On Tue, 19 Aug 2008 12:36:30 -0700, Jim Lux
wrote:

Richard Clark wrote:
On Tue, 19 Aug 2008 09:17:01 -0700, Jim Lux
wrote:

But your search for small, broadband antennas puts you bump-up against the
principle "small - broadband - efficient, choose any two".
But actually, that's not the principle.. The actual limits have to do
with the ratio of stored energy vs radiated power in the antenna (Q, in
the energy storage sense, not in the "resonant circuit" sense).

Hi Jim,

Different meanings of Q? The measure of Q may vary according to
arbitrary usage: the choice of SWR limits to define bandwidth which
infers Q. Some choose 2:1, classic Q would go further. Either way,
and for either quantitative result, the meaning of Q remains
essentially the same.


Actually, the Q you use is not the actual definition, which is the ratio
of the energy stored in the system vs the energy lost per cycle (or
possibly per radian, so there's factor of 2pi in there)

Hi Jim,

If you will note, my example was a measure, not a definition of Q.
The half power points bandwidth compared to the center frequency is a
classic computation of Q. As I point out, the choice of 2:1 SWR, not
being half power points, is arbitrary but in no way diverges from the
sense of Q (and could be extrapolated anyway).

The articles that talk about size, efficiency, and Q, use this
definition (Energy stored in near field vs energy radiated away).
(e.g. the papers by Chu, Harrington, etc.)

Use "which" definition? You offer what appear to be two, and that is
one too many.

For a single LCR tuned circuit with reasonably high Q, it just happens
that the BW/CF works out to the same thing (because it's a quadratic
equation that determines both..)


Chu, 1948, defines Q = 2*omega*mean electric energy stored beyond input
terminals/(power dissipated in radiation) (page 1170). he goes on to
say,"We have computed the Q of an antenna from the energy stored in the
equivalent circuit and the power radiated, and *interpreted it freely*
as the reciprocal of the fractional bandwidth." (my emphasis added) " To
be more accurate, one must define the bandwidth in terms of allowable
impedance variation or the tolerable reflection coefficient over the
band. For a given antenna, the bandwidth can be increased by choosing a
proper matching network. The theoretical aspect of this problem has
been dealt with by R.M. Fano."

Harrington, 1965, considering directional antennas (Chu dealt with
omnis) defines Q as = 2*omega*W/Pin [eq 54], which is slightly different
than Chu. W is either We or Wm (the energy stored in the E or H field
respectively), and Pin is the input power to the array. He goes on to
say,"If the Q is large, it is related to the frequency bandwidth of the
array as follows. Consider the array to be resonated by a suitable
reactance network at the frequency of interest, omega/sub/r. Define the
ferquency bandwidth of the array in the usual manner to be the
fractional frequency increment between the 0.707 points on the
normalized input |Z|,
beta = deltaOmega/Omega/sub/r. [eq 56]

If the Q is high (say Q10) then we have the relationship:
Q approx= 1/beta [eq 57]"

Hence the two definitions? What I see are computational models for
different systems which arbitrarily restrain loss to inhabit their
model or exclude it. Q can deteriorate considerable if you open the
definition to include more components going back to the power supplies
to the finals.

Note well the "approximately equal" and the "resonated by a suitable
reactance network"


SWR bandwidth is something totally different of course..

But again, if you pick the right SWR value for your bandwidth
measurement, then a dipole antenna is modeled pretty well by a single
pole LCR resonant, so the math works out conveniently the same.

Once you start straying away from antennas that can be modeled as a
single LCR, the "bandwidth" vs "Q" relationship goes away.

A good example would be some forms of phased arrays with non-reciprocal
devices.

This example needs a sub-example: Non-reciprocal devices? I don't see
how this will alter the concept of Q.

Antennas with multiple resonances would be another case
(except if you are only working over a restricted range, where the
single resonance in your range is approximated well by a single LCR)

All antennas have multiple resonances and it is a classic
differentiator between themselves and LCR circuits. I see no example
here that is not already offered by my original post.

Let's simply return to the quote I responded to:
On Tue, 19 Aug 2008 09:17:01 -0700, Jim Lux
wrote:
(Q, in the energy storage sense, not in the "resonant circuit" sense)
embodies two explicit definitions of Q:
1. Q for energy storage, and
2. Q for "resonant circuit."
The measurement of Q might deviate by the result offered, but
conceptually energy storage and resonance are inextricably congruent.
If by resonance you are suggesting ONLY the peak frequency, then it
follows there is(are) some other frequency(ies) that are not resonant
(a tautology), and Q follows by exactly the same relation and degree
as it does for energy storage.

If there is any suitable distinction of Q in an antenna, then it is
that the degradation of Q is a benefit to the antenna, IFF the
substantial portion of R is radiation resistance. If the benefit of
filtering (phasing) due to resonance were not an issue, then a Q of 1
would be the Holy Grail of antenna design. Even in the design of the
finals stage in a tube amplifier, Terman teaches us that the final's
tank should NOT have an excessive working Q (beyond 10-15); hence a
high Q is NOT beneficial. Q and efficiency are a slippery topic when
you try to tie them together.

73's
Richard Clark, KB7QHC