cancel
Showing results for 
Search instead for 
Did you mean: 

ADSL - Causes of Sudden Large Changes of SNR???

jamesbee
Hooked
Posts: 5
Registered: 03-12-2013

ADSL - Causes of Sudden Large Changes of SNR???

This is bog-standard BT-supplied ADSL max over a 3 mile rural phone line. It works well - except for sudden large changes (4 - 6dB steps up or step down) of the downstream SNR. The steps in SNR started around a year ago (about the time FTTC equipment began to be installed at the exchange). Steps take place perhaps once a day on average at seemingly random times, but with a bias towards late evening. Sometimes there is a second step reversing the initial step change after around an hour. The attached RouterStats plot shows a typical "step event" that occurred last night.
I have done the usual diagnostics and remedies:
- tried two modems/routers (an old Netgear and a Billion 8800nl),
- checked line for noise,
- filtered mains incoming,
- switched off houshold mains powered stuff,
- connected modem direct to to incoming phone socket
- etc.
It's not a show-stopper as the lower SNR state is good enough for a 2.5 or 3Mb/s profile, but I would very much like to get to the bottom of the problem. Anyone suggestions as to the cause? Have you encountered something similar (and how did you fix it)?
2 REPLIES
ian007jen
Rising Star
Posts: 392
Thanks: 4
Fixes: 2
Registered: 06-09-2007

Re: ADSL - Causes of Sudden Changes of SNR???

Probably crosstalk...the jump from 6 to 12 may be a neighbour switching off their router for the night, you are loosing about 2 Mbps synch speed.
Do the times match up with that?
jamesbee
Hooked
Posts: 5
Registered: 03-12-2013

Re: ADSL - Causes of Sudden Large Changes of SNR???

I don't think the pattern fits a router being switched on and off as it is very irregular for a daily routine - sometimes dips for half an hour, sometimes for a day or more.
Yes, the SNR change is equivalent to a loss of around 2Mb/s. I have no experience of router-to router crosstalk - would it really be that large an effect unless there was some fault condition?