There's no facility on the 8200 to "tweak" the SNR either directly or via the config file.

I didn't get any joy either, using a Planet VC-230N. The Planet VDSL2 modem-router has the same Metanoia MT2301/3301 chipset that powers the Draytek 2750 / 2850.

The web UI options for tweaking TSNRM (and line rates) in CPE mode are greyed out when synced with a Huawei DSLAM.

IIRC, the Kitz contributor was Ixel, who reported that he couldn't directly change the data rates with his Fritz!Box (7390). Only the TSNRM was tweakable, and only tweakable in a downwards direction. [1]

And, worse yet. Just like the "X-Factor brat" who couldn't help pulling funny faces (see Daily Mail [2] !), when the wind blew, things just stuck!

Ixel found himself irrevocably banded on dramatically lower data rates than he initially enjoyed. As StephenTodd has already lamented, it's all part of The Plan!

cheers, a

[1] http://forum.kitz.co.uk/index.php/topic,12398.msg233...
[2] http://www.dailymail.co.uk/tvshowbiz/article-2322403...

edit: wrong way round..

Edited by deleted (Sun 19-May-13 22:18:07)

luckily I havent upset DLM doing that test.

I of course undid the change. But yes whilst the fritz can adjust the snrm the affects of the FTTC DLM system on a tweaked CPE is unpredictable and if the DLM sees it as a fault condition it may change line settings in reaction making things worse.

I think you guys know I dont like the new FTTC DLM in my eyes its worse than the adsl versions, as it seems openreach are taking the position we are in control, we know best and tough luck if you dont like what we give you.

In reply to a post by Chrysalis:

as it seems openreach are taking the position we are in control, we know best and tough luck if you dont like what we give you.

I think that is the way that DLM is going to have to go...

That's because the future direction is far more than just DLM applied to one single line. It is far more than trying to get one line at the maximum speed, and hang everyone else.

No more "I'm alright, Jack". It will includes spectrum management over multiple lines. Reducing power where it can, to make room for other subscribers. Altering frequency usage for some people, so more distant users get a reasonable look-in.

Eventually, vectoring plays its part too - but it looks like one aim there is to reduce power rather than a massive increase in speed.

In such circumstances, there is no option other than to be fully in control. No one user can know better what is needed for all lines.

Very sensible comment. I wonder if you could explain it to some of the users on the BT Infinity forum, though?

In reply to a post by simon194:

There's no facility on the 8200 to "tweak" the SNR either directly or via the config file.

Thanks. I did find the user manual a little later and a conspicuously absent facility as you say.

I agree when it comes to power masking.

But not when it comes to noise margins and interleaving levels.

they both two different bits of management.

Power masking configuration to make lines play nice with each other although its a poor brother to full vectoring.

Interleaving/banding to control fault calls and engineer callouts.

On power and noise (and by "power", I'm not restricting things to just the power masking):
The problem is that, once you have brought power into the equation, it seems that *everything* is related to it. Once you have the aim of reducing power (either to reduce the impact on other lines, or just to be green), every variety of DLM can be determined to have a subtly different effect on power.

For example, when using a fixed 6dB margin without variation. Under this system, you could be forcing those on very good lines to transmit at higher power than necessary. They could get the same speed, at lower power, using a 3dB margin - and have less crosstalk impact on other lines.

A better DLM/DSM process might be changing both the noise margin and the power settings at the same time.

An even better one might be making those decision differently for different parts of the spectrum, for different lengths of lines.

The Broadband Forum's document on various DLM/DSM/DQM techniques gives a good idea, as it tends to describe the interplay between methods, and describes the impact on power.

On FEC/interleaving:
This is a totally different issue, though obviously it reduces throughput for the same notional power setting.

It is interesting to see that SIN 498 requires a modem to support PHY re-transmission, alongside this statement:

Note: Openreach do not currently enable PHY Layer retransmission as defined in
G.998.4 but expect to do so in due course. It is therefore strongly recommended
that the CPE modem should support PHY layer retransmission to minimse errors and
retrains.

Note: Openreach do not currently support Virtual Noise (VN) requirements as defined
in G.993.2. This SIN will be updated if at a later date Openreach decide to offer
this feature.

Perhaps FEC and interleaving are to be consigned to the scrapheap.

The only downside is that interleaving had a quantifiable effect on sync speed, reducing the throughput expectations of the susbcriber. PHYR isn't quite so quantifiable.

the question is have you ever seen snrm or interleaving levels adjusted by DLM as part of managing a batch of lines? probably not, I think those 2 variables only get adjusted to manage a single lines instability.

Edited by Chrysalis (Wed 22-May-13 12:50:11)