OT MF interference and ADSL {Was: Slow radio downloads - a bit off topic}

Dave Liquorice allsorts at howhill.com
Wed Nov 19 13:43:29 PST 2014


On Wed, 19 Nov 2014 16:46:02 +0000, Colin Law wrote:

> I know RF interference at night can cause a resync to a slower speed
> but in my experience it will not automatically resync in the morning
> again, but will stay at the slower speed, so it will not need to slow
> down again the next night.  Several days of continuous good signal is
> needed before it will resync to a higher speed.  This explains why if
> you get some temporary interference on the line which causes it to
> slow down (an extended thunderstorm for example) then it can be
> several days before it comes back up to full speed again.

I think you might be confusing the sync rate and the BRAS rate.

The sync rate is what the modem and DSLAM (in the local exchange) negotiate 
and constantly monitor and adjust up/down in response to line conditions. 
This can be simply the diurnal noise level change, an Openreach person 
fiddling somewhere and a poor connection crackling as it waggles or as you 
say a thunderstorm. I switch the modem off for the half hour or so that a 
thunderstorm takes to pass so the bursts of noise don't force repeated 
resyncs at low speeds. If that happens the DSLAM will notice fairly quickly 
and signal the network to change the the BRAS rate.

The BRAS rate is the maximum rate that your ISP can feed your (virtual) 
line. There is no buffering within the network so if you feed data in faster 
than it can get out you end up with packet loss, retries and reduced through 
put.

If the BRAS gets knocked back and the modem/DSLAM negotiate a higher rate it 
can take the system up to 72 hours to notice the discrepancy and reset the 
BRAS to something just below the sync rate. They have tweaked the algorithm 
since the early days and now if there is a big difference between the sync 
and BRAS rates the system will notice quicker than a small difference. The 
BRAS rate used to be in 500 kbps steps but I think these days it's far less 
granular.

> I would have thought some electrical event that happens morning and
> evening and knocks it out each time is more likely to explain the OPs
> problem.

It could well be, as I said the evening break being half an hour before 
sunset strikes me as a bit early. But if you are listening to MF radio that 
is about the time the sky wave starts produce interference. I don't know how 
big the diurnal noise level change is on a line with say several hundred 
meters over head. Ours (only 50 m ish overhead) has about 6 dB change in 
noise level.

-- 
Cheers
Dave.





More information about the get_iplayer mailing list