Losses can get quite high in a common mode choke in a line that is
mismatched to it's load. Count on destructive failure with high power.
Yes, the average chokes embodied in the average amateur market
matching networks (a.k.a., antenna "tuner") are pretty bad. In most
cases, they aren't baluns, but just transformers. I have
successfully melted many in my decades past learning days
(which never ends, thank heaven). Losses present themselves as heat.
That is why I have stacked cores in my common mode chokes to handle
the induced magnetic currents within the cores (I made all my measured
data
available on the NANOVNA Users google.io group a couple of months
ago). Losses are also why I finally went to AWG #12 conductor.
Key down in RTTY at 1 kW output from the Ameritron AL-82 on 40-meters
into the antenna / feedline combination for 5-minutes yields a
temperature rise of less than
10-degrees above ambient. The measured raw impedance in the shack at
7.0 MHz
is 1161 - j 1110. With the 3-stacked 2.4" 31 material core in line (Zo
of
the core measures 77.5 ohms) the impedance presented to the matching
network
measures 17.2 - j 170. I've also measured the common mode
losses in the chokes using the HP 8753C. All come in less than
0.5 dB through 17-meters. I don't care much for anything above that
or in general, 20-meters.
My CMCs are constructed using either 240-31
> (2.4-inch OD) or 400-31 (4.0-inch OD)
I really object to the 240 and 400 nomenclature, which was conceived
several decades ago by unscrupulous vendors to sell standard Fair-Rite
products to hams at obscene markups.
I am not a fan of it either and usually don't use it. This time
around I decided to conform. Well, not quite as I noted the
outside diameter as well.
material and are wound in a bifilar
> manner using AWG #12 Teflon insulated stranded conductors.
Do you mean as a 2-wire transmission line, or as a transformer?
"Bifilar" implies a magnetically coupled transformer.
Please, absolutely no transformers!!!! A transformer has no
electrical or physical operation for balancing the differential
mode outputs properly. If the load (antenna / feedline combination)
is perfectly balanced, they work as assumed. However, any set
of conductors deployed in the amateur 'back yard' is highly likely
to be unbalanced. A transformer is not the correct circuit element
for properly addressing this inherent unbalance. A transformer
is n o t a balun!! Yes, the common mode chokes
are wound as a parallel conductor transmission line
on the cores. I have never considered "bifilar" to imply
a transformer. Maybe in some low frequency applications,
but not in this shack at radio frequencies.
I do not use
> coaxial cable wound on these cores due to constraints the long doublet
plus
> feedline places on impedances in the shack (7.000 MHz measures 3.6 - j 62
> with the coaxial choke inline).
A closely spaced insulated 2-wire line has Zo on the order of 100 ohms.
I have measured several of my bifilar wound common mode chokes
for Zo (characteristicimpedance or 'shock' impedance). They range
from 77.5 to 95.9 Ω. The Vp (propagation velocity) on the few I
measured
for that parameter come in around 0.595.
The CMC choke was placed between the
> output of the homebrew L-network matching network and the input to the
> parallel conductor feedline in the shack.
The most effective location for a common mode choke (Please use WORDS,
not abbreviations) is at the antenna's feedpoint, because noise current
on the feedline is coupled to the antenna because the feedline becomes
part of the antenna without the choke).
Yes, I appreciate that for common coaxially fed dipoles. However,
every installation
is unique to a greater or lesser extent. I am not using coax and have
rather extensively
measured my system using the HP 8753C. Since I can not attach
anything, I have the data,
but am unable to share. The noise current is coupled common mode from
the
feedline onto the antenna with coax or open wire This problematic
noise current resides on the
outside of the coaxial shield only. Even with parallel conductor
line, it couples common
mode onto *both* conductors and equally (common mode) onto the antenna
conductors.
In that respect, using parallel conductor transmission line, the choke
placement between
the output of the matching network and the feedline is just as
effective as being placed
at the antenna feedpoint. The same can n o t be written for
coaxial feed as the inner conductor
is isolated from the common mode currents.
The results are documented in
> the attachment. Let the data speak for itself.
Since we can not attach anything, too bad....., I'll summarize: The
measured reduction
in broadband noise (common mode) with the choke in line was three
S-Units on the Icom
7300. Since each S-Unit on the 7300 is 3 dB, the noise reduction
measured a solid 9 dB.
I have verified the 3 dB / S-Unit using the HP 8648C signal generator
for both the 7300 and 7610.
In addition, Icom designed the radios to perform at 3 dB / S-Unit.
That makes sense as it represents power,
not voltage as the old standard addressed. The modern standard of
measuring this
parameter is in dBm, or power, not voltage, both in a 50 ± j 0
system. True, this looses
27 dB on the bottom end of the S-Meter between S-4.5 and S-0.
However, I'd rather deal with power
(which S-Parameters represent) than voltage. Power in dBm has become
the standard.
So much for that as it's a whole different subject.
>
> And I should add there are good and sound engineering reasons for this
> noise getting onto the feedline, but not the subject (could be rather long
> with math) of this email.
Little math is required, only an understanding of fundamental concepts.
EMC guru Henry Ott takes pride at breaking complex problems into simple
concepts, and is a master at it. In a three-day seminar I attended in
2005, there was almost no math at all.
True, but I've found that if any math is used, the average amateur's
eyes glaze over and he is gone to the ozone.
Very few amateurs have the technical basics to digest the Ott and
certainly the Howard Johnson courses. None
I have taken all of Henry Ott's courses in addition the four-day
Howard Johnson course and have
both his book - autographed. I well understand the concepts involved
and have the instruments to measure the
appropriate parameters.
73, Jim K9YC
Dave - WØLEV
On Fri, Apr 9, 2021 at 6:04 PM Jim Brown <jim@audiosystemsgroup.com> wrote:
> On 4/9/2021 10:19 AM, David Eckhardt wrote:
> > I see we can attach something to email within this group.
>
> Sorry, we cannot.
>
> > So, one afternoon on a whim, I made a test first with the CMC inline and
> > then w/o CMC on my parallel conductor transmission line to / from my
> > 450-foot long double.
>
> Losses can get quite high in a common mode choke in a line that is
> mismatched to it's load. Count on destructive failure with high power.
>
> My CMCs are constructed using either 240-31
> > (2.4-inch OD) or 400-31 (4.0-inch OD)
>
> I really object to the 240 and 400 nomenclature, which was conceived
> several decades ago by unscrupulous vendors to sell standard Fair-Rite
> products to hams at obscene markups.
>
> material and are wound in a bifilar
> > manner using AWG #12 Teflon insulated stranded conductors.
>
> Do you mean as a 2-wire transmission line, or as a transformer?
> "Bifilar" implies a magnetically coupled transformer.
>
> I do not use
> > coaxial cable wound on these cores due to constraints the long doublet
> plus
> > feedline places on impedances in the shack (7.000 MHz measures 3.6 - j 62
> > with the coaxial choke inline).
>
> A closely spaced insulated 2-wire line has Zo on the order of 100 ohms.
>
> The CMC choke was placed between the
> > output of the homebrew L-network matching network and the input to the
> > parallel conductor feedline in the shack.
>
> The most effective location for a common mode choke (Please use WORDS,
> not abbreviations) is at the antenna's feedpoint, because noise current
> on the feedline is coupled to the antenna because the feedline becomes
> part of the antenna without the choke).
>
> The results are documented in
> > the attachment. Let the data speak for itself.
> >
> > And I should add there are good and sound engineering reasons for this
> > noise getting onto the feedline, but not the subject (could be rather
> long
> > with math) of this email.
>
> Little math is required, only an understanding of fundamental concepts.
> EMC guru Henry Ott takes pride at breaking complex problems into simple
> concepts, and is a master at it. In a three-day seminar I attended in
> 2005, there was almost no math at all.
>
> 73, Jim K9YC
>
>
> _______________________________________________
> RFI mailing list
> RFI@contesting.com
> http://lists.contesting.com/mailman/listinfo/rfi
>
--
*Dave - WØLEV*
*Just Let Darwin Work*
_______________________________________________
RFI mailing list
RFI@contesting.com
http://lists.contesting.com/mailman/listinfo/rfi
|