[Elecraft] [K3] Issue with HI CUR on 12m

[Jim Brown]

On 10/13/2020 8:12 AM, John Oppenheimer wrote:
> Hi Jim,
> 
> For high power Balun Applications, most critical parameter is Ferrite
> core loss which causes core heating. Core loss, parallel resistance
> (Rp), can be computed from complex permeability data from Fair-Rite.

Hi John,

While I respect your engineering chops, you clearly don't understand how 
ferrite common mode chokes work. They DEPEND on the loss component 
coupled from the ferrite core, which has a broad peak around resonance 
of the winding. Fair-Rite's #31 material is MnZn, which exhibits two 
resonances. One is formed by the inductance of the winding and the 
capacitance between turns; the second is a dimensional resonance that 
results from standing waves within the core. #43 material is NiZn; these 
materials do NOT exhibit dimensional resonance, because loss is low at 
the frequencies where it would occur.

At HF, the two resonances in multi-turn chokes on #31 cores combine to 
provide a much broader resonance, comparable to what happened in 
stagger-tuned IFs. This has two implications. First, the broader 
resonance allows the choke to cover more bands.

Second, and more important, ferrite cores have wide manufacturing 
tolerances that cause the resonances to shift in frequency. The broader 
impedance curve formed by chokes on #31 material allows a designer to 
specify chokes that will work within a specified frequency range with 
that tolerance variation; it is NOT possible with the much narrower 
impedance curve of chokes wound on #43 or #52 material.

My "choke cookbook" is the result of first characterizing nearly 200 #31 
cores of a given size, selecting cores at the tolerance limits for each 
size, winding and measuring more than a thousand chokes on these 
"limits" cores. The cookbook is based on chokes working with all of 
those "limits" cores.

After seeing recommendations from a ham in the UK for #52 material, I 
bought 40 of these cores over a period of about four weeks, splitting 
the order between four franchised vendors (that is, ten from each 
vendor). I characterized those cores, selected cores at limits, and 
wound chokes using that ham's recommendations. I could not reproduce his 
results -- the tolerance variation moved their resonance away from their 
intended operating range.

I learned about dimensional resonance from a classic engineering book on 
ferrite applications by E. C. Snelling that a colleague found for me in 
the engineering library of the U of Chicago, where he was on faculty. 
Snelling's work is considered "the bible" by engineers working in mfg 
and application of ferrites. It's reference 7 in the AES paper, and is 
also referenced in the tutorial.

These concepts are discussed in detail in an AES paper from 2005 and in 
a tutorial I wrote for hams in 2007, the latter updated several times 
over the years.

http://k9yc.com/AESPaperFerritesASGWeb.pdf
http://k9yc.com/RFI-Ham.pdf

The "cookbook" is here. http://k9yc.com/2018Cookbook.pdf

73, Jim K9YC