VDSL2: The cost of PSD masks for legacy ADSL

I recently found an Australian document from their equivalent of NICC. It describes the (technical & non-technical) things needed in order to roll out FTTN (their name for our FTTC) instead of the current FTTH-based plan for NBN Co: WC58 INDUSTRY PAPER ON FTTN AND VDSL2 REGULATION (see attachment 1)

What is interesting about their deployment is that they very much have in mind that the NBN rollout is a full replacement for legacy broadband - so they discuss the transition from today (including exchange-based ADSL and SLU) to the future (barring SLU, and allowing for FTTN cabinets mingled with basement FTTN in MDU).

For the transition, they note that VDSL2 will need PSD masks to prevent interference with exchange-based ADSL (and will need to avoid vectoring over those frequencies too). Very much the same as we do here.

The cost of applying those PSD masks, and the alien crosstalk on those frequencies, appears to be an "up to" 15Mbps drop in speeds.

Because the problem is at the lowest frequencies, I'd say that is a drop in speed for *every* FTTN line. Today's Ofcom report indicates rural FTTC averages 29Mbps; an extra 15 would be a significant jump.

If we want BDUK to reach a higher percentage of people with superfast speeds, wouldn't an easy answer be to phase out exchange-based ADSL, including LLU?Especially in rural areas?

Shouldn't we at least have the discussion going on somewhere?

already known, I posted in many previous threads that area's with long E sides have huge penalties on vdsl sync speed for the power cutback. Meaning distance to exchange does actually still affect vdsl sync speeds.

I estimated my own line ironically at around 15-18mbit loss.

I was told by my install engineer his opinion of my starting 110 attainable was when I first got installed the psd masks were NOT applied and dozens of adsl lines were knocked out when I got activated, so according to him my first drop of sync speed to around 90 was when the PSD masking was added and not a new installed however there was also a new install that day so not conclusive. This engineer has been great to me he even still rings me to ask how line is going.

Last night my PSD mask appeared to change, the characteristics of the line changed, my upstream attainable went down, modem reports much higher downstream signal power and my downstream attainable went up by 10mbit and I now am about 4mbit of a max sync speed.

Also to point out long lines are not a rural exclusive thing, there is short lines in rural areas as is long lines in urban and suburban areas, so why limit phasing out of adsl to rural areas only?

Edited by Chrysalis (Tue 15-Apr-14 15:55:52)

In reply to a post by Chrysalis:

already known, I posted in many previous threads that area's with long E sides have huge penalties on vdsl sync speed for the power cutback. Meaning distance to exchange does actually still affect vdsl sync speeds.

I estimated my own line ironically at around 15-18mbit loss.

I'd known about the impact of PSD masks (and hence of E-side length) for a long time, but not really the quantity of the impact.

I hadn't seen you mention your initial installation experience, but it makes some kind of sense. Interesting to see the same kind of value on your line too.

I recently found updates to SIN 498 that added tests for VDSL2 modems. From the way they use the line profiles (as DLM would do), it suggested that long E-sides have less impact to speeds than short ones.

Last night my PSD mask appeared to change, the characteristics of the line changed, my upstream attainable went down, modem reports much higher downstream signal power and my downstream attainable went up by 10mbit and I now am about 4mbit of a max sync speed.

Interesting. I guess you'd see quite a lot from the graphs from bald_eagle's logging scripts - at least the SNR and bit-loading graphs.

sadly no as his current version is broken on my system, I grabbed a graph using dslstats instead but ideally want bald eagle graphs so I got better comparison.

hopefully he can fix today.

I've often wondered why they didn't allow the cabinets to do adsl as well, and then everybody's broadband whether ADSL or VDSL comes from the cabinet's. Obviously there's money involved, people may not upgrade to VDSL if their broadband was better, and also cabs would need to be bigger, but going forward it makes sense. They could of even just replaced the PCP, doing away with the link cable distance, space allowing for installation of bigger cabs of course.

that certianly would be better but one reason I guess is cost and it arguable also devalues the premium feel of vdsl.

eg. in my area 70mbit to 6mbit is a huge gulf.

move the adsl to the cabinet tho and it starts syncing at 16mbit+ instead of 6mbit and vdsl will be harder to justify by the peeps connected.

an idea is they seperate adsl connections into seperate pair bundles.

Edited by Chrysalis (Tue 15-Apr-14 20:56:01)

I wonder if there will ever be a town with close to full migration from adsl to vdsl?,
Is it possible the fttc concept would struggle fulfilling the capacity required?
Any idea on numbers of cabinets required to fully supply all premises in an example town with vdsl?
Obviously it would knacker the current LLU configuration also?