Radio-controlled clock at the Lille Flandres railway station

At the Lille Flandres railway station, on December 31 2017, I spotted this digital clock.

The interesting thing is that it has an LED at the side, flashing in a one-second rhythm. Upon closer inspection, the flashes were sometimes short (0.1 seconds) and sometimes long (0.2 seconds), and there was no flash in the 59th second. That very much suggests that this is a radio-controlled clock, with the LED indicating the bits as they come in, as both the German and the French longwave time-signal transmitter use such a format for transmitting the time code.

Carefully analysing a movie I made of it (view it on youtube), and knowing that a short flash is a 0 and a long flash a 1, leads to the following sequence of received bits:

010000100110001111101001111010000-00010010000001000010101101010000100110001

The first bit is for the 26th second (because that's where I happened to start the movie). The - indicates the 59th second, where there's no flash.

Let's try to decode it using the format used by the German DCF77 transmitter on 77.5 kHz:

00 second           10                  20                  30                  40                  50                59
|                   |                   |                   |                   |                   |                 |
                                                    0 1 0 0 0 0 1 0 0 1 1 0 0 0 1 1 1 1 1 0 1 0 0 1 1 1 1 0 1 0 0 0 0 -
0 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 1 0 1 0 1 1 0 1 0 1 0 0 0 0 1 0 0 1 1 0 0 0 1
                              | | | | | | -------------   -----------   ----------- ----- --------- ---------------
                              | | C C | |    minutes         hours          day    weekday  month        year
                              | | E E | |       56             8             31       7      12           17
 abnormal transm. operation --' | S T | |
     summer time announcement---' T   | |
       leap second announcement ------' `---- always 1

It looks as though the received bits could match DCF77's time encoding. However, according to https://www.dcf77logs.de/logs/DcfLog_20171231/dcf DCF77's first 15 bits in this minute were 011000010101111, which clearly does not match the received bits here. (These bits encode encrypted weather information.)

Next try the format used by the French transmitter on 162 kHz:

00 second           10                  20                  30                  40                  50                59
|                   |                   |                   |                   |                   |                 |
                                                    0 1 0 0 0 0 1 0 0 1 1 0 0 0 1 1 1 1 1 0 1 0 0 1 1 1 1 0 1 0 0 0 0 -
0 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 1 0 1 0 1 1 0 1 0 1 0 0 0 0 1 0 0 1 1 0 0 0 1
  | | hhhhhhh             | | | | | | | | -------------   -----------   ----------- ----- --------- ---------------
  | | tomorrow publ.hol.--' | | | C C | |    minutes         hours          day    weekday  month        year
  | | today public holiday -' | | E E | |       56             8             31       7      12           17
  | |       to be ignored ----' | S T | `--- unused, always 1
  | | summer time announcement--' T   `----- unused, always 0
  | |
  | `---- negative leap second announcement
  `------ positive leap second announcement

Yes, this matches! The time data is correct, and the "tomorrow is a public holiday" bit is correctly set (indicating that January 1 is a public holiday). Furthermore, the bits marked hhhh represent the number 18, which should be the number of 1s in positions 21 through 58 (for error detection); indeed, there are 18 1s in those positions.

So, here we indeed have an example of a clock using the time code from France's 162 kHz megawatt time code transmitter. See http://www.pa3fwm.nl/technotes/tn15d.html for some more background about that transmitter...

I do wonder what kind of circuitry they use to convert the transmitter's phase modulation in real-time into such nice short and long flashes. That's much easier to do with DCF77's amplitude modulation.

P.S.: I noticed this clock while walking back from "platform 0" which this railway station has. Starting a count at zero is quite usual in computer science and programming, but very unusual among railway station designers...