Flag and Pennant
info on the flag antennas
Listening test and other observations
How to make your own Flag
How to best match your Flag
the right core material
By George K0FF
antenna mail list, November, 2000
My suggestion is to use the exact measurements and techniques described
by K6SE, Earl Cunningham, for the Pennant transformer.
There is some variation in the permeability factor in the
FT-140-43 core and I'm working with the factory on that right now.
You would expect around 950 uH from 10 turns spaced on the form,
but actual samples are coming in at higher values. In a cross sample
of almost 100 cores that I've tested, I've seen 2 distinct groups,
but within the group, the tolerance is close. It seems that the
batch to batch mix is variable though, and I will get that resolved
Back to the original design (using 26ga enamel wire):
FT-140-43 core 8T primary, 28 (for 75) or 34 (for 51) Ohm feedline.
Make sure the windings both have the same "sense" or "hand",
in otherwords, make sure they point the same way on the core.
Freq 8T (56uH) 28T (672 uH) 34T (991 uH) 12.25:1 ratio= 18:1 ratio=
918 Ohms 918 Ohms R*L R*L R*L 500 kHz 177 Ohms 2111 Ohms 3113 Ohms
1000 kHz 355 4223 6227 1800 kHz 639 7600 11200 3500 kHz 1242 14780
These number have been rounded off a little, but it matters not.
All that is important is:
A. Balance (and sense)
C. Primary initial impedance at the design lowest design
The inductance remains the same (more or less) across the spectrum
for which the core is designed. Most cores are good way below the
broadcast band. Of course the reactance (R*L) changes according
to the frequency. Changing the turns count (as long as the ratio
remains constant) will only shift the operating range. As long as
your desired frequency is still within the operating range it will
work the same, but again I suggest go with Earl's design so we are
all reading off the same sheet of music as far as the transformer
As far as the Bev transformers are concerned, I continue
to recommend the DC connected transmission line type over the others.
I do agree that an isolated shield can (possibly) improve noise
coupling, but the DC shunt in the grounded design seems far safer
and I have not had trouble with the noise coupling at all here.
Pennant Transformers according to Earl's design will be available
shortly from Array Solutions. They are in a gasket cast aluminum
box with a brass weep hole, and have a stainless steel rope eye,
and gold plated binding posts. The coil is wrapped in fiberglass
tape, and dipped in liquid polystyrene for stability.
The Beverage transformers according to my own design will
be there also, and have Cast aluminum housing, gasket sealed, groundrod
clamp and aux. brass ground screw, DC grounding of the antenna wire
with a Kynar fusible link, and an air gap, choke shunt to ground,
gas tube protected, and on the coax side it is DC blocked,and drained.
Taps allow 50 or 75 Ohm feed and 4:1 9:1 and 16:1 winding ratio
selection. Fiberglass wrapped coil with a 140 size core for low
loss, dipped in liquid polystyrene. Teflon connector and gold plated
Gary Breed, K9AY: 43 material is commonly used in EMI filtering
applications where looser manufacturing tolerances may be acceptable.
It could be the cores you have were made to meet a spec of "guaranteed
minimum" mu of 850. I haven't seen this variation with Fair-Rite
cores, but other companies (e.g. Steward) make ferrites solely for
EMI filters, to the same general spec as 43, but with a disclaimer
that they are not characterized for transformer use. Your experience
points out the need for measurements.
Your experiments are interesting, and I have done some myself.
My 9:1 xfmrs use 1.5 inch dia. 43 material toroids -- 8 trifilar
turns of #24 telephone wire put the lowest loss region from about
2 to 5 MHz. Loss is no more than 0.25 dB.
I have also wound these as a single continuous winding with a tap
1/3 the way from ground. This is easier for a non-technical ham
to wind and although the loss and mismatch are worse, no one is
likely to hear the difference.
For isolated windings, I use two 1/2" dia x 1/2" long 77
material (mu = 2000) cores taped together binocular fashion. 9T/3T
works OK impedance-wise for a 9:1 XFMR, but the loss is about 0.4
dB. I also have used these cores for 2:1 xfmrs (7:5 turns ratio)
in RX array combiners.