Gain and
Efficiency, The Big Myth
by Tom Rauch,
W8JI
It may be time to call attention
to the fact that gain or efficiency has nothing to
do with receiving system performance, so long as the
background noise or QRM level is high enough to establish
noise floor in the receiver. One of the most common
and misleading myths is that any system gain increase,
either through decreased feedline loss or reduced
matching transformer loss, somehow improves reception.
This is NOT true, unless the receiving system is limited
by internal noise.
I have examples on my website
dispelling some common myths. Receiving ability is
tied directly to pattern directivity, not absolute
efficiency, signal level, or "gain".
Stacked Beverages
A gain increase in Eznec (or
some other modeling program) by adding parallel
lossy antennas, without mutual coupling effects,
will NOT represent increased receiving performance.
One example of this is paralleling a number of lossy
Beverages with small spatial separation. Neglecting
feedline and coupling losses, when the number of
close-spaced Beverages doubles the absolute signal
level increases 3dB. This signal level increase,
however, is not accompanied by an increase in S/N
ratio from the antennas unless spacing is significant
compared to the length of each antenna! The reason
for this effect is simply explained. Efficiency
doubles while the pattern does not change.
Reception is improved
only when the ratio of response in the desired direction
to the sum of responses in ALL other directions
is increased! You can find the correct number by
finding average gain of an antenna, and subtracting
the absolute gain at the angle and direction of
the desired signal. I call this the RDF (receiving
directivity factor) of an array (more).
Reduced Feedline/Transformer Loss
A recent article indicates the
best transformer designs somehow are related to
impedance matching and coupling loss. Practical
experience and theory disagree strongly with this
contention.
In a receiving system, SWR between
the receiver and antenna is unimportant except as
it increases effective system energy transfer losses.
If the receiver hears a reasonable noise increase
from propagated noise, changes in loss or matching
will not affect reception. Most modern receivers
are in the area of -130dBm sensitivity at narrow
bandwidths, and even a modest preamplifier with
4dB NF (very easy to obtain at HF or lower) will
allow -123dBm signal power levels to be 10dB out
of system noise.
As an example, at my quiet rural
location a push-pull high dynamic range amplifier
(DX Engineering RPA-1) allows 1.8MHz propagated
noise to limit lower threshold with only a NON-impedance
matched wire of ten feet connected to the amplifier's
low impedance input terminal. Matching would substantially
increase signal level, but would NOT improve S/N
ratio.
My System Examples
My system consists of around
30 Beverages and two arrays of phased verticals
with a total of twelve vertical elements. Most Beverages
are around 800 feet long, and are in co-phased pairs
spaced 350 feet broadside. The directivity of these
Beverages is up around 3dB from single antennas
on any frequency where the spacing is 5/8th wavelength
or more, although best usable performance centers
between 1.8 to 4.0 MHz.
The verticals are 20 feet tall,
resonated with an inductor/hat combination, and
coupled through an intentional series loss resistance
of almost 75 ohms to my 75-ohm feedlines.
Total feedline length is over
2500 feet in these systems, before the signal arrives
at any amplifying device. The cables are mixtures
of flooded (direct burial) F11 (RG-11) and RG6 CATV
cables, as well as some .625" hard line (1000
feet long) for the main trunks.
Despite this long feedline length
and the intentional matching loss on the verticals,
my system clearly limits on external noise. Why
then would anyone worry about any modest (under
5 or 10 dB) transformer loss in a matching transformer?
Certainly I don't, and I'm miles from any noise
sources.
By far the largest worry is unwanted
signal ingress into the feedlines, through common
mode currents. In effect the shield can act like
part of the antenna system if the cable is not correctly
balanced for currents at each end. It isn't a matter
of double-shielded cables; it is a matter of eliminating
common mode coupling. Even a poorly constructed
cable has negligible signal ingress through the
shield at HF when the shield and center carry exactly
equal and opposite currents. Make a poor shield
connection anywhere in the system will cause currents
to become unequal or incorrectly phased, and allow
even a quad-shielded cable to suffer from unwanted
signal ingress (check for more).
Summary
Dedicated DX'ers are well advised
to beware the pitfalls of considering absolute signal
level important, especially if such an increase
comes with a penalty of reduced directivity. Reduced
directivity through poor common-mode signal rejection
in a feed system may not appear as an increase in
absolute noise level. You may not know you have
hurt things at all.
With any listening antenna, it
is an excellent idea to isolate the antenna as much
as possible from the feedline shield. This is true
even when the increased isolation results in a slight
reduction in system gain or system signal levels.
Also beware of gain increases
in modeling programs. A gain increase may not be
related to an increase in directivity, and when
NOT related to increased directivity (or if your
receiver is so poor its internal noise sets the
system noise figure) the gain increase is useless.
You can find more details of
this at my website.
Good DX'ing,
Tom W8JI
published on DXing.info
on July 13, 2003
  
|
