A Low Cost RF Impedance Analyzer
Actually, it shouldn’t be too complicated to build and it should make a nice
weekend project as it is straightforward
and there is no power supply needed.
Now, refer again to the impedance
analyzer circuit schematic shown in
Figure 6. Only a few parts are required
but they must be selected carefully for
best results. D1 through D3 are
1N5711 Schottky diodes. It is best if
they are matched. I bought 10 diodes
which were packaged on a tape reel
and chose three diodes that were
placed together. Of course, this doesn’t guarantee the diodes are matched,
but it may improve the odds. Similarly,
I bought 20 pieces of 49. 9 ohm, 1/4
W, 1% resistors and carefully measured
each one. I found one that was 50
ohms and used that for Rm.
I also found several resistors that
were close to 49. 8 ohms which were
then used for R1 and R2. It is important
that these two resistors be matched in
value. I also matched the three
capacitors C1 through C3, 390 pF, 5%
ceramic capacitors which should be
COG type. I also purchased some
12. 5, 25, 100, and 199 ohm, 1/4 watt,
1% resistors for use in calibration.
Shown on the diagram are a
coupling capacitor C4, a switch SW1,
and filter components R3, R4, and C5.
Tolerances of these components are
not critical. R4 may or may not be
required, depending on the input
impedance of your DVM. This is
discussed in more detail later on. SW1
is used to select the voltage for
measurement by the DVM. Finally,
you should choose the appropriate
input and output connectors for the
type of measurement you will make,
be it a component or transmission line.
■ FIGURE 7. Photo of
breadboard setup of the
Signal Generator and
A sine wave signal generator
with 50 ohm output impedance that
produces a reasonable level is needed.
Signal generators are usually rated in
dBm output (for 50 ohm systems) which
can be converted to peak volts, as needed. For example, 13 dBm output into 50
ohms produces 1V RMS,
or about 1.414 volts peak.
Although the generator will
not always see 50 ohms,
this is still a convenient
reference point. Because of
the diode nonlinearity,
there is a lower limit of generator output
where it becomes very difficult to compensate accurately for the voltage drop.
This occurs somewhere around 13 dBm.
I don’t recommend using a generator
with an output below this value.
Unfortunately, some generators
“max” out at 10 dBm (0.707 VRMS or
one volt peak) and would produce
marginal results. Several generators in
my shack have that limitation. On the
other hand, many generators can produce much more. My Wavetek Model
81, 50 MHz generator can produce up
to 24 dBm (or five volts peak voltage).
Some generators produce a DC
offset voltage which can cause the
diode sampler to have an error. My
Wavetek — although it was set for 0
volts offset — actually produced 6 mV
offset, and while that doesn’t sound
like much, it was enough to cause
errors. I cured the problem by
placing capacitor C4 in series with the
generator to block the DC voltage.
A QRP rig is probably over-powered
for use as a signal generator for this application, so an attenuator must be used to
bring the level down below 24 dBm.
Attenuation of transmitters is beyond the
scope of this article so you will need to
consult an appropriate radio handbook.
Digital voltmeters or multimeters
are very common today. A basic DC
accuracy of 0.5% or better is needed.
Normally, when accuracy is specified
in this manner it means 0.5% of span.
So, if your voltmeter range or span is,
say, two volts, then an error of 0.5% of
two volts (plus or minus 10 mV) is still
Of course, it doesn’t mean you
will necessarily see that large of an
error. Many of the very low cost
units are definitely not appropriate as
they are rated above 0.5%. Some
economical units can be found with an
accuracy better than this. Unfortunately,
a few meters may proclaim high
accuracy but it may well be salesmanship on their part. Caveat emptor!
Multimeters can have quirks. As a
case in point, an auto-ranging digital
multimeter, with a stated accuracy of
0.3%, was obtained on the Internet for
around $19. With the range fixed, it
produced acceptable results but when
in auto-range mode, produced errors as
high as 40% on low voltages. Why? As it
turns out, for the lower DC range, the
meter’s input impedance changes from
10 Meg-Ohms to over 100 Meg-Ohms.
This causes a different loading on the
circuit in Figure 6, with resulting higher
voltages measured on the low range.
This can be remedied by fixing the range
so that it is not allowed to auto-range to
the lowest span. This seems to work best
with meters having higher counts.
Even high-end meters like the
Protek 608 have different input
impedances for different ranges. For
example, on the 500 and 2,500 mV
ranges the input impedance is greater
than 1 Giga-Ohm. On the 5.0 to 5,000
volt ranges, the input impedance
ranges from 10 to 10. 5 Meg-Ohms.
If the input impedance is extremely high, looking at a capacitive load
like C5 in Figure 6 may cause
problems, so resistor R4 was added. It
may not be needed in some cases, and
if omitted, will allow a higher voltage
to be measured. Note, too, that some
DVMs have a low input impedance
and are designed for special applications. An input impedance of at least
10 Mohms is needed here to prevent
over-loading the diode voltage samplers. Be sure to check your meter for
February 2008 43