AVR single chip controllers AT90S, ATtiny, ATmega and ATxmega DCF77 - A primer on resonance and the ribbon cable loop antenna
DCF77 - A short primer on resonance and the ribbon cable loop antenna
If you want to build your own receiver, the first question is: which antenna circuit to
select and how large shall the coil and capacitor be?
Now, the answer to that is very simple: take what you like, but be aware which consequences
Single loop antenna
To start with the most simple construction: your coil has one winding and has a diameter of
1 meter. Relevant: This coil has no other components such as a ferrit core, just the one
wire. To make it mechanically stable, you can either use a 20 mm steel or iron pipe
or you build some wooden or plastic frame to keep the loop straight. Why iron or steel and
not copper or aluminium? Now, at these low frequencies of DCF77 the HF are majorly magnetic
fields and only minorly electric fields. And magnetic fields need a material that can sense
magnetic fields. If you would take copper or aluminium your antenna circuit would miss the
strong magnetic field of DCF77 and would only sense the electric fields. Together with many
noise that are majorly electric fields, from switches, switched power supplies, and so on.
So this is receiving not what you want. So better take an iron or steel pipe for your
DCF77 loop antenna.
Your magnetic loop now has an inductivity of 2 Micro-Henry (µH). Smaller loops
have less, loops with more windings than one have higher inductivities (see below). The
first thing we have to know is what resistance has this loop against the DCF77's magnetic
waves? The formula to calculate this is:
ZL = 2 * Π * f * L = 2 * 3.141592654 * 77500 * 0.000,002 = 0.974 Ω
This resistance is NOT Ohm's resistance, which could be much smaller than 1 Ω,
if your pipe is made of a 2 mm material with a diameter of 20 mm. Or higher, if
you make your loop out of 0.1 mm iron wire. No, Z instead of R is the resistance that
your loop has against alternating currents with a frequency of 77.5 kHz. We will see
later on that more windings have a very much higher Z than our one-winding loop.
The loop is in resonance as such at around 8 MHz frequency, because it has a parasitic
capacitor of 186 pF. But 8 MHz is not the frequency you want to receive. To get
the loop down to a resonance frequency of 77.5 kHz (103-fold smaller), you'll need a
capacitor, so that your loop produces a resonance circuit at that frequency. How large has
this capacitor to be?
Now, simple: its capacitive resistance, which is
ZC = 1 / 2 / Π / f / C
has to have the same value than your inductive resistance. So, if we have a L of 2 µH,
a ZL of 0.974 Ω and a frequency of 77,500 Hz, our capacitor is
C = 1 / 2 / Π / f / ZL, or
C = 1 / 2 / 3.141592654 / 77500 / 0.974 = 0.000,002,108 F or 2.108 µF
Our single loop now gets resonance to DCF77 if we add this capacitor to the loop's endings.
If you think that the resonance circuit has a combined coil and capacitor resistance of
0.974 / 2, just because the coil's inductivity resistance and the capacitor's capacitive
resistance are parallel, you are not correct. The fact that your two components are in
resonance increases their combined resistance. This increase factor is called Quality
factor and is
This quality factor depends from lots of things, namely from the Ohm resistance of the
coil. It can as well be as large as 100, for lower quality LC's it would be around 40.
So, the LC's resistance can be up to 50 Ω instead of the 1 Ω of
the L's and the C's resistance only.
The Q factor is also resüponsible for the bandwidth of the LC circuit. Within the
f1/2 = fcenter / Q
the resonance is reduced by a factor of two. A Q of 100 at 77.5 kHz means that
f1/2 is 775 Hz broad, at 77.5-0.375=77.113 and at
77.5+0.375=77.888 kHz the signal strength drops to half of the signal at the
If you need to calculate the resonance frequency from a given L and C, you'll need
the following formula:
ZL = ZC, or
2 * Π * f * L = 1 / 2 / Π / f / C, or
f2 = 1 / 4 / Π2 / L / C, or
f = 1 / 2 / Π / √(L * C)
The frequency therefore drops by the radix of L and C, not linearly.
More loops - even more inductivity
What if we add another loop to this. If you would expect that this also doubles the
inductivity: this is incorrect. It grows with the square. A two-loop coil of 1 m
diameter is already at 8 µH.
But as our inductivity rose by a factor of 4, the needed capacitor is four times smaller
now to be resonant at 77.5 kHz. This also increases the inductance of the L and C
to 3.9 Ω and increases the inductance of the LC already to 200 Ω,
if the Quality factor would be 100.
We can add even more loops to that. A ten-loop coil is at 200 µH, and
a 100-loop coil is already at 14.6 mH. To bring such a large 100-winding-loop coil
to resonance with DCF77 you'll need a capacitor of 290 pF only, ideal for a Medium
Wave variable capacitor of 365 or 500 pF.
You can calculate the inductivity from the windings for a 1-m-diameter loop by the
L (H) = 0.000,002 * w2, or
w = √ (L / 0.000,002)
So, for a 1 mH air coil with a diameter of 1 m we need √(0.001 / 0.000,002)
= 22 windings. (Reference to below: Take a 26-wire ribbon cable and leave four
What does this increase in inductivity mean? Now, the ZL as well as the
ZC increases. If our L is 1 mH, the ZL at 77.5 kHz
ZL = 2 * Π * 77,500 * 0.001 = 487 Ω
If in resonance, the ZC is at the same value, with a quality factor of
40 or 100 the ZLC is up to 100 kΩ. If we add a pre-amplifier,
its input resistance should be high enough to not interfere with the LC's selectivity.
Otherwise the pre-amp would not only decrease the amplitude, but would also decrease
the Q of the LC circuit and, by that, would also increase the bandwidth of the LC,
its selectivity. So, better make the input resistance of the pre-amp high enough.
The ribbon cable loop antenna
To get practical: I built a 26-wire-loop to study its properties. Why 26? I used a
90 cm long piece of ribbon cable with 26 wires, which I pressed onto two IDC
ribbon cable sockets. On a small 55-by-67-mm PCB I soldered two 26-pin box headers
for that, in a distance of 50 mm to each other.
The 90 cm ribbon cable and the 5 cm distance between the IDC sockets yield
a circular loop of a diameter of 302 mm. With the following parameters
0.09 mm2 wire area, 0.339 mm wire diameter,
1.27 mm wire distance, distance between outer diameter of wires =
an inductivity of 319 µH should result, according to DF7SX's internet
calculator. That would require a capacitor of 13.2 nF for 77.5 kHz.
To test this I used a 16.0 nF Styroflex capacitor that I had at hand, and
fed a rectangle via a resistor of 10 kΩ to the LC circuit. I found
resonance at 65.04 kHz, from which an inductivity of 374 µH
results. The LC's resistance is obviously 3.33 kΩ, while the inductive
and capacitive resistance is 153 Ω, from which a quality factor Q of
43.6 results. The drop in quality is resulting from the measured 5.7 Ω
of Ohm's resistance, due to the more than 23 m long wire of the loop. A Q
of 43.6 means that the LC has a bandwidth of +/-0.89 kHz, which is a rather
If we exchange the sides of the IDC plus by exchanging the two sides of the ribbon
cable when pressing it into the IDC female plug, we'll get this connection scheme
for the ribbon-cable-loop-antenna. It is a little bit simpler to solder, but works
in the same way like the other option.
In general: increasing wire lengthes increases Ohm's resistance linearly,
increasing wire diameters increases Ohm's resistance by the square of the diameter.
The same applies roughly for the reduction of Q.
Nevertheless, I plan a 1 m diameter loop with 50-wire ribbon cable as well, but still need
some time. I'll report on the results later on.
Measuring the properties of the 26-winding ribbon cable loop
For measuring the loop's properties I built a test oscillator with a 4011 NAND.
This produces a strong oscillator signal. With 10 MΩ the oscillation
was too weak if a large-size coil had too high losses, so I reduced it to
When measuring coils with a higher inductivity, it can be recommendable to use
a smaller capacitor than 16 nF, e. g. 1 nF, so that the oscillation
does not stop.
With the capacitor of 16 nF the 26-windings-loop oscillated at a frequency
of 661 kHz. With the capacitor removed, the loop oscillated at
The first approximation yields an impedance of 362.341 µH for the
LC circuit. Calculating the capacitor for the frequency without the fixed
capacitor arrives at a parasitic capacitor of 60.5 pF. This is not very
large and in an acceptable range.
Repeating the calculation of the inductivity with the parasitic capacitor leads
to a slightly different inductivity of 360.976 µH. Additional
approximations do not change the resulting inductivity any more, so the
parasitic capacitor of this loop is at 60.7 pF.
Practical reception of DCF77 with the 26-windings-ribbon-loop
Bringing the 26-loop into resonance with a 10 + 1 nF capacitor
showed nothing on my scope: the signal strength was well below 1 mV HF. So
I had to add an amplifier first. I chose an OpAmp amplifier and tested a 741 and
a CA3140 for that. Both worked as designed. The trim potentiometer allows to
increase the gain of the OpAmp.
This amplifier can also be used as a remote driver for the loop antenna circuit,
for operation in some distance to the receiver. The DC operating voltage is
supplied via the inner wire of a shielded cable and is decoupled by the
33 mH coil and the two capacitors of 10 µF and 100 nF.
I adjusted the gain to roughly 100 and this is how the DCF77 signal looks like
(scope on 20 mV per division). So, the sensitivity of the loop is rather limited,
as compared to a ferrite antenna: roughly 0.3 mV HF. But: at least the
resistance is rather low, well below 10 kΩ, so the antenna circuit
can directly drive a silicon transistor amplifier and does not need an FET driver
This is such an NPN amplifier with a gain of roughly 200.
This amplifier can, like the one with the OpAmp above, also be used as a remote
driver for the loop antenna circuit, for operation in some distance to the receiver.
The DC operating voltage is supplied via the inner wire of a shielded cable and is
decoupled by the 33 mH coil and the two capacitors of 10 µF and
Extending the 26-loop to a cross antenna
Of course I had to test whether it is possible to build a crossed antenna with those
loops, to get rid of the directional properties and problems of the DCF77 reception.
That meant another 30 cm diameter loop, two additional IDC connectors (male and
female) and extending the wooden holder in the middle by another element - crossed.
I fixed the ribbon cable with two additional pins to the wooden cross.
The design, here photographed in the evening sun, is indeed impressive. Something
that an electronic and radio freak loves to have in his living room. If he declares
to outsiders that this receives an atomic clock in every direction, guests may be
That meant to add two additional male 26-pin IDC plugs and to connect the additional
two sides by soldering those with enamed copper wire. This time I decided that this
second loop is changing the sides, so that the ribbon cable on pin 1 on the left
arrives at pin 26 on the right.
This is the base plate that holds the wooden double-cross with the four IDC connectors.
These are the results of the measurements with the CMOS oscillator described above.
The two loops are nearly identical to each other, as regarding their inductivities
as well as their frequencies without a fixed C. It is interesting that combining
the two loops by connecting them serially leads approximately to the sum of their
inductivities. It is like both loops were not at all interacting, the increase of
inductivity is only 1.8%. This is completely the opposite to the cross antenna with
a ferrite core: there the inductivity of the cross rises much higher, the two windings
are coupled much closer than here.
Also interesting: the parasitic capacitor of the combination is not combined but
decreases by a factor of almost two, if compared to a single loop. This is obviously
an advantage of the crossed loop and makes it a better receiver than the single-loop
Now: this cross antenna, equipped with two 2n2 and one 1n5 capacitors, brings already
a visible result on my oscilloscope, even without a pre-amp. Here it is:
Overview over ribbon cable loops
This summarizes the results of the experiments. To the left, the different ribbon
cables, their dimensions and their connexion variants are listed. Next are the
values that were calculated with DF7SX's webtool. To the right the values measured
in practice are given.
Relevant differences are only for the serially connected cross antennas: the
inductivity is roughly the sum of the two single loops, their inductivity does not
increase by the square of the windings. Both antenna loops in crosses are not
Again: practice is the master: build it and you'll be surprised.