U.S. Television Polarization

On Wed, 4 Dec 2019 20:02:05 -0800 (PST), Kevin Neilson
<kevin.neilson@xilinx.com> wrote:

>On Wednesday, December 4, 2019 at 2:27:03 PM UTC-7, Steve Pope wrote:
>> Kevin Neilson  <kevin.neilson@xilinx.com> wrote:
>>=20
>> >On Tuesday, December 3, 2019 at 7:28:50 PM UTC-7, Randy Yates wrote:
>>=20
>> >> > http://sbe.org/handbook/fundamentals/RF/RF-TV_Antennas.pdf
>> >> >
>> >> > According to this (mostly section 5.2.2) NTSC was originaly horizont=
>ally
>> >> > polarized at the transmitter (for about thirty years until the 70's)=
>,
>> >> > thus explaining all the horizontal TV receiver antennas.  FCC
>> >> > then allowed an equal amount of power to be radiated with vertical
>> >> > polarization, such that stations could double their power and
>> >> > send a circularly polarized signal.  Usually right-handed.  And the
>> >> > horizontally polarized receivers did not lose signal.
>>=20
>> >I would think circular polarization wouldn't take more power:  the
>> >transmitter is still only using one polarization; it's just that the
>> >polarization angle increases continuously.  But then a
>> >horizontally-polarized antenna would only be able to receive half the
>> >power, so maybe the transmit power has to be doubled so that the
>> >horizontally-polarized rx antenna gets enough power?
>>=20
>> I wasn't clear in the above -- it's that the regulator (FCC) permitted
>> a doubling of power if the transmission was converted from horizontal
>> to circular polarization.  =20
>>=20
>> I would guess that in at least the ~10 or so major markets TV transmitter=
>s=20
>> were at or near their maximum regulatory power limit, and that the=20
>> shift to circular (with power doubling) did in fact increase the
>> reach to some customers.
>>=20
>> Steve
>
>Circular polarization sounds better--the reflections can still be filtered =
>out (since the reflection changes the handedness) and the horizontal orient=
>ation of the rx antenna doesn't matter.  But it would probably be worse if =
>you are transmitting circularly-polarized signals to a horizontally-polariz=
>ed antenna, because the horizontally-polarized antenna won't be able to fil=
>ter out the reflections.
>
>This all makes me think of OAM:  Orbital Angular Momentum modulation.  It w=
>as a fad for a bit, but I haven't heard much of it for a while.

A good equalizer will integrate multipath, though, and realize the
"multipath gain" from the power in the reflected signals.  If a
matched circular-polarization antenna is being used, the potential 3dB
increase in rx power might be more than the multipath gain, but it
would probably depend on the environment.

Using high-gain antennas in general is sometimes a mixed bag for this
reason, too, the rx gain goes up by the increased gain of the antenna,
but potentially at a loss of the multipath gain from the paths outside
of the high-gain antenna beam.   If you have no equalizer or a limited
equalizer (e.g., restricted delay spread), then the high-gain antenna
may be good net gain.

Another thing to keep in mind is that in multipath even linearly
polarized signals get rotated at potentially random angles, so even if
the tx is horizontal, the rx will be getting whatever LOS horizontal
energy plus whatever randomly-rotated relfected energy.   Many TV
receivers are non-LOS, (e.g., indoor), so a rabbit ear antenna will
often wind up with all kinds of wacky angles for best reception of a
particular transmitter.

On Wed, 4 Dec 2019 12:00:35 -0800 (PST), "Richard (Rick) Lyons"
<r.lyons@ieee.org> wrote:

>Hi Randy.
>I called my local television station and spoke to an "Engineer." He said their transmitted television signal is "horizontally polarized with a 17-degree tilt."
> 


Just redefine the axes and all is good.   ;)

On Tue, 03 Dec 2019 13:08:54 -0500, Randy Yates
<randyy@garnerundergroundinc.com> wrote:

>Hey Folks,
>
>I did some googling on this and couldn't get a satisfying answer. 
>
>1. What polarization is used to transmit?
>
>2. What polarization is used to receive? 
>
>The answer to question 2 would seem to be horizontal since that is the
>orientation of the antennas I've seen. But most broadcast TV antennas
>are (I believe) vertical.
>
>If I've got it right, then why do the tx/rx polarizations differ? I
>would think this arrangement would lose signal.
>-- 
>Randy Yates
>Embedded Linux Developer
>111 West Main Street, Suite 201, Garner, NC 27529
>http://www.garnerundergroundinc.com

This, from Feb 2018, sez most Tx are still horizontal.   Mobile or
multipath receivers need to be able to receive whatever rotated
reflections they get in order to integrate the reflected power in the
equalizer, but an outdoor LOS antenna should do fine with horizontal
polarization.

https://www.tvtechnology.com/opinions/circular-elliptical-polarization-for-tv

On 12/06/2019 02:34 PM, Eric Jacobsen wrote:
> Many TV
> receivers are non-LOS, (e.g., indoor), so a rabbit ear antenna will
> often wind up with all kinds of wacky angles for best reception of a
> particular transmitter.

Thank you. I've been wondering why I have to do strange things. Wish an 
root top or even in attic placement was feasible.

Eric Jacobsen <theman@ericjacobsen.org> wrote:

>A good equalizer will integrate multipath, though, and realize the
>"multipath gain" from the power in the reflected signals.

This reminds me of a question I've had percolating for a while now.

What is a good definition of multipath gain?

I've tended to define the multipath gain (or loss, if it is negative)
as the improvement in sensitivity, when using an optimal or agreed-upon
good equalizer, for the multipath situation relative to a non-multipath
situation, when both these two situations have the same total power
presented at the receiver's antenna or antennas.

(Side note: sensitivity should be defined with respect to an 
agreed-upon performance metric.)

The pushback I've received against this definition is the "same total
power" qualifier.  With this qualifier, for some systems, there is 
often an multipath impairment / loss;  and one can measure how well
a given equalizer algorithm is dealing with the impairment.

Without that qualifier in the definition, there can be multipath gain 
nearly always, leading to general statements that multipath improves 
performance.  

Maybe either definition is good, but they can lead to very different
summary statements about how well a system is doing.

Steve

On Sat, 7 Dec 2019 22:16:40 +0000 (UTC), spope384@gmail.com (Steve
Pope) wrote:

>Eric Jacobsen <theman@ericjacobsen.org> wrote:
>
>>A good equalizer will integrate multipath, though, and realize the
>>"multipath gain" from the power in the reflected signals.
>
>This reminds me of a question I've had percolating for a while now.
>
>What is a good definition of multipath gain?
>
>I've tended to define the multipath gain (or loss, if it is negative)
>as the improvement in sensitivity, when using an optimal or agreed-upon
>good equalizer, for the multipath situation relative to a non-multipath
>situation, when both these two situations have the same total power
>presented at the receiver's antenna or antennas.
>
>(Side note: sensitivity should be defined with respect to an 
>agreed-upon performance metric.)
>
>The pushback I've received against this definition is the "same total
>power" qualifier.  With this qualifier, for some systems, there is 
>often an multipath impairment / loss;  and one can measure how well
>a given equalizer algorithm is dealing with the impairment.
>
>Without that qualifier in the definition, there can be multipath gain 
>nearly always, leading to general statements that multipath improves 
>performance.  
>
>Maybe either definition is good, but they can lead to very different
>summary statements about how well a system is doing.
>
>Steve

Yeah, like SNR, there's no universal or "best" definition. 

FWIW, I just think of it as the gain realized from integrating the
multipath energy compared to not doing so.

In that sense, one definition for an LOS signal might be the ratio of
the reflected energy to the direct ray energy, which would be more
"potential" multipath gain than actual, depending on whether an
equalizer could capitalize on it or not.

Another definition might be the realized gain from turning an
equalizer on compared to not turning it on and just synchronizing on
the strongest path.

Both of those definitions should provide "gain" from multipath, or at
least when the eq doesn't hose you for whatever reason.   So a
high-gain antenna reduces multipath gain by reducing the amount of
multipath energy from the directions in which it has reduced
sensitivity.  It may or may not be made up for by increasing the gain
in the direction in which it is pointed, but usually in multipath
environments the antenna gain can't be relied on completely since some
multipath energy will be lost (assuming the system could have
exploited it).

It's also something that isn't always quantifiable, but in some cases,
with some specific channel models that include the appropriate
directions of the rays it might be.   A specific delay spread profile
might also be split into the direct and reflected portions to compute
a ratio, but it would only be good for that specific delay spread.

So, yeah, there's not gonna be complete agreement on how to quantify
it, but many will agree that it exsits and what sorts of things make
it better or worse.

On Friday, December 6, 2019 at 12:34:38 PM UTC-8, Eric Jacobsen wrote:

(big snip)

> Another thing to keep in mind is that in multipath even linearly
> polarized signals get rotated at potentially random angles, so even if
> the tx is horizontal, the rx will be getting whatever LOS horizontal
> energy plus whatever randomly-rotated relfected energy.   Many TV
> receivers are non-LOS, (e.g., indoor), so a rabbit ear antenna will
> often wind up with all kinds of wacky angles for best reception of a
> particular transmitter.

Yes.

I suppose there is a rare case of someone a long distance from a
transmitter over ground so flat that it is LOS with no reflections.
(Though there could be reflections off neighbor's antennas.)

It happens that I can see some transmitting antennas out my living
room window, and have attic antennas pointing in that direction.

But there are other transmitters around, and some still come though.

An antenna rotator is convenient, in that one can rotate for maximum
signal, independent of actual transmitter direction.