Downloads
Download the Download FreePC files in and the detector board layout png
file in zipped forma:
schottkydedtector080309
Find updates at
www.projects.cappels.org
Starting to do a little work at 330 MHz, I decided that my existing
field strength meters were not adequate for for my needs. The 330 MHz
project came to a stop, and I set out to put together a field strength
meter with a moving needle indicator, with higher sensitivity than the
others I have built, and the ability to indicate RF in the 2.4 GHz
band.
Before getting too far into the description, I would like to apologize
to those who design real field strength meters and acknowledge that
this is device indicates relative field strength from moment to moment.
It is not calibrated and I have certainly not gone to any pains to
assure a flat frequency response, or even a characterized frequency
response for the detector. Having said that, this is a pretty sensitive
indicator, and there are provision to add a tuned circuit to the input
so as to give the device some selectivity. I will continue to make
loose use of the term "field strength meter" in this web page to aid
web searches for this kind of instrument. Such are the conditions of
our time. I kindly request the reader's kind
indulgence in this matter.
Overview
External
Connections
and controls
When an antenna is attached, (or even if not, in the presence of a
strong RF field) the field strength meter has a moving coil meter to
indicate relative field strength. The box holds a Schottky diode
detector, a DC amplifier, and a meter display. If the parts used
in construction are the same values shown in the
schematic, the meter shows full scale ("+3db" on the surplus store VU
meter movement) when the output of the Schottky diode detector is
between 1.6 millivolts and 160 millivolts, depending up the setting of
the gain control. Since opamps are used as amplifiers, the range of the
gain control can be modified easily. The actual sensitivity in terms of
field strength is primarily a function of the antenna system, and that
is well beyond the scope of this web page.
An RCA connector with which you can attach an antenna (as in the
photograph near the top of the page), of if desired signal
preconditioning circuits such as a filter and/or preamp. I expect that
for most of my low power uses, a short whip antenna or no antenna at
all will be sufficient.
A stereo headphone jack that provides a way to connect the amplified
detector output to a digital voltmeter, chart recorder, oscilloscope,
or other instrument, and also allows the application of an offset
voltage from an external device. I had imagined adding an external
logger with an automatic offset adjust function.
An offset knob allows correction for drift in the Schottky detector as
well as offsets created by interfering signals. A coarse offset
control on the circuit board allows for compensation for the mismatch
between components.
A power switch and pilot light complete the set of user controls.
The meter operates from the power of two AA cells. With new Zinc
Chloride cells installed, it draws only about 3 milliamps, including
the LED when the meter reads full scale, and about 3.5 milliamps when
the meter reads zero, so it seems that the use of an external power
supply would not be worthwhile.
The Circuit
The field strength meter is a broad band temperature compensated biased
Schottky diode detector with an offset adjustment, followed by a buffer
(U1A), amplified by a single gain block (U1B), which drives the meter
driver (U1C). A integrated circuit voltage reference is used to
generate stable +1.25 and -1.25 volt reference supplies and an opamp
(U1D) provides a ground reference which is at half the voltage of the 3
volt battery. The circuit blocks will be discussed in turn below.
Temperature
Compensated Biased Schottky Diode Detector
The detector board takes its input from short wires soldered directly
to
an RCA audio connector.
The Schottky diode detector is built on a single sided fiberglass
circuit board that is separate from the amplifier board. This was
mainly because the detector board needed to be close to the RCA input
connector, which would be on the front panel. With the presence of the
other connector and the controls on the front panel, there was no
room for a larger circuit board to accommodate the DC amplifier.
Whip antennae require a ground structure in order to work. The
large copper areas on the detector board, as show in the picture above,
are intended to improve the efficiency of a short whip antenna at GHz
frequencies. Being in a mostly unshielded enclosure, the meter
can detect strong GHz range signals picked up on its internal leads,
without the need for an external antenna.
Advantages of using a printed circuit board for the detector over using
other construction methods, are that it allows the use of surface mount
parts, which not only provides for small physical dimensions which
minimize parasitic reactances, but also provide for good thermal
coupling of both of the diodes to the circuit board. Keeping the diodes
at the same temperature is important in keeping temperature caused
drift as small as possible.
Schottky diode D1 is the RF detector. Diode D2 is a diode of the same
type, and it provides an offset voltage that tracks the forward voltage
drop of D1. R2 and R3 provide 650 nanoamps of bias current. Half the
current is drawn through detector D1 and bias resistors R5 and R6, and
half the current is drawn through D2. Each diode is lightly biased at
about 330 nanoamps to improve sensitivity and linearity for small
signals.
Both D1 and D2 are dual diodes, but I only used one diode in each
package.
The use of 1k resistor, R2, on the cathode of D2, with current
from offset pots R17 and R18 through their respective series resistors
(R7 and R8), allow for an offset adjustment of about plus and minus 10
millivolts. The unadjusted offset of my detector was about 1.5
millivolts. This was without any effort to match the diodes for forward
drop. The 2 Megohm bias resistors are physically mounted on the
amplifier board, and are shown in the schematic above to improve
understanding of circuit operation. The purpose of R3 is to work
with C4 to attenuate any noise that might be picked up in the wiring
harness.
There was an unsettling effect just after the meter was assembled. With
no external antenna attached, with a strong signal from the right side
of the meter (the side on which D1 was mounted), the meter reading
increased, but when the signal came from the left side of the meter,
the meter reading decreased -went negative. The effect was removed by
soldering a 680 pf capacitor directly across D2. The 680 pf capacitor
is
not shown in the photograph above.
The DC offset corrected output of the detector drives the input of a
variable gain amplifier.
Amplifiers
And Meter Driver
The amplifier was built on a piece of pre punched phenolic board that
has one copper pad around each hole. This results in a circuit that is
easy to assemble, and fairly easy to repair or change. A holder for two
AA batteries sits on the board. A five pin single row connector, on the
left side of the board, makes DC connections to the Schottky diode
detector board through a four wire harness and a 14 pin dual row
connector on the right side of the board makes DC connections to the
indicators, connectors and controls on the front panel through a 12
wire harness.
Although not shown in this version of the schematic, the bias resistors
for the diode detector (R3, R4, R5, and R6) are all mounted on the
amplifier board with the idea that they will track each other thermally
better that way.
In the photograph above, the two blue colored resistors near pin 1 of
the quad opamp were removed in order to reduce the gain of the circuit.
The amplifier board sits in the bottom of the enclosure.
The opamp used in this instrument is a Texas Instruments TLC274C. It
was chosen because of its ability to operate with low power supply
voltages, its fairly wide output voltage range, and its very good
offset and drift specifications. There are better opamps around for
this application, but this is the best one I had on hand. You can do
well with a better opamp, but I do not suggest trying this with a less
capable opamp.
The signal from the detector, which we can treat as DC, is buffered by
voltage follower U1A. R9 is intended to compensate for input bias drift
in U1A and C6 is intended to assure that there would not be excessive
lag in the feedback through R9 because of stray capacitance and the
input capacitance of U1A's inverting input. Upon reviewing the
schematic, I see that if I had a resistor closer to 1.5 Meg Ohms, it
would have provided a better approximation of the input resistance seen
by the noninverting input of U1A. C6, The capacitor across R9 is
a low leakage polypropylene type.
The second stage is the adjustable gain stage, U1B. With the component
values shown in the schematic, this amplifier has a gain that ranges
from 1X to 148X. I went through my collection of surplus store
100k pots and found that they measured between 60k and 87k, so I used
the 87k pot, which gave me an actual gain range of 1X to 137X. This
meter is not calibrated, so parts tolerance is not much an issue.
The output of this stage is applied to the meter driver and to the
stereo headphone jack. When the meter is properly zeroed, the
voltage at this point will go negative with respect to ground.
The meter driver serves two purposes - it limits the maximum amount of
current that can pass through the meter movement, and putting the meter
movement in the feedback path of an inverting amplifier, instead of to
ground, assures that the current to drive the meter movement flows from
the battery terminals through the amplifier stages and not through the
artificially generated ground. Keeping the meter movement current out
of
the ground circuit minimizes the chance of incidental feedback through
the grounds that could result in oscillation.
The diodes in the meter driver circuit allow separate paths for
positive and negative outputs. This is because I want to limit how hard
the needle can hit each "pin" that limits its travel. With a signal
into the instrument, the output of U1C is positive, and D3 conducts.
The maximum current is limited by the voltage output capability of U1C
minus the forward drop of D3, and the divided by the series resistances
of the meter movement and R14.
When the output of U1C swings negative, as when the offset is being
adjusted, and particularly at first setup when the offset adjustment
has a range of several times full scale, the most of the feedback
current is supplied through D4 and the current through the meter is
limited to the diode drop of D4 divided by the series resistance of the
meter movement and R15. I felt it necessary to allow the meter to
operate below zero to make it easier to adjust the offset.
Here (above) is an alternate way to use the first amplfier, U1A.
Instead of using it as a unity gain buffer, you can make it into a gain
stage. As shown here, it has a gain of abaout 100, which is probaly way
too much sensitivity for many applications. For a gain of 10X change
the 100k resistor to 10k.
By the way, I used metal film resistors in many parts of the circuit
for stability. The gain stability is not particularly important, but
thinking that maybe someday, I would calibrate this as an AC voltmeter
at the lowest gain, in which case, the gain pots' temperature
coeficient would not affect circuit gain). It is a really good idea,
however, to use very stable resistors for R4, R5, and R6 to minimize
offset drift.
Find updates at
www.projects.cappels.org
The Power
Supply
Everything is powered by two AA cells. In theory, they need to maintain
nearly 3 volts in order for the circuit to work, but in real life, the
one I built worked find when I used old batteries that provided only
2.5
volts. A more robust design would use three cells for a nominal 4.5
volt
power supply. Changing from two cells to three cells can be done
without any circuit changes, but it would be better to increase the
values of R20 and R23 to reduce current drain.
The reference diode D6, LM335Z, is biased by two equal resistors
connected to the positive and negative battery terminals. This places a
2.5 volt reference midway between the two potentials. The output of the
resistive divider made R21 and R22, is in half way between the negative
and positive terminals of the 2.5 volt drop across D6. Voltage follower
U1D buffers this voltage, which I call "ground". The voltages cross the
anode and cathode of the reference diode are -1,25 volts and +1.25
volts
with respect to ground, and these voltages are used to provide bias for
the circuitry. The battery voltage provides power directly to the quad
opamp.
Construction Consideration
The instrument is housed in a plastic box designed for such purposes. A
metal box would have been quite a bit better. I imagine that if the
detector circuit were built in a small RF tight box with all the bias
and control leads decoupled, a plastic box would have been almost as
good as a metal box. Whatever could have been, the reality is that I
built it in a plastic box because I don't have any suitable metal boxes
available at the time, and I am more interested in using this
instrument than building it, so waiting until after my next trip to the
big city (Bangkok) to buy a metal box was not a good choice for
me.
In the picture above, you can see the front panel with the
controls, connectors, and indicators mounted on it. A piece of copper
clad printed circuit board is held in place with silver colored tape
under where the detector board will be mounted. The piece of
copper clad board helps shield the back of the detector, and it is
connected in four places to the ground area on the detector board.
Harnesses, one captive to the detector board, and one captive to the
front panel connector, indicators, and controls, are made of #26
stranded wire, antique lacing cord, tied about every 1 cm and at bends
and breakouts, holds the harnesses in shape.
The meter movement is held in place with liberal amounts of
cyanocarylate super glue backed up with more than enough hot melt glue.
I found that as I put my hand on the back of the instrument and then
moved it away, it would change the reading substantially. This effect
was greatly reduce by adding a little shielding to to the case under
the main circuit board. The wire seen in the photograph above plugs
into the amplifier
board and connects to the amplifier board ground. The shield, made for
a piece of copper clad fiberglass board is held in place by friction
between the edges of the boards and the molded-in standoffs in the
bottom of the plastic box.
Looking
back
Its only been a week since I built this, but I can already see that the
sensitivity is about right for my kind of use. I seems to work well at
detecting the low amplitude filed around the antenna of
a
crystal controlled low power FM transmitter that I built a few
years ago. When held next to a piece of wire connected to the
output of my
MAX038-based
function generator, set to 12 kHz, the meter displays the relative
amplitude of the filed from that! Taht's pretty low. It also senses the
RF field around my WIFI base station. So, it came as no surprise when I
was able to moved the around the microwave oven in my kitchen while it
was boiling a cup of water and get an idea of where the leakage from
the door was maximum -at the edges, and minimum -right in front of the
door. So far, I am happy with the instrument.
HOME
(More Projects)
Contents ©200
9 Richard Cappels All Rights Reserved.
Find updates
at
www.projects.cappels.org
First posted in March, 2008.
Expanded,
updated April 1, 2009, and April 19 2009.
You can send email to me at
projects(at)cappels.org. Replace
"(at)" with "@" before mailing.