[TowerTalk] Roller inductor sources

[Tom Rauch]

> I must respectfully disagree with Tom Rauch about the
concept of a balun on
> the output side of a tuner vs the input side. I currently
have 4 tuners that
> I have basically fried due to legal limit power and
mismatches on 450 ohm
> line. These 4 tuners all had output baluns that saturated
and got hot, but
> this is not always the failure point. Mostly HV failures
at switches or
> caps.

Hi Chuck,

Heating is almost never due to saturation. That's another
myth that is repeated over and over. Heating is almost
always due to loss tangent in the core and the voltage
applied across the windings. The winding and core combo
looks like a certain resistance in parallel with the
winding's complex impedance. When the voltage gets too high
across that resistance that largely represents the core's
loss tangent, the resistive heating can exceed  the core's
ability to dissipate heat. It's like placing a resistor
across the load rather than saturation.

Moving the balun will only help if we are using a VOLTAGE
balun and when the tuner has a symmetrical balanced network
after the balun. Of course we all know (or should know) a
voltage balun is the worse possible balun to use in a random
antenna application.

On the other hand if it is a current balun, heating will
remain essentially the same. Some balanced tuners use a
current balun and a flating T or L network. This includes
some commercial tuners.

The general claim or idea moving the balun to the input
always helps because balun SWR is less really has to be
stopped. It is a misleading idea.

A properly designed current balun would likely survive on
the tuner output without undue heating, a place where it
would be almost impossible to make a voltage balun work.

> Only one tuner I have found works great in all 450 ohm
line mismatch
> situations. This is a Palstar 1500 BAL. It has 2 roller
inductors and a
> balun on the input. The inductors, in series with each
side of the output of
> the balun will match a 450 ohm line with high reactance
very well. It has a
> switchable cap to move it to the input side of the
inductors or output side
> for HI-Z or LOW-Z situations.

You disagreed, but then you describe a tuner that has a
balanced network (which I made an exception for)! The
problem is this.... the Palstartuner might have less balun
heating but that's because the balun does very little if
anything to balance the system.  If the network is a
balanced network the balun on the input can be eliminated
with very little change in the system. What we have with a
balanced network is really just a reasonably stiff fixed
voltage source working against ground for each feeder
conductor. This is great if the reactances in the tuner all
track (so voltages are equal) and the feedline (including
the antenna) is nearly perfectly balanced.

It also sounds like you are describing an L or Pi network.
**If** it is an L network or pi network, the network will
offer significantly restricted matching range compared to
similar cost components in a T network.

(The ideal tuner would be a ground independent floating
current source. If we use all those expensive parts, why not
just make a link-coupled floating network and have the
best?)

My statement regarding this type of tuner is one would be
better off to simply save the money required to build the
balanced network, and put a portion of that money into
building a proper current balun on the tuner output. The
result would be less cost for a better product.

> Basically the tuner completes the impedance transformation
so the balun is
> not stressed and will not saturate. Keep in mind I am
talking about
> situation like a shortened 80 meter doublet loading.

We often hear people use the term saturation, but saturation
generally has nothing to do with it. The reason the balun
runs cooler with a symmetrical network is the balun does
almost nothing at all, and the balun has less voltage
impressed across the equivalent core loss resistance under
some load conditions.

What people have been doing with tuners is using a very poor
type of balun (voltage balun), failing to understand why it
heats or what is needed to improve it, and assuming moving
the balun fixes all the problems.   It would be much better
to understand what causes the heating and fix the real
problem.

I'm not saying moving the balun in a true balanced network
tuner won't change heating of a balun, especially with a
voltage balun. I'm only saying it is useless to move a
current balun, which is the best type of balun, to the front
of a conventional T or L network. It is also much cheaper to
fix the balun, rather than buying components to build a
balanced network....and we will have a better overall system
at the same time we save money.

I can build anything I want for myself, and I use
conventional T networks with good 1:1 choke (current) baluns
on the output. When I have a really high line voltage at the
tuner that taxes the limits of a balun or tuner design, I
either adjust the feeder length or use a link coupled
network. That's because I don't like to waste my parts or
time.

As for the tuner examples, Dentron's idea of a good balun
was a 4:1 voltage balun using a few turns of wire on a red
(-2) mix core. I really can't think of a worse balun to use
on a tuner. As I recall a Nye Viking did nearly the same,
and I know MFJ did early on.

If I was allocated even 1/4 of then additional money spent
on building a symmetrical network tuner, I could have fixed
the baluns in any of those tuners and produced a tuner with
superior balance over most load and frequency combinations.

73 Tom