A Frankfurt/Mainflingen DC77 atomic clock interface

The 77.5kHz transmitter in Mainflingen near Frankfurt gets its timing signal from four caesium clocks and hence provides a very accurate time signal. The 50kW transmitter is more powerful than the one near Rugby and can be received over a 2000 km wide area. Recently many cheap clocks have become available that make use of this transmitter. I adapted one of them so that the signal can be decoded by a RiscPC. The project is described below. It can be attempted by anyone who has the clock, a screwdriver and a soldering iron and the total cost should be no more than 15 UKP.

More information about the atomic clock transmitter in Mainflingen can be found on two websites:

Official page

The Funkuhr page

People with a multi-lingual background or German speakers will get most out of these pages, though I think the official page has an English translation.

The clock The radio clock I used was bought from Lidl's supermarket - one of their special offers, but I've seen the same model advertised elsewhere. It is a rather large and ugly square beast with the biggest LCD numerals I have ever seen. It is in fact a pretty good clock that can be adjusted for any offset and even shows the phases of the moon. It costs about 13 UKP, which compares favourably with the price Maplins charge for a Rugby clock module, which is more than 20 UKP plus another fiver for the antenna. My project does not damage the original clock, which can be used as normal - it simply adds a computer output.

Here is how to proceed: Remove the two batteries and open the case by removing 8 small screws from the back of the case. Two of the screws are shorter than the others, so remember where they came from. The plastic of the case is rather soft and the screw-posts are easily damaged. The inside contains three PCBs: the real time clock, the switch module and the radio module. The ferrite rod antenna is easily made out and two wires lead from it to the radio module shown on the right. Notice the four connections in the top right corner - if they are missing you have the wrong type of clock and you should think forgiving thoughts, close the case instantly and not read any further.

The radioTCO is the clock signal. It goes high/low once a second except in second 59. The period of the low pulse determines whether a one or a zero is transmitted. 100ms low is a zero, 200 ms low is a one. The high state measures about 1.5V, which is too low to trigger a TTL circuit, hence you can't connect the radio directly to the user port.

PON is a control signal which turns the radio on or off. The standard digital clock of the unit only checks the atomic time every five minutes as well as when the reset button is pressed. At all other times it switches the radio OFF to save battery power. We of course need the radio ON all the time, hence this control wire has to be disconnected from the real time clock and connected permanently to 0V. This reduces battery life considerably, so use rechargeable batteries if you can.

GND is 0V return and VDD is the 3V supply voltage.

Interfacing this circuit to the RiscPC is quite simple, assuming you have a user port fitted. Basically TCO - the clock pulse - must be connected to PB0 and the 0V line needs to be connected to the the computer's 0V line. The wire connected to PON needs to be cut and permanently connected to 0V, so that the digital clock can't disable the radio. Because the clock uses C-MOS chips and the user port uses TTL chips, a small interface has to be fitted. Do not try to connect the radio directly, it will not work.

I've used a half of a 4011 C-MOS NAND gate - simply because I had one available. You could also use any other type of logic gate. You must use C-MOS (4000 series) gates because the signal at TCO does not go high enough to drive TTL logic. This logic gate draws its supply from the userport's 5V line. I've dropped this via a diode to 4.4V. This lowers the 'high' voltage at the input to the NAND-gate so that the 3V C-MOS circuitry can drive it. The switching threshold of C-MOS circuits is half the supply voltage, hence the diode.

I have also used one gate to drive an LED which is mounted on the right hand side of the case. This LED will flash in time with the clock pulse and as it draws its supply from the computer, it is a handy indication that the computer is connected to the clock.

The actual Radio clock will operate whether the computer is connected or not - my circuit is simply an addition. Below is the circuit diagram of the interface and also a block diagram of the complete system, I hope it is clear enough.

Circuit diagram
The next photo shows the back of the radio PCB. The red wire is soldered to the back of the TCO terminal. Two wires are soldered to the ground terminal, one of them is the cut PON lead. This makes sure that the radio is always turned on. The wires are quite hard to see.

Three PCB's

The other two photographs are obviously the front and the back of the circuit built on vero board. The diode is mounted vertically, with the track cut between the pins. The resistor is the current limiter for the LED about 270 Ohm. The 5V supply is soldered directly to the anode of the 1N4001 diode. If you run into any trouble or need more advice please feel free to contact me.

The final photograph shows the assembled circuit.

Notice the LED, simply inserted into the case via a suitable drilled hole. Also notice the cut wire leading to the radio module. It is folded over and soldered to the 0V connector at the back of the radio module. At the very bottom left is the white three wire cable which connects to the user port. I have mounted the interface circuit with a small screw which is one of the two that held the radio module in place.

Assemble the case carefully - it's a bit fiddly. Insert the batteries, connect to the computer and line the clock up so that the antenna points to Frankfurt. In Britain this would be in an east-west direction. The receiver is sensitive to interference from computer equipment, so use a long cable and try out various locations. I found that my Canon printer produces most interference and I have to switch it OFF when I use the clock. If all goes well, the LED should begin to flash. Download my program "AtomZeit", which you will find on the PD page of this site and you should see the signal arriving. The decoding scheme is explained in the Help file of the program. If it doesn't work, feel free to give me a shout, but please, don't blame me for any damage you might have done!

The finished jobThe clock hasn't changed much, apart from the glowing eye and a single cable. But you can now study the signal as it arrives and set your computer's time at the press of a button.

Return to interfacing index
Back to the start
Tudor with sign