AVR single chip controllers AT90S, ATtiny, ATmega and ATxmega DCF77 cross antenna
1 DCF77 cross antenna
The German date and time standard transmitter signal of
DCF77 can be received
all over Europe, but signal strength is strongly depending from the
antenna direction. To get independent from the direction this cross
antenna has been developed and tested.
I tested two versions of the antenna. Both consist of two coils on two
different ferrite rods. In the first version, those two ferrite rods
are mounted in an angle of 90°. In the second version those are
mounted in an angle of 45°. By angling, one of the two rods always
has reception, no matter what angle the DCF77 transmitter has towards
the antenna rods, even if the other rod is completely misaligned. The
signal of the sum of both coils is never zero, the misaligned coil just
does not add to the total signal.
The second version, with 45°, has been develloped because two
horizontal 10 cm rods do not fit into many plastic casings. In
small casings (e. g. with 5 cm) the first rod can be placed
on the bottom while the second rod can be placed on the top of the
casing, both having 45° offset to each other. The nearer the
angle of both towards 90° they come, the smaller is the angled
That is how the signal strength varies in different angles, for a
single rod/coil and for two rods/coils in 90 and 45° direction.
In the cases with two rods the reception strength is never zero
but varies between 0.5 and 0.7 (90°) resp. 0.92 and 0.35 (45°).
To test the two rods a little further, I have given them different
windings. For the 90° version I covered 45% of the 10 cm
ferrite rod with copper enameled wire (0.255 mm) in single layer
fashion (just because the two rods then can be tied together in a
90° angle without too much interference between the two coils).
That meant 110 windings for each coil. In the second version
I covered the complete rods with single-layer wire, leaving
0.5 cm on both ends uncovered. That meant approximately
350 windings for each coil.
Tying the two coils together and angling them has interesting effects
on the inductivity of the single coils and on their sum. In the
90° case the inductivity of the crossed coils is smaller than
the sum of the inductivity of both, while in the 45° case their
inductivity increases. If, in the second case, their mounting in a
larger distance (e.g. 10 cm between both) is chosen, this effect
will be much smaller than with both ends tied together.
The 90° antenna (left picture) is mounted like shown. First
the rod is shrinked with a piece of plastic cover to get some
distance between the rod and the wire. The 10 cm rod is then
covered with 110 windings of 0.255 mm copper wire over 45%
of its length and both wire ends are fixed with plastic tape. The
two rods are then tied together with two crossed cable ties so
that the angle is approximately 90°. The inner ends of the
two coils are soldered together.
The 45° antenna is mounted similarly but the whole rods are
covered with wire (except 5 mm on both ends). For each coil
one needs approximately 13 m of copper wire. Both ends are
fixed with a piece of shrink sleeve. Both ferrite rods/coils are
mounted on a blank 100-by-100 mm epoxy plate and fixed with
cable ties. The near ends of both coils are soldered together.
1.2 Measuring the coils
The coils have been measured with two different methods:
with a FET and a variable capacitor equipped grid dip meter,
with a CMOS oscillator.
1.2.1 Measuring results with a grid dip meter
With my grid dip meter I installed the coil and varied the
FET oscillator with the 2*365 pF variable capacitor.
From previous experiments with fixed inductivities a
capacity of 200 pF (in full capacity) resp. 23 pF
as smallest capacity has been determined.
The two 45° coils oscillated with 217.64 resp. with
221.75 kHz under full capacity, resulting in inductivities
of 2.67 resp. 2.58 mH. The sum of both would then be
1.2.2 Measuring results with a CMOS oscillator
This schematic was used to determine the inductivity in a
different method. Measuring with this resulted in significantly
larger inductivities of 3.87 resp. 3.79 mH, which would
result in a sum of 7,66 mH.
Measuring the 45° coils tied together resulted in a
significantly higher inductivity: 9.58 mH. That's what
you get from nearing the coils in an angle.
1.3 Buffer stage
This is the schematic for the buffer stage (the 45° version).
The antenna circuit is formed with the stacked coils and a
capacitor of 330 pF. The signal goes to the gate of a
N-FET (any N-FET type can be used). The drain and the source
of the N-FET are connected to two resistors of 1k (to the
filtered operating voltage and to ground, the HF is coupled
with two 1nF capacitors (ZC = 2.5 kΩ)
to the amplifier stages (symmetric output).
Note that the 90° version needs a larger capacitor of
2.7 nF due to the smaller inductivity of the coils.
The buffer stage with the FET is necessary to protect the sensitive
properties of the LC resonance circuit. The large coil above has
an inductivity of 9.58 mH. That means that the coil has, at
77.5 kHz, an inductive reactance of ZL =
2 * π * f * L of 4,66 kΩ. If the coil is in resonance
with the capacitor, the reactance of the LC circuit is by a factor
of Quality larger than this, the circuit has more than
466 kΩ. That means that the resonance curve is very narrow,
suppresses nearby noise sources and the sensitivity is very high.
Hence, it would be not a good idea to attach a stage with a lower
resistance to it. This would seriously drop the LC circuit's high
sensitivity and would broaden the resonance curve. The FET stage
does not amplify, but only keeps the high entry resistance and
provides a reduction of the resistance at its output. The high
quality of the LC circuit on its input is protected and kept.
1.4 Frequency adjustment
The resonance frequency of the cross antenna and the 330pF
capacitor can differ slightly (temperature, iron in the near
field, etc.). Therefore two varactor diodes are attached to
the antenna circuit, both in reverse direction (anti parallel).
I have used two of the three diodes in a TOKO KV1235Z, use
of other types such as BB112 (double diode) is possible. The
diodes should at least have 100 pF at 0.7V (medium wave
Depending from the AFC voltage (0 to 5 V) half of the
capacity of the regulating varactor lies parallel to the
antenna circuit. This allows for a sensitive regulation of
the resonance frequency and adjustment to 77.5 kHz.
You can use a potentiometer, a trim resistor or a digital
PWM to adjust that. Because the varactor diodes are operated
in reverse direction, no current is drawn from the diodes.
This is approximately the capacity of the KV1235Z diode versus
the reverse voltage applied. As the original curve in the available
datasheets looks a little bit weird, I have interpolated it with
a polynome (see the calculation sheet "FET-RX" in the
LibreOffice Calc file.
So do not expect this to be correct.
This is the capacity of the varactor diode and the resonance
frequency of the cross antenna versus the AFC voltage applied.
Two combinations are considered here:
a large coil of 9.58 mH and a small fixed cap of
330 pF (red curve),
a small coil of 1.5 mH and a large fixed cap of
2.7 nF (violet curve).
The range that the combination of the large coil and the small
fixed cap allow is fine, but the small coil with the large fixed
cap covers only a very small range. Note that two of those
varicap diodes are anti-parallel, so their capacity is halved.
In this case we can apply the varicaps a little different to
enlarge the range: we put three of them in parallel and reduce
the fixed cap to 2.2 nF. The orange curve in the diagram
shows that the range is now comparable to that with the large
1.5 Properties of the cross antenna
By adding two coils with only one capacitor in the antenna
circuit a phenomenon occurs that has to be accounted for in
frequency adjustment: both coils have a combined inductivity
as if they were one but each coil has its own additionally.
This is approximately half of the combined inductivity and
produces its own resonance. If the capacitor is larger,
this second resonance can be reached. In the 90°
case this second resonance cannot be reached because the
varactor diodes do not have enough capacity. But in the
45° case, with a large inductivity and a smaller
capacitor, the varactor diodes can well reach this second
resonance points. In order to not stick to this second
resonance point (with its single direction property) the
voltage of the varactors should always start from +5V
downwards, even if the signal strength is larger with the
larger capacitor (e. g. when one of the coils is in
perfect direction towards the transmitter).
This second resonance could have been avoided if both coils
get their own (larger) capacitor. But that would make
frequency adjustment via AFC more complicated because one
needs two PWMs for AFC or a small fixed capacitor exactly
compensating the difference of the inductivity of each coil.
Practice has shown that this antenna is very selective. While
my energy saving lamp transmits at roughly 80 kHz, and
with that strong signal confusing commercially available DCF77
receivers so that they do not work in less than 50 cm
distance to the lamp, the cross antenna is not sensible for
this has to do with the varactor diodes that allow exact
resonance to 77.5 kHz (maladjustment to the lamp's
frequency shows a much stronger signal there), or whether
the bandwidth of the LC circuit is indeed that narrow
(has to be, otherwise the 80 kHz would still come in
even if adjusted to 77.5 kHz), or whether
commercially available DCF77 receivers have no N-FET
buffer stage but couple the signal to a transistor, by
this reducing the high resonance resistance of the LC
circuit and increasing its bandwidth, or whether
those do not have an AFC to exactly adjust the
frequency, and can well be far away from 77.5 kHz,
can not be determined exactly, but this effect alone is a
good argument for having a home-brewed receiver instead of
the cheap mass ware.
The cross antenna is very insensitive to direction changes. The
amplitude drop down to 0.35, when in maximum misalignment, is
simply compensated by a small change in the AGC voltage that
regulates the gain of the receiver.
Because I do not own a mechanical compass (and the one in my
Android mobile is a useless equipment here because the term
"North" is a very wide field for that equipment)
I am not able to provide exact directional data on the cross
two antenna versions. Sorry for that.