In a recent discussion with a fellow ham about where to use a RigExpert antenna analyzer for tuning an antenna for minimum SWR, we couldn't decide if "antenna only" or "feedline+antenna with the analyzer at the transmitter end". For example, if I tune my antenna for minimum SWR right at the antenna itself ... that's perfection, right? But if I then analyze the feedline coax PLUS the antenna, the "whole antenna system" ... and realize that I need to tune the antenna again for minimum SWR (as seen by the radio end), then that tune would result in the highest power efficiency 'into the air'. But wouldn't that be a de-tune of the antenna itself? I'm confused on where I should connect the analyzer, and whether or not the "de-tune of the antenna" is real or best or not? Can any of you, smarter than me, explain and explain what I should do? BTW, calling different antenna manufacturer tech support lines gave us different answers ... one said to tune at the antenna, the other said tune from the transmitter end of the coax.
Tuning at the antenna will give you the most accurate antenna characteristics, but a good VNA can cover up some of the inaccuracies from tuning at the end of the coax instead.
However, you have to realize that the coax is also part of the antenna system. You shouldn't have to tune the coax, but you can.
If SWR is lower at the end of the coax, that's normal loss in the coax. If it is higher, the coax is acting like part of the antenna and a choke to stop common mode current will help.
If there is common mode on the coax, that means the coax is radiating and part of the antenna. You can tune the coax by making it a multiple of a quarter wavelength (corrected for velocity factor) long. However, if you do this, you have to realize the coax is radiating like part of the antenna which can be undesirable and possibly dangerous. If the coax is not tuned, this will raise your SWR. A balun will prevent (or at least reduce) common mode current and (hopefully) make the coax length irrelevant.
Some antennas (like the carolina windom) intentionally allow part of the coax to radiate by careful placement of the balun (or unun). However, you don't want the coax in your shack radiating, as that adds extra RF exposure, and the possibility of RF burns.
You are asking two questions, tune the antenna, or tune the coax/antenna system for low SWR. Or tune for best performance/tune for lowest SWR. They aren't mutually exclusive, but pretty hard to achieve in practice for anything but one specific frequency.
A feedline 1/2 wavelength long, (electrical wavelength, = physical length x velocity factor (.66 for most coax)) reflects the impedance it is terminated at. 50 ohms in, 50 ohms out. 75 ohms in 75 ohms out. A dipole antenna in free space ( no ground effects like capacitance or wire resistance) is 73 ohms. So if you had a dipole, @ a specific frequency, terminated into a 1/2 wavelength (electrical) long 75 ohm coax, when you trimmed it to SWR 1:1 the dipole impedance would be 75 ohms, which would be the sum of (Radiation resistance and ohmic resistance). (73+2ohms ish) So there is no magic 1:1 over more than a few KHz.
A good 1/2 wave antenna will give you better than 1:2 at the resonant frequency, (This broadening of the 1:1 is because of ohmic resistance and other stuff) As your antenna gets closer to earth, the capacitance of wire to earth causes the antenna impedance to drop to something between 30-70 ohms (height, ground conductivity, slope of wire, etc are variables). This is all at one frequency.
If your coax is odd 1/4 electrical wavelength long, it will be a very high impedance at the antenna. This will be seen as a impossible to match value for a dipole. I wish I had a table to refer you to of never use lengths of coax for a multiband dipole. Maybe others can chime in for that. There are tables/graphs of antenna impedance vs height above ground to use. I'll caution that most of those do not take into account local ground conductivity, so add +/- 20% to your error analysis if you are calculating something. Moderately over long, or over short antennas do not effect the radiation resistance much, that part of the resistance that is caused by energy loss through radiation, what does happen is over long or short causes capacitive or inductive loading on the antenna, which added or subtracted from the capacitive loading of the antennas proximity to the ground, causes a value of something that will not be 75 ohms, which is what your 1/2 wave coax wants to see. If you are using 50 ohm coax, you can match this by lowering your antenna wire below the 1/4 wavelength above ground distance, or shortening your antenna below the calculated value. So when you worry about achieving SWR 1:1, you are subtracting or adding the mismatch of your antenna impedance, from the mismatch of your coax impedance at one specific set of conditions.
It's a mugs game to try to get this prefect 1:1 SWR at any place but this one perfect frequency with a dipole antenna. As long as your trans-match or radio loads your antenna to a 3:1 match, you'll do ok.