Topband
[Top] [All Lists]

Re: Topband: 73 mix cores

To: "Mike Waters" <mikewate@gmail.com>
Subject: Re: Topband: 73 mix cores
From: "Tom W8JI" <w8ji@w8ji.com>
Reply-to: Tom W8JI <w8ji@w8ji.com>
Date: Fri, 28 Nov 2014 13:13:10 -0500
List-post: <topband@contesting.com">mailto:topband@contesting.com>


Fair-rite also states that multi-aperture ferrite cores (like this one)
have greater bandwidth vs. a single-hole core made of the same material. I
cannot explain that.


Increased bandwidth comes from minimizing the wire outside of the core window.

When I was experimenting with multiple phased loops and reversible Beverages in the early 1970's, prior to Fair-rite tooling that core, I made my own binocular cores from stacks of shield beads. I initially used a string of beads in a single line, but over half of the wire length in the windings was outside the core material. Even back then, I realized over half the conductor length was outside the core where the wire did nothing good and everything bad.

I thought it would be better to use side-by-side rows, so only a small fraction of wire length was outside the core material.

I also used a circle formed from large shield beads (and still do) in high current chokes, where the high current demands large conductors. It's easy to build a 100 ampere or larger choke using what looks like cores parked like a circle of wagons, and obtain wide bandwidths with very high current chokes. This also works for baluns and matching transformers when conductor size is large.


I can verify that the losses in any single-core transformer that I've ever
made from that core, regardless of the Z ratio, was less than 1 dB. (0.9 dB
was an extreme case.) And I've measured losses as low as 0.2 dB. ON4UN in
his latest book wound transformers on that core with even less loss, IIRC.

I've been using systems like that for at least 40 years, and I always measure things.

I used a stack of three once when I experimented with transmitting on a
Beverage. Amazingly, the core never approached the Curie temperature even
feeding it with 800 watts of CW and heavily processed SSB. Photos of that
core test are at
www.w0btu.com/files/misc/Binocular%20core%20power%20vs%20temp%20test/ .
I cannot find the measured loss on that core at the moment, but it was
certainly low enough (~0.4 dB, IIRC) that it never exceed 225 degrees F
even with 850w key-down, with no air movement and a layer of heat shrink
which reduced heat radiation. And I'm using that same transformer on the
input of my swamped-grid amplifier. It barely gets warm even with 100w CW.

If power levels or impedances are high enough, you'll sometimes have to move to a core with higher Q over the operating range. This narrows bandwidth a bit, but greatly reduces heat.

This is a major issue with baluns, which are always a compromise of several things. If you use a system with extreme choking impedances, even 10,000 ohms, cores can become hot enough to reach curie temperature. If voltage across a winding is 1000 volts, easily obtained with a kilowatt and common antennas, a 10k ohm choke would have 100 watts of core heat. At high power, we can't "impedance" ourselves out of a poorly planned feedline and antenna combination.

High power baluns that run the coolest are often baluns that are not even acting like baluns!

If we need an extreme impedance, we probably have a poorly planned system. A poorly planned system with an extreme choke or balun will not work as well as a properly planned system with a modest impedance balun or choke, so it is most valuable to learn how this stuff really works as a *system*.

Some systems, most notably line isolators, give advice contrary to what is best for the system.

73 Tom
_________________
Topband Reflector Archives - http://www.contesting.com/_topband

<Prev in Thread] Current Thread [Next in Thread>