cancel
Showing results for 
Search instead for 
Did you mean: 

Target noise margin reduction

maranello
Pro
Posts: 1,267
Thanks: 200
Fixes: 2
Registered: ‎11-01-2008

Target noise margin reduction


Has anyone been fortunate enough to have had an unsolicited target noise margin reduction?
Prior to an electrical supply fault last Sunday, my router had been in sync continuously for over 5 weeks, and d/s noise margin fluctuated between -0.5 and +1.1 against a target of 9dB. Up until the end of June I had been on a 6dB target, but there seemed to be a few random disconnections in the latter half of June, which I'm guessing accounted for the increase.
I was hoping that the BT systems would have recognised the stability of my connection over the past weeks and reduced the target back to 6dB. Can anyone give feedback on how long a period of stable connection has resulted in a reduction?
My other car isn't a Ferrari
16 REPLIES 16
James
Grafter
Posts: 21,036
Thanks: 5
Registered: ‎04-04-2007

Re: Target noise margin reduction

I've seen it happen before.  It is meant to take 28 days though.
maranello
Pro
Posts: 1,267
Thanks: 200
Fixes: 2
Registered: ‎11-01-2008

Re: Target noise margin reduction


Thanks James.
28 days - if only!
Do you happen to know what is the stability criteria to allow a reduction?
My other car isn't a Ferrari
James
Grafter
Posts: 21,036
Thanks: 5
Registered: ‎04-04-2007

Re: Target noise margin reduction

I believe it's multiple drops within a 15 minute timeframe.
What's your normal uptime?
maranello
Pro
Posts: 1,267
Thanks: 200
Fixes: 2
Registered: ‎11-01-2008

Re: Target noise margin reduction


Multiple drops within a 15 minute timeframe will give me a reduction in target SNR?
Not a chance then, as my last period of uptime between disconnections was 930 hours. The drop was caused by electrical power failure. The previous disconnect was manually initiated, as was the one prior to that.  I haven't had a random loss of sync since 20th June.
I suppose I should think myself lucky that the connection is a stable as it has been, following the problems last year, but when I'm only getting 1200-1800kbps anyway due to high attenuation it would be good to be able to squeeze as much as I can out of my connections potential, and not be restricted by BT's DLM.
My other car isn't a Ferrari
James
Grafter
Posts: 21,036
Thanks: 5
Registered: ‎04-04-2007

Re: Target noise margin reduction

Sorry - I meant that that is what causes an increase.
What I'm saying is that if that doesn't happen within the 28 day window, your SNR should get reduced.
maranello
Pro
Posts: 1,267
Thanks: 200
Fixes: 2
Registered: ‎11-01-2008

Re: Target noise margin reduction

Quote from: Jameseh
if that doesn't happen within the 28 day window, your SNR should get reduced.

Which is what prompted my query. Does the DLM simply look at disconnections over the 28 day period, or does it also monitor SNR margin and/or errors.

My other car isn't a Ferrari
James
Grafter
Posts: 21,036
Thanks: 5
Registered: ‎04-04-2007

Re: Target noise margin reduction

I think it's just disconnections.
maranello
Pro
Posts: 1,267
Thanks: 200
Fixes: 2
Registered: ‎11-01-2008

Re: Target noise margin reduction

Which in my case it appears the DLM doesn't work as it should, and I suspect this is the same for many others.
Thanks James for confirming my understanding.
So, back to the original query
Quote from: maranello

Has anyone been fortunate enough to have had an unsolicited target noise margin reduction?.....
Can anyone give feedback on how long a period of stable connection has resulted in a reduction?
My other car isn't a Ferrari
baldrick
Grafter
Posts: 128
Registered: ‎07-08-2007

Re: Target noise margin reduction

I used to switch my router off at night to try and save some energy. Without warning the S/N ratio shot up to 15 dB where it appeared to be permanently stuck. Eventually I gave up on the tree hugging front and left the router on 24/7. It then eventually recovered back to 6dB.
This makes me sceptical about this 15 minute period. I was only switching my router on and off once a day and it screwed up. Alternatively, It also left me wondering whether the router was putting out rubbish on the line as the power supply slowly died during switch off, which the exchange was interpreting as a line problem. This scenario could support the 15 minute theory.
Oldjim
Resting Legend
Posts: 38,460
Thanks: 787
Fixes: 63
Registered: ‎15-06-2007

Re: Target noise margin reduction

One reason I don't switch my router off at night is that the following evening after a morning switch on the error rate jumps alarmingly until the bit swapping sorts itself out.
If that happens every day I can see the automatic systems thinking - high error rate increase the default noise margin.
mapletree
Grafter
Posts: 644
Registered: ‎28-07-2007

Re: Target noise margin reduction

Quote from: Jameseh
I believe it's multiple drops within a 15 minute timeframe.

Does this mean more than one drop within a 15 minute timeframe will reset the 28 day "counter"?
maranello
Pro
Posts: 1,267
Thanks: 200
Fixes: 2
Registered: ‎11-01-2008

Re: Target noise margin reduction


Well I have to admit that a reduction of target SNRM does happen apparently without intervention. After weeks of monitoring SNR and resyncs, I finally got bored with the process as the noise margin wasn't changing by more than +/-0.5dB, and with very few random resync events over the monitoring period. But I did notice over the weekend what appeared to be slightly faster response, and that my PN profile had increased from 1250 to 1500kbps. Checked router stats which indicated that line rate had increased from 1408 to 1856kbps with a d/s noise margin of 5.9dB. It appears that there was a resync at 8.00am on 22/09/09 at which point the target SNR must have been 6dB.
Ok, it's not a spectacular improvement, but for my 63dB line I'll take what I can get.
My other car isn't a Ferrari
James
Grafter
Posts: 21,036
Thanks: 5
Registered: ‎04-04-2007

Re: Target noise margin reduction

@ Mapletree - I believe so.
nozzer
Hero
Posts: 3,298
Thanks: 676
Fixes: 3
Registered: ‎04-08-2009

Re: Target noise margin reduction

Quote from: Jameseh
Sorry - I meant that that is what causes an increase.
What I'm saying is that if that doesn't happen within the 28 day window, your SNR should get reduced.

Strange...    I've had forced resyncs where the system reduces the target SNRM at around 17 day intervals (ie just into the IP Profile "preservation" window). The target seems to have reduced by 1.5dB intervals each time. The next one is due around the 8th Oct and should take me to max sync speed.  Cheesy