antenna-theory.com

[wolberine]

A lot of the videos on youtube talk about how we can change input impedance by moving the feed point. However, when we match the antenna with lumped elements, is the increase in radiated power due ONLY to reducing mismatch losses? Will we change the current or voltage distribution on the antenna or are they permanently fixed by the shape of the radiating structure?

[bigSteve]

If we used lumped element matching (caps and inductors), I believe we are simply reducing mismatch loss.

For sure, the inductors and capacitors will also introduce some loss, because they will always have some associated resistance with them.

I don't think we change the voltage/current distribution. Certainly, for a wire dipole, the current distribution is fixed and can't be changed by the impedance matching.

Now, if we use impedance matching away from the feed, as in a PIFA where we capacitively load it at one end, that does change the distribution which enhances radiation for small antennas.

But I don't think we change the distribution by matching at the feed.

[wolberine]

So we can say lumped elements reduce mismatch loss - I can see this.

Is this to say we cannot really improve antenna efficiency unless we physically change the structure to lower/raise the input impedance to see ~50ohm?

If so, is there a good rule of thumb for acceptable input impedances?

[bigSteve]

Yes, physically change the structure to improve efficiency, or change materials to reduce dielectric or ohmic losses.

There are no rules of thumb for input impedances. It would depend strongly on the application, which is to variable for antennas. If someone gave a rule it would be objected to in many applications.

[wolberine]

If we cannot really tell the radiation efficiency from the input impedance, why should we bother moving feed and/or tap points to make a 50 ohm match between source and load? Shouldn't we really move the feed/tap points to make a resonant and efficient antenna and match the source to load with lumped components?

[bigSteve]

That line of reasoning won't work. Why make a dipole a half-wavelength long if we can just use a lambda/100 dipole and cancel out the reactance?

If you take an antenna that is poorly matched, you will need large inductors and capacitors to match it. The losses in these components will hurt your efficiency.

More importantly, you're bandwidth will go to zero if you use a lot of impedance matching, particularly if you are matching wildly mismatched impedances.

In addition, if you feed in the wrong spot you can't physically match an open or short or anything purely reactive.

Finally, intelligent feed point selection is critical for multiband antennas.