[opendtv] Re: The "real" problem with OFDM in the U.S.
- From: "Manfredi, Albert E" <albert.e.manfredi@xxxxxxxxxx>
- To: "OpenDTV (E-mail)" <opendtv@xxxxxxxxxxxxx>
- Date: Thu, 17 Mar 2005 13:48:13 -0500
Craig Birkmaier wrote:
> Yes overlaps are a necessity; this is especially
> true for SFNs where there are multiple synchronized
> transmitters. But overlaps between markets SHOULD
> be avoided or at least minimized.
I'll accept that, if it's the FCC mandating such a
policy. That makes sense. It doesn't make sense to
expect individual OTA broadcasters to want this of
their *own* coverage pattern, but it makes good
sense for OTA broadcasters to demand that the FCC
impose such policy to protect their markets from
flagrant intrusion. Whew. One point of possible
agreement.
Now comes the logical disconnect you keep falling
into:
> > Thank you. Now, if you were to deploy small sticks
> > in an SFN for Baltimore TV stations [which requires
> > ~60 mile radius of coverage, at least in all
> > directions except to the SW], to achieve your
> > desired super-sharp contours, just exactly how many
> > towers do you expect you would need to cover a 50-60
> > mile radius with the kind of sharp coverage contours
> > you want?
>
> Probably four synchronized mains around the core
> Baltimore market, and a handful of translator/gap
> fillers in the population pockets that are at the edges
> of the coverage area.
The four taller sticks need to get to most of the area,
and would create a vague pattern. The edges would have to
be created by an enormous number of small sticks. Because
if you look at a map, or if you actually go to these
places, there are no parts with zero population. The
"pockets" of population are small clusters of homes and
individual farms. No "pockets" in reality. So in order to
minize overlap, you'd have to create a sharp contour
while providing uniform coverage inside the market.
The circumference of a circle with radius 60 miles is 377
miles. So a very marginal design could use low enough
power from your 4 main towers, to keep that signal from
straying too far, then a series of perhaps 30+ smaller
towers, spaced 10 miles apart. And to provide ubiquitous
coverage, these 34+ towers would all have to be
synchronized, *even if* they were COFDM. The reason
being, the bulk of the area is covered by those 4 main
towers which must be more than just one GI-equivalent
round trip delay time apart from one another *and* from
the smaller edge sticks. But their signal will be strong
enough where the edge sticks are to create interference,
unless synchronized.
(The idea of creating the sharp pattern with on channel
*repeaters* will not work well. These OCRs work well to
fill signal strength gaps in a larger pattern, and where
their signal drops off and is overwhelmed by the main
signal again, before creating interference. To create
tight patterns and ubiquitous coverage, the small sticks
at the edge will be operating in low signal zones. They
will have to be synchronized such that in contours of
equal power density between two sticks and between each
stick and the main SFN, symbols will be in phase.)
This is what *all* existing DTT installations want to
avoid. Huge, synchronized SFNs.
The French don't look for sharp patterns, but they do
want to create uniform coverage with low power. So they
use possible tight SFNs in the major markets, then low
power translators beyond the low power SFN coverage. So
in terms of using frequencies per given area, you need
more with this approach, not less.
The Germans use just the main SFN and its vague pattern.
The US, Australian, and UK DTT systems use single big
sticks, with varying amounts of power. But vague
coverage patterns result. These schemes are very
similar to our own, in terms of efficient use of
spectrum and in terms of (not) creating sharp contours
of coverage. The difference, if any, is coverage area
per big stick. Lower power means less area per stick and
more sticks to cover the same area.
The common denominator is that no one is creating SFNs
with dozens of synchronized sticks, as you would require
to fulfill your coverage ideas. That's why I hear your
words but see no basis for them.
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.
Other related posts:
