In reply to a post by Chrysalis:

I think his solution is banding.

If I remember right when I read one of BT sin documents years ago in the early FTTC days was something in there that isps can request a line to be banded. This would artificially keep the SNRM high providing the capability of the line doesnt significantly regress.

Not in any SIN I've ever read. Confident that can't be done.

ISP's can request banding on a line but have no way to remove it or increase it? That would be silly.

The FTTC DLM is entirely automated. No ISP has any control of any specific settings other than to choose the policy (Speed, Standard or Stable).

Icaras is describing something very out of the ordinary, arranged for him (an OpenReach engineer) by the OpenReach DCoE for a problem line.

No way ISP's can do this kind of thing.
An ISP like AAISP would have been on this immediately if fixed profiles were possible.

To add to my comment to Icaras a few days ago

In reply to a post by j0hn83:

There is a 4th DLM profile often overlooked or unknown to many.
This is named "Custom" and AIUI can be set to a fixed setting like you suggest.
Documentation for ISP's tells them the Custom DLM profile should only be used on test lines.

This was introduced as a possibility in August 2018

https://www.openreach.co.uk/orpg/home/updates/briefi...

Where a CP utilises dedicated test lines the use of DLM may be inappropriate. We are therefore allowing CPs to request that DLM is deactivated and a fixed profile is placed on test lines.

I can't see why they would brief ISP's that they can now fix profiles on test lines if it was always possible on customer lines.

I'd be amazed if Icaras's example wasn't using the method described above which is supposed to be limited to test lines.

In reply to a post by j0hn83:

Still never seen DLM jump from 6dB to 3dB though, always in 1dB increments at least 24 hours apart.

We moved to IDNet on 11/09/18 and went to the no G.INP 6dB regime and a sync speed of 62459. On 13/09/18 G.INP was enabled and our sync went up to 72233. We had an uplift to on 20/09/18 to 75314 but my records don't show the SNR. The router was then powered down from 22/09/18 to 29/09/18 when it reconnected at 74917 - again no record of the SNR. We then saw a silent re-connect on 10/10/18 when the sync speed jumped to 79230 and my notes show a new SNR of ~3dB.

That looks a little different to a 6 > 5 > 4 > 3 progression but I haven't got any evidence to prove the case.

There is things that isp's miss tho. Also asking for manual banding on lines complicates things, its cheaper for an isp to not have that as part of their policy.

But with all that said I dont know if it still applies now, the sin document I read was in the first 1-2 years of FTTC rollout, and since those days a lot of things have changed.

In reply to a post by j0hn83:

In reply to a post by Chrysalis:

I think his solution is banding.

If I remember right when I read one of BT sin documents years ago in the early FTTC days was something in there that isps can request a line to be banded. This would artificially keep the SNRM high providing the capability of the line doesnt significantly regress.

Not in any SIN I've ever read. Confident that can't be done.

ISP's can request banding on a line but have no way to remove it or increase it? That would be silly.

The FTTC DLM is entirely automated. No ISP has any control of any specific settings other than to choose the policy (Speed, Standard or Stable).

Icaras is describing something very out of the ordinary, arranged for him (an OpenReach engineer) by the OpenReach DCoE for a problem line.

I was the Openreach engineer! I wasn�t the customer.

Isn't that what I said?

Icaras is describing something very out of the ordinary, arranged for him (an OpenReach engineer) by the OpenReach DCoE for a problem line.

Apologies! I read your post as �by an Openreach engineer�.

In reply to a post by Icaras:

They didn�t introduce a 3dB default, but they introduced the possibility of a 3dB margin about 3 years ago in some cases. Why do you think they didn�t? I remember reading about it at the time.

I stand corrected. Time flies.