Joerg said:
A rather huge log-periodic on a mast.
Bingo. Are you *SURE* it's really a log periodic antenna or some
kludge that only looks like one? Duz it look like one of these?
<
http://en.wikipedia.org/wiki/Log-periodic_antenna>
I modeled a few of the Radio Shock TV antennas with 4NEC2 to see what
they were doing. The results were awful. On some channels, it had
more gain out the back of the antenna than in the forward direction.
Impedance matching was awful and varied wildly across the frequency
range. Side lobes were all over the place. There were nulls in the
forward direction on some channels.
Under ideal conditions, a real log periodic has a forward gain of only
5 to 7dBi. Howver, that's about 6dBi across the entire frequency
range which is the point of such a broadband antenna. Having such a
low gain, it also has a rather wide forward beamwidth. My guess is
30-50 degrees at -3dB points horizontally. That's NOT what you want
if you want to avoid multipath.
I live in the Santa Cruz mountains. Before I gave up and went with
DirecTV satellite TV, I used a mast, rotator, and tangle of antennas
on the roof. Multipath from the nearby mountains was a serious
problem with Radio Shack antennas. Even the one station that was line
of sight had reflection problems.
So, I started to play with antennas. My first solution was a simple
bow tie antenna with a flat barbeque grill reflector. That worked
very well for UHF, but was too small for the VHF channels. It's main
advantage was that it was very broadband, few side lobes, and
excellent f/b. However, it didn't have much gain (about 4dBi).
Looking for something better, I decided that single channel yagi
antennas were the only way. The commerical versions are:
<
http://www.blondertongue.com/media/pdfs/catalog_classes/reception/bty.pdf>
However, I built my own. I had 4 different dipoles, pointed at the 4
transmitter locations. Gain was about 10dBi for VHF and 12dBi for
UHF, a substantial improvement over broadband antennas. Each antenna
had a really ugly homemade Dual Gate MOS FET RF amplifier at roof
level, with a combiner feeding the TV. I removed the rotator as it
was un-necessary with one antenna per transmitter/channel. Fine
aiming required a 20ft fiber glass pole to bang on the boom in one
direction or other.
The results were a dramatic improvement in reception signal strength
and quality. There were still a few ghosts visible but nothing as
horrible as with previous antennas. I was happy for about a year,
when a large fir tree branch managed to mangle 3 out of 4 antennas. As
I was recovering from surgery at the time, and was in no condition to
do anything major, I ordered DirecTV and gave up.
It looks ugly. Lots of signal strength but lots of selective fading that
already happens when a freighter lines up for final into Mather Field.
Fedex, Lufthansa Cargo, UPS, you name it. It's a major hub by now, used
to be a military base.
That's "frequency selective fading" and tends to be difficult to fix
unless you have a very very very narrow front beamwidth with no side
lobes and no ground bounce. If you have the possibility of
reflections from behind, also good f/b ratio. If these are difficult,
you might want to consider dual diversity antennas (about 10
wavelengths apart), with dual tuners and a diversity switch. If the
signal level to one antenna falls due to frequency selective fading,
the odds are good that the other antenna will still have a usable
signal.
Nope. Politics is the major problem. If the FCC would have approved
European COFDM modulation instead of crappy 8VSB, we would not have as
many multipath problem.
<
http://en.wikipedia.org/wiki/8VSB#8VSB_vs_COFDM>
The FCC has several chances to listen to sane technical comparisons,
but elected to follow the path paved by the US patent owners.
We have a clear shot through a narrow Hwy 50 gap but it is so small that
echoes from that ridge already come back in.
That's going to be very difficult to reduce multipath if both the
incident and reflected signals are coming from so close a direction.
Diversity reception might be the only workable band-aid. However, in
my limited experience with such multipath, that doesn't happen as much
as one would think. The reflected path ghost delay is usually so
small, that it's only visible as a slight smear to the right. The
reflections that generate the really visible ghosts, usually come from
the back, and sometimes from the sides. A good antenna pattern will
eliminate those.
