VK3AQZ-RFPM1-power-meter-kit-V1

Page | 1
VK3AQZ RF POWER METER KIT MODEL RFPM1
Introduction and description.
An instrument to measure RF power, with a reasonable degree of accuracy, and over a wide frequency
range, is a valuable tool for radio amateurs, electronic hobbyists and students.
The main reason for the development of the RFPM1 power meter was to enable the design and testing
of radio amateur receivers and transmitters without the necessity to purchase an expensive
commercial instrument. Performance of home brew equipment can be improved if an accurate means
of measuring RF levels is available.
In recent years a number of publications have appeared using the low cost, but
reasonably accurate ANALOG DEVICES AD8307 logarithmic amplifier/detector
IC.
The AD8307 converts amplitude variations of an RF input voltage to a DC output
voltage which varies as the logarithm of the input voltage changes. It does this from
near DC to over 500 MHz, and, over a 92dB input range, with a reasonably high degree of accuracy
(+-1dB). A logarithmic relationship between input and output means that a simple linear display
device can be used to obtain input signal level changes directly in decibels.
The AD8307 ANALOG DEVICES datasheet contains additional information about the IC including a
comprehensive discussion on the theory of log amplifiers. Please refer to these datasheets for further
information.
Regarding the application of this IC to measure RF power, one of the simplest, but very useful designs
was published in QST, June 2001, by Wes Hayward and Bob Larkin titled “Simple RF - Power
Measurement”. See reference 1.
Following that article, quite a number of circuits and adaptations of the original design have been
published in various magazines, and on web pages. Some with minor additions, and others with more
complexity, utilising microprocessors, or computers, to display the resulting measurement of RF
voltage and power in a 50 Ωload.
When considering the use of an RF power meter for radio amateurs and hobbyist, it is most likely to
be used on an intermittent basis, and not constantly, as might be the case in an RF Design Engineer’s
laboratory.
With this in mind, the design finally settled on had to be as low a cost as possible, easy to construct,
minimal use of complex integrated circuits, and simple to align and use. Portability and battery
operation was also a very desirable feature –particularly for the measurement of RF field strengths
around an amateur station antenna.
In the initial stages, various published circuits were tried, and many failed to meet the above
requirements. Eventually the design chosen was a variation of the original QST design.
The QST design used a simple analogue meter movement and a chart to convert the AD8307 output
to a power reading. After building and using the various designs, it became evident that the analogue
meter was a very important part of the design when adjusting tuned circuits. However a digital display
VK3AQZKITS
AD8307 SMD

Page | 2
was also useful for measuring amplifier and filter responses. A modification by Bob Kopskie in an
article titled “An Advanced VHF Wattmeter” in QEX May/June 2002, added a simple LCD meter to
provide 1dB/mV readout. This was found to be quite a simple but effective solution. See reference 3.
So the final design consists of both an analogue readout and a simple digital readout.
During the evaluation of various circuits with digital displays, it was found that some of the designs
using microprocessors were unsatisfactory. Some designs produced jumps in the reading at various
levels whilst others just hung if the signal amplitude varied suddenly. Some computer versions simply
failed to work on my PC, or just gave incorrect answers due to rather poor software implementation of
the A to D conversion. One annoying aspect was the response time of the designs. The simple LCD
meter has a reasonably quick update time which can show small variations in level due to movement
of test leads and so on. However in the case of some microprocessor designs, the software was either
too slow or smoothed so as to avoid these fluctuations. Side by side measurements confirmed the
problem hence the reason for avoiding some of these software derived A to D solutions.
And some designs gave readouts to 2 decimal points of a dB when the overall accuracy of the chip is
considerably less than that.
As a result of these tests, a low cost 200mV meter was finally chosen as the digital display portion of
the RF power meter.
Commercially available, low cost, LCD voltmeters have well established, bug free, and smooth
hardware A to D converters, low power consumption, and generally low spurious RF output.
In the later case, it was found that some LED type meters produced considerable RF emissions which
proved quite difficult to reduce. Similarly, some of the microprocessor designs were also
unsatisfactory due to high levels of radiated RF. You will notice when you have completed the kit,
that the detector probe is very sensitive to RF around the probe.
The RF signals emitted by some of the devices I tested were so high that the dynamic range of the
meter was reduced by a significant amount. It took considerable shielding and effort to reduce such
noise. On the other hand, the LCD versions were considerably less noisy.
The final design of an RF power meter suitable as a low cost kit, with minimal complications, consists
of a shielded metal RF detector head containing the AD8307 chip (the RFPMH) connected to a
separate meter unit containing the analogue and digital displays, buffering, and power supply circuitry
(the RFPMM).
Frequency of operation is from 5 kHz to 500 MHz. The AD8307 IC has a falling response around the
300 MHz to 500 MHz region. In order to flatten the response out to 500 MHz, the RF head includes a
simple RLC input network which reduces low frequency signals by around 3dB to 4dB.
The QST design incorporated this addition. See reference 1.
Because of the addition of the input RLC network, the low signal response, or sensitivity, is also
reduced by around the same amount. So with the RLC network, the overall input power range is from
-70dBm to around +16dBm. The input impedance is 50 Ωset by a terminating resistor in the RF head.
The RFPM1 has provision for 2 switched RF detector heads which I found useful for measuring
inputs and outputs as well as front to back ratio measurements on antennae.
The RF detector head to meter unit cable carries only DC signals so ordinary shielded
audio cable can be used. The use of a long audio cable simplifies the measurement of antenna front to
back ratios, or in situations remote from the meter unit.

Page | 3
RFPMH –RF head circuit description.
The circuit of the RF head is shown below. As mentioned previously,
the input circuit contains a 51.1 Ω 1% terminating resistor which is in
parallel with the input network resulting in 50 Ω input impedance.
R2, C1 and L1 form a high frequency boost network which helps to
flatten the UHF response of the unit. The input impedance of the
AD8307 is 1.1k in parallel with 1.4pf. The 470 ohm resistor, R2, in
conjunction with the input R of 1.1k forms a simple voltage divider
which serves to lower the sensitivity at low frequency. C1 starts to
bypass R2 at UHF and in conjunction with L1, forms a small UHF boost
circuit thereby flattening the overall response of the RF head.
RFPMH - RF detector head circuit
RFPMH detector head
RFPMH detector PCB.
The SMD parts are on the copper side, but values printed on the component side

Page | 4
The AD8307 IC is a surface mount version along with a number of components connected to it. The
surface mount components offer better performance at UHF. The surface mount version of the
AD8307 is also considerably lower in cost and more readily available than the 8 pin DIL version.
Also, double sided PC board was tried but the ground plane turned out to be one large UHF bypass
capacitor resulting in poor UHF performance. Some may wonder about this as it is common to use
double sided material for RF circuits. The original QST design used “air supported” wiring and this
was found to produce better UHF performance. So the RFPMH PCB is single sided which appears to
be a good compromise between ease of construction and performance.
The use of surface mount components requires some care in assembly. 1206 size components are used
which are not too difficult to mount.
Referring to the AD8307 datasheet, the intercept point and slope can be
adjusted by the addition of some variable resistors.
VR1 is a 20 turn trimpot which can be used to set the intercept point. VR1
varies the voltage on pin 5 from around 3 volts to 4.99 volts. Referring to the
graph in FIG 9 of the AD8307 datasheet, this represents an intercept point
variation from -96.5dBm at 3 volts to -82.9dBm at 4.99 volts at 10 MHz.
However, in this design the inclusion of the input UHF boost circuit causes
the intercept point to shift.
Measurements taken on the prototype indicated that at an input level of -20dBm, the DC output
voltage was around 1.3 volts as against the published value of around 1.7 volts. In other words, the
traces on the graph in Fig 9 all fell by around 0.4 volt. This was for a voltage of 4.0 volts on pin 5.
In the RFPM1, the intercept is normally set to give
the lowest reading with a 50 ohm termination
across the input. However, some users may require
a different intercept point depending on their
application. For this reason, the trimpot can be
adjusted through a small hole in the RF head
enclosure. It can also be used to match the RF
heads when 2 are being used. There is a spread in
characteristics between individual chips and the
boost component values, which results in different DC outputs for a given input level. The intercept
adjustment can be used to help match 2 RF heads into the one meter unit. Initial setting for VR1 is
fully clockwise. Measurements on 2 prototype probes show the variation in intercept points.
VR2 is also a 20 turn trimpot which can be used to adjust the slope of the log output. Without VR2,
the nominal slope is 25mV/dB. The slope for the RFPM1 is adjusted for 20mV/dB which simplifies
the meter circuitry and matches some of the other designs published.
VR2 is adjusted using the VK3AQZ CAL1 calibrator, or an accurate signal generator. Initial setting is
11k between pin 1 and 2 (measured in circuit).
A signal of -20dBm is fed into the unit and the DC voltage on the calibration terminals of the meter
unit are noted using an accurate DVM. The signal is then dropped by 10dB and a second reading
noted. The calibration terminals on the meter unit are scaled to 10mV/dB so a change of 10dB should
result in a DVM voltage change of 100mV.
So VR2 is adjusted till a 100mV change is achieved for a 10 dB input level change.
The CAL1 uses a 5 MHz crystal oscillator module to provide a square wave test signal of known
value and easily measured with a digital multimeter. The CAL1 calibrator has a toggle switch which
can be used to select -20dBm and -30dbm. This is used to set the slope.
AD8307 SMD
Slope and Intercept adjustments.

Page | 5
Using 5 MHz simplifies the calibrator circuitry. Also, at this frequency, an oscilloscope can also be
used to set the input voltage from a signal generator fairly accurately.
However, one must note that the AD8307 can have small variations with frequency.
Also, the waveform is very important. The AD8307 is designed to respond to sine wave shapes and is
therefore most accurate with sine wave inputs. However, a square wave can also be used and it has
been found that at an input level of around -20dBm, a square wave of half the peak to peak amplitude
of a sine wave gives the same DC output with a high degree of accuracy. This characteristic behaviour
was documented by Bob Kopskie in a QEX article, Jan/Feb 2004 titled “A Simple RF Power
Calibrator”. See reference 2.
The peak to peak amplitude of a square wave can be measured with a DVM meter. So a square wave
fed into the AD8307 can be used to calibrate the AD8307. The CAL1 calibrator makes use of this
unusual response to a square wave. It should be noted that this relationship only holds around an
input level of -20dBm to -30dBm. The AD640 also exhibits the same behaviour. Fig 23 in the AD640
datasheet shows a graph of the deviation of the square wave to sine wave relationship at other levels.
Because of this, you cannot use a square wave to test the linearity of the AD8307 over its full range of
90dB. The use of a square wave will give misleading results making it look as if the chip is not very
linear. Because of this, the CAL1 calibrator also produces a 5 MHz sine wave which is adjustable.
Used in conjunction with an accurate stepped
attenuator, one can check the linearity of the meter
over a much wider range. The CAL1 sine wave is
generated by passing the square wave through a low
pass filter. Although this is reasonably simple, it does
mean that some harmonic energy can still be part of the
waveform making this only a moderately accurate
method. A good clean sine wave from a signal
generator, of known amplitude, is the best way to test
the linearity of the power meter.
However the CAL1 calibrator is a good compromise
between cost and performance.
It should be noted that the attenuation at UHF is not the same as at VHF in some attenuators. Leakage
across attenuator components can be much higher at UHF due to the intrinsic capacitance of standard
off the shelf resistors. The AQZ attenuators use surface mount resistors but these are still not as good
as the specially designed resistors such as those used in HP professional attenuators. These resistors
have a small shield foil wrapped around the body and small tabs are adjusted in the factory to ensure
non-reactive 50 ohm impedance is maintained over a wide frequency range.
With a high bandwidth RF chip, like the AD8307, an imbalance in the harmonic content of a signal
reaching its input can produce misleading results. For example, an attenuator which produces a 10dB
drop at 5 MHz may not produce a 10 dB drop at, say, the 10th harmonic of 50 MHz. This imbalance
effectively alters the waveform and the AD8307, responding to the higher level 50 MHz harmonic,
may appear not to be as linear as predicted in the data sheet. Unless one has accurate test equipment
and excellent test loads and attenuators, the performance of the AD8307 may seem to be not as good
as the datasheets. This is not a fault of the AD8307 but rather a lack of good test gear capable of
working accurately up to 500 MHz with minimal reactive impedance problems and resonances.
The RFPM1 is a very wideband device so please bear that in mind when calibrating the RFPM1.
Referring to Fig 1, pin 3 of the AD8307 IC has an optional 100nf capacitor (C8) connected to ground.
Using a switched attenuator