[opendtv] Cookie cutter SFN

From: "Manfredi, Albert E" <albert.e.manfredi@xxxxxxxxxx>
To: "opendtv@xxxxxxxxxxxxx" <opendtv@xxxxxxxxxxxxx>
Date: Wed, 30 Mar 2011 20:00:40 -0500

By the way, possibly some might have missed that the Ericsson filing posted by
Mark Aitken:

http://fjallfoss.fcc.gov/ecfs/document/view?id=7021034720

described exactly the sort of sharply defined SFN scheme that we talked about
many times before. One where you need a dense mesh of towers to create the
sharp boundary contours, permitting the reuse of the same SFN frequencies in
adjacent markets.

The valid discussion, and potential for contribution to spectral efficiency,
got lost in all the charts and graphs, most of which (IMO) did not help make
their case at all. The one true potential for spectrum savings could come from
allowing adjacent markets to share the same SFN frequency. At least, in theory.

Here's the quote from the filing that hints at this possibility, if only ever
so briefly:

"MBMS supports MBSFN12 area-specific reference symbols so that an advanced
receiver can estimate the channels to the desired and interfering MBSFN
transmitters. This can be exploited for receiver algorithms cancelling the
interference from other MBSFN areas. ATSC does not provide transmitter- or
DTS-specific reference or training symbols. Cancelling interference from
transmitters not belonging to the DTS is therefore difficult."

But you can't squeeze blood from a turnip. If you do overlap the coverage of
each SFN, you have to give room for the symbols of each SFN, within each time
slot. If you want to use the same frequency for adjacent SFNs without wasting a
lot of spectral efficiency in each one, in practice you will be requiring the
receivers at the edges of the SFN to aim their antennas. This is not so big a
deal, if the spacing between transmit antennas is only 1.25 miles (2 Km). And
they had previously stated that the scheme assumed 30' masts and 10dB gain for
receive antennas anyway.

But as we discussed before, the signal of US TV markets has to overlap
considerably, in crowded corridors. So reusing the same SFN frequency will not
save anything, unless you're going to accept no significant overlap. And
constantly fussing with the cell towers at the edges of the contour, to
accommodate new housing developments and the like.

Anyway, this would be an example of the dense mesh I've postulated before. I'd
say 2 Km spacing is dense.

This Motorola white paper, toward the end, shows a graph of the waste you would
create to make overlapping SFNs (same frequency). Motorola suggests not
supporting this feature.

http://www.motorola.com/web/Business/Solutions/Industry%20Solutions/Service%20Providers/Wireless%20Operators/LTE/_Document/Static%20Files/6834_MotDoc_New.pdf

Bert

----------------------------------------------------------------------
You can UNSUBSCRIBE from the OpenDTV list in two ways:

- Using the UNSUBSCRIBE command in your user configuration settings at
FreeLists.org

- By sending a message to: opendtv-request@xxxxxxxxxxxxx with the word
unsubscribe in the subject line.

[opendtv] Cookie cutter SFN

Other related posts: