Really got to go now and dont have time to search, but if someone wants to look further into Active RFI cancellation because I believe there may even be 'RFI profiles' that can be loaded
Here http://books.google.co.uk/books?id=m77kZl71gysC&pg=PA400&lpg=PA400&dq=active+rfi+cancellation+adsl&source=bl&ots=wx6uTaqkS4&sig=GebI7HlJ00d1o43-ED9H5PaosyY&hl=en&sa=X&ei=TdygUZS-EqWo0wW89oHwAw&ved=0CDYQ6AEwAQ#v=onepage&q=active%20rfi%20cancellation%20adsl&f=false
** Health Warning ** Not exactly bed-time reading!
Yea gawds... ur not kidding
I just started to have a look, but I saw all the equations and a lot of unavailable pages... and since it is my bedtime reading on the ipad.. it didnt exactly inspire me to read it Im afraid.
Instead i'm going to summarise something that I can recall from several years ago when I first went on Be* and no-one at the time really knew much about the missing gap in tones, which is when I researched it.
Because it was so long ago, I could be wrong, but I will expand a little on my above post to a basic summary of how I understood it to work. Im using my own words as Im not an electronics person so Im happy to be corrected with correct phrases if any of what Im saying isnt quite correct.
I seem to recall the theory behind active RFI cancellation was that when RFI was detected at a particular tone, that frequency was not used. By cancelling that sub-channel and taking it out of use, the effects of the noise was less likely to bleed over to the neighbouring frequencies. Taking the tone out of use results in some loss of bit loading, but should makes up for it with a cleaner neighbouring tones.
The downside of constantly monitoring for active RFI cancellation is that it can be put pressure on the modem processing power and cause delay.
So enter a kind of hybrid -
if RFI is found to be affecting certain tone(s)*,
then apply
one of a set of pre-defined masks. These masks are up to the dslam owner to define... ie BTw or the LLU provider... so they can cancel 'sets' of known problematic frequencies.
*I'm guessing that this would have to be done during the sync process, but it means once the mask is applied then the modem doesnt have to work as hard by actively monitoring each tone.
Im not saying this is what is happening... but it
could explain what some users are seeing. It wouldnt be too hard for BTw to get their mits on known profiles of problematic frequencies.
Over to someone else to do the technical jargon... I need sleep.. g'nite.