Originally Posted by MikeySoft
Wouldn't having a 90 degree angle between the two phase lines when they cross minimize any interference between the two? Why is it recommended to keep a space between them?
You aren't trying to "minimize interference". Indeed, specific interactions between the two wires are desired. The parallel wires form a transmission line. Transmission lines have near magical powers.
They require constant spacing. Any discontinuity tends to result in reflections and other weird effects that can sometimes be beneficial but are usually detrimental if they weren't deliberate.
Some examples of transmission lines:
- Open air balanced transmission lines (as used here)
- Twin lead
- Ethernet, USB, firewire, RS-422/485, etc.
- High speed signals connecting components on PC motherboards
Two parallel wires, with or without a ground plane, or one wire with a ground plane or shield make a transmission line if they are long relative to the highest frequency in use. If the distance is shorter than 1/6 of a wavelength, they can be treated as a lumped sum capacitance for digital signals but longer than that and they are a transmission line; for analog signals, effects may be significant even at 1/6 wavelength.
A short wire, relative to the wavelength, has, more or less, the same voltage at every point on a wire. For a transmission line, the signal travels down the wire and different voltages are simultaneously present at different points on the wire. If you send a +/-1V signal down a transmission line one wavelength long (with no reflections or other oddities) you will have +1V at one point on the wire and -1V on another and every voltage in between simultaneously.
Imagine a pond or other still body of water. Now draw an imaginary line segment across the surface of the pond with two endpoints. Drop a pebble at one end and watch how the waves travel down that line. If you put a large solid object in the water along the line, you can see reflections occurring that will act somewhat like reflections from discontinuities in the line. If you imagine that the line passes through two fluids with different viscosities, say water and oil that have magically kept separate sideways instead of the usual vertical separation, that is roughly equivalent to having two different impedances along the line. The wave energy from the water doesn't smoothly transfer to the oil (some does, some doesn't); the energy that doesn't transfer is reflected.
Oversized washers on the line act as a capacitor on the line and create an impedance mismatch that can cause reflections. At certain key points (i.e. where the antenna elements connect and at the feedpoint), the capacitance could be good or bad, depending on whether it improves the impedance match of the elements at that point or makes the match worse. Crossing the wires makes a portion of the impedance line have much lower impedance which creates a discontinuity. This is usually bad.
A transmission line generally wants to have a constant impedance along the entire length of the line (which means, among other things, constant spacing and the same interactions with external conductors or dielectrics along the entire length), be driven by a source that matches the characteristic impedance of the line and be terminated by a resistor that matches the characteristic impedance. Any exception causes trouble unless you are specifically exploiting the specific weirdness that results to your benefit.
Avoid any non-uniformities in transmission lines, including non-uniformities in objects in their vicinity. Do not change any parameter of a transmission line including: wire diameter, wire spacing, wire length, wire insulation, metallic objects in their immediate vicinity, and non-conductive objects in their immediate vicinity. Immediate vicinity would be several times the spacing. An exception is shielded cables, such as coax and twinax, which tolerate objects in their vicinity. Another other exception is irregularities deliberately introduced by someone who understands transmission lines. If a transmission line is driven with its characteristic impedance (usually not true at most frequencies on an broadband antenna, unfortunately) and is also terminated in its characteristic impedance, then the length of the line has little effect except for some loss proportional to length. Otherwise, you have frequency dependent effects that change with the length of the wire. This is one reason an amplifier at the antenna can improve things because the coax cable is driven properly at all frequencies.
TC 9-64 (link posted earlier) has a description of how a transmission line works.
As a random example of the wierdness of transmission lines, I once had some network cards that had suffered lightning damage on a long bus topology network. Many hundreds of feet of cable connected computers in one or two buildings (don't remember how many at that time). Reflections were caused by the damaged cards. As a result, two PCs couldn't talk to each other even though the same PCs could talk to other PCs. The reflections from a 3rd (damaged) computer were canceling out the signals those specific PCs saw form each other. A 10volt signal was being annihilated! Took a while to figure that one out.