[TowerTalk] BALUN revisited again

[Jim Brown]

On Sat,4/9/2016 10:03 AM, Jim Thomson wrote:
> Date: Sat, 9 Apr 2016 00:29:21 -0700
> From: Jim Brown <jim at audiosystemsgroup.com>
> To: towertalk at contesting.com
> Subject: Re: [TowerTalk] BALUN revisited again
>
> On Fri,4/8/2016 10:13 PM, Jim Thomson wrote:
>> ##  say what ?   AFIAIK, he uses  FOUR type 31 cores for the CMC-230-5K ,
>> 2 x cores  for each assy.
> Could be -- can't tell from the photos. That helps with power handling,
> and it also adds L, which lowers the resonance.
>
>>     8000-11,200 ohms of RS is nothing to sneeze at. Its
>> rated at 5 kw pep out.   It wont blow up with 1.5 kw cxr.  Its been tested. 2-30 mhz
>>
>> ##  The BL2006-3k  is a 14-54 mhz balun.  Single core.   5400- 8600 ohms of RS
>> is again, nothing to sneeze at.   Rated for 3 kw pep out.   It  wont blow up on 20-6m,
>> with 1.5 kw cxr.
> There's a serious fallacy here about power ratings for chokes that use
> the resistance at resonance (which is the right way to do it).
> Dissipation in this kind of choke depends STRONGLY on the common mode
> voltage, which in turn depends strongly on the antenna system, INCLUDING
> the feedline and its length.
>
>> ##  Your  single  core overheated on 80m.   His BL2006-3k  Is rated for 14-54 mhz.
>> Its not intended to be used on 80m!
> Operating frequency is only one factor. The choke I fried was more turns
> to move the resonance down so that it would cover 80M, and it provided
> about 5K ohms resistive. But that antenna SYSTEM needed at least twice
> that choking Z to run US legal limit.
>
> Dissipation in a choke is the SUM of the differential mode power and the
> common mode power. N6BV wrote an excellent piece that ran in QST a
> couple years ago on the issue of differential mode dissipation in chokes
> with severely mismatched antennas (he predicted disaster), and ran it
> past me first. I told him that his analysis of differential mode was
> right on, but that he had failed to consider common mode, which makes
> matters even worse!
>
> Another thing that bothers me about these chokes is that they are in
> what appear to be sealed enclosures, which can limit the ability of the
> assembly to conduct heat away from the choke, reducing its power handling.
>
> The key to power handling is to understand that power is I squared R,
> and that the choking Z must be high enough to reduce the current so that
> dissipation is not excessive. Because power is increasing (or
> decreasing) as the square of the current, you CAN outrun your tail by
> making the choking Z high enough, because the power is dropping twice as
> fast as the current. :)
>
> The differential mode dissipation must be modeled using transmission
> line techniques -- a Smith Chart or N6BV's TLW (free with the ARRL
> Antenna Book) will get you right answers. We must model the common mode
> part of the problem by adding the equivalent circuit of the choke(s) to
> an NEC model, where the choke is added to a single conductor the length
> of the feedline running between the feedpoint and ground (or wherever
> the coax is connected). Remember -- the equivalent circuit of a ferrite
> common mode choke is NOT best described as a series R + jX, but rather
> as a parallel resonant circuit, where L is the mid-to-low frequency
> inductance, R is the Z at resonance, and C is the stray C that resonates
> with L to establish the resonant peak. That simple circuit works for #43
> and for #31 above about 7 MHz. For #31, it's more complicated than that,
> because it has both a circuit resonance and a dimensional resonance that
> give it a double humped response (like a stagger tuned IF), which is
> what gives #31 its extended bandwidth.
>
> Much of this is developed in the tutorial stuff on my website.
>
> 73, Jim K9YC
>
> ##  Points very well taken.   Ok, what about increasing power handling
> by designing /winding the choke slightly different.  IE:  more ferrite cores
> and less turns...assuming  this new config has the same or close  RS as
> your cook book designs. .

That's a big part of why the coax chokes have a lot of cores -- 
dissipation is split between more cores.  It's also part mechanical -- 
we need to move the resonance down to where we want it, we want to use 
big coax, inductance is proportional the length of the core surrounding 
the coax, so we get 5X the L by using 5 cores instead of 1. For the same 
reason, the "biggest clamp-on" is roughly worth 3 cores. We can also 
tweak the resonance by squeezing turns closer together to lower Fr, and 
spread them to raise it. I showed some photos of that.

> I looked at your charts and graphs,and
> tried to extrapolate between the various  cook book designs. X turns of 213-U
> wound on  Y  cores.    Sure, more cores would be required..due to the N2
> effect being less pronounced, but with more cores used, there is more thermal mass
> because of the increased number of cores used.

Exactly.

I started investigating the bifilar-wound chokes in response to the guys 
in the UK who couldn't get what they wanted from Farnell (their version 
of Newark, who is the Fair-Rite distributor there), or would have to pay 
wildly inflated prices for the cores.  What I learned by 2010 is in that 
RFI-Ham.pdf   I've since used single core bifilar chokes on 80/40 fans, 
mostly with success, but ONLY if there was a second choke on same line 
to take part of the common mode voltage. When I changed the coax and 
forgot to add the lower choke is when that choke fried.

> ##  I don’t think there is much in the way of any wiggle room on your cook book designs
> though,  you appear to have nailed it  across the spectrum.   For some applications,
> that must cover a huge chunk of spectrum,  stagger  tuning 2  or even 3 of your cook book designs,
> may well work.   Another thought here is for say a monoband application, like my 80m rotary dipole,
> 40m yagi, and 20m yagi,  perhaps  another approach would be to  optimize a balun for each band,
> but build 2-3 identical assys  for each band.   IE:  2-3 baluns, each optimized for the same ONE band,
> then wired in series to increase RS.  On paper, this should /might work ?

Yes, on all counts. BUT -- each of those chokes is a parallel resonant 
circuit, and, in effect, they are stagger tuned.  In the overlap between 
their resonances, the XC of the lower freq choke will cancel with the XL 
of the higher freq one. But the resistances continue to add.

> ##  Can we safely assume that if  2-3  identical assys  are all in series,  and all in the same nema box,
> that we can simply total the RS  up ??    IE: 5 k ohms of RS per  assy, and  15K  if  3 used ?

If they're identical, yes, they will add, and they don't need to be in 
the same enclosure -- remember my advice to NOT enclose them to improve 
heat transfer. I did exactly that with the first serious stress test 
around 2007 -- I used one choke as the end insulator for a vertical 
dipole fed from the end, where the coax shield is the bottom half of the 
dipole. Discussed in the slide show, with photos. With one choke, it 
fried melted the coax about ten minutes into a CW rag chew.  Adding a 
second choke doubled the choking Z and split the dissipation between the 
two chokes, and that was enough.

There's another example of this in this slide show description of an 
antenna W6GJB and I designed for CQP and FD.  We planned to run 600W 
into it. Three chokes turned out to be overkill. The coax was RG8X 
inside the tube and down to the ground, then RG213 to the operating 
position.

http://k9yc.com/80M-FDVertical.pdf

> ##  After looking at ur cook book designs, at a given freq /band, I see know other way to increase RS, except
> to   series several  identical assys.   Any thoughts  on this ?

I think we see it the same way, Jim.  What I did back around 2007-8 was 
to build one choke at a time, measure it, and go systematically from 
there to end up with designs that got where I wanted to go. You can see 
that in the curves, where I varied the number of turns and cores to find 
the most effective (and cost effective) combinations. BTW -- between the 
slide show and the RFI-Ham.pdf  there is some development of the 
dissipation issue.

The way I study the behavior of the chokes in an antenna system is to 
first determine the parallel equivalent circuit, then plug them into the 
common mode circuit.  As you see from that, determining L, C, and R are 
nothing more complicated than the curve fitting I learned in EE school 
50+ years ago.

Other thoughts about putting them in an enclosure.  One reason mfrs do 
that is so that they can pot them and make them both more rugged 
mechanically, but also harder to see what's inside. DX Eng has always 
been paranoid about reverse engineering, to the extent that they don't 
include schematics with stuff they sell, and they won't sell you parts 
to repair their stuff. Another reason is that they think hams want it in 
an enclosure.  The only reason I can think of for an enclosure is to 
prevent issues with the cores cracking with freeze/thaw cycles. I can't 
test that here along the Pacific coast, but in the eight years my stuff 
has been on the internet with more than a million download in the first 
five years, no one has ever told me that had been a problem.

73, Jim K9YC