kHz, in dB).
- Pink line: CAL=50, approx 5km full notch, trough at 0.5MHz, like @BE1's
- Blue line: CAL=40, approx 4km, full notch, trough 0.6-0.8MHz
- Light blue line: CAL=30, approx 3km, full notch, trough 1.1-1.2MHz
- Green line: CAL=20, full notch, trough 1.5-2MHz,
- Yellow line: CAL=10, partial notch at 2MHz
- Red line: CAL=0, outside the exchange, no notch

When looking at the variations, the CAL=20 notch looks worst to me, but I could be wrong - I hate trying to interpret logarithmic graphs.

My notch looks very much like CAL=25ish, but the distance could be as little as 1.4km ... but I'm pretty sure the lines are only 0.4mm copper, which will skew things.

[Black Sheep]

From our internal docs for the Planning teams ....... taken at 300kHz 1dB is approx. 87mtrs @ 0.5 Cu conductor. 

[WWWombat]

And, ta-da  , A CAL of 50 at (87m per dB) maps to 4350m. Which is what @BE reckoned his distance was.

Now my rule of thumb isn't as easy to remember!

[loonylion]

I do wonder why people on ADSL/2/2+ don't get migrated to ADSL/2/2+ provisioned from the cab. The DSLAMs are capable of doing it, it would likely reduce fault callouts and everyone would benefit, the ADSL/2/2+ people from having a shorter copper loop and less crosstalk, and the VDSL2 people from not needing such severe power cutback masks.

[Ronski]

Perhaps because less people would switch to fttc, and every adsl connecton wold be one less port fttc customers. I think they may be heading in this direction with the lower price packages mentioned recently, but then cabs will become full a lot quicker causing problems.

[WWWombat]

1. There is, of course, the issue that they never provisioned the FTTC cabinets with enough space for ports for every single broadband line.

Adding a second cabinet is feasible, but it has adverse effects on vectoring ... if that were ever to be deployed.

2. BT might want to migrate users to their own equipment in the cabinet, but you can bet that none of the LLU operators do.

3. And the killer - until every single line has been migrated, LLU operators included, there is nothing to be gained at all.

This is because the power masks that currently apply to VDSL2 would also apply to any cabinet-based ADSL/2+ services too. We know that the intention of the power masks is to reduce VDSL2 power (in ADSL spectrum) such that the VDSL2 signal level generated becomes identical to the exchange-based ADSL/2+ signal level that passes by.

Obviously, if you reduce the cabinet-based ADSL/2+ signal level so that it is identical to the exchange-based ADSL/2+ signal level, the consequences are clear: The cabinet-based signal will be indistinguishable from a passing exchange-based signal. You would get no extra range, and no extra speed.

(This latter thing is why I don't believe that the lower 18/2 package won't be a cabinet-based ADSL product).