Towertalk
[Top] [All Lists]

[TowerTalk] Measuring balun/choke impedance, power dissipation, etc.

To: <towertalk@contesting.com>
Subject: [TowerTalk] Measuring balun/choke impedance, power dissipation, etc.
From: ccc@space.mit.edu (Chuck Counselman)
Date: Thu Jul 24 14:22:48 2003
"Tod - Idaho" K0TO wrote:
>I have been curious as to the experimental setup, including test 
>equipment, that folks are using to make balun measurements. I figure 
>I will see if there is a common (or nearly common) method and then 
>use that to make my measurements.....


Good questions.  Although you asked them off-list, I'm answering 
on-list to prompt others to answer and to comment on-list, for my 
benefit as well as others'.

As with, e.g., an automobile, several independent characteristics of 
a balun are potentially important; and the perceived importance of 
any particular characteristic varies radically from one user to 
another.  Therefore, different folks use more or less care, and 
different methods, to measure/test different characteristics.  IMO 
these differences are worth remembering as you read responses to your 
questions.  As with all advice, "Consider the source."

My responses are those of an HF operator who uses one and the same 
transmission line with one and the same antenna to transmit 
legal-limit power and to receive on all ham bands (and MARS 
frequencies) from 1.8 through 29.7 MHz, except that I transmit less 
power on 160 meters.  My "one and the same transmission line" is 
partly 50-ohm coax and partly 600-ohm open-wire line.  It includes 
two coaxial common-mode chokes ("1:1 current baluns"), one 4:1 
Guanella ("current") balun transformer, and a lumped-L-C 
impedance-matching network.  My antenna is a 102-ft. doublet. 
Another peculiar aspect of my situation is that my coax runs 70 feet 
through the interior of my house, where its common mode is relatively 
closely coupled to AC-power and other wiring.

Characteristics that seem important to me include:

1. Common-mode choking impedance.

2. Ability to transmit power in the desired, transmission-line, mode.

3. Ability to handle power incident in the generally undesired, common mode.

4. Relation between input and output impedances in the desired, 
transmission-line, mode.

Every one of these characteristics is a function of frequency.  It is 
essential to measure them over the full range of frequencies that 
matters to you.

Taking these characteristics in order now,...


1. Common-mode choking impedance.

If both the input and the output transmission-line interfaces are 
unbalanced, as in a coaxial-cable common-mode choke, I measure the 
impedance between the outer- or shield-conductor terminals at 
opposite ends of the device under test.  Doing so requires connecting 
these opposite ends to the closely-spaced terminal-pair of my 
impedance bridge/meter.  I use the shortest possible lengths of 
insulated #18 wire for these connections.

My impedance bridge/meter is an Autek model VA-1, which costs about 
$200 new and is mostly adequate for this purpose.  Probably its worst 
limitation here is its inability to measure an impedance whose 
magnitude exceeds about 1 k ohms.  A professional-grade instrument 
would be better.  Surplus/used HP, GR, and other respected-brand-name 
instrumentation is available.

I do _not_ like MFJ's instruments which cannot determine the sign of 
the reactance (the imaginary part of the impedance).  IMO it is too 
difficult to understand what's wrong with a device without knowing 
the sign of the reactance.  I have a Palomar noise bridge, which 
_can_ determine the sign of the reactance, but I don't like it for 
this purpose, either.  Using it is much more tedious, and it's also 
less precise than the Autek VA-1.

Whatever instrument you use, always _check_ it by measuring 
lumped-element resistors, capacitors and/or inductors of 
independently known values, and/or by looking at a short- and an 
open-circuit, both directly and through known, short, electrical 
lengths of known-low-loss line.  The impedance at the input of such a 
transmission line is easy to calculate.  Suitable computer programs 
are widely available, or you could use a Smith chart.


2. Ability to transmit power in the desired, transmission-line, mode.

Unless you transmit high power and/or operate with a high SWR on the 
transmission line, it's enough to estimate the loss of the device by 
shorting its output and measuring the SWR at its input.  Tables for 
inferring the loss from the the SWR are given in most radio/antenna 
handbooks; suitable computer programs are also widely available; or, 
_if_ you understand transmission lines(!), you can do the calculation 
yourself.  Strictly speaking, the relation between loss and input SWR 
depends on whether the loss comes from resistance in the conductors 
or from conductance in the dielectric of the transmission line, and 
the standing-wave pattern in the line, which in turn depends on 
whether you've shorted or opened the output of the device; but for HF 
you'll do well enough by just shorting the output.

Probably most of the SWR bridges or directional wattmeters in use by 
hams work well enough in this application, but I strongly advise 
_checking_ any instrument by measuring SWR or reflected-power ratio 
looking into a short-circuit and an open-circuit, both directly and 
through short lengths of known-low-loss line.

If you transmit high power and/or operate with high SWR on the 
transmission line, you should worry about heating and dielectric 
breakdown.  IMO there's no substitute for testing with high current 
and with high current.  If you're like me, you don't own dummy loads 
with the necessary impedances and power-dissipation abilities, and 
you don't own a power amplifier with the output current and voltage 
capabilities, that you'd really like for such testing; so you'll have 
to improvise.  To test your balun at high current, short its output 
and drive its input as hard as you can safely and appropriately.  To 
test your balun at high voltage, open-circuit its output and drive 
its input as hard as you can safely and appropriately.  With a 
directional wattmeter, observe the forward power in the line between 
your transmitter/PA and the device under test.  (The reverse power 
will be nearly as great, unless your device is awfully lossy, in 
which case you have more fundamental problem.)  Calculate the current 
at the short-circuited output of the device, or the voltage at its 
open-circuited output, from transmission-line equations.  Again, 
suitable computer programs are widely available.


3. Ability to handle power incident in the generally undesired, common mode.

Terminate the device with its intended load impedance, using an 
appropriate dummy load; short one side of the device's balanced 
output, or the "hot" side of its unbalanced output, to the "cold" 
side of its unbalanced input; and drive the input as hard as you can 
safely and appropriately, in light of your 
transmitted-power/current/voltage measurements.  [Refer to 2. above.] 
Start with the input power low, and increase it in steps of, say, 3 
dB.  At each step, observe the input forward/reverse power and SWR 
immediately, monitor it/them continuously, and QRT _quickly_ if you 
sense anything untoward -- to protect your transmitter/PA/etc.

Otherwise leave the power on for one minute, then SHUT IT OFF(!!), 
and THEN immediately _feel_ the device for heating and _smell_ it for 
overheating.  If you detect nothing untoward, then reapply the power 
for five minutes, SHUT IT OFF(!!), and feel and smell the device 
again.  Repeat at (say) 3 dB higher power until you've reached a 
sufficient level.  At this level you may wish to run a test for 
substantially more than five minutes.  "Sufficient" can mean that the 
device feels or smells too warm to proceed; it may mean that this is 
all your transmitter, coax, dummy load, or whatever, is good for; or 
it may mean that it's enough for your intended use(s).

The device under test may fail from dielectric breakdown quickly or 
only after heat has damaged its dielectric.  Heating may occur in a 
conductor, in a dielectric, and/or in a ferromagnetic material such 
as ferrite.  Dielectrics may slowly decompose, or slowly soften and 
slowly flow like viscous fluids.  A ferromagnetic material will lose 
its ferromagnetism and its loss may skyrocket when its so-called 
Curie temperature is reached.  Thermal runaway may occur through 
various mechanisms, and may surprise you by destroying your device 
very quickly, before you suspected trouble.


4. Relation between input and output impedances in the desired, 
transmission-line, mode.

In principle a device may be ideal in the sense of having infinite 
common-mode impedance, zero transmission loss, and infinite 
power-handling ability both in transmission-line and in common mode, 
yet it may not present the desired input impedance when its output is 
properly terminated.  E.g., it may have a large equivalent shunt 
susceptance.  Or it may not have the desired impedance-transformation 
ratio.

So, in addition to everything else, you must measure its input 
impedance with its output properly terminated.  I use the 
aforementioned Autek model VA-1 bridge.  You could use a 
professional-grade instrument, but IMO the Autek is plenty good 
enough for ham use.  I don't care whether the SWR seen by my 
transmitter is 1.05 (as the Autek can resolve) or 1.01, or 1.1 or 
even 1.5, as long as it doesn't change unexpectedly.  Probably most 
of the SWR meters, directional power meters, antenna analyzers, noise 
bridges, etc., in common use by hams are adequate for this purpose.


In closing, let me repeat that every characteristic I've mentioned is 
frequency-dependent, and sometimes very sharply frequency-dependent 
(when resonance occurs).  Therefore it's important to test throughout 
the relevant frequency band.  This doesn't mean that you must measure 
everything at 100-kHz intervals from 1.8 to 29.7 MHz.  Measure at, 
say, octave intervals; consider the trends that you see; and proceed 
accordingly.

73 de Chuck, W1HIS
<Prev in Thread] Current Thread [Next in Thread>