Changes in router stats - what happened?

This questions probably shows my basic lack of understanding of ADSL...

On Monday my router stats were:
Attainable line rate : 9752 kbit/sec
Attainable Atm rate : 8672 kbit/sec
Used line rate : 8600 kbit/sec
Fast used Atm rate : 0 kbit/sec
Interleaved used Atm rate : 7616 kbits/sec
Rel. capacity occupation : 88
Noise Margin : 10 dB
Line attenuation : 22 dB
Output Power : 19 dBm

currently they are:
Attainable line rate : 8152 kbit/sec
Attainable Atm rate : 7168 kbit/sec
Used line rate : 7476 kbit/sec
Fast used Atm rate : 0 kbit/sec
Interleaved used Atm rate : 6560 kbits/sec
Rel. capacity occupation : 92
Noise Margin : 13 dB
Line attenuation : 22 dB
Output Power : 19 dBm

I thought that an increased Noise Margin was a good thing, so why does
it (apparently) go with a decreased Attainable line rate?

Could it be related to a power cut that we had here in between taking
the above two stats?

Thanks,

Kit

"Kit" <(E-Mail Removed)> wrote in message
news:310320070923598696%(E-Mail Removed)...
>
> This questions probably shows my basic lack of understanding of ADSL...
>
> On Monday my router stats were:
> Attainable line rate : 9752 kbit/sec
> Attainable Atm rate : 8672 kbit/sec
> Used line rate : 8600 kbit/sec
> Fast used Atm rate : 0 kbit/sec
> Interleaved used Atm rate : 7616 kbits/sec
> Rel. capacity occupation : 88
> Noise Margin : 10 dB
> Line attenuation : 22 dB
> Output Power : 19 dBm
>
> currently they are:
> Attainable line rate : 8152 kbit/sec
> Attainable Atm rate : 7168 kbit/sec
> Used line rate : 7476 kbit/sec
> Fast used Atm rate : 0 kbit/sec
> Interleaved used Atm rate : 6560 kbits/sec
> Rel. capacity occupation : 92
> Noise Margin : 13 dB
> Line attenuation : 22 dB
> Output Power : 19 dBm
>
> I thought that an increased Noise Margin was a good thing, so why does
> it (apparently) go with a decreased Attainable line rate?
>
> Could it be related to a power cut that we had here in between taking
> the above two stats?
>
> Thanks,
>
> Kit
..
The answer will be here somewhere.. enjoy.
..
http://freespace.virgin.net/adsl.test/doc2.pdf

In article <(E-Mail Removed)>, Alan
<(E-Mail Removed)> wrote:


> > I thought that an increased Noise Margin was a good thing, so why does
> > it (apparently) go with a decreased Attainable line rate?

> The answer will be here somewhere.. enjoy.
> .
> http://freespace.virgin.net/adsl.test/doc2.pdf

Thanks for that.
It is very useful in aiding my basic understanding, but having searched
through for everything that mentions noise it still seems to me that
increased Noise Margin should be a 'good thing'. So I'm still not sure
why I get a decreased Attainable line rate with an increased Noise
Margin.

Kit

In message <310320070923598696%(E-Mail Removed)>, Kit
<(E-Mail Removed)> writes
>
>This questions probably shows my basic lack of understanding of ADSL...
>
>On Monday my router stats were:
>Attainable line rate : 9752 kbit/sec
>Attainable Atm rate : 8672 kbit/sec
>Used line rate : 8600 kbit/sec
>Fast used Atm rate : 0 kbit/sec
>Interleaved used Atm rate : 7616 kbits/sec
>Rel. capacity occupation : 88
>Noise Margin : 10 dB
>Line attenuation : 22 dB
>Output Power : 19 dBm
>
>currently they are:
>Attainable line rate : 8152 kbit/sec
>Attainable Atm rate : 7168 kbit/sec
>Used line rate : 7476 kbit/sec
>Fast used Atm rate : 0 kbit/sec
>Interleaved used Atm rate : 6560 kbits/sec
>Rel. capacity occupation : 92
>Noise Margin : 13 dB
>Line attenuation : 22 dB
>Output Power : 19 dBm
>
>I thought that an increased Noise Margin was a good thing, so why does
>it (apparently) go with a decreased Attainable line rate?
>
>Could it be related to a power cut that we had here in between taking
>the above two stats?
>
>Thanks,
>
>Kit
Yes its related to the power cut. When the power came back, the router
restarted and negotiated a speed setting with the exchange.

A lower speed will always produce a better noise margin (for a given set
of line conditions) so ADSLMax adjusts the speed upwards until the noise
margin comes down to a preset value (normally about 6db). Noise margins
vary depending on external conditions, interference etc. If the noise
margin drops too low, the modem disconnects and then reconnects at a
lower rate. The idea is that the system attempts to find a stable speed
that still works when the noise conditions on your particular line are
at their worst.

Try switching the router of for a minute or so, then switching it back
on. It may reconnect at the slightly higher speed (with a lower noise
margin) - OTOH it might go lower.

Mine usually reconnects after a reboot at about 6800, but will then
drift down over the next few days (with losses of sync occasionally
logged) to about 6000. That's the Max stable speed.
--
Peter R Cook

Kit wrote:

> In article <(E-Mail Removed)>, Alan
> <(E-Mail Removed)> wrote:
>
>
>> > I thought that an increased Noise Margin was a good thing, so why does
>> > it (apparently) go with a decreased Attainable line rate?
>
>> The answer will be here somewhere.. enjoy.
>> .
>> http://freespace.virgin.net/adsl.test/doc2.pdf
>
> Thanks for that.
> It is very useful in aiding my basic understanding, but having searched
> through for everything that mentions noise it still seems to me that
> increased Noise Margin should be a 'good thing'. So I'm still not sure
> why I get a decreased Attainable line rate with an increased Noise
> Margin.

Because lower speeds can work with lower absolute signal levels, so the
margin [difference between actual noise and worst-case noise for your
speed] is higher.

--
<http://ale.cx/> (AIM:troffasky) ((E-Mail Removed))
11:31:02 up 1 day, 3:17, 2 users, load average: 0.59, 0.63, 0.60
Yes. I'm just guessing.

In article <knI$(E-Mail Removed)>, Invalid
<(E-Mail Removed)> wrote:

> A lower speed will always produce a better noise margin (for a given set
> of line conditions) so ADSLMax adjusts the speed upwards until the noise
> margin comes down to a preset value (normally about 6db). Noise margins
> vary depending on external conditions, interference etc. If the noise
> margin drops too low, the modem disconnects and then reconnects at a
> lower rate. The idea is that the system attempts to find a stable speed
> that still works when the noise conditions on your particular line are
> at their worst.

Ah!
Now I understand.
Thanks for that very clear explanation...

Kit