Balun at input or output of tuner

[Dean Straw]

Balun at Input or Output of Antenna Tuner?
Dean Straw, N6BV (Senior Assistant Technical Editor, Retired) December 12,
2011
 
I have been lurking on the Elecraft Reflector monitoring the animated
discussion about where to place a balun -- at the input or the output of a
tuner. I was going to jump into the discussion after I read comments
claiming that a balun at the input of an unbalanced tuner feeding a balanced
transmission line (and balanced antenna) simply "doesn't work." On Dec. 8
Alan Bloom, N1AL, wrote such an elegant analysis that I didn't have to.

If I might paraphrase Alan, the stress across the balun due to unwanted
common-mode current is the same whether the balun is placed at the input or
at the output of an unbalanced antenna tuner network. Other investigators
have come to the same conclusion and have argued that the complexity of
"floating" the antenna tuner components doesn't warrant installing the balun
at the input of the antenna tuner.

As Alan pointed out, when a balun is placed at the input of the tuner, the
differential-mode currents/voltages are well controlled, since the tuner's
input impedance when tuned is 50 ohms. When the balun is placed at the
output of the tuner, the differential-mode currents/voltages are determined
by whatever impedance the load happens to be at the shack-end of the
transmission line feeding the antenna. And that impedance can range very
widely.
 
For my own curiosity, I wanted to get a handle on the amount of loss
incurred by the differential-mode loss (i.e., the loss "inside the coax")
for a common-mode choke constructed with coaxial cable. (Google Jim Brown,
K9YC's excellent paper "A Ham's Guide to RFI, Ferrites, Baluns, and Audio
Interfacing" Revision 5a, 5 Jun 2010. The section on transmitting baluns
starts on page 25. Photos of typical coaxial CM transmitting chokes appear
on page 29.)

By the way, like Jim Brown, K9YC, I prefer the name "common-mode choke"
rather than the term "balun," whether it is used at the input or at the
output of an unbalanced tuner feeding a balanced transmission line and
balanced antenna. I shall use the abbreviation "CM choke" in the rest of
this paper.

I assumed that the CM choke consisted of three feet of RG-213 wound through
ferrite torroids of the appropriate material and size to achieve K9YC''s
target common-mode impedance of 5000 ohms. At that choke impedance the
effects of unwanted common-mode current would be low and the effect of
differential-mode loss could be explored from a conventional
transmission-line perspective.

A quick review: a CM choke exhibits two main modes: (1) an impedance to
common-mode currents flowing on the outer surface of the coaxial shield and
(2) the effect of differential-mode currents/voltages - that is,
matched-line loss and the additional loss due to SWR of a three-foot length
of RG-213.

My tool of analysis is my own TLW program, but there are other programs out
there that can do the analysis. I assumed that the load presented to the
output terminals of the common-mode choke balun at the output of the tuner
was 3000 + j0 ohms at 28 MHz. This amounts to an SWR of 60:1.
 
TLW says that the total transmission-line loss in the three feet of RG-213
under these conditions is 0.373 dB. This is comprised of 0.034 dB of
matched-line loss if the RG-213 were matched and an additional loss due to
the 60:1 SWR of 0.339 dB. A total loss 0.373 dB doesn't sound like much, but
when you pump 1500 W into such a tuner this translates into a lot of watts
in a confined space. For 1500 W into the input of the CM choke, 123.5 W must
be dissipated safely in the three feet of RG-213. This is 41.2 W per foot
and at that level the RG-213 would get very warm and could even melt,
especially if the choke were confined in a small box with no circulation of
cooling air.

Times Microwave has a convenient calculator on their web site, where the
power-handling capability of various types of coaxes may be calculated. 100
feet of RG-213 is rated to handle 2.02 kW at 28 MHz with a 1:1 SWR. For a
matched line there would be no "hot spots" along the line due to SWR. The
average power-handling capability of RG-213 is thus 2020 W divided by 100
feet, or 20.2 W per foot.
 
Interestingly enough, this short length of coax transforms the 3000 + j 0
load to 1.72 - j 47.3 ohms at the output terminals of the tuner, making the
task of the antenna tuner a little more difficult at such a low resistance
level.
 
If this 3000-ohm load were presented to a tuner with the CM choke at the
input, the loss would be the matched-line loss only of 0.034 dB, which is a
power loss of 11.7 W for 1500 W input to the tuner.
 
The same kind of TLW analysis can be done for a low-impedance load of 3 + j
0 ohms, for an SWR of 16.67:1. For a CM choke balun at the output of a tuner
this results in an additional loss due to SWR of 0.402 dB and a total loss
of 0.436 dB. This is a power loss of 143.3 W in the three feet of RG-213,
meltdown city again.

        While we're at it, let's look at the potential loss due to line
losses at a CM choke balun placed in the wrong place in an antenna system.
Assume the common scenario where a balanced antenna is fed with open-wire
transmission line to a 1:1 common-mode choke balun located at the shack
window. From the balun at the window the ham uses, say, a 20-foot section of
RG-213 to the antenna tuner (which in this case is an unbalanced tuning
network). Assume again that the CM choke balun uses three feet of RG-213
wound on the appropriate ferrite donuts to achieve the target common-mode
impedance of 5000 ohms so that common-mode currents are choked off properly.

        The total length of RG-213 is now 23 feet. Again, we'll present the
balun at the windowsill with a load of 3000 ohms. The overall
differential-mode loss in 23 feet of RG-213 is 4.534 dB, nearly 4 dB worse
than connecting the open-wire line directly to a tuner with a CM choke balun
at its output! Ouch, that's a lot of wasted power.

73, Dean, N6BV

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Jim AB3CV]

Dean

Nice analysis of the issues. K9YC also presents baluns done with wire pair
rather than coax, RG213 or otherwise. Might not they be a better solution
than coax when constructing a CM choke for driving windowline?

Jim ab3cv
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Jim AB3CV]

Poor choice of words: first I fell into the "balun" habit rather than CM
choke on the first usage. Second: "bifilar wound CM choke" would have been
a better description for the alternative mentioned.

Jim ab3cv
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Eric Swartz - WA6HHQ]

Also note that in the case of the KPA500, we will be matching SWRs of
10:1 or less, which results in less loss for the short coax choke mentioned.

73, Eric

www.elecraft.com


On 12/12/2011 2:00 PM, Ron D'Eau Claire wrote:

> I would avoid coaxial cable in any situation where there can be a large
> impedance mismatch because of just the losses mentioned. Even though low
> power might not burn up 3 feet of coax, that's still a chunk of power thrown
> away.
>
> The ideal CM choke at the OUTPUT of a tuner feeding a balanced line would be
> made of open wire line. The old air-core chokes did just that but, to have
> the choking impedance needed, they are physically B-I-G, especially if used
> on 80 or 160 meters. Some look like 'steampunk' art with huge coils
> occupying a whole wall, not unlike what one sees in some photos of early
> spark transmitting stations.
>
> A parallel pair of wires wrapped on suitable toroid cores would be my first
> choice today, but never with coax between the choke and the matching
> network. With a little ingenuity one can make a very clean, neat open wire
> line installation running across a ceiling using small insulators attached
> to the ceiling if the operating position is far from the point of egress to
> the outside.
>
> 73,
>
> Ron AC7AC
>
> -----Original Message-----
>
> Dean
>
> Nice analysis of the issues. K9YC also presents baluns done with wire pair
> rather than coax, RG213 or otherwise. Might not they be a better solution
> than coax when constructing a CM choke for driving windowline?
>
> Jim ab3cv
>
> ______________________________________________________________
> Elecraft mailing list
> Home: http://mailman.qth.net/mailman/listinfo/elecraft
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:[hidden email]
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[P.B. Christensen]

> This is 41.2 W per foot
> and at that level the RG-213 would get very warm and could even melt,
> especially if the choke were confined in a small box with no circulation
> of
> cooling air.

If a balun of like reactance were instead wound in bifilar fashion using two
parallel wires, rather than coax, wouldn't "additional loss due to mismatch"
be less when compared to coaxial-wound types, assuming that the bifilar
winding choke offers lower matched and mis-matched loss, much like open
lines?  However, the bifilar winding is not really like a wide
conductor-separated transmission line so, perhaps mismatch losses are less
than the same length of RG-213, but more than open or balanced lines.  Most
of the coaxial-wound baluns I've seen of the W1JR type, have been designed
where the output Z much more closely matches the characteristic Z of the
coax used for the winding (e.g., CM chokes used as line isolators between
the transmitter and amp, or monoband dipoles at the feed-point).

I've never seen a coaxial-wound balun at a tuner output (within the tuner),
only bifilar types -- although I know so-called "remote baluns" exist that
do use coaxial turns.   Your example of the "remote balun" shows some
startling, but realistic loss when using a tuner and traditional wire
antennas for multiband HF operation.  If the losses are really that bad,
then it's the tuner that should be remoted, not just the output balun.  So,
it would be an interesting exercise to look at the mis-matched losses under
the two Z extremes when using a bifilar type and use this as a closer
approximation of loss for current choke/baluns placed inside the tuner, at
the output terminals.

Paul, W9AC

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Eric Swartz - WA6HHQ]

I meant to say KAT500..
73, Eric

www.elecraft.com


On 12/12/2011 2:05 PM, Eric Swartz - WA6HHQ, Elecraft wrote:
> Also note that in the case of the KPA500, we will be matching SWRs of
> 10:1 or less, which results in less loss for the short coax choke mentioned.
>
> 73, Eric
>
> www.elecraft.com
>
>
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[KEN-3]

FWIW, where there is 1 db of loss in a matched line, a 2:1 SWR adds an additional 0.2 db loss, a 4:1 SWR  adds about 0.8 db loss.  A 10:1 SWR results in an additional 2.5 db loss.

Ken WA8JXM

On Dec 12, 2011, at 5:00 PM, Ron D'Eau Claire wrote:

> I would avoid coaxial cable in any situation where there can be a large
> impedance mismatch because of just the losses mentioned. Even though low
> power might not burn up 3 feet of coax, that's still a chunk of power thrown
> away.

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Ignacy]

Dean,
Nice analysis.

One can take it a step further. Assume your 3 foot RG-213 balun that dissipates about 8% of power or about 40W for 500W input. It would contain some 36 beads and would weight perhaps 2 pounds.

An input balun with identical choke impedance would be a single bead with 6 turns of RG-174. One foot of RG-174 has 1% loss at 30 MHz and 0.25% at 2 Mhz. Also RG-174 is rated 200W average at 30 Mhz and 750W at 2 MHz. Most likely it would carry 500W intermittent on 10m when matched.

In this case, an output balun would weigh 36 times more, be 36 times more expensive, and have 8 times higher losses. Mind this is for extreme 3000 Ohm load.

I would not use an input balun for large stationary setup because stray capacitances and switching would make it less efficient and cumbersome. I would use an output balun in portable situations because it is so much lighter, has so low losses, and one can have very large common impedance by adding extra turns.

Ignacy, VK2/NO9E

[Jim Brown-10]

On 12/12/2011 2:06 PM, Paul Christensen wrote:
> However, the bifilar winding is not really like a wide
> conductor-separated transmission line so, perhaps mismatch losses are less
> than the same length of RG-213, but more than open or balanced lines.

I have not attempted to measure the Zo of the bifilar wound chokes I've
built using #12 and #14 THHN, but Jerry Sevick, in the last of his
books, did wind some using exactly that method and that wire, and he
says the Zo of those he wound were about 100 ohms. As I often do, I
wound mine and studied them before seeing Jerry's work, so it's sort of
like independent researchers coming to similar conclusions. What I have
not seen (so far) in Jerry's work is the concept of common mode Z and
resonance as a useful thing.

The REALLY NICE thing about the bifilar chokes I've wound is that they
hit the target 5K ohm CM impedance over the entire HF ham bands -- that
is, 80M to 10M, or with a few more turns, 160M to 17M -- with a single
#31 core and very inexpensive wire.  The secret sauce is simply how the
the dielectric constant of the insulation meshes with the rest of the
design.  Another nice thing about these chokes is that they can handle a
lot of voltage and current.  I've patched them in line with my 1.5kW
Titans and the antenna tuner (that is, matched load, no common mode
voltage), run them keydown for a while, and they barely get warm to the
touch.

Yes, chokes like these would be a great choice for use at the feedpoint
of ANY HF antenna, whether fed with coax or parallel conductor line.  
The primary issue to deal with is how to rig them so that they are
secure mechanically.  I've not played with that yet, and mechanical
design is not my strong suit. :)  There are photos of the chokes I
measured in the RFI tutorial.

73,
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Jim Brown-10]

On 12/12/2011 5:10 PM, Jim Brown wrote:
> Yes, chokes like these would be a great choice for use at the feedpoint
> of ANY HF antenna, whether fed with coax or parallel conductor line.

One VERY important exception that I forgot to mention is that common
mode chokes can be destructively overheated by running high power into
badly unbalanced antennas. Off-center fed antennas like Windoms can
place VERY high common mode voltages across common mode chokes. The only
solution I know of is to use multiple chokes in series on the feedline
IMO, this sort of antenna is a poor choice in today's world, where local
RF noise is made worse by pickup on the feedline.

A few years ago, I investigated this by  modeling the common mode
voltage, and the resulting heat dissipation, in a common mode choke at
the feedpoint of a 40M dipole whose feedpoint was moved off center in
increments of 3 ft, at 1.5kW, fed by a half wavelength of coax (67 ft).  
At this worst case feedline length, you don't have to go very far off
center to produce a lot of heat in the choke.  The results are
summarized in a table in a Power Point for a presentation I've done for
several ham clubs.  One of these bifilar chokes would be OK on a Windom
at the 600W level produced by the KPA500, but could fail at max legal
power. See page 43 of

http://audiosystemsgroup.com/CoaxChokesPPT.pdf

73, Jim Brown K9YC
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Dean Straw]

Jim Brown said:
Mon, 12 Dec 2011 17:10:36 -0800
 
> I have not attempted to measure the Zo of the bifilar wound chokes I've
> built using #12 and #14 THHN, but Jerry Sevick, in the last of his
> books, did wind some using exactly that method and that wire, and he
> says the Zo of those he wound were about 100 ohms.

This is a useful data point. (I've got to rummage through my library to find
the Sevick book.)

I used a bifilar wound CM choke at the input of the ARRL high-powered tuner
described in late editions of "The ARRL Antenna Book." It had 12 bifilar
turns of #10 AWG Formvar wire on a 24-inch diameter OD Type 43 core.
(Nowadays I'd probably use a more optimal Type 31 mix.) In testing the input
balun (aka CM choke) 1500 W of RF at 29.7 MHz was applied for 60 seconds.
The #10 wire in the balun got warm to the touch (after the RF was shut off!)
but the core remained cool, as it should when there are no common-mode
currents, only differential-mode current in the bifilar-wound transmission
line.

Now, #10 wire is roughly the same size as the inner conductor used in
RG-213. On 10 meters the majority of loss in the bifilar transmission line
wound around the torroid will be I-squared-R conductor loss, rather than
additional dielectric losses that come into effect in the VHF and UHF
regions. So, I then assume that the matched-line loss in the bifilar-wound
transmission line is the same as that for RG-213 at HF so that I can do
computations using TLW.

I then used the "User-Defined Transmission Lines" capability in TLW as
follows: Frequency = 28.0 MHz; Matched-Line Attenuation, dB/100 Feet =
1.142, Velocity Factor = 0.95; R0 = 100 ohms; Computed X0 = -0.698 ohms.
Again, a total length of three feet is assumed for the bifilar-wound
transmission line.

For a 3000 + j 0 load, TLW reports additional line loss due to SWR (which is
30:1) of 0.416 dB, a power loss in the balun  of 137.0 W for a 1500-W
transmitter. This level of dissipation in a physically small package will
result in catostrophic destruction when the balun is placed at the output of
the tuner.

For a 3 + j 0 ohm load, the SWR is 33.33:1, and the total line loss is 0.449
dB, amounting to 147.3 W dissipation in the balun -- again, this amount of
power in the CM choke balun would surely destroy it. The use a a
bifilar-wound transmission line instead of RG-213 has resulted in a slightly
greater susceptibility to catosphrophic destruction at low-impedance loads
when the balun is placed at the output of the tuner.

For a 5 + j 0 load (a 10:1 SWR), the total line loss is 0.274 dB, which for
1500 W is 91.7 W for 1500 W input, or 30.6 W for 500 W RF input. This would
be about the limit of safe operation for a CM choke balun placed at the
output terminals of an antenna tuner.

73, Dean, N6BV


______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Jim Brown-10]

On 12/13/2011 10:43 AM, Dean Straw wrote:
> I used a bifilar wound CM choke at the input of the ARRL high-powered tuner
> described in late editions of "The ARRL Antenna Book." It had 12 bifilar
> turns of #10 AWG Formvar wire on a 24-inch diameter OD Type 43 core.

Jerry said that Zo for this sort of wire and winding style was about 50
ohms.  I wound some that way, and they acted like they were close to 50
ohms (as observed by SWR measurements) when inserted between TX and tuner.

73, Jim K9YC
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Joe Subich, W4TV-4]

> I used a bifilar wound CM choke at the input of the ARRL
> high-powered tuner described in late editions of "The ARRL Antenna
> Book." It had 12 bifilar turns of #10 AWG Formvar wire on a 24-inch
> diameter OD Type 43 core. (Nowadays I'd probably use a more optimal
> Type 31 mix.) In testing the input balun (aka CM choke) 1500 W of RF
> at 29.7 MHz was applied for 60 seconds. The #10 wire in the balun
> got warm to the touch (after the RF was shut off!) but the core
> remained cool, as it should when there are no common-mode currents,
> only differential-mode current in the bifilar-wound transmission
> line.

Moving this discussion away from the tuner and to the feedpoint of the
antenna ... I would never use a bifilar wound CM choke with a high HF
antenna.  Years ago I tried to use a well known, third party high power
"balun" on a triband antenna with a reputation for blowing its OEM
(fuse) balun.  That attempt was spectacularly unsuccessful on 15 meters
where the 90-100 Ohm Zo of the bifilar winding coupled with a "line
length" of slightly over 12 feet transformed the normally benign 50
Ohm SWR of the antenna into something that was poor across the entire
band.

With an antenna supporting more than three bands, it is likely that
the transformer effect would impact at least one band!

73,

    ... Joe, W4TV


On 12/13/2011 1:43 PM, Dean Straw wrote:

>
> Jim Brown said:
> Mon, 12 Dec 2011 17:10:36 -0800
>
>> I have not attempted to measure the Zo of the bifilar wound chokes I've
>> built using #12 and #14 THHN, but Jerry Sevick, in the last of his
>> books, did wind some using exactly that method and that wire, and he
>> says the Zo of those he wound were about 100 ohms.
>
> This is a useful data point. (I've got to rummage through my library to find
> the Sevick book.)
>
> I used a bifilar wound CM choke at the input of the ARRL high-powered tuner
> described in late editions of "The ARRL Antenna Book." It had 12 bifilar
> turns of #10 AWG Formvar wire on a 24-inch diameter OD Type 43 core.
> (Nowadays I'd probably use a more optimal Type 31 mix.) In testing the input
> balun (aka CM choke) 1500 W of RF at 29.7 MHz was applied for 60 seconds.
> The #10 wire in the balun got warm to the touch (after the RF was shut off!)
> but the core remained cool, as it should when there are no common-mode
> currents, only differential-mode current in the bifilar-wound transmission
> line.
>
> Now, #10 wire is roughly the same size as the inner conductor used in
> RG-213. On 10 meters the majority of loss in the bifilar transmission line
> wound around the torroid will be I-squared-R conductor loss, rather than
> additional dielectric losses that come into effect in the VHF and UHF
> regions. So, I then assume that the matched-line loss in the bifilar-wound
> transmission line is the same as that for RG-213 at HF so that I can do
> computations using TLW.
>
> I then used the "User-Defined Transmission Lines" capability in TLW as
> follows: Frequency = 28.0 MHz; Matched-Line Attenuation, dB/100 Feet =
> 1.142, Velocity Factor = 0.95; R0 = 100 ohms; Computed X0 = -0.698 ohms.
> Again, a total length of three feet is assumed for the bifilar-wound
> transmission line.
>
> For a 3000 + j 0 load, TLW reports additional line loss due to SWR (which is
> 30:1) of 0.416 dB, a power loss in the balun  of 137.0 W for a 1500-W
> transmitter. This level of dissipation in a physically small package will
> result in catostrophic destruction when the balun is placed at the output of
> the tuner.
>
> For a 3 + j 0 ohm load, the SWR is 33.33:1, and the total line loss is 0.449
> dB, amounting to 147.3 W dissipation in the balun -- again, this amount of
> power in the CM choke balun would surely destroy it. The use a a
> bifilar-wound transmission line instead of RG-213 has resulted in a slightly
> greater susceptibility to catosphrophic destruction at low-impedance loads
> when the balun is placed at the output of the tuner.
>
> For a 5 + j 0 load (a 10:1 SWR), the total line loss is 0.274 dB, which for
> 1500 W is 91.7 W for 1500 W input, or 30.6 W for 500 W RF input. This would
> be about the limit of safe operation for a CM choke balun placed at the
> output terminals of an antenna tuner.
>
> 73, Dean, N6BV
>
>
> ______________________________________________________________
> Elecraft mailing list
> Home: http://mailman.qth.net/mailman/listinfo/elecraft
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:[hidden email]
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html
>

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Alan Bloom]

> Now, #10 wire is roughly the same size as the inner conductor used in
> RG-213. On 10 meters the majority of loss in the bifilar transmission
> line wound around the torroid will be I-squared-R conductor loss,
> rather than additional dielectric losses that come into effect in
> the VHF and UHF regions. So, I then assume that the matched-line
> loss in the bifilar-wound transmission line is the same as that for
> RG-213 at HF so that I can do computations using TLW.

Is that a valid assumption?  I thought that much of the loss in coax is
due to the dielectric loss of the insulation.  That implies that the
bifilar winding should have less loss than coax.

Alan


On Tue, 2011-12-13 at 10:43 -0800, Dean Straw wrote:

> Jim Brown said:
> Mon, 12 Dec 2011 17:10:36 -0800
>  
> > I have not attempted to measure the Zo of the bifilar wound chokes I've
> > built using #12 and #14 THHN, but Jerry Sevick, in the last of his
> > books, did wind some using exactly that method and that wire, and he
> > says the Zo of those he wound were about 100 ohms.
>
> This is a useful data point. (I've got to rummage through my library to find
> the Sevick book.)
>
> I used a bifilar wound CM choke at the input of the ARRL high-powered tuner
> described in late editions of "The ARRL Antenna Book." It had 12 bifilar
> turns of #10 AWG Formvar wire on a 24-inch diameter OD Type 43 core.
> (Nowadays I'd probably use a more optimal Type 31 mix.) In testing the input
> balun (aka CM choke) 1500 W of RF at 29.7 MHz was applied for 60 seconds.
> The #10 wire in the balun got warm to the touch (after the RF was shut off!)
> but the core remained cool, as it should when there are no common-mode
> currents, only differential-mode current in the bifilar-wound transmission
> line.
>
> Now, #10 wire is roughly the same size as the inner conductor used in
> RG-213. On 10 meters the majority of loss in the bifilar transmission line
> wound around the torroid will be I-squared-R conductor loss, rather than
> additional dielectric losses that come into effect in the VHF and UHF
> regions. So, I then assume that the matched-line loss in the bifilar-wound
> transmission line is the same as that for RG-213 at HF so that I can do
> computations using TLW.
>
> I then used the "User-Defined Transmission Lines" capability in TLW as
> follows: Frequency = 28.0 MHz; Matched-Line Attenuation, dB/100 Feet =
> 1.142, Velocity Factor = 0.95; R0 = 100 ohms; Computed X0 = -0.698 ohms.
> Again, a total length of three feet is assumed for the bifilar-wound
> transmission line.
>
> For a 3000 + j 0 load, TLW reports additional line loss due to SWR (which is
> 30:1) of 0.416 dB, a power loss in the balun  of 137.0 W for a 1500-W
> transmitter. This level of dissipation in a physically small package will
> result in catostrophic destruction when the balun is placed at the output of
> the tuner.
>
> For a 3 + j 0 ohm load, the SWR is 33.33:1, and the total line loss is 0.449
> dB, amounting to 147.3 W dissipation in the balun -- again, this amount of
> power in the CM choke balun would surely destroy it. The use a a
> bifilar-wound transmission line instead of RG-213 has resulted in a slightly
> greater susceptibility to catosphrophic destruction at low-impedance loads
> when the balun is placed at the output of the tuner.
>
> For a 5 + j 0 load (a 10:1 SWR), the total line loss is 0.274 dB, which for
> 1500 W is 91.7 W for 1500 W input, or 30.6 W for 500 W RF input. This would
> be about the limit of safe operation for a CM choke balun placed at the
> output terminals of an antenna tuner.
>
> 73, Dean, N6BV
>
>
> ______________________________________________________________
> Elecraft mailing list
> Home: http://mailman.qth.net/mailman/listinfo/elecraft
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:[hidden email]
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html
>

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Dean Straw]

Joe:

Right on -- certain unnamed baluns had a quite reputation as being RF fuses.

But as the suject title above says, I'm still talking about the pros and
cons of placing a CM choke balun at the input or at the output of an
unbalancing antena tuner to feed balanced lines. Both positions are valid
ones, and like most engineering matters there are tradeoffs to both
approaches. Some tradeoffs involve significant smoke and flames... !

73, Dean, N6BV

-----Original Message-----
From: Joe Subich, W4TV [mailto:[hidden email]]
Sent: Tuesday, December 13, 2011 11:21 AM
To: Dean Straw
Cc: [hidden email]; [hidden email]
Subject: Re: [Elecraft] Balun at input or output of tuner


> I used a bifilar wound CM choke at the input of the ARRL high-powered
> tuner described in late editions of "The ARRL Antenna Book." It had 12
> bifilar turns of #10 AWG Formvar wire on a 24-inch diameter OD Type 43
> core. (Nowadays I'd probably use a more optimal Type 31 mix.) In
> testing the input balun (aka CM choke) 1500 W of RF at 29.7 MHz was
> applied for 60 seconds. The #10 wire in the balun got warm to the
> touch (after the RF was shut off!) but the core remained cool, as it
> should when there are no common-mode currents, only differential-mode
> current in the bifilar-wound transmission line.

Moving this discussion away from the tuner and to the feedpoint of the
antenna ... I would never use a bifilar wound CM choke with a high HF
antenna.  Years ago I tried to use a well known, third party high power
"balun" on a triband antenna with a reputation for blowing its OEM
(fuse) balun.  That attempt was spectacularly unsuccessful on 15 meters
where the 90-100 Ohm Zo of the bifilar winding coupled with a "line length"
of slightly over 12 feet transformed the normally benign 50 Ohm SWR of the
antenna into something that was poor across the entire band.

With an antenna supporting more than three bands, it is likely that the
transformer effect would impact at least one band!

73,

    ... Joe, W4TV


On 12/13/2011 1:43 PM, Dean Straw wrote:

>
> Jim Brown said:
> Mon, 12 Dec 2011 17:10:36 -0800
>
>> I have not attempted to measure the Zo of the bifilar wound chokes
>> I've built using #12 and #14 THHN, but Jerry Sevick, in the last of
>> his books, did wind some using exactly that method and that wire, and
>> he says the Zo of those he wound were about 100 ohms.
>
> This is a useful data point. (I've got to rummage through my library
> to find the Sevick book.)
>
> I used a bifilar wound CM choke at the input of the ARRL high-powered
> tuner described in late editions of "The ARRL Antenna Book." It had 12
> bifilar turns of #10 AWG Formvar wire on a 24-inch diameter OD Type 43

core.
> (Nowadays I'd probably use a more optimal Type 31 mix.) In testing the
> input balun (aka CM choke) 1500 W of RF at 29.7 MHz was applied for 60
seconds.

> The #10 wire in the balun got warm to the touch (after the RF was shut
> off!) but the core remained cool, as it should when there are no
> common-mode currents, only differential-mode current in the
> bifilar-wound transmission line.
>
> Now, #10 wire is roughly the same size as the inner conductor used in
> RG-213. On 10 meters the majority of loss in the bifilar transmission
> line wound around the torroid will be I-squared-R conductor loss,
> rather than additional dielectric losses that come into effect in the
> VHF and UHF regions. So, I then assume that the matched-line loss in
> the bifilar-wound transmission line is the same as that for RG-213 at
> HF so that I can do computations using TLW.
>
> I then used the "User-Defined Transmission Lines" capability in TLW as
> follows: Frequency = 28.0 MHz; Matched-Line Attenuation, dB/100 Feet =
> 1.142, Velocity Factor = 0.95; R0 = 100 ohms; Computed X0 = -0.698 ohms.
> Again, a total length of three feet is assumed for the bifilar-wound
> transmission line.
>
> For a 3000 + j 0 load, TLW reports additional line loss due to SWR
> (which is
> 30:1) of 0.416 dB, a power loss in the balun  of 137.0 W for a 1500-W
> transmitter. This level of dissipation in a physically small package
> will result in catostrophic destruction when the balun is placed at
> the output of the tuner.
>
> For a 3 + j 0 ohm load, the SWR is 33.33:1, and the total line loss is
> 0.449 dB, amounting to 147.3 W dissipation in the balun -- again, this
> amount of power in the CM choke balun would surely destroy it. The use
> a a bifilar-wound transmission line instead of RG-213 has resulted in
> a slightly greater susceptibility to catosphrophic destruction at
> low-impedance loads when the balun is placed at the output of the tuner.
>
> For a 5 + j 0 load (a 10:1 SWR), the total line loss is 0.274 dB,
> which for
> 1500 W is 91.7 W for 1500 W input, or 30.6 W for 500 W RF input. This
> would be about the limit of safe operation for a CM choke balun placed
> at the output terminals of an antenna tuner.
>
> 73, Dean, N6BV
>
>
> ______________________________________________________________
> Elecraft mailing list
> Home: http://mailman.qth.net/mailman/listinfo/elecraft
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:[hidden email]
>
> This list hosted by: http://www.qsl.net Please help support this email
> list: http://www.qsl.net/donate.html
>

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[David Herring-3]

Aloha Jim, Joe, Alan, Dean and the rest of the reflector,

I'm thoroughly enjoying reading this thread.

I'd like to ask a few questions.

Mine is a 100 watt station, but with aspirations of getting a KPA-500 eventually. I have a Palstar BT1500A Balanced L Antenna Tuner feeding a vertical dipole fed with 450 window line.  The antenna is crafted from one continuous length of window line, the dipole being formed by splitting the last 30-some-odd feet of the window line and attaching it to a 40 foot fiberglass pole. This antenna is used only for 30 meters and higher. I believe Palstar puts their balun at the input of this tuner.  I have a fairly small amount of RF in the shack.

I read one of you advocate the use of a CM choke at the feed point of an antenna, even if it's being fed with a balanced feed line.  I had never considered that, so the notion of a cm choke with a balanced feed line is a new one to me (but seems logical enough). If I understand correctly, that will minimize common mode currents on the feed line, just as it would with coax, and that may very well help to minimize the little bit of RF in the shack I seem to have.

Then there was discussion of placing one at the output of the tuner, but then I read about the one at the tuner output being heat stressed and prone to fail ( ? )  

My initial thought was if I'm guarding from common mode currents, maybe one would want to put a CM choke at both the feed point and the tuner output.

My questions are basically these:  would it be advisable to use CM chokes at both positions? (for me, the one at the tuner output would likely be outside, thus electrically 5 feet from the physical tuner)  What about the heat dissipation and stress on the tuner output side CM choke - it sounds like a show stopper to me but maybe you can provide some perspective on this?  If in this case only one cm choke is necessary or recommended, are we able to come to some consensus as to which position is best, at least in general?

Again, I'm thoroughly enjoying this thread and thanks a lot for sharing this with us.

73 & Aloha,

Dave
AH6TD


______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Guy, K2AV]

Jerry Sevick used a monster T400A-2 toroid, #2 powdered iron, 4 inches in
diameter, intended to be the 4:1 up transformation for the back of high
power tuners. (Sevick, Understanding Baluns.... 2003, CQ Communications, pp
60-61)  I have run these some times with brick-on-key 1500 watts and never
managed to get heat.  I've never personally managed to construct anything
that would stress one of these.

I'm currently using a 17 turn trifilar winding on a T400A-2 as a 4:1
isolation transformer (not a balun, no direct connection between primary
and secondary) feeding the 90 ohm base of my 160m 3/8 wave inverted L plus
folded counterpoise to 360 ohm "450" window line (Wireman #554).
 Particularly with the significant capacitive reactance of the
counterpoise, I was definitely expecting this would put some serious heat
on the core QRO, and maybe invalidate the concept, but I have gone 15 min
QRO BOK, immediately walked out to the base, and the core was stone cold.
 The whole thing seemed cold. There was a little bit of condensation
visible inside the teflon tubing beforehand, and the BOK did not cause it
to evaporate.  I really don't know why it didn't heat up, but I'll take it.
 Thing is a killer ant.

So I'm thinking if you put up Jerry's 20 turn bifilar on a T400A2 as a
Ruthroff balun and slap it on the back of a tuner, that you're going to be
very hard pressed to warm it up with ordinary stuff.

73, Guy.

On Tue, Dec 13, 2011 at 1:43 PM, Dean Straw <[hidden email]> wrote:

>
> Jim Brown said:
> Mon, 12 Dec 2011 17:10:36 -0800
>
> > I have not attempted to measure the Zo of the bifilar wound chokes I've
> > built using #12 and #14 THHN, but Jerry Sevick, in the last of his
> > books, did wind some using exactly that method and that wire, and he
> > says the Zo of those he wound were about 100 ohms.
>
> This is a useful data point. (I've got to rummage through my library to
> find
> the Sevick book.)
>
> I used a bifilar wound CM choke at the input of the ARRL high-powered tuner
> described in late editions of "The ARRL Antenna Book." It had 12 bifilar
> turns of #10 AWG Formvar wire on a 24-inch diameter OD Type 43 core.
> (Nowadays I'd probably use a more optimal Type 31 mix.) In testing the
> input
> balun (aka CM choke) 1500 W of RF at 29.7 MHz was applied for 60 seconds.
> The #10 wire in the balun got warm to the touch (after the RF was shut
> off!)
> but the core remained cool, as it should when there are no common-mode
> currents, only differential-mode current in the bifilar-wound transmission
> line.
>
> Now, #10 wire is roughly the same size as the inner conductor used in
> RG-213. On 10 meters the majority of loss in the bifilar transmission line
> wound around the torroid will be I-squared-R conductor loss, rather than
> additional dielectric losses that come into effect in the VHF and UHF
> regions. So, I then assume that the matched-line loss in the bifilar-wound
> transmission line is the same as that for RG-213 at HF so that I can do
> computations using TLW.
>
> I then used the "User-Defined Transmission Lines" capability in TLW as
> follows: Frequency = 28.0 MHz; Matched-Line Attenuation, dB/100 Feet =
> 1.142, Velocity Factor = 0.95; R0 = 100 ohms; Computed X0 = -0.698 ohms.
> Again, a total length of three feet is assumed for the bifilar-wound
> transmission line.
>
> For a 3000 + j 0 load, TLW reports additional line loss due to SWR (which
> is
> 30:1) of 0.416 dB, a power loss in the balun  of 137.0 W for a 1500-W
> transmitter. This level of dissipation in a physically small package will
> result in catostrophic destruction when the balun is placed at the output
> of
> the tuner.
>
> For a 3 + j 0 ohm load, the SWR is 33.33:1, and the total line loss is
> 0.449
> dB, amounting to 147.3 W dissipation in the balun -- again, this amount of
> power in the CM choke balun would surely destroy it. The use a a
> bifilar-wound transmission line instead of RG-213 has resulted in a
> slightly
> greater susceptibility to catosphrophic destruction at low-impedance loads
> when the balun is placed at the output of the tuner.
>
> For a 5 + j 0 load (a 10:1 SWR), the total line loss is 0.274 dB, which for
> 1500 W is 91.7 W for 1500 W input, or 30.6 W for 500 W RF input. This would
> be about the limit of safe operation for a CM choke balun placed at the
> output terminals of an antenna tuner.
>
> 73, Dean, N6BV
>
>
> ______________________________________________________________
> Elecraft mailing list
> Home: http://mailman.qth.net/mailman/listinfo/elecraft
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:[hidden email]
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html
>

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Jim Brown-10]

On 12/13/2011 11:48 AM, Alan Bloom wrote:
> Is that a valid assumption?  I thought that much of the loss in coax is
> due to the dielectric loss of the insulation.  That implies that the
> bifilar winding should have less loss than coax.

This is a very common misconception, and it is VERY wrong below UHF for
nearly all practical transmission lines that aren't defective (for
example, a wet dielectric). If you do the math, you see that below UHF,
the loss is virtually ALL due to copper (taking skin effect into account
for both conductors).  There's an excellent paper by Frank Witt in one
of the ARRL Antenna Compendiums (which Dean also edited) showing that
window line exhibits significant dielectric loss at HF when it gets wet.

You can see the equation for coax on datasheets for Times LMR coax types
on their website, with the equation for each cable type reflecting the
physical constants for that particular cable. There are two terms, one
for copper loss, the other for dielectric loss. Measured data for a few
cables that I've measured track those computed curves, and if you put
them into a spreadsheet and plot the two terms vs frequency, you can
clearly see which terms are contributing.  I suspect that they are also
used in Dean's TLW program. Right, Dean?

73, Jim K9YC
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[gm3sek]

Jim Brown wrote:

>On 12/13/2011 11:48 AM, Alan Bloom wrote:
>> Is that a valid assumption?  I thought that much of the loss in coax is
>> due to the dielectric loss of the insulation.  That implies that the
>> bifilar winding should have less loss than coax.
>
>This is a very common misconception, and it is VERY wrong below UHF for
>nearly all practical transmission lines that aren't defective (for
>example, a wet dielectric). If you do the math, you see that below UHF,
>the loss is virtually ALL due to copper (taking skin effect into account
>for both conductors).

Much of the confusion arises from the advertising for newer types of
coax that have lower loss than a similar solid PE equivalent. The
improvement is touted as being due to "low loss foam dielectric" when
that simply isn't true.

The reduction in loss is almost entirely due to increase in the diameter
of the center conductor (because that conductor has the largest current
density and hence the highest skin effect losses). The foam dielectric
is merely something that *has* to be used to compensate for the thicker
center conductor, in order to keep the same characteristic impedance.

In all the coaxial cables we know in amateur radio, dielectric losses
only begin to become important at frequencies above 1GHz.


--

73 from Ian GM3SEK
http://www.ifwtech.co.uk/g3sek
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Jim Brown-10]

On 12/13/2011 2:47 PM, Ian White GM3SEK wrote:
> The reduction in loss is almost entirely due to increase in the diameter
> of the center conductor (because that conductor has the largest current
> density and hence the highest skin effect losses). The foam dielectric
> is merely something that*has*  to be used to compensate for the thicker
> center conductor, in order to keep the same characteristic impedance.

Exactly!

Jim
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html