[Elecraft] Balun for KAT2

[Charles Greene]

At 04:46 PM 1/17/2002 -0500, Bill Coleman wrote:

Some comments below:

>On 1/12/02 9:54 PM, Don Wilhelm at w3fpr@peoplepc.com wrote:
>
> >I may be opening a can of worms here because there are a lot of 'balun
> >believers' out there, but I would start with no balun at all.
>
>There's a lot of bad infomation about baluns out there.
>
> >It WILL work,
>
>Yes, it will.
>
> >but there is a possibility that you may get some 'RF in the shack' -
>
>Which is exactly what any balun is designed to prevent -- RF currents
>flowing on the outside of the coax.
>
> > should
> >that happen, I would recommend an RF choke on a short section of coax (use
> >big coax for best efficiency).
>
>The size of the coax doesn't matter, but if you coil coax, be sure to
>respect the minimum radius of the type of coax.
>
> >You could use a coil of coax, but even
> >better is the W2DU type of balun (a bunch of ferrite cores over the coax
> >shield) - make your own with 12 to 20 FT-50-43 cores slipped over RG213
> >coax.  You need more cores at lower frequencies because it is the choking
> >inductance that does the job.
>
>Coiled coax, W2DU baluns, or toroidial core current baluns (Guanella or
>W1JR types), all work on the same principle. So long as there is
>sufficient choking inductance, they work.
>
>There's a few caveats. First, you must have sufficient inductance, but a
>coiled coax balun can have TOO MANY turns. The distributed capacitance
>can make the coil resonant on the higher frequencies. Lots of turns tends
>to move that resonance lower, which is exactly what you don't want.
>
>6-8 turns on a 4-6" form is sufficient for 10-30 MHz. 9-12 turns on the
>same form works for 3.5-25 MHz. 16 turns for 1.8-7.0 Mhz.
>
>Second, the ferrite core baluns have a wider frequency range, but the
>characteristics of the core come into play. This usually isn't too much
>of a problem for QRP operating, but QRO operators need to beware they
>don't heat the cores.
>
>Third, the W2DU baluns depend on the size of the coax, type of material
>and the total length of the beads. The size of the beads isn't really
>important. Typically, a 7-30 MHz balun can be constructed with about a
>12-18 inches of beads. For lower frequencies, you need more inductance --
>hence you need to double or triple the length.
>
>Bead material is important. Type 43 material has maximum impedance at 200
>MHz, so it isn't very appropriate for HF baluns.

The 200 MHz pertains to coils, not to baluns; type 43 material works very 
well on HF baluns.  I have some type 43 material current baluns that work 
well from 700 KHZ through 30 MHz.  The balun made with type 43 material is 
somewhat more lossy than the powdered iron cores ones and the type 77 
material even more so, but losses still are reasonable for low power 
operations.

>(However, I have such a
>balun using FB1024-43 beads on my A3S -- works great) You're better off
>using type 77 material for an HF balun. You'll end up needing fewer
>beads. Power handling may be a problem for QRO operators with type 77
>beads, but it isn't an issue for QRP.
>
>Forth, Guanella and W1JR baluns are just as effective as W2DU baluns, and
>can be made a lot cheaper. A FT240-77 core cost $9.00 from Amidon,
>whereas the beads required for a W2DU balun are closer to $18.00. If you
>buy commercial baluns, however, the W2DU types tend to fit well with your
>typical dipole installations, and can be lighter.
>
> >The baluns made by winding bifilar (or tri-filar) turns on a core can be
> >very lossy when used on a feedline having input impedances far away from the
> >25 to 300 ohm impedances that these baluns are usually designed for and they
> >are not very happy with reactive loads.
>
>This is wrong on three points.
>
>First, I believe you are talking about voltage-type baluns. These
>typically use powdered-iron cores instead of ferrite. Such cores have
>1/100 of the permeability of ferrite cores. These cores are not suitable
>for making current baluns.
>
>Second, not all toroidial baluns are the voltage-type. The Guanella
>design uses a bifilar winding for a 1:1 balun. The voltage-type 1:1
>requires a trifilar winding. The W1JR consists of a short length of coax
>(usually a very small diameter type -- often teflon insulated -- RG-142,
>RG-303, RG-316) wound 3-9 turns on the core. There's no electrical
>difference between the W2DU, Guanella or W1JR balun.
>
>Third, none of the discussed baluns (with the exception of the coaxial
>coil) performs well in the presence of reactive loads. Such loads can
>cause core saturation, heating and additional loss, unless accounted for
>in the design of the balun.

Core saturation is a factor in coils, but is not a factor in current baluns 
if the balanced current is maintained.  The currents in the coiled 
transmission line are equal and opposite, and the net flux is zero.  I have 
done extensive testing of baluns into reactive loads with high SWRs.  The 
loss increases, but the balun will work well into a SWR of 5:1.  The losses 
are primarily dielectric and eddy current losses.  I^2 R losses contribute 
just a small amount to heating.  The core stays cool until heated by the 
wires.  The wires get hot as that is where the loss is.  Core saturation is 
another of the misinformations commonly expressed about baluns.  Yes, you 
can cause core saturation, but the other losses will probably destroy the 
balun first.  One phenomena I have noticed during loss testing is the run 
away condition.  If you apply high power to a balun working into a reactive 
load, the balun wire will get hot.  That increases the losses, causing even 
more heating.  That condition continues until the wires in the balun burn 
up.  Another factor to watch out for is excessive voltage in the 
balun.  Many low power baluns use
only enamel insulated wire wound on the core.  The enamel insulation has a 
break down voltage of about 500 volts, and worse case with a 5:1 SWR you 
can get nearly 500 volts in the balun.  Most core materials have a varnish 
insulation which also has a breakdown voltage of about 500 volts.  However, 
the wire can have nick in the insulation where it passes over the 
core.  Type 77 material has no varnish insulation and you need to be 
careful when using it for a balun including winding a coax like RG-316 
around it.  A higher power core can be wound with teflon tubing over the 
wires, or wrap the core in glass tape.  I have also successfully used 
teflon insulated, silver plated stranded wire for baluns.

>However, you are correct that current-type baluns provide better balance
>to the load than voltage-type, regardless of the reactance present. The
>ARRL Handbook no longer provides descriptions of voltage-type baluns.
>Current-type baluns are recommended.

Sevich devotes an entire chapter in his book "Building and Using Baluns and 
Ununs."  to baluns for antenna tuners and recommends voltage baluns, 
sorta.  He goes to great pains to flatten response of the balun used by 
McCoy in his original transmatch.    However, there are 2 problems with the 
voltage balun used in the antenna tuner.  As the as the frequency is 
increased, the phase shift in the transmission coiled transmission line is 
no longer able to sustain a 4:1 ratio and the high frequency performance 
falls off.  The second problem is a result of the first.  In an attempt to 
extend the high frequency range, turns are removed from the balun until at 
the lower frequency there is barely enough inductance to prevent the 
primary current from flowing in the balun, decreasing the current to the 
load and increasing losses.  This condition becomes worse when working into 
a high impedance load.  As the load impedance increases, more primary 
current flows in the balun as the impedance of the balun to choke the 
primary current doesn't change and becomes a smaller ratio of the total 
impedance.  I made a bunch of voltage baluns, but finally gave up on 
them.  All the baluns I made would cover 160 through 15 or 80 through 10, 
but not both.  BTW the T184 size is a good compact powdered iron core to 
use if you are into voltage baluns, as it has an inductive index higher 
than any core smaller than 4 inches.

>Bill Coleman, AA4LR, PP-ASEL        Mail: aa4lr@arrl.net
>Quote: "Not within a thousand years will man ever fly!"
>             -- Wilbur Wright, 1901

Nice chatting with you.
73, Chas, W1CG