Telechargé par Francis Barthe


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
The AD8307 converts amplitude variations of an RF input voltage to a DC output
AD8307 SMD
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
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 Ω load.
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
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
Page | 1
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 Ω set 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 | 2
RFPMH – RF head circuit description.
RFPMH detector head
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 detector PCB.
The SMD parts are on the copper side, but values printed on the component side
RFPMH - RF detector head circuit
Page | 3
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.
AD8307 SMD
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
Slope and Intercept adjustments.
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.
Page | 4
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.
Using a switched attenuator
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.
Page | 5
This capacitor can be added to improve the low signal, low frequency log law conformance.
It becomes part of the internal offset circuitry. It can be left off if the unit is mainly used above 1MHz
or it can be increased to 1uF for very low frequencies. Should you wish to go below 5 KHz you will
also need to increase the input coupling capacitor C2. Most designs do not use a capacitor on pin 3.
RFPMH case and optional connectors.
The RF head is housed in a small split case aluminium enclosure.
The enclosure is designed to accept a number of different RF connectors with 25mm flanges. This
includes SO239, N type, and wide flange BNC connectors. Select the connector that suits your setup.
The SO239 is convenient if the usage is mainly for HF. The BNC is a good compromise. However
you need to be aware that there are 2 sizes of BNC connector. The video BNC connectors are
nominally 75 ohm whilst the RF BNC connectors are 50 ohm. As you may be aware, the impedance is
a ratio of the inner to outer diameter so the centre pin of a 75 ohm connector is thinner than a 50 ohm
connector. Plugging a 75 ohm male into a 50 ohm socket may result in a loose connection and vice
versa for the opposite combination. Unfortunately suppliers of these connectors often ignore this
subtle difference and as a result, some connections become unreliable due to this mix up.
The BNC connectors supplied by me are purchased as 50 ohm RF types. However, I have received a
mixture at times from suppliers. So if you wish to use BNC connectors for this instrument please bear
that in mind. I use BNC as they are used on most of my gear. They can be used at VHF and UHF, are
conveniently small, and are low cost. For serious VHF and UHF use, I would recommend type N.
The enclosure itself is a split unit. Because of this, you may experience some RF leakage through the
gap. However, this particular case is low cost and very convenient to use and quite a good
compromise. In some critical usage, you may use adhesive backed copper foil inside the case to help
reduce the leakage. In addition, the cable input may require better filtering, such as the use of feed
through capacitors in an extended section at the rear of the case, or a ferrite choke.
Finally, the RF head connects to the meter unit via shielded balanced flexible microphone cable. This
cable is quite robust and well shielded. A 4 pin microphone type connector is used as the interface.
These connectors are readily available and are sufficiently robust. Extension leads can be easily made
up using these connectors for applications where the RF head is remote from the meter unit.
The RF head PCB contains a small robust metal clamp to ensure a reliable connection is maintained.
Page | 6
RFPMM – Meter unit circuit description.
The circuit of the RFPMM meter unit is shown below.
The RF head connects to the meter unit via 4 pin microphone
style connectors SK3 and SK4.
Provision is made for 2 RF heads which can be selected on
the rear panel with a slide switch.
The slide switch selects the RF head power feed and the
AD8307 output signal.
RFPMM meter unit
The selected signal feeds the RFPMM printed circuit board
via a 4 pin XH2.54 connector.
The RF head DC signal slope at this point in the circuit is
Op amp IC3A is configured as a non inverting unity gain buffer. The output of IC3A is split 3 ways.
One of the splits goes to a resistive voltage divider consisting of R14 and R15. This divider reduces
the slope to 10mV/dB. The output of this divider goes via CN6 to the 2mm banana sockets on the rear
panel. This output can be used to calibrate the RF head independently of the 2 meters in the unit. This
output also contains any modulation on the detected RF signal. It can also be used to feed an
oscilloscope as part of a spectrum analyser. Please note that the logarithmic response of the AD8307
will produce a log version of any amplitude modulation of the RF signal. R14 and R15 are 1%
resistors in order to maintain a high accuracy.
Page | 7
Please note that the 2:1 divider feeding the 2mm banana sockets can introduce and additional error
due to the tolerance variations in the resistor values. Better accuracy can be obtained if you measure
the output of the AD8307 directly at the output of IC3A before the 2:1 resistors. At this point the
reading will be 20mV/db.
The second split from IC3A feeds IC3B which is configured as a 3
times amplifier producing a slope of 60mV/dB. This level is
required to feed the analogue meter which has a full scale
sensitivity of around 1mA.
Trimpot VR4, in series with the meter, can be adjusted through the
front panel to calibrate the needle of the meter. The same meter is
also used to measure the internal battery voltage via selector switch
M1 Analogue meter
The analogue meter was found useful when adjusting tuned
circuits. The direction of movement is easier to pick up during the adjustment. Due to the low cost
movement used, the meter does not have good balanced ballistics and tilt compensation so the dial
accuracy is around +- 2dB. Changing the angle of the unit can change the reading. However it is quite
useful for tuning adjustments and quick signal variations. In addition, the magnetic field is not linear
over the arc of the scale. This causes the scale accuracy to vary from meter to meter. This occurs
mainly around the -60 to -80db area. Below -70db, the reading maybe up to 5 dB out. Also another
factor affecting the tracking of the 2 meters is the inclusion of the 3x gain Op amp in the analogue
meter signal path. At low levels, the Op amp offset adds an error. For example a 1.6mV output offset
can add 1dB error in the analogue meter reading. Use of a higher quality rail to rail Op amp will
reduce this error.
Accuracy can also depend on the intercept point voltage. Changing the intercept point shifts the
analogue meter readings in the -60 to -70dB area. Measurements made on 2 prototypes for 3 levels of
intercept illustrate this problem. These tests are shown in attached spreadsheets. The scales on the
analogue meters were made for an intercept setting of 4.5 volts on pin 5. You may want to set the
intercept voltage close to this value for a better accuracy in the analogue meter at low levels.
The analogue meter is also used to measure the internal battery voltage.
A 6.8 volt Zener diode is used to expand the battery voltage reading on the meter scale.
VR5 can be adjusted via the front panel to calibrate the battery voltage meter scale.
The unit will work down to around 7.5 V DC.
The third split from IC3A goes to an adjustable voltage divider consisting of R16 and VR6.
This in turn feeds a 200mV LCD digital panel meter.
The 20mV/dB slope needs to be reduced to 1mV/dB so that the panel meter reading is in dB.
The negative terminal of the panel meter is biased at around
85mV so that 0dBm reads “00.0” on the display. The panel meter
will read a positive number when the + input of the meter has a
voltage higher than the – terminal. When the + terminal has an
input lower than the – terminal, the meter will display a negative
number. For example, an input of 65mV, on the positive terminal
will display -20.0 if the negative terminal is biased at +85mv.
The exact amount of bias depends on the exact DC voltage at the
output of the AD8307 for a 0dBm signal, and the adjustable
divider setting. VR3 varies the bias on the negative terminal of the
meter and can be adjusted from the front panel to give a correct reading for a corresponding input on
the + input.
M2 Digital meter
Page | 8
The bias is derived from a 5 volt regulator. Without the regulator, changes in battery voltage
causes the meter readings to change.
One problem with most of these panel meters is that the supply to the meter has to be independent of
the supply feeding the circuit under test. There are some models available that can share the same
supply and some RFPM1s will have these meters. However the availability of these meters is
unreliable so provision has been made to use the more common type of meter. This is achieved by
using the isolated DC/DC converter chip, IC5. The use of a separate supply also removes some issues
with the bias circuit.
Something to bear in mind with the digital display is the apparent accuracy. The display gives
readings down to 0.1dB. The meter itself has an accuracy of +- 0.5% and +-1 count so there is
a measurement error of +-1mV over the 200mV range plus a display count error of +-0.1mV.
This means that the reading could be wrong by as much as 1.1dB irrespective of the AD8307
specification of +-1dB. In some situations that could be significant. However, as is the case with
some commercial instruments, if a calibration signal is applied to the meter of known value, and the
meter readouts are adjusted to that calibration signal, then the accuracy becomes that of the calibration
signal for the measurements being performed at that time. As mentioned, even expensive laboratory
instruments have a calibration facility for exactly the same reason. As well as the intrinsic error within
the meters and IC, there are additional variations possible due to external factors such as humidity and
temperature. The RFPM1 does not have any temperature compensation. Although the AD8307 is
relatively temperature stable, other sections of the circuits may exhibit some variation with
Again, providing the calibration signal is temperature compensated, the readings can be calibrated at
the time of measurement to help reduce these errors.
When using the instrument for comparing, say, input and output values or small amplitude changes,
the absolute values may not be exact, but the variation, or comparison, can still be quite accurate over
the measurement range and better than +- 1 or 2 dB.
In this situation the test is looking for a relative change and this can be of the order of a fraction of a
dB and quite accurate. The absolute readings maybe a 1dB or 2dB out but the change observed can be
quite small and valid. In many cases, a lot of testing involves relative measurements only.
On the other hand, an absolute error of 1dB can represent a
RFPMM internal view
significant number of watts when testing a high power RF amplifier.
Please bear in mind that the RFPM1 cannot perform to the same
level as an expensive laboratory instrument. Considering its low
cost, it achieves quite a good result for amateur usage.
The remainder of the meter unit contains a battery pack which can
take 6xAA batteries. Alternatively a standard 9 volt battery can be
The power consumption is around 50mA for the meter unit and
10mA for each RF head.
RFPMM PCB sections: Main PCB, Rear PCB, DVM PCB, and LED PCB
Page | 9
Kit contents and tools.
Before starting to assemble the kit, check that all the components have been supplied.
Refer to the “Component list” for details.
Assuming all the components are there, and correct, you can commence the assembly process.
You will need the following tools:
For the SMD components in the RF head, you will require a soldering iron with a fine tip,
a pair of fine tweezers, a small fine pointed pick (toothpick or spike) for holding down the
components, and some fine solder. Small dobs of glue can also be used to temporarily hold
components in place. A piece of solder wick will be handy for removing excess solder.
Soldering temperature should be around 190 degrees C. This is just above the Tin Lead Eutectic point
for 60/40 solder. However most non industrial irons cool significantly once the solder melts, and in
order to maintain the solder melt, you may need to set the dial as high as 300 degrees or so, depending
on the size of the joint and size of the tip. Very small tips have low thermal inertia and cool rapidly
during the solder melt phase. So higher initial temperatures are needed.
For the remainder of the kit, a soldering iron with normal tips, a pair of fine tipped side cutters, small
and medium sets of pliers, a set of cross head and flat blade screwdrivers of various sizes, a 3mm nut
driver, an alignment tool set with a small flat blade for trimpots. Size #1 and #00 cross head
screwdrivers for the small screws on the microphone plug body.
A 4.5mm diameter drill for winding L1, or a rod of any material of the same diameter.
A cutter for stripping the audio cable covering. A small heat gun for heat shrink.
Some self closing Tweezers would be handy.
A 12 V PC fan running from a battery or plug pack maybe suitable for blowing away fumes if the
soldering rate is low enough.
4.5mm drill or rod for winding L1
Test equipment
Page | 10
You may wish to commence construction of the RF Head.
Refer to the circuit diagram and PCB layouts.
Assembling the RFPMH detector head
Commence with the PCB components.
Solder a 3mm nut to the underside of the PCB at the connector end. This is used for
holding the PCB in the lower half of the case. Use a screw to hold it in place during
Solder the SMD components next.
IC1 – AD8307
R1 = 51R1, R2 = 470R, C1 = 15pf, C2, 3, 4, 4A, 7, 8 all = 100nf
Low profile parts: R3 = 4R7, R4 = 100R, R5 =1k, R6 = 51k 1%, R7 =33k
Mono ceramic capacitors: C5, 6 all = 100nf
IC2 = 78L05
VR1 = 50k trimpot. Location of screw to match hole in case – see PCB layout.
You can set VR1 fully clockwise as a starting position. Push down flat on PCB.
VR2 = 50k trimpot. In circuit resistance between pin 1 (screw end) and pin 2 (middle pin)
measured on prototype at 11k in circuit. You can preset the trimpot to this value to start.
Fit L1 = 1 turn on 4.5mm diameter 0.5 to 0.7mm diameter wire.
LED1 = 3mm Blue LED. Flat on body towards edge of board. Position the top of the LED
about 10mm to 12mm above the PCB.
Fit 6 off PCB pins for RF IN, RF GND, TP (test point), +V, OP, GND (cable connections).
Attach clamp with 3x10mm dome screws, washers, nuts (on top) but leave loose.
That completes the PCB section.
Page | 11
Continue with the case assembly.
Attach a piece of adhesive copper foil piece to the bottom of the case at the connector end.
Fold a short section over the edge of the case for clamping under the RF connector flange.
The foil is very thin so try not to tear it. Use a second layer if necessary. Push a hole in the
foil where the hole used for holding the PCB in place is located. Attach a second piece on the
side of the case opposite the screw with the copper facing the connector flange and clamped
under it. A short piece of wire will be soldered on this foil to connect this to the nearby PCB
RF GND pin. The adhesive side will be on the case.
The RF connector can be attached to the bottom half of the case using 3x10mm screws.
Try to keep the screws straight so as not to cross thread the screw. The thread in the case is
delicate and easily damaged. Do not use force. Tighten just enough to hold the connector.
Prepare the ends of the rubber coated audio lead for attaching to the PCB. The leads will be
short. The clamp will clamp over the rubber. Carefully separate the braid fibres and twist to
form a tight lead. Tin the ends and attach to the 3 pins on the PCB.
Red to +V, White to OP, and the braid to GND. See Video.
Fit a grommet to the case rear plate with the hole in it and pass the other end of the audio
cable through the grommet. Do not forget to carry out this step before attaching the 4 pin
connector at the other end.
Fit the PCB into the case and fix it using a 3x10mm countersunk screw into the nut on the
PCB at the RF connector end.
Using the small countersunk screws supplied with the case, attach the grommet plate to the
case bottom. You may notice the grommet is off centre. The lowest portion goes to the
bottom so that the cable is lined up with the clamp. Try to keep the screws straight and leave
slightly loose at this stage. Leaving the screws slightly loose allows easier alignment of the 4
screws when the upper case section is attached. The same applies at the RF connector end.
You can now attach the 4 pin microphone plug at the other end of the cable as per the circuit
details. To remove the connector shell, undo the tiny countersunk screw using a #00 size
cross head screwdriver and twist a small amount. There is a keyway which guides the shell
onto the connector body but it is offset slightly. Pulling straight without the twist will not
Solder the wires, reattach the shell and mini screw then do up the clamp at the rear.
A #1 cross head screw driver is used here.
It is a good idea at this stage to check the continuity, and any shorts, from the connector to the
PCB whilst the RF head upper case is not attached.
Assuming the audio cable is wired correctly and checked, you can now attach the upper
section of the case. The trimpot and LED holes should all line up. You can now attach the
labels on the non split faces.
That completes the RFPMH detector probe.
Page | 12
Assembling the RFPMM unit
Commence with the main PCB components.
Fit low profile parts first. R8 to R18, D1, D2, C10 to C14, and wire link 1. Refer to parts list
for values.
Fit SK7 which is the 8 pin IC socket for IC3. Check correct orientation.
Fit the 2 pin and 4 pin XH2.54 connectors. Check PCB drawing for correct orientation.
If you put them in the wrong way round, they can be a pain to remove without damage.
Fit C9 = 220uf electrolytic capacitor. Check orientation on PCB.
Fit 78L05 regulator IC4, and IC5, the VB1212LS-1W (or B1212) DC DC converter.
Fit trimpots.
Fit SPDT PCB pin toggle switches. Clean the pins with a small file or blade to improve
soldering. Solder quickly. The plastic material in the switches may melt if you take too long.
VR3 = 100R
VR4 = 5k
VR5 = 5k
VR6 = 2k
Prototype measured 20 ohm in circuit.
Prototype measured 4.4k in circuit.
Prototype measured 3k in circuit
Prototype measured 1.1k between Pin 1 (screw) and pin 2.
That completes the main PCB section.
Assemble the rear connector panel.
Solder a cleaned 3mm nut in the small hole at the bottom. Use a screw to hold it in place
during soldering.
Attach the 4 pin microphone sockets to the rear detector socket PCB. Tighten the large nuts
but be careful not to strip the thread. Use a plug to hold the socket in line while tightening.
The sockets have a key which matches the keyway in the plug. Orient the key so that it faces
the case top. This makes it a bit easier to determine the plug orientation when plugging in.
Temporarily attach the slide switch with one or two 3x10mm csk screws. This is to allow you
to wire the mic sockets to the switch before mounting to the rear case. These screws will be
used in a later step to hold the PCB to the case. Some slide switches may need nuts and
washers. Some are tapped 3mm.
Solder short wires between mic socket pins and switch tabs as per the relevant
diagram. Pin 1 of each socket wires to left and right switch tags of upper pole.
Use red insulated wire. Pin 2 of each socket to left and right switch tags of lower
pole. Use yellow insulated wire. Pin 3 of each socket to copper on PCB (ground).
Use black insulated wire.
The wires on the 4 pin plug going to the main PCB can be attached later.
That completes the rear connector PCB section.
Page | 13
Assemble the DVM panel.
Solder a cleaned 3mm nut in the small hole at the top. Use a screw to hold it in place.
Fit a red and a black 2mm banana sockets and washers in the 2 larger holes. Tighten the nuts.
That completes the rear DVM PCB panel.
Attach front panel label to lid. There are 4 holes for the handles and 2 smaller holes which
hold the label in place. Fit a 3x 10mm, cut down to 8mm, black dome screw, washer and nut
in the top hole. The cut down bolt is needed in order to clear one of the XH connectors on the
PCB. The handles are best fitted last.
The upper middle screw is also fixed in place along with the LED PCB.
Fit the 2 meters in place. The digital meter clips in. The analogue meter has white
rotating clamps at the side of the bezel. There are small pips in the bezel sides which mate
with small holes in the clamp. The pips lock the clamp in place. One end of the white
clamps has been filed down to allow for the thicker plastic front panels. The filed down end is
identified with a black mark.
You can also solder the LED onto the small LED PCB as shown in the video.
Bend the LED legs so as shine down into the analogue meter bezel.
Before fitting the main PCB onto the front panel, plug in all the 2 pin and 4 pin XH2.54 leads
as per the wiring drawing.
These are: Battery supply, Analogue meter, Digital meter (4 pin). Rear connector panel
(4 pin), DVM panel, LED panel.
At this stage you should check for short circuits on the +V rails. both the 8.7V and 5V rails.
Plug the uPC358 IC into the IC socket.
Remove the first nut from the switch shafts, leaving the star washer and second nut on the
shaft. Push the switches through the mounting holes on the front panel. Attach the second nut
so as to hold the switches on the panel and do up the nuts firmly but not so tight as to strip the
threads. The trimpot screws should all line up. If not, then check their orientation on the PCB.
You can now fit the rear panel and DVM panel. The screws holding the slide switch are
removed first leaving the switch floating. These screws are used to hold the rear panel in
place along with a 3x10mm nickel dome screw in the lower hole with the nut soldered to the
Page | 14
The various leads can be tidied up with zip ties. Consult the drawings in the manual and the
video for more detail. Solder the various leads to the terminals and PCBs.
Solder the red and black wires from the battery snap to the corresponding colour wires on the
XH2.54 lead going to CN3 (the battery input). Use heat shrink over the joins.
The battery holder can be loaded with 6xAA batteries and clamped in place with the metal
clamp and a 3x10mm countersunk fixing bolt, washer, and nut. A piece of self adhesive
rubber can be placed inside the clamp. Turn the power toggle switch off prior to doing this.
Do not plug the snap in yet.
Check for any shorts between the red and black battery wires with a meter. If there is a low
resistance then something is incorrectly wired, or the wrong 2 pin lead has been selected.
At this stage, I would recommend that the one of the leads to the analogue meter be left
unsoldered in case the wiring is wrong.
Once you are satisfied there are no short circuits on the battery lines, you can switch on the
battery and check that the LED and LCD light up. Some LCD meters have a small black
button which turns the backlight off.
Select the battery test position on the front panel toggle switch marked RF – Batt. and
measure the voltage across the 2 wires going to the analogue meter. You should get around
2.5 volts DC. Then attaching the loose wire to the terminal on the back of the meter you
should read around 140mV (see video).
This step is as a precaution in case you have a fault or mistake.
Now attach the handles with the 4mm dome screws and washers. The lid can be screwed
down now.
There are also a pair of angled metal stands which clip into the slots at the back. This tilts the
meter making it easier to read.
There are also 4 self adhesive rubber feet which can be attached to the rear for use when the
meter is lying flat.
That completes the RFPMM unit.
Page | 15
At this stage do not plug the detector head in.
Commencing with the meter unit:
Turn the power switch off, and with a multimeter, check for short circuits or very low
resistance on the +8.7 V rail of the meter unit. Note this rail may be as high as 9.3v with fresh
batteries. It should be around 0.15 V to 0.35 V below that of the battery voltage due to the
voltage drop across D1.
The bottom of the PCB is easily accessed. The pole (middle terminal) of SW1 is the +8.7 V
rail. Referring to the photo of the bottom of the meter PCB can help with testing.
In order to avoid damage to the analogue meter, it is wise to check that the zener diode is in
the right way. It’s function is to drop the 8.7 V down by 6.8 V. If it is in the wrong way, the
voltage drop will only be the forward drop of an ordinary diode which is around 0.6 V.
Switching on the battery test in this situation would put too high a voltage on the meter and
destroy the coil. Use a multimeter switched to “diode” test. With the + terminal of the
multimeter on the +8.7 V rail, and the – terminal on the track going from D2 diode
anode to VR5 trimpot centre terminal, the reading should be a high resistance or no reading
at all.
Reversing the leads should show a normal silicon diode drop of around 0.7 V.
Do not skip this test.
Check you have around 9 V DC at the battery holder snap terminals. Operating the front panel
switch should bring on the LED over the meter and the LCD display should light up. If your
unit has an LCD with a small button switch, push it down to turn the blue backlight on.
Page | 16
Check the 5 volt regulator. The junction of R18, C13, and pin 1 of the 78L05 should
read +5 V with respect to earth. If not, check orientation of the 78L05 and for shorts
on the +5 V sections of the circuit.
(You may have already performed these tests during the last stages of the assembly.)
The analogue meter should read nearly zero. And the digital meter will probably read
around -80 to -88. If not, you can adjust the M2 bias trimpot to around -80.
You will trim the bias later.
Measure the actual battery voltage at the battery holder, or on pin 2 of CN3.
Switch the RF – Batt. switch to Batt. and adjust the battery cal trimpot for a reading
the same as the multimeter. Estimate the needle position between the major
voltage marks. This represents the actual battery voltage. With a fresh battery it may
read nearly full scale. After some use the battery voltage will fall to 9V or lower and
that will enable you to set the battery calibration trimpot more accurately.
Measure the voltage on Pin 1 of the RF detector head socket, SK3. The slide switch
should be slid towards this connector. It should read around +8.7 V. This test is to
ensure that you have wired the switch and connector correctly. +8.7 V incorrectly
wired to pin 2 instead, would damage the AD8307 output circuit. Slide the switch
across to the other RF detector socket and check Pin 1 for +8.7 V.
For now, that completes the initial checks on the meter unit. Assuming there are no
solder bridges or wiring errors, we can now check the RF head before plugging it in.
Check for any short circuits on the +V rail of the RF head. This can be done by
putting the ohm meter between pin 1 and 3 of the 4 pin microphone plug. Or you can
remove the top half of the RF detector case and measure at the cable entry pins.
Assuming there are no wiring or component errors, you can plug the RF detector head
cable into one of the meter unit sockets and select that socket with the slide switch.
Turning the unit on, should bring up the LED on the RF detector.
If not, check that the 78L05 in the RF head is working.
You should get around 5 volts on either side of R3, the 4R7 ohm resistor, with respect
to earth.
Also, there should be a small increase in the reading on the analogue meter.
At this stage, the digital meter may be way off.
So adjust the M2 bias trimpot for a reading around -73
Page | 17
We now need to calibrate the unit. You will need a clean sine wave signal of known level.
The frequency can be 5MHz. But you can use whatever frequency you may have.
You can also use a standard signal generator and an accurate oscilloscope to measure the
peak to peak value of the sine wave. A 0dBm sine wave has a peak to peak amplitude of
632.3mV. A level of -10dBm is a good place to start. A -10dBm signal across 50 ohm has a
peak to peak amplitude of 200.5mV. However, if using an oscilloscope, a 0dBm signal
feeding a 10 dB pad might be easier to set up.
Alternatively, you can use the CAL1 calibrator or a homemade version. Refer to the Bob
Kopskie article for his version of the calibrator. Reference 2.
Adjust the intercept trimpot for around 4.5 volts as measured at the TP test point near
L1 coil.
At this point in time you may want to consider the theory behind log amplifiers.
You may wish to consult the AD8307 datasheet for an explanation of the intercept point and
slope in log amplifiers.
For a discussion on the intercept point, refer to Appendix A.
Before adjusting the slope, we can rough set the 2 meters for a -10dBm signal.
With a -10dB input signal, adjust the “M1 Cal” trimpot so that the analogue meter
reads -10dBm. This will change after we set the slope to 20mV/dB.
Adjust the M2 Cal trimpot for a reading of -10.
We can now adjust the slope to 20mV/dB.
For this adjustment we need an input signal we can switch between 2 accurately
known levels in dB.
The most common level used for this adjustment is -20dBm and -30dBm
This is around the middle of the linear transfer function.
The CAL1 calibrator puts out a 5MHz square wave whose output level can be set
accurately using a digital multimeter. The square wave can be switched between 2
levels. Each of the levels is set by a trimpot accessible from the front panel.
The facility to switch between -20dBm and -30dBm using the toggle switch assists in
setting the slope.
The AD8307 slope is adjusted by VR2 in the RF head.
The unadjusted slope is nominally 25mV/dB.
The trimpot allows this to be reduced to 20mV/dB.
Page | 18
To set the slope, a DVM is plugged into the 2mm banana sockets at the rear of the
meter unit. This output is a scaled 2:1 so that a 10dB change in input level results in a
100mV change in DVM reading.
On the prototype, with my setting of intercept, I found that -20dBm gives a reading of
around 630mV. And a -30dBm signal gives a DVM reading of 530mV.
So adjust VR2 so that the DVM reading changes by 100mV for a10dB input change.
This adjustment tends to take a few goes. The DVM readings change for both the
-20dBm level and the -30dBm level. So there is a bit of tail chasing going on.
I found that tending to “over adjust” by 5mV at a time produces a convergence of the
2 readings. During adjustment, mentally keep in mind that you are looking for a
change of 100mV. It does not matter what the first digit is, as long as the last 2 digits
are the same for both levels. The aim is to get these 2 last numbers the same.
Mentally note whether the difference is getting smaller or larger. If the difference is
getting larger, the slope is getting higher, so turn the trimpot counter clockwise.
If the difference is getting smaller, adjust the trimpot clockwise to increase it.
Having set the slope, the analogue and digital meters can be more accurately set.
With a -20dBm signal, adjust the M1 and M2 CAL trimpots to read -20dBm.
The setting of the digital meter can be a bit more complex. The signal to the digital
meter needs to be 1mV/dB. Please note that the digital meter is factory set so that a
100mV DC input produces a reading of 100. Most of these meters have a gain trimpot
somewhere at the rear. It is important to know if the factory setting is correct before
one can use it to measure an input voltage accurately. Most will be correct but I have
come across units where 100mV might read, say, 97 or 102, or similar. So these
meters need to be set up so that accurate 50mV, 100mV and 150mV inputs actually
read those values. In other words, the input gain and slope of the meters is correct.
This can be set by testing the meter independently of the RFPMM unit.
A stable low voltage DC is required (i.e. a 1.5 volt battery) and an accurate digital
multimeter. A pot or some resistors are then used to form a voltage divider which
produces the 3 levels above as read on your digital multimeter. The meter’s trimpot is
then adjusted to confirm the meter accuracy and the slope is correct. A voltage
reference IC, in conjunction with a divider made using 0.1% resistors can also be
used to produce an accurate 100mV test voltage.
Please note that an accuracy of +-0.1mV at 100mV is required for verifying the
accuracy of the LCD meter.
Also note: If the LCD meter does not read 100mv when a 100mV DC signal is fed
into it, the readings on the LCD meter will not match the analogue meter scale.
The error will mainly be at the low signal end.
For example, if the input signal is -70dBm then the LCD meter should read -70.0.
Page | 19
If however, the LCD meter reads 75mV instead of 70mV, then when you look at the
analogue meter you will see a difference between the 2 meters of 5dB. The high
signal end at 0dBm will appear to match since at 0dBm, the LCD meter will have a
zero voltage difference across it’s input terminals due to the bias voltage on the
negative input terminal. And the gain error will be zero at zero input. So the LCD
meter matches the analogue meter at 0dBm. But at -70dBm it will be 5dB different.
This can mislead you when checking the tracking of the 2 meters.
So if the 2 meters do not track, check that the LCD meter has the correct gain.
The LCD supplied with the kit will be checked and set. So you can assume it will be
correct. However if the 2 meters do not track, check that the LCD meter is still set
Moving on, M2 can be adjusted by using 2 signal levels such as -20dBm and -30dBm
to set the correct slope for the digital meter.
With an input of -20dBm you would have set the LCD meter to read -20.0.
Now feed a -30dBm signal into the meter. If the slope of the LCD meter is 1mV/dB,
the reading will fall to -30.0.
However, it may fall to only -29dB. This means the signal path to the LCD meter has
a gain less than 1mV/dB. The purpose of the M2 Cal trimpot is to increase or decrease
the LCD meter gain slope.
So if the LCD meter changes by less than 10 for a 10dB change, then the gain is too
low and you need to increase the M2 gain by turning the M2 Cal trimpot clockwise.
This will change the reading at -20dBm. However ignore that for now and keep
adjusting M2 Cal until the difference in the LCD meter display is 10 for a 10dB
change. Once you have that, the slope or gain of the M2 signal path is 1mV/dB.
You can now correct the reading with the M2 bias trimpot. The bias trimpot slides the
reading up or down without changing the slope. So in the above example you might
end up with an LCD meter reading of -18.5 with a -20dBm input, and – 28.5 with a
-30dBm input. So now adjust M2 bias till the meter reads -20 with a -20dBm input.
The DVD video showing the testing and calibration of the RFPM1 can be used to
assist with this setup.
Once the slope of the RF power meter has been set, the calibration is finished.
However you may want to check the overall linearity of the meter.
To do this you need a very clean sine wave which can be accurately adjusted or
switched in level.
A good quality stepped attenuator can be very helpful here.
The AD8307 has a very wide bandwidth. This is great for a power meter but it makes
the testing of it’s linearity more complex.
The square wave output of the CAL1 calibrator cannot be used to measure the
linearity of the meter. The accurate relationship between the sine wave response and
the square wave response only holds over the range of -20dB to -30dB. This is evident
Page | 20
when you study the graphs for the AD640 which is also a log amp with the same
characteristics as the AD8307.
The other problem I found when using a square wave through a switchable pad, is the
flatness of the attenuator. Even the good quality HP attenuator, had frequency
response variations for each step. For example, it may give an attenuation of 20dB at
10 MHz but not at 400 MHz. Normally with a narrow band power meter or tuned
detector, this is not a problem. However, a detector that responds to all harmonics of a
10MHz square wave all the way up to 500 MHz will see a different distribution of
frequency energy and give a result which may appear to be too high or too low.
Certainly the 50th harmonic of a 10MHz signal will be quite low. However, there may
be parasitic resonances as well as uneven responses which collectively act to
distort the waveform and produce strange results.
We tested this behaviour by filtering signals into the AD8307 and noting the
difference with and without filters. And one must note that with the AD8307 IC
having a dynamic range of 86dB, even these filters need to be excellent.
For this reason, the CAL1 calibrator also produces a 5MHz sine wave.
A 2 section low pass filter at the output of the square wave crystal module
is used to produce the sine wave.
This is not ideal, but it is considerably less costly than a fully fledged sine wave signal
generator with a very pure output.
The sine wave level in the CAL1 is variable from around -40dBm to +13dBm.
This enables one to get a reasonable idea of linearity over most of the range of the RF
power meter. Used in conjunction with a good set of attenuators, you can test the
power meter over it’s full range.
You can obtain a cleaner sine wave from the CAL1 calibrator by passing the output
through an additional 2 stage low pass filter. This test reveals that the linearity is
actually better than what you see with just the filter in the CAL1.
As these filters are not too difficult to make at 5MHz, the builder can try this also.
If you have access to a laboratory signal generator and attenuators, you will be able to
check the performance to a greater degree of accuracy.
With that level of precision you may then be able to pickup other linearity and
response variations such as those caused by strays within the RF probe housing,
temperature, humidity, regulator voltage fluctuations, etc.
However, for amateur radio work, and considering the cost, the RFPM1 accuracy is
quite satisfactory and a very useful instrument.
The tendency, in most articles I have read, is to believe the AD8307 is neither linear
nor flat in its frequency response.
Page | 21
However, unless one has “perfect” signals and attenuators all the way from near DC
to 500 MHz, then one cannot say what the cause of any non linearity in the
response of the AD8307 IC is.
Stray L and C at the input of the AD8307, bypass component self resonance, pieces of
metal used as an attempt to shield the input connector. etc., can also have significant
effects on the response – and often overlooked as a source of error by those testing
their version of the power meters.
Lastly, voltage measurements at the end of the calibration taken on the prototypes are
shown on test tables for the RFPMH and the RFPMM. Your voltages may differ
slightly due to component spread and accuracy of the slope adjustment.
If using 2 probes, you will need to adjust the intercept point so that both probes give
very similar readings on the meters. You will most likely be able to match the 2
meters from around -50dBm to +10dBm. However, due to differences between the 2
types of meters when responding to low level signals mixed with noise, the 2 meters
may not match. The difference could be as much as 5dB.
When matching 2 probes use one as the reference then adjust the second unit to
match. Adjust the meters etc so that -20dBm is displayed on the meters. Then adjust
the second probe to match. However they may not match at low levels.
The table below shows the result of matching 2 prototype probes. Note that in order to
obtain a match at -20dBm as read on the LCD meter, the intercept adjustment pin
voltage on the 2 prototypes varies by as much as 800mV. The low signal readings do
not match. The middle column shows the results when both probes have the same
intercept of 4.30 volts. Note the difference in the Vcal readings.
Page | 22
The following tables show measured component voltages on prototypes which
may help you to diagnose problems after construction. In this case, the intercept was
set to maximum.
Page | 23
Notice in the above measurements, the Vcal voltage shows an error of only 0.3 or 0.4 dB
over the range of 80dB (there is only a 3mV to 4mV change per 10dB change).
Also, the voltage across the analogue meter changes by around 1.4mV/dB or 1.5mV/dB.
For the analogue meter to track, the intercept pin voltage should be around 4.5 volts.
However that may not be possible if you have 2 probes, In that case set the primary probe to
4.5 volts and the second probe intercept point to match the LCD meter readings.
Page | 24
APPENDIX A: intercept point discussion.
Briefly any amplifier can be described by a graph or equation which tells you what the output
signal level will be for a given input signal level. A graph is a quick and convenient way of
predicting the output level for a given input level.
This graph, showing the output voltage for a given input voltage, is called the voltage transfer
function. Usually the vertical axis is the output voltage and the horizontal axis is the input
If the transfer function for a log amplifier is drawn on a graph with the X axis evenly
graduated as the log to the base 10 of the input voltages, then the graph of output voltage
versus the log of the input voltage turns out to be a straight line.
See Fig 21 in the AD8307 datasheet for a picture of an ideal log amp graph.
The AD8307 input/output function is a pretty good straight line for over 80 dB of input
signal. From the AD8307 transfer graph you can find the output voltage for any input sine
wave signal.
The intercept point is defined as that point on the transfer
function which passes through the X axis and Y = zero.
For the AD8307, the intercept point is that input voltage,
Vx, where Vout = 0.
Because the output voltage of the AD8307 is a linear
function of the log of the input, a meter with a linear
scale, connected to the output, can be used to measure
the input level in dB volts, or as a power level in dBm, if
you are measuring the input across a 50 ohm resistor.
Effect of lowering or increasing
the intercept point, Vx
In a real log amp, at the very low end of the input signal range,
the transfer function starts to bend away from a straight line as it is not possible to maintain a
straight line all the way down to zero volts in a practical amplifier. The intercept point is
really only a projection of that straight line where it would have passed through the axis, had
it been an ideal log amp.
An important thing to notice is that if you lower the intercept point (make Vx smaller), the
graph moves sideways towards the vertical axis and an input voltage will gross the transfer
graph higher up. So a given input signal now gives more DC out. So lowering the intercept
point gives more output! - sounds strange but we are talking about the log amp behaviour
This is the same effect as if you had added gain ahead of the device. So a lower intercept
appears to increase the sensitivity of the detector. A 0dBm input will produce a higher DC
output as the intercept point is lowered.
If you could continue to lower the intercept, eventually the 0dBm signal would cause the DC
output to hit the amplifier’s rail voltage and the output would stop rising.
Page | 25
In other words we have lowered the headroom of the meter and it’s ability to handle strong
signals. Alternatively, increasing the intercept point lowers the graph and improves the high
signal handling ability. But the input signal cannot go higher than the input stage can handle.
In the AD8307, the intercept trimpot adds a voltage offset which raises or lowers the intercept
The trimpot in the RFPMH adjusts the voltage on pin 5 from 3 volts to 4.99 volts.
At 3 volts, the intercept is at its lowest, and at 4.99 volts at its highest.
In the RFPMH and other designs with adjustable intercept, the voltage is normally set at 4.99
volts, or maximum. In designs with no trimpot, the intercept point is the same as if around
4.90 volts was there. Most designs do not use a trimpot here. However, in the RFPM1 it is
used to match 2 RF heads to the one meter unit.
Where you operate your AD8307 is really up to you. It is something you can play around
with. In the end, the meter has an overall dynamic range of around 86dB and moving the
intercept point up or down serves to either clip the high signal end or the low signal end.
It’s your choice where you want it. I operate it at a high intercept point in line with some
other designs. But I do change it if I want a better look at certain signal levels. In particular
those near the low end bend in the curve. I can readjust the meters with the front panel
trimpots so 0dBm still reads 0dBm on the scales. The calibrator is used for this adjustment.
Please note that intercept points that give a low DC output with low level signals may result
in the analogue meter not tracking the LCD meter around the -70dB area due to effects such
as Op amp offsets, etc. This will vary from meter to meter due to spread of characteristics and
is difficult to correct in a kit design (it would require the production of unique meter scales
for each kit). In this case, the LCD meter is the one to use for your measurements.
One further comment about the AD8307 intercept point is that the effective intercept point
depends on the input waveform. Referring to the AD8307 log amp notes, a square wave input
signal will give an output which is 6 dB higher than that of a sine wave of the same RMS
And Gaussian noise will give an output which is around 0.5dB higher.
In effect, the intercept point is different for each of these waveforms.
See AD8307 data sheet page 12 for more information and the discussion regarding the
response of a log amp to the instantaneous input signal amplitude variations.
Page | 26
W. Hayward, W7ZOI, and R. Larkin,
W7PUA, “Simple RF Power Measurement,”
QST, June 2001, pp 38-43.
R. Kopski. K3NHI. A Simple RF Power Calibrator”
QEX, Jan/Feb 2004 pp 51-54
R. Kopski, K3NHI, “An Advanced VHF
Wattmeter,” QEX, May/Jun 2002, pp 3-8.
Analogue Devices website:
AD8307 data:
Some AD applications documents:
Owen Duffy has made available his loop design calculator on his website.
Loop calculator:
The spreadsheet version is stored here:
First issue 16/3/2014
Updated: 11/6/2014 Added links to Owen Duffy website.
Updated: 24/6/2014 Changed the wording in the first sentence at top of page 8.
30 Regency Drive, Wonthaggi, Victoria,
Tel: +61 03 5672 4607
as above
Email: [email protected]
Page | 27