On Wed, 3 May 2017 07:47:12 +1000, Clifford Heath
wrote:

It seems as though it should be straightforward to feed power
up the co-ax (ala masthead amplifiers) to drive a pair of isolation
amps/buffer before a combiner. That could give you a lot more isolation,
without needing to go to relays.

Clifford Heath.

Nope. The lack of isolation is NOT in the combiner. Assuming a
reasonable 75 ohm termination on all ports, a good combiner can
deliver 20 to 50dB isolation over the VHF-UHF TV band. I could
replace the splitter/combiner with a 6dB resistive power
divider/combiner, which has 6dB of isolation, and get the same
problems at a 3dB lower signal level.
Resistive power divider/combiner:
https://www.microwaves101.com/encyclopedias/resistive-power-splitters
A proper splitter:
http://802.11junk.com/jeffl/crud/CATV-splitter.png
Good splitter/combiner and total crap splitter/combiner:
http://802.11junk.com/jeffl/crud/CATV-splitters.jpg

Anyway, the problem is the isolation between antennas. If both
antennas can pickup the same signal, from the same station, the
signals are going to add or cancel depending on the phase and
amplitude. If there is a phase delay between these two signals, they
will act much exactly as if there was a multipath problem. ATSC 8-VSB
has some limited protection against multipath, but I wouldn't count on
it. The symptoms manifest themselves as everything that could
possibly trash a DTV signal. Fluttering, Stuttering, freezing,
pixellation, rainbow color light shows, signal loss, etc. As I
mentioned in a previous rant, the degree of cancellation and
impairment is frequency dependent, which means I can't easily use a
simple phase shifter to remedy the situation. I found some reference
which suggested that one can see ghosts on a DTV with multipath. I
never saw it or if they were there, they were buried under the light
show and pixellation.

Someone is sure to ask why then do two stacked antennas work?
https://www.google.com/search?q=tv+antenna+stacking&tbm=isch
These work because both antennas involved are looking at the same
station, which produces the same signal level at the same phase at
each antenna. Therefore, they can safely be combined, where the two
signals add in phase, and therefore produce 3dB more effective anenna
gain.

Incidentally, I originally started by building a DPST PIN diode switch
which selected which antenna was active by which coax cable had DC
supplied to it. However, I made a stupid mistake and couldn't get it
to function correctly. Since I was burning time and loosing money on
this particular customer, so I took the easy way out and used a pair
of spare latching relays and two push buttons from an HF antenna tuner
project.

Bottom line: RF is magic.
--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558