Roy Lewallen wrote:
Hm, I'll have to think about that one a bit. I checked out the Elecraft
info and it appears that the balun is wound on a single binocular core,
not two separate cores. You could make a 4:1 current balun that way if
you could manage to keep the flux from one pair of windings from
coupling to the other, but I don't know right off how you'd do it. I
suspect that some sort of complex winding scheme might make it work. But
if the one you're trying to measure really has two separate cores, I
suspect that measuring the Elecraft might not be a good test of the method.

Roy Lewallen, W7EL

Hi Roy,

I have read your comment about the requirement for two cores for many
years, and I am still puzzled. In the ideal case the balun action is
performed by transmission line transformers, and the flux induced in the
core by the differential current is minimal. The core primarily affects
frequency response and common mode choking action.

According to his books, Jerry Sevick has designed, analyzed, built, and
measured many single core 4:1 current baluns. He never mentions the
problem you allude to.

What is happening here?

73,
Gene
W4SZ

On Thu, 16 Nov 2006 19:04:39 GMT, Cecil Moore
wrote:

Is the inductive side of resonance of no use in
that type of balun?

Its resonate at about 1.58 MHz. Whatcha going to use it for? Broadcast
AM radio?

Danny

Danny Richardson wrote:
Cecil Moore wrote:
Is the inductive side of resonance of no use in
that type of balun?

Its resonate at about 1.58 MHz. Whatcha going to use it for? Broadcast
AM radio?

I wasn't talking about that particular balun just that
particular method of construction. If a balun like that
one were made resonant at, say 7 MHz, would it still be
useful at 4 MHz? at 2 MHz? In other words, what is the
shape of the choking curve below resonance for that
particular method of construction?
--
73, Cecil, http://www.qsl.net/w5dxp

Gene Fuller wrote:
Roy Lewallen wrote:

Hm, I'll have to think about that one a bit. I checked out the
Elecraft info and it appears that the balun is wound on a single
binocular core, not two separate cores. You could make a 4:1 current
balun that way if you could manage to keep the flux from one pair of
windings from coupling to the other, but I don't know right off how
you'd do it. I suspect that some sort of complex winding scheme might
make it work. But if the one you're trying to measure really has two
separate cores, I suspect that measuring the Elecraft might not be a
good test of the method.

Roy Lewallen, W7EL


Hi Roy,

I have read your comment about the requirement for two cores for many
years, and I am still puzzled. In the ideal case the balun action is
performed by transmission line transformers, and the flux induced in the
core by the differential current is minimal. The core primarily affects
frequency response and common mode choking action.

According to his books, Jerry Sevick has designed, analyzed, built, and
measured many single core 4:1 current baluns. He never mentions the
problem you allude to.

What is happening here?

I'm not sure myself, Gene. I've wound single core 4:1's, I've bought at
least 2 of them. I've also been told variously they don't work, and
can't even be made.

Just what IS happening here?

- 73 de Mike KB3EIA -

Oop, looks like I done hit de tar baby.

Gene Fuller wrote:

I have read your comment about the requirement for two cores for many
years, and I am still puzzled. In the ideal case the balun action is
performed by transmission line transformers, and the flux induced in the
core by the differential current is minimal. The core primarily affects
frequency response and common mode choking action.

It's true that the flux induced by the differential current is minimal
in a transmission line transformer. If, in fact, the balun is wound with
coax as I usually make them, the flux induced by the differential
current is zero because the differential field is entirely inside the
shield.

And it's true that the core affects common mode choking action. The
common mode choking action is exactly what the function of a balun is --
that's how it effects equal and opposite (purely differential mode)
currents on the two transmission line conductors. In other words, that's
how it produces balance and prevents feedline radiation.

So the flux from the common mode component of the current is the very
thing we need to look at when we evaluate baluns.

According to his books, Jerry Sevick has designed, analyzed, built, and
measured many single core 4:1 current baluns. He never mentions the
problem you allude to.

I don't know which edition you have. I have only the first and second; I
believe there was a third. In the first edition, it was clear that
Sevick didn't understand the purpose of baluns or how they accomplished
it. So I discount anything he wrote about baluns in the first edition.
But he figured it out some time between the first and second editions,
so there's some credible information in the second. A number of the 4:1
baluns he shows are indeed wound on a single core. And he dutifully
reports their frequency responses when terminated in a correct impedance
load (seldom the case in a real antenna application). But nowhere do I
see any measurement of their ability to do the primary job of a balun:
to balance the feedline currents. So there's no way of knowing whether
any or all of the many "current" baluns he constructed actually perform
the function of a current balun.

A way to qualitatively check to see if a balun is really acting like a
current balun is to connect a load across the output. Short one terminal
of the output to the "ground" terminal of the input, and measure the
impedance looking in. Then disconnect that short and short the other
output terminal to the "ground" terminal of the input and measure it
again. You can do this with an antenna analyzer or impedance bridge. A
perfect current balun will show the same, correct impedance in both
tests. A perfect voltage balun will show very different impedances.
There should be a way to get a quantitative measure of the longitudinal
(common mode) impedance from this test, but I don't have time to dig it
out just now.

I'd like to add that not long ago I encountered a claim of a 4:1 current
balun on a single core. It was indeed -- it was a conventional, as
opposed to a transmission line, transformer, although it was wound in a
way that made it resemble a transmission line transformer. And it had a
fairly decent bandwidth, although a transmission line transformer is
inherently much broader and doesn't couple differential flux into the
core. So a 4:1 current balun can be constructed on a single core. But I
don't know how a 4:1 transmission line current balun can.

What is happening here?

A whole lot of things could be happening. If an antenna is perfectly
symmetrical, a voltage balun will do the same job as a current balun.
For an explanation why, see
http://eznec.com/Amateur/Articles/Baluns.pdf. So a lot of the "current
baluns" might well be voltage baluns. I seldom see anyone actually test
the current balancing properties of baluns, and without this, there's no
telling what's been constructed.

I haven't looked carefully at Sevick's single core 4:1 "Guanella"
baluns. They might be voltage baluns, or they might be conventional
transformers -- bifilar windings don't automatically create a
transmission line transformer. I don't see any test he ran to verify
that they're functioning as current transformers or for that matter as
transmission line transformers. In a transmission line transformer, the
differential current always has a transmission line path to follow. Some
connections can split the wires carrying a differential current apart,
in which case you no longer have a transmission line transformer.

If you know of a topology which produces a transmission line 4:1 current
balun, I'd be very interested. You could use the same simple analysis
method used in the balun article to show that it really is what you think.

Mike Coslo wrote:

I'm not sure myself, Gene. I've wound single core 4:1's, I've bought
at least 2 of them. I've also been told variously they don't work, and
can't even be made.

Just what IS happening here?

You probably made 4:1 voltage baluns, which worked well enough in your
applications. Where a voltage balun is really a bad idea is when your
antenna system isn't symmetrical. In an effort to balance the voltages
on the two output terminals, it creates unbalanced currents. The
difference between them becomes a common mode feedline current. This is
also explained in the balun article.

.. . . Here Uncle Remus paused, and drew a two-pound yam out of the ashes.

"Did the fox eat the rabbit?" asked the little boy to whom the story had
been told.

"Dat's all de fur de tale goes," replied the old man. "He mout, an den
agin he moutent. Some say Judge B'ar come 'long en loosed 'im - some say
he didn't. I hear Miss Sally callin'. You better run 'long."

Roy Lewallen, W7EL

Hi Roy,

Sorry about making such a mess. That tar is really hard to clean up.

FYI, I was referring mostly to Sevick's 4th edition of Transmission Line
Transformers, published in 2001. As you mention, he has undergone a
considerable transformation (pun intended) since his earliest books.

Sevick does indeed recommend a two-core approach for a 4:1 Guanella
balun if the load side is not balanced. As you know, he generally avoids
the use of the terms "current balun" and "voltage balun".

My question still holds, however. In a typical situation the "balanced"
side, generally the antenna, is only partially unbalanced, not simply
shorted to ground on one side. In this generic case would you expect the
single core Guanella balun to perform poorly with slight or modest
imbalance?

If both sides of the transformer are completely unbalanced it is not
clear why one would expect a balun of any type to be useful. What is
needed in that case is impedance transformation, not balancing.

73,
Gene
W4SZ

Roy Lewallen wrote:
Oop, looks like I done hit de tar baby.

Gene Fuller wrote:

I have read your comment about the requirement for two cores for many
years, and I am still puzzled. In the ideal case the balun action is
performed by transmission line transformers, and the flux induced in
the core by the differential current is minimal. The core primarily
affects frequency response and common mode choking action.

It's true that the flux induced by the differential current is minimal
in a transmission line transformer. If, in fact, the balun is wound with
coax as I usually make them, the flux induced by the differential
current is zero because the differential field is entirely inside the
shield.

And it's true that the core affects common mode choking action. The
common mode choking action is exactly what the function of a balun is --
that's how it effects equal and opposite (purely differential mode)
currents on the two transmission line conductors. In other words, that's
how it produces balance and prevents feedline radiation.

So the flux from the common mode component of the current is the very
thing we need to look at when we evaluate baluns.

According to his books, Jerry Sevick has designed, analyzed, built,
and measured many single core 4:1 current baluns. He never mentions
the problem you allude to.

I don't know which edition you have. I have only the first and second; I
believe there was a third. In the first edition, it was clear that
Sevick didn't understand the purpose of baluns or how they accomplished
it. So I discount anything he wrote about baluns in the first edition.
But he figured it out some time between the first and second editions,
so there's some credible information in the second. A number of the 4:1
baluns he shows are indeed wound on a single core. And he dutifully
reports their frequency responses when terminated in a correct impedance
load (seldom the case in a real antenna application). But nowhere do I
see any measurement of their ability to do the primary job of a balun:
to balance the feedline currents. So there's no way of knowing whether
any or all of the many "current" baluns he constructed actually perform
the function of a current balun.

A way to qualitatively check to see if a balun is really acting like a
current balun is to connect a load across the output. Short one terminal
of the output to the "ground" terminal of the input, and measure the
impedance looking in. Then disconnect that short and short the other
output terminal to the "ground" terminal of the input and measure it
again. You can do this with an antenna analyzer or impedance bridge. A
perfect current balun will show the same, correct impedance in both
tests. A perfect voltage balun will show very different impedances.
There should be a way to get a quantitative measure of the longitudinal
(common mode) impedance from this test, but I don't have time to dig it
out just now.

I'd like to add that not long ago I encountered a claim of a 4:1 current
balun on a single core. It was indeed -- it was a conventional, as
opposed to a transmission line, transformer, although it was wound in a
way that made it resemble a transmission line transformer. And it had a
fairly decent bandwidth, although a transmission line transformer is
inherently much broader and doesn't couple differential flux into the
core. So a 4:1 current balun can be constructed on a single core. But I
don't know how a 4:1 transmission line current balun can.

What is happening here?

A whole lot of things could be happening. If an antenna is perfectly
symmetrical, a voltage balun will do the same job as a current balun.
For an explanation why, see
http://eznec.com/Amateur/Articles/Baluns.pdf. So a lot of the "current
baluns" might well be voltage baluns. I seldom see anyone actually test
the current balancing properties of baluns, and without this, there's no
telling what's been constructed.

I haven't looked carefully at Sevick's single core 4:1 "Guanella"
baluns. They might be voltage baluns, or they might be conventional
transformers -- bifilar windings don't automatically create a
transmission line transformer. I don't see any test he ran to verify
that they're functioning as current transformers or for that matter as
transmission line transformers. In a transmission line transformer, the
differential current always has a transmission line path to follow. Some
connections can split the wires carrying a differential current apart,
in which case you no longer have a transmission line transformer.

If you know of a topology which produces a transmission line 4:1 current
balun, I'd be very interested. You could use the same simple analysis
method used in the balun article to show that it really is what you think.

Mike Coslo wrote:

I'm not sure myself, Gene. I've wound single core 4:1's, I've bought
at least 2 of them. I've also been told variously they don't work, and
can't even be made.

Just what IS happening here?

You probably made 4:1 voltage baluns, which worked well enough in your
applications. Where a voltage balun is really a bad idea is when your
antenna system isn't symmetrical. In an effort to balance the voltages
on the two output terminals, it creates unbalanced currents. The
difference between them becomes a common mode feedline current. This is
also explained in the balun article.

. . . Here Uncle Remus paused, and drew a two-pound yam out of the ashes.

"Did the fox eat the rabbit?" asked the little boy to whom the story had
been told.

"Dat's all de fur de tale goes," replied the old man. "He mout, an den
agin he moutent. Some say Judge B'ar come 'long en loosed 'im - some say
he didn't. I hear Miss Sally callin'. You better run 'long."

Roy Lewallen, W7EL

Gene Fuller wrote:
Hi Roy,

Sorry about making such a mess. That tar is really hard to clean up.

FYI, I was referring mostly to Sevick's 4th edition of Transmission Line
Transformers, published in 2001. As you mention, he has undergone a
considerable transformation (pun intended) since his earliest books.

I'm glad to hear that. The first edition sold a lot of copies and
contained a lot of misinformation. The second was better, and hopefully
undid a little of the damage the first did. Maybe the net result was
neutral by the time the fourth was published.

Sevick does indeed recommend a two-core approach for a 4:1 Guanella
balun if the load side is not balanced.

That's one of those improvements over the first and second editions.

As you know, he generally avoids
the use of the terms "current balun" and "voltage balun".

Yes, I think that's mostly because they're the terms I coined and used.

My question still holds, however. In a typical situation the "balanced"
side, generally the antenna, is only partially unbalanced, not simply
shorted to ground on one side. In this generic case would you expect the
single core Guanella balun to perform poorly with slight or modest
imbalance?

I'm sorrry, I don't know, because I haven't looked carefully at his
"Guanella" balun to see if it's a current balun, voltage balun, or
something in between. As I pointed out in my balun article many years
ago, if an antenna is perfectly symmetrical, voltage and current baluns
perform equally. If it isn't, the voltage balun can make things even worse.

If both sides of the transformer are completely unbalanced it is not
clear why one would expect a balun of any type to be useful. What is
needed in that case is impedance transformation, not balancing.

Have you read the article at
http://eznec.com/Amateur/Articles/Baluns.pdf? If you have but still
think that, then I certainly failed to communicate.

For the benefit of readers who have read the article, I didn't mention a
second source of imbalance: mutual coupling to the antenna. This can
occur if the feedline isn't run completely symmetrically relative to the
antenna. And, interestingly, there are situations where this is
occurring that inserting a balun can actually increase the common mode
current. That's a story for another article some day. But first I've got
to somehow get people to read the first one.

Roy Lewallen, W7EL

Roy Lewallen wrote:


Have you read the article at
http://eznec.com/Amateur/Articles/Baluns.pdf? If you have but still
think that, then I certainly failed to communicate.

For the benefit of readers who have read the article, I didn't mention a
second source of imbalance: mutual coupling to the antenna. This can
occur if the feedline isn't run completely symmetrically relative to the
antenna. And, interestingly, there are situations where this is
occurring that inserting a balun can actually increase the common mode
current. That's a story for another article some day. But first I've got
to somehow get people to read the first one.

Roy Lewallen, W7EL

Hi Roy,

Yes, I have a copy of your article, and I have actually read it. I think
I even understand it. I took a fresh look just before I made my first
post in this thread. The schematic for Sevick's Guanella 4:1 balun is
exactly the same as your figure A3-4.

The only question I have is regarding the issue of two cores vs. one
core. I have no doubt that grounding one of the high impedance leads can
lead to improper choking action with a single core. In the completely
balanced case the single core should work perfectly. So what happens in
a more typical case where the antenna is in an unbalanced environment,
but not fully shorted to ground on one side?

There are a number of designs out there for single core 4:1 "current
baluns", all essentially the same, including the Elecraft balun. I have
not tried to perform any measurements, so I have no idea whether these
designs typically work or typically fail. That was the foundation of my
original comment and question.

73,
Gene
W4SZ

Gene Fuller wrote:

Yes, I have a copy of your article, and I have actually read it. I think
I even understand it. I took a fresh look just before I made my first
post in this thread. The schematic for Sevick's Guanella 4:1 balun is
exactly the same as your figure A3-4.

Then I'm puzzled by your statement that if an antenna is completely
symmetrical, no balun is necessary. The very first two figures in the
article and accompanying text show this isn't true. Of course, if you
have a symmetrical antenna AND the feedline is placed symmetrically AND
you have a truly balanced tuner, then there will be no common mode
current and a balun won't do anything. Maybe that's what you meant.

The only question I have is regarding the issue of two cores vs. one
core. I have no doubt that grounding one of the high impedance leads can
lead to improper choking action with a single core. In the completely
balanced case the single core should work perfectly. So what happens in
a more typical case where the antenna is in an unbalanced environment,
but not fully shorted to ground on one side?

It's absolutely true that the test I proposed is a worst case, much more
severe than any realistic application. But I'm sorry, I don't believe
there is a "typical" case -- there are much too many variables involved.
In many situations, a balun isn't used at all and the result is entirely
satisfactory. Others take rather heroic effort to keep the common mode
current tolerable.

There are a number of designs out there for single core 4:1 "current
baluns", all essentially the same, including the Elecraft balun. I have
not tried to perform any measurements, so I have no idea whether these
designs typically work or typically fail. That was the foundation of my
original comment and question.

Sorry, I've never seen or measured one. It would be great if someone
would run some balance tests and see how effective it is.

By the way, there seems to be a pervasive myth that a 4:1 balun or
transformer used in an antenna system actually effects a 4:1 impedance
transformation. You can make one look wonderful by terminating it with
the design impedance (typically 200 ohms resistive) and looking into the
input. But anyone wanting to have a real eye-opening experience should
measure the impedance of a multi-band antenna on several bands, then
look at the impedance at the input of that "4:1" transformer when
terminated with the actual impedance it sees. That's not to say the
transformer doesn't do something, or even that it might present an
impedance that's more (or less) favorable for a tuner to match. But
you'll find that the transformation ratio is often not only far from
4:1, but the transformer also introduces considerable reactance. Whether
this helps or hurts when it comes to matching is something of a crap shoot.

Roy Lewallen, W7EL

Roy Lewallen wrote:
Gene Fuller wrote:

Yes, I have a copy of your article, and I have actually read it. I
think I even understand it. I took a fresh look just before I made my
first post in this thread. The schematic for Sevick's Guanella 4:1
balun is exactly the same as your figure A3-4.

Then I'm puzzled by your statement that if an antenna is completely
symmetrical, no balun is necessary. The very first two figures in the
article and accompanying text show this isn't true. Of course, if you
have a symmetrical antenna AND the feedline is placed symmetrically AND
you have a truly balanced tuner, then there will be no common mode
current and a balun won't do anything. Maybe that's what you meant.

Roy,

If I said that no balun is necessary then I apologize. I certainly did
not intend to say or imply such a thing. I do believe, however that a
single core balun would serve well if the antenna is balanced by design
and by its environment.

The question is how the single core balun performance degrades as the
antenna becomes less balanced through, for example, an asymmetrically
placed feedline. It seems difficult to analyze that problem, and
experimentation would be required.

73,
Gene
W4SZ