II have covered all the math in the last posting, that says a 180 degree signal will only give a almost 1 dB subtraction or less, again this is acceptable.
Two identical, equal amplitude signals, 180 degrees apart and summed together will cancel each other. This is a manifestation of the same principle that's used to design multielement phased directional arrays. Phase and space things right, and you'll cancel signal in some directions, and add in others. Make it all random, and you end up with a random pattern. This is exactly what AM broadcasters are doing with their 3,4, and 5 tower directional systems. Only, their phasing is anything but random.
Your "math" by the way, is exhibiting a fundamental gap in your understanding of how that combiner with 20 db isolation works. That only means that a signal into the input of port 1 will be 20 db down at the input of port 2, and visa versa. Since there is minimal isolation between ALL input ports, to the common output port, and insertion loss is identical, that 20 db isolation figure is not applicable. It's a simple fact that 180 degree out of phase signals applied to the inputs will result in no signal at the output. Depending on the direction of the incoming wave front hitting the antennas, the signals from all the antennas will arrive at various phase angles, and randomly add and subtract, creating an antenna pattern that would more resemble a bug splat than an omni-directional pattern.
The aircraft and marine industries use this method, having two or more antenna feed a combiner to have more than one antenna feed one or more radios.
You're making a couple of apples and oranges comparisons here. In an aircraft, the upper and lower antennas are isolated, with substantially different coverage patterns. It's also very rare in general aviation, actually. Being a pilot, and having avionics experience, I've NEVER seen multiple antennas combined on anything from a Cessna 150 to a King Air. I HAVE seen antenna splitters used, though. I've never seen VHF antennas combined on a boat, either.
Are you sure you're not thinking about cases with multiple radios, multiple antennas, where each radio gets it's own? That's the way every installation in boats or aircraft that I've ever seen or worked on is done.
Aircraft use a upper antenna and lower antenna feeding a combiner to allow antenna coverage to allow communications when in high pitch, roll, and yaw positions. They rely on the isolation factor not having the out of phase signals cancel each other, its works very well.
Yes, that COULD work ok on an aircraft. See my comment above. Even then, there could be some pattern weirdness. Again, it's not a common technique on general aviation aircraft. Possibly on fighters that DO experience extreme pitch and roll angles, but I'd bet the antenna patterns and placement on the aircraft are very carefully considered. Not my field of expertise though, so I couldn't really say. But the point is moot... you're not trying this on an aircraft, are you.
My modified Scantenna and Discone Antennas feed a two port combiner, and a single coax runs down to another two port combiner that feeds two scanner radios, it also works very well. For the garage radios, one needs to keep an ear on the goings on when working ah.
Well, if it works for you, that's great. But I bet you would be shocked at how weird the pattern looks if you were to plot it. As long as signals are in saturation, a scanner user has no idea whether he's 20 db up or down because of a null or lobe.
I have some large diameter aluminum tubing that I will use to help broadband the VHF Aero Band, and VHF High Band Ground Plane Antennas. Or maybe even use the small 3 or 4 inch dia coffee cans. But I am still a few months away from putting up the 4 antenna array until the new air condition and heater system is put in.
Seems like an awful lot of work for what amounts to questionable benefits. But if it's fun, that's the point. For systems that are not fun, and seem more like work, I would NEVER attempt such a combining system.
Just for clarification, you keep mentioning feeding more than one radio with this. I hope you understand that my objection is NOT in feeding more than one radio. That's fine. It's combining the multiple antennas with overlapping frequency and pattern coverage that's the problem.
But go ahead. *shrug* Build it. Me? I know enough not to. My discones work just fine, thank you.