[opendtv] Re: Seeing Ghosts on a Single Frequency Network

  • From: "Allen Le Roy Limberg" <allimberg@xxxxxxxxxxxx>
  • To: <opendtv@xxxxxxxxxxxxx>
  • Date: Wed, 9 Feb 2011 13:44:28 -0500

I believe the intended question is:

What is the advantage of using two 6 MHz wide channels in a checkerboard MFN
(multiple-frequency network) vs. just using a 12 MHz wide channel and using
longer guard intervals,
slower symbol rates, and more error correction in an SFN?

The redundant coding used in COFDM uses up frequency spectrum in order to
avoid the need for the receiver to have good adaptive channel-equalization
filtering to counter the effects of multipath reception.  Indeed, where
strong echoes (ghosts) cause substantially complete cancellation of energy
in portions of the RF channel as received, well-designed COFDM should be
able to recover information that would be very difficult or impossible to
recover if the transmitters employed single-carrier modulation.

When two or more transmitters in an SFN transmit similar signals essentially
concurrently from different locations, there will be zones of overlapping
coverage in which there are strong echoes (ghosts) that cause substantially
complete cancellation of energy in portions of the RF channel as received.
Even at the edges of these zones where one signal is much stronger than the
other, the co-channel interference is apt to reduce carrier-to-noise ratio
(CNR) and to "color" the noise spectrum..  COFDM can cope with a severe
co-channel interference problem much better than single-carrier modulation.
However, as noted above, COFDM uses up frequency spectrum in order to be
able to overcome substantially complete cancellation of energy in portions
of the RF channel as received.

A strength of COFDM is that the same measures taken to combat problems with
multipath reception help overcome the problems of co-channel interference
associated with SFNs.  However, this strength is not a valid rationale for
using SFNs.  Note that the cancellation of received signal owing to
co-channel interference in zones of overlapping coverage in an SFN harms a
COFDM transmission, as well as a transmission using
single-carrier-modulation  (e.g., 8-VSB).
When two or more transmitters in an MFN transmit similar information
essentially concurrently from different locations, the receiver receives
their respective signals independently without co-channel interference from
each other.  These independent RF signals can be down-converted to baseband
separately to "de-rotate" each of them.  The resulting independent baseband
signals are coherent in phase, having the same zero-frequency carrier, so
they can be combined constructively rather than with random phasing that is
as apt to result in them being combined more or less destructively as being
combined more or less constructively.  The combining of the independent
baseband signals may be done before data-slicing, or code combining can be
employed after data-slicing.  Channel-equalization filtering may be done
individually on the independent baseband signals or may be done on the
combined baseband signals before data-slicing.  The essential point is that
the independent RF signals can be processed without suffering the losses
associated cancellation of received signal owing to co-channel interference.

SFNs simply do not conserve frequency spectrum; they waste it because
portions of the spectrum are less useful owing to co-channel interference in
zones of overlapping coverage.  Well-designed COFDM may be able to recover
the data lost in zones of overlapping coverage in an SFN owing to
cancellation of signal energy in portions of the frequency spectrum.
However, in zones of overlapping coverage the co-channel interference will
reduce SNR for COFDM signals, as well as for single-carrier-modulation
signals.

MFNs avoid wastage of the frequency spectrum caused by co-channel
interference at the receivers in zones of overlapping coverage.  So, assured
audience can be increased in the zones of overlapping coverage.  It seems to
me that pushing SFNs, rather than frequency diversity, is harmful to
broadcasters' interests and to consumers' interests.  It certainly seems to
encourage the FCCto grab spectrum away from broadcasters without giving much
thought to the process.

The following argument can be made against MFNs.  If MFNs are used, a
receiver must then have the capability to down-convert two signals, so two
front-end tuners are required.  MFNs can be operated using
iterative-diversity, however, as well as frequency-diversity.  One set of
transmitters transmits Program X in Slot A of a prescribed time interval and
transmits Program Y in Slot B of the prescribed time interval.  The other
set of transmitters transmits Program Y in Slot A of a prescribed time
interval and transmits Program X in Slot B of the prescribed time interval.
A single frequency-agile front-end tuner can receive either Program X or
Program Y on a time-division-multiplex basis.  The earlier received version
of the selected program will have to be delayed for combination with the
later received version of the selected program, of course.  So, the receiver
better to utilize transmissions from an MFN will still be more complicated
than the receiver to utilize transmissions from an SFN.  (Note that insofar
as transmitting is concerned, this sort of operation can be very close to
that posited by John Shutt.)

The iterative-diversity can help overcoming short burst noise caused by
electrical interference generated nearby the receiver.  This can help
justify the expense of the memory used to delay the earlier received version
of the selected program for combination with the later received version of
the selected program.

Also, a receiver with two front-end tuners may become commonplace in any
case.  Such a receiver is better for channel surfing or for switching back
and forth between RF channels.

There are likely to be IF shielding problems in a receiver with two
front-end tuners that use the same intermediate frequency during down
conversion procedures.  Problems with IF amplifier interaction are avoided
in a receiver with a single frequency-agile front-end tuner, but
automatic-gain-control (AGC) design is much trickier.

Proposals for wider than 6 MHz UHF channels may not be received kindly by
receiver designers, especially those specializing in front-end tuner design.
In the days that LC filtering was used for selectivity, obtaining a
reasonably flat response over 6 MHz while maintaining skirt selectivity
against adjacent channels was a challenge.  Acoustic delay line filtering is
used in modern IF amplifier designs, I believe.  First conversion to an
above-band IF may help in getting reasonably flat very-wide-bandwidth
response with good skirt selectivity.  Very-wide-bandwidth receivers are apt
to have more problems excluding noise and interference.  From the viewpoint
of considering a system overall, I am concerned whether some of the
super-wideband "utility" proposals being floated are supported solidly by
currently known receiver design techniques.

Al

----- Original Message ----- 
From: "John Shutt" <shuttj@xxxxxxxxx>
To: <opendtv@xxxxxxxxxxxxx>
Sent: Tuesday, February 08, 2011 6:41 AM
Subject: [opendtv] Re: Seeing Ghosts on a Single Frequency Network


> What is the advantage of using two 6 MHz wide channels in a checkerboard
SFN
> vs. just using a 12 MHz wide channel and using longer guard intervals,
> slower symbol rates, and more error correction in an SFN?
>
> John
>
> ----- Original Message ----- 
> From: "Manfredi, Albert E" <albert.e.manfredi@xxxxxxxxxx>
>
> > Qualcomm needed to provide coverage over a very large area with just Ch
> > 55. They had no choice but to use an SFN. But look at the complexity of
> > their network, and look at the lack of coverage area, compared to what
> > they could have done with a checkerboard of frequencies.
> >
>
>
>
>
> ----------------------------------------------------------------------
> 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.
>

 
 
----------------------------------------------------------------------
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:

  1. Home
  2. opendtv
  3. Archive
  4. 02-2011