Print

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

[Weaver]

 I assume that I could never use the highest tones just immediately below 1.1 MHz anyway, hence Burakkucat's helpful suggestion that I artificially limit my modems to ADSL2 not auto = ADSL 2+, which I did, on all three modems.

[HPsauce]

Btw, how is the overall attenuation reported by typical xDSL modems derived mathematically, that was the other half of my question? A straight summation, or weighted in any way?

I recall, back in "the mists of time" various discussions on that and (as noted in another post) "vendor specific" was as close as anyone got, with answers varying from "whatever the reading was on one particular tone" to "an average of some tones".
Whatever it is, you can bet it's at best a quick and dirty calculation. 

[roseway]

"Quick and dirty" is my bet too. My Zyxel VMG8324-B10A doesn't report overall attenuation values in the CLI, but it does report them in the GUI. The values reported in the GUI don't appear to have any logical connection with the per-band attenuation figures - in the GUI the upstream attenuation is reportedly higher than the downstream, but the per-band figures indicate the opposite.

[Weaver]

Has anyone read anything about the pre-existing setup in the LR-VDSL trial areas before the trial started ?

I'm wondering if they put cabs in for the first time in order for the trial to go ahead. (Presumably not, unless they wanted to get a feel for economics of the whole thing too.) I wonder if there were EO lines at the one extreme, or at the other, whether they simply tweaked one or more existing cabs.

[ejs]

The initial LR-VDSL pilot in Isfield was done there because the Isfield exchange never got any ADSL, so there is no ADSL for the LR-VDSL to cause interference to.

[WWWombat]

I think the Lewis cabinet were enabled as part of BDUK prior to LR-VDSL.

Just for the record, this seems to be a chart of results from Isfield...

[NewtronStar]

I see a graph with no titles on the X and Y now if i take Y as Kbps and the x as the distance from a users cabinet 0 to 3500 meters it looks incorrect 

[burakkucat]

Looking at a page which shows that graph, the immediate preceding text is --

Today Selley confirmed that their original Proof of Concept trial with up to 200 lines in Isfield was able to deliver an average increase in download speeds of 13Mbps to locals, which is a nice boost to have (sorry for the lack of labels on the chart below, but that’s how it was presented and the change is still evident).

(My emboldening.)

[NewtronStar]

Cheers BC a Graph without X & Y titles is useless like a toilet roll holder without the loo roll 

[burakkucat]

The more I look at that graph, the more faults I find . . .

Notice that there is a caption beneath the plot which claims that the blue spots are "Pre LR-VDSL Down Stream (MB/s)" and the yellow spots are "Post LR-VDSL Down Stream (MB/s)".

Putting aside two pedantic observations: (1) Surely it is LR-VDSL2 and (2) The yellow spots are not "Post" anything, they are "With LR-VDSL2", I find the units for the Y-axis, MB/s (i.e. megabytes per second), very difficult to accept. Take the example on the 5 line of the X-axis. The blue spot is under the 10000 line of the Y-axis; let's call it 8000. Now add the stated units, so we now see 8,000 megabytes per second. Huh?    The corresponding yellow spot is midway between the 20000 and 30000 lines of the Y-axis; let's call it 25000. Now add the stated units, so we now see 25,000 megabytes per second. Double huh?   

That graph is pure nonsense, put together by some overpaid advertising entity which has been employed by Beattie Bellman. 

[Weaver]

Thank god - when I looked at that ‘graph’, I thought I was going mad.

[NewtronStar]

Looking at that Graph I would be better off on this LR-VDSL than what I am currently on it should increase my DS sync rate from 39000 kbps to 52000 Kbps

but lets say if vectoring was enabled it could be possible but that's just a dream on a non BDUK cabinet 

[burakkucat]

After further consideration of that "graph", the units for the Y-axis must surely be kbs (i.e. kilobits per second) but the units for the X-axis could be . . . 

As for the cabinet through which N*Star's circuit is connected, at least it is equipped with a Huawei SmartAX MA5603T MSAN and not an ECI Hi-FOCuS M41 Mini-Shelf MSAN.

[WWWombat]

I assumed the y-axis was indeed Kbps, and that the x-axis was "line #". i.e. it shows the results for 33 different lines.

I re-ordered the information along the x-axis, to present it in order ... under the assumption that this was a meaningful possibility.
I added trend lines, but that is more dubious ... we can guess that the re-ordered lines have some relationship with distance, but we don't know it to be linear.

[WWWombat]

Looking at that Graph I would be better off on this LR-VDSL than what I am currently on it should increase my DS sync rate from 39000 kbps to 52000 Kbps

but lets say if vectoring was enabled it could be possible but that's just a dream on a non BDUK cabinet 

The gains in LR-VDSL2* come from 3 sources, making it hard to quantify the outcome:
- Vectoring can give a big gain, but only if the line already suffered from crosstalk. No disturber => no gain.
- Removal of CAL could give a gain of up to 8Mbps for cabs that are ~2km from the exchange, but little gain for those outside the exchange or 5km away. But almost every line on the cab ought to be able to gain the same amount from this.
- The increase in total aggregate power probably helps lines with access to more tones. i.e shorter lines

The lines that failed to improve are interesting. Does that mean they didn't suffer from crosstalk beforehand? Or is it a pointer that the line has a fault, and can't access the gains that are possible?

Vectoring ... That reminds me. Watch out for a different post.

* - Copyright b*cat