"Antonio Vernucci" wrote in message
...
I have recently mounted a 6-meter long Yagi. The driven element is a
(non-folded) dipole whose impedance is brought up to 200 ohm by
significantly shortening it (so introducing a high series capacitive
reactance) and using a (rather small) hairpin to resonate the residual
reactance. The system is fed by a 4-to-1 balun (200-to-50 ohm) made of a
half-wavelength cable. The SWR at resonance is almost perfect, but the SWR
response vs. frequency is very sharp, this meaning that the system has a
high Q. I am wondering whether such a 200-ohm feed technique could cause
non negligible losses caused by the higher current circulating in the
hairpin (due to its low reactance as well as to the high voltage on the
high-impedance feed point, that is 200 ohm + reactance of the shortened
driven element). 73 Tony I0JX

I wonder what you consider narrow bandwidth. 500 KHz seems to be about par
for a decent 6 m beam.

Tam/WB2TT

I wonder what you consider narrow bandwidth. 500 KHz seems to be about par
for a decent 6 m beam.


This antenna shows an SWR of 1.7 at 150 kHz below the resonant frequency, where
the SWR is just 1, so I consider it narrow. Another antenna I have, also using
an hairpin match but at 50 ohm instead of 200 ohm, is by far broader. On both
antennas I have about 100 feet of low-loss 1/4" Andrew hardline, so the SWR is
only slightly influenced by cable loss.

Talking with the manfacturer, he told me that he preferred rasing the antenna
impedance up to 200 ohm, instead than to 50 ohm, because a 4:1 cable balun can
me more easily realized than a 1:1 balun. But doing so the feed system Q factor
increases quite a lot, causing a significant bandwidth reduction. Moreover ohmic
losses increase, due to the higher circulating currents (for a given RF power),
but I am not able to predict whether such extra losses are significant or can be
disregarded for all practical purposes.

73

Tony I0JX

"Antonio Vernucci" wrote in message
...
I wonder what you consider narrow bandwidth. 500 KHz seems to be about
par for a decent 6 m beam.


This antenna shows an SWR of 1.7 at 150 kHz below the resonant frequency,
where the SWR is just 1, so I consider it narrow. Another antenna I have,
also using an hairpin match but at 50 ohm instead of 200 ohm, is by far
broader. On both antennas I have about 100 feet of low-loss 1/4" Andrew
hardline, so the SWR is only slightly influenced by cable loss.

Talking with the manfacturer, he told me that he preferred rasing the
antenna impedance up to 200 ohm, instead than to 50 ohm, because a 4:1
cable balun can me more easily realized than a 1:1 balun. But doing so the
feed system Q factor increases quite a lot, causing a significant
bandwidth reduction. Moreover ohmic losses increase, due to the higher
circulating currents (for a given RF power), but I am not able to predict
whether such extra losses are significant or can be disregarded for all
practical purposes.

73

Tony I0JX
I don't think it is the balun. I measured a 4:1, 1/2 wave 6m balun made of
LMR240 and got a 2:1 SWR bandwidth from 40 to 60 MHz. This was with an
MFJ-269 and about a foot of 50 Ohm coax. Load was a 1/2 W 180 Ohm resistor
(measured high).

Tam/WB2TT

Tony I0JX
I don't think it is the balun. I measured a 4:1, 1/2 wave 6m balun made of
LMR240 and got a 2:1 SWR bandwidth from 40 to 60 MHz. This was with an MFJ-269
and about a foot of 50 Ohm coax. Load was a 1/2 W 180 Ohm resistor (measured
high).

You are right, it is not a balun problem. It is a problem due to the reasons I
had explained in my post, which had nothing to do with the balun.

73

Tony I0JX.

On Apr 7, 8:36 am, "Antonio Vernucci" wrote:
I wonder what you consider narrow bandwidth. 500 KHz seems to be about par
for a decent 6 m beam.

This antenna shows an SWR of 1.7 at 150 kHz below the resonant frequency, where
the SWR is just 1, so I consider it narrow. Another antenna I have, also using
an hairpin match but at 50 ohm instead of 200 ohm, is by far broader. On both
antennas I have about 100 feet of low-loss 1/4" Andrew hardline, so the SWR is
only slightly influenced by cable loss.

Talking with the manfacturer, he told me that he preferred rasing the antenna
impedance up to 200 ohm, instead than to 50 ohm, because a 4:1 cable balun can
me more easily realized than a 1:1 balun. But doing so the feed system Q factor
increases quite a lot, causing a significant bandwidth reduction. Moreover ohmic
losses increase, due to the higher circulating currents (for a given RF power),
but I am not able to predict whether such extra losses are significant or can be
disregarded for all practical purposes.

73

Tony I0JX

I'm trying to find a good way to wrap my mind around this "general"
problem. I'm not at a point I'm really comfortable with yet, but I
offer the following ideas as "food for thought."

First, the matching is being done essentially with an "L" network (or
rather the balanced version of an "L" network), where there is a load
resistance (the resistive part of the feedpoint impedance, which
includes radiation resistance and element loss resistance reflected to
the feedpoint), the series capacitive reactance of the feedpoing
element, and a shunt inductive reactance, provided by the hairpin.

Because shortening the driven element causes a decrease in resistance
and an increase in capacitive reactance, it's possible to find a
length that allows matching to any of a wide range of resistances.
But the higher the resistance to which you match, the shorter you need
to make the element and the lower the feedpoint resistance. The ratio
of feedpoint resistance to matched resistance determines the loaded Q
of the matching network; as you make the matched resistance higher,
the loaded Q goes up rather quickly. If you know the loaded Q and the
unloaded Q of the hairpin, you have a good handle on the amount lost
to heat in the hairpin: if the hairpin Q is two times the loaded Q,
half the power is dissipated in the hairpin, for example.

However, unless you build an antenna with a very low feedpoint
resistance at resonance, there almost certainly won't be an efficiency
problem: the reactance changes quickly enough with changes in driven
element length that the resistance won't drop much by the time you
reach enough reactance to get a match to 200 ohms. It appears that
the loaded Q of the match to 200 ohms for your case will be less than
4. I would think unless you really messed up badly, the hairpin
unloaded Q should be well in excess of 100, and if that's the case,
the power lost in the hairpin would correspond to well under 0.1dB
signal level change.

On the other hand, I'm surprised by the comment from the manufacturer
about difficulties making a 1:1 balun. I have had good luck using
ferrites and/or self-resonant coils of feedline and/or coils of
feedline specifically resonated with additional capacitance. A 4:1
balun from a "hairpin" of 1/2 wave of coax, arranged symmetrically, is
easy enough to make, but I would not rule out using a 1:1, if it has
advantages for you.

Cheers,
Tom

Hello Tom,

First, the matching is being done essentially with an "L" network (or
rather the balanced version of an "L" network), where there is a load
resistance (the resistive part of the feedpoint impedance, which
includes radiation resistance and element loss resistance reflected to
the feedpoint), the series capacitive reactance of the feedpoing
element, and a shunt inductive reactance, provided by the hairpin.
Because shortening the driven element causes a decrease in resistance
and an increase in capacitive reactance, it's possible to find a
length that allows matching to any of a wide range of resistances.
But the higher the resistance to which you match, the shorter you need
to make the element and the lower the feedpoint resistance. The ratio
of feedpoint resistance to matched resistance determines the loaded Q
of the matching network; as you make the matched resistance higher,
the loaded Q goes up rather quickly. If you know the loaded Q and the
unloaded Q of the hairpin, you have a good handle on the amount lost
to heat in the hairpin: if the hairpin Q is two times the loaded Q,
half the power is dissipated in the hairpin, for example.

However, unless you build an antenna with a very low feedpoint
resistance at resonance, there almost certainly won't be an efficiency
problem: the reactance changes quickly enough with changes in driven
element length that the resistance won't drop much by the time you
reach enough reactance to get a match to 200 ohms. It appears that
the loaded Q of the match to 200 ohms for your case will be less than
4. I would think unless you really messed up badly, the hairpin
unloaded Q should be well in excess of 100, and if that's the case,
the power lost in the hairpin would correspond to well under 0.1dB
signal level change.

All OK. I however reckon that, due to the parasitic elements effect, the
radiation resistance of the driven element (before shortening it) would be in
the order of 20 ohm. So, impedance gets brought up by a factor of 10 or so.

On the other hand, I'm surprised by the comment from the manufacturer
about difficulties making a 1:1 balun. I have had good luck using
ferrites and/or self-resonant coils of feedline and/or coils of
feedline specifically resonated with additional capacitance. A 4:1
balun from a "hairpin" of 1/2 wave of coax, arranged symmetrically, is
easy enough to make, but I would not rule out using a 1:1, if it has
advantages for you.

His argument is that, for high-power operation (say 1500W), it is more
convenient for him to provide a quarter-wavelength 4:1 balun made of RG-142
teflon cable than a sealed box with a 1:1 coil on an ironpowder toroidal core. I
mentioned him that, just using some extra length of RG-142 cable, he can easily
build a 1:1 balun, and hence design the antenna matching system for 50 ohm
instead of 200 ohm.

I hope he will listen to me, because that antenna is really narrowband!

73

Tony I0JX

On 7 abr, 23:15, "Antonio Vernucci" wrote:
Hello Tom,



First, the matching is being done essentially with an "L" network (or
rather the balanced version of an "L" network), where there is a load
resistance (the resistive part of the feedpoint impedance, which
includes radiation resistance and element loss resistance reflected to
the feedpoint), the series capacitive reactance of the feedpoing
element, and a shunt inductive reactance, provided by the hairpin.
Because shortening the driven element causes a decrease in resistance
and an increase in capacitive reactance, it's possible to find a
length that allows matching to any of a wide range of resistances.
But the higher the resistance to which you match, the shorter you need
to make the element and the lower the feedpoint resistance. The ratio
of feedpoint resistance to matched resistance determines the loaded Q
of the matching network; as you make the matched resistance higher,
the loaded Q goes up rather quickly. If you know the loaded Q and the
unloaded Q of the hairpin, you have a good handle on the amount lost
to heat in the hairpin: if the hairpin Q is two times the loaded Q,
half the power is dissipated in the hairpin, for example.

However, unless you build an antenna with a very low feedpoint
resistance at resonance, there almost certainly won't be an efficiency
problem: the reactance changes quickly enough with changes in driven
element length that the resistance won't drop much by the time you
reach enough reactance to get a match to 200 ohms. It appears that
the loaded Q of the match to 200 ohms for your case will be less than
4. I would think unless you really messed up badly, the hairpin
unloaded Q should be well in excess of 100, and if that's the case,
the power lost in the hairpin would correspond to well under 0.1dB
signal level change.

All OK. I however reckon that, due to the parasitic elements effect, the
radiation resistance of the driven element (before shortening it) would be in
the order of 20 ohm. So, impedance gets brought up by a factor of 10 or so.

On the other hand, I'm surprised by the comment from the manufacturer
about difficulties making a 1:1 balun. I have had good luck using
ferrites and/or self-resonant coils of feedline and/or coils of
feedline specifically resonated with additional capacitance. A 4:1
balun from a "hairpin" of 1/2 wave of coax, arranged symmetrically, is
easy enough to make, but I would not rule out using a 1:1, if it has
advantages for you.

His argument is that, for high-power operation (say 1500W), it is more
convenient for him to provide a quarter-wavelength 4:1 balun made of RG-142
teflon cable than a sealed box with a 1:1 coil on an ironpowder toroidal core. I
mentioned him that, just using some extra length of RG-142 cable, he can easily
build a 1:1 balun, and hence design the antenna matching system for 50 ohm
instead of 200 ohm.

I hope he will listen to me, because that antenna is really narrowband!

73

Tony I0JX

Hello,

Imagine 10 Ohms radiation resistance + capacitive component after
shortening the open dipole radiator. When you convert that to 200 Ohms
via a parallel inductance (hairpin), the Q factor of such a network is
about 4.4. At 50 MHz that will result in a bandwidth (VSWR=2) of 8
MHz.

A 200 Ohms to 50 Ohms coaxial balun will have a lower Q factor. So the
combination of balun and L-network will certainly have a useful
bandwidth 340 kHz.

Maybe the radiation resistance is less (you can derive that from the
hairpin inductance) and/or the antenna is by nature (very) narrow
band.

Best regards,

Wim
PA3DJS
www.tetech.nl
remove abc from the address.

On Apr 7, 4:02 pm, Wimpie wrote:
On 7 abr, 23:15, "Antonio Vernucci" wrote:



Hello Tom,

First, the matching is being done essentially with an "L" network (or
rather the balanced version of an "L" network), where there is a load
resistance (the resistive part of the feedpoint impedance, which
includes radiation resistance and element loss resistance reflected to
the feedpoint), the series capacitive reactance of the feedpoing
element, and a shunt inductive reactance, provided by the hairpin.
Because shortening the driven element causes a decrease in resistance
and an increase in capacitive reactance, it's possible to find a
length that allows matching to any of a wide range of resistances.
But the higher the resistance to which you match, the shorter you need
to make the element and the lower the feedpoint resistance. The ratio
of feedpoint resistance to matched resistance determines the loaded Q
of the matching network; as you make the matched resistance higher,
the loaded Q goes up rather quickly. If you know the loaded Q and the
unloaded Q of the hairpin, you have a good handle on the amount lost
to heat in the hairpin: if the hairpin Q is two times the loaded Q,
half the power is dissipated in the hairpin, for example.

However, unless you build an antenna with a very low feedpoint
resistance at resonance, there almost certainly won't be an efficiency
problem: the reactance changes quickly enough with changes in driven
element length that the resistance won't drop much by the time you
reach enough reactance to get a match to 200 ohms. It appears that
the loaded Q of the match to 200 ohms for your case will be less than
4. I would think unless you really messed up badly, the hairpin
unloaded Q should be well in excess of 100, and if that's the case,
the power lost in the hairpin would correspond to well under 0.1dB
signal level change.

All OK. I however reckon that, due to the parasitic elements effect, the
radiation resistance of the driven element (before shortening it) would be in
the order of 20 ohm. So, impedance gets brought up by a factor of 10 or so.

On the other hand, I'm surprised by the comment from the manufacturer
about difficulties making a 1:1 balun. I have had good luck using
ferrites and/or self-resonant coils of feedline and/or coils of
feedline specifically resonated with additional capacitance. A 4:1
balun from a "hairpin" of 1/2 wave of coax, arranged symmetrically, is
easy enough to make, but I would not rule out using a 1:1, if it has
advantages for you.

His argument is that, for high-power operation (say 1500W), it is more
convenient for him to provide a quarter-wavelength 4:1 balun made of RG-142
teflon cable than a sealed box with a 1:1 coil on an ironpowder toroidal core. I
mentioned him that, just using some extra length of RG-142 cable, he can easily
build a 1:1 balun, and hence design the antenna matching system for 50 ohm
instead of 200 ohm.

I hope he will listen to me, because that antenna is really narrowband!

73

Tony I0JX

Hello,

Imagine 10 Ohms radiation resistance + capacitive component after
shortening the open dipole radiator. When you convert that to 200 Ohms
via a parallel inductance (hairpin), the Q factor of such a network is
about 4.4. At 50 MHz that will result in a bandwidth (VSWR=2) of 8
MHz.

A 200 Ohms to 50 Ohms coaxial balun will have a lower Q factor. So the
combination of balun and L-network will certainly have a useful
bandwidth 340 kHz.

Maybe the radiation resistance is less (you can derive that from the
hairpin inductance) and/or the antenna is by nature (very) narrow
band.

Best regards,

Wim
PA3DJSwww.tetech.nl
remove abc from the address.

Yes, I had similar thoughts, but a bit different. First, I think it's
safe to say that if, at resonance, the driven element presents about
20 ohms at the feedpoint, shortening the D.E. only a little will give
enough capacitive reactance to allow the hairpin match to 200 ohms.
Even if the D.E. looks like 5 ohms, the "L" network match still gives
a 3dB bandwidth of 8MHz at a 50MHz center, or 3MHz 1.5:1 SWR
bandwidth. I think we need to look somewhere else for the answer to
the narrow bandwidth. My working hypothesis at the moment is that
it's in the antenna, or perhaps rather in the combination of antenna
and matching network. Note that the calc for the L match assumed a
constant capacitance, but the antenna will not, in general, look
anything like a constant C even over a fairly narrow frequency range.
I just ran EZNEC on a frequency-scaled version of the 14MHz 5 element
Yagi included in the sample files, with the D.E. slightly shortened to
allow a decent hairpin match to 200 ohms at the design center
frequency. I did a frequency sweep, 49 to 51 MHz, in 0.25MHz steps.
Over that range, the equivalent series capacitance changes from 59pF
at the low end to 138pF at the high end, and at least by NEC2's
prediction, the impedance changes especially quickly around 51MHz--
both reactive and resistive parts. 50.75MHz: 10.3-j31.76; 51MHz: 3.91-
j22.56, quite a large percentage change in 250kHz. Having the
effective series capacitance change that quickly will cause the
matching network to behave very differently than it would with a
capacitance element that is fixed.

Is it possible to lengthen the D.E., causing it to present an
inductive reactance at the feedpoint, and match that (to 200 ohms, or
to 50 ohms) with a shunt capacitance? That may work better, giving a
broader SWR bandwidth. The resistive component should be higher,
further lowering the Q, and I suspect the reactance won't change so
quickly with frequency. I don't have time at the moment to compare
the two, but may this evening. It may also be possible to raise the
resistive part to 50 ohms and match to 50 with a series capacitance.

Cheers,
Tom

K7ITM wrote:

Is it possible to lengthen the D.E., causing it to present an
inductive reactance at the feedpoint, and match that (to 200 ohms, or
to 50 ohms) with a shunt capacitance? That may work better, giving a
broader SWR bandwidth....

Good, constructive suggestion, Tom. Kudos for putting in a bit
of design/analysis effort on this rather than just shooting from
the hip. This is definitely a weird way of driving a yagi. It makes
me yearn for the old TV-antenna schemes that used folded
dipoles for the driven element, suitably split-up between upper
and lower wires so as to give a 300-ohm terminal impedance even
with the resistance-lowering effects of the reflector and directors.

Jim, K7JEB

Over that range, the equivalent series capacitance changes from 59pF
at the low end to 138pF at the high end, and at least by NEC2's
prediction, the impedance changes especially quickly around 51MHz--
both reactive and resistive parts. 50.75MHz: 10.3-j31.76; 51MHz: 3.91-
j22.56, quite a large percentage change in 250kHz. Having the
effective series capacitance change that quickly will cause the
matching network to behave very differently than it would with a
capacitance element that is fixed.

That is exactly the point! It would not be correct to calculate bandwidth on the
basis of the Q factor at resonance and assuming that the capacitive antenna
reactance is equivalent to that of a fixed capacitor.

Today I have discovered another shortcoming of that antenna. After raining cats
and dogs, the antenna resonant frequency gets lowered by about 130 kHz due to
the influence of the wet terrain. That is really a lot if you consider that,
after making very accurate measurements with a Bird wattmeter, the antenna
bandwidth is only 100 kHz at 1.4 SWR!

I am considering to re-build the driven element for 50-ohm match, by using a
longer driven element and a 1:1 balun. However it will not be easy to find the
optimum situation because there are two variables to be adjusted, that is the
driven element length and the hairpin length.

Also, I am not too sure on to which extent using a longer driven element would
influence the antenna radiation pattern.

Any comment?

73

Tony I0JX