Hacker News new | past | comments | ask | show | jobs | submit login
DCF77 (wikipedia.org)
67 points by mahathu 3 months ago | hide | past | favorite | 37 comments



Meta-comment: These are some of my favorite HN submissions - a direct link to a Wikipedia article with no other context, but one that is apparently interesting enough for it to make the front page. I enjoy the process of figuring out just what it was that other people find so fascinating, although it can be a dangerous rabbit hole to go down!


I thought the same, so my quarantine project was a site which collects good Wikipedia articles and their respective discussions posted on HN. To make browsing more structured, it also tags every article with topics, in this case "Germany", "Time" and "Radio Stations".

If you find yourself one day without a good Wikipedia article to read, you may want to check it out: https://www.mostdiscussed.com/


This is really cool, nice work! It's funny, as I was typing my original comment, I was thinking to myself "some kind of aggregator for these links would make a great side project..." - and turns out you were way ahead of me!


At the risk of sounding ignorant to anyone with more knowledge of radio technology than myself: I recently purchased a radio-controlled alarm clock and was highly fascinated by how these devices get their time (and date) information accurately encoded through amplitudes and frequencies of radio waves from some radio antenna hundreds of kilometres away. I'm young enough to have grown up with the internet and have a decent understanding of the different network stacks in use today. But radio communications and their various applications is still such a fascinating topic to me!


Because of the long wavelength and high powers, the radio waves can be picked up pretty well by small ferrite aerials ('loopsticks'). In most cases, all the receiver has to do is detect the presence/absence of a signal; the transmissions start and stop every second, with variations in the starts and stops encoding the time (mostly it's necessary to receive a full 60s of transmissions, to get 60 bits of data that can be decoded to give the time).

It's one of the easiest radio signals to pick up, a cardboard box with a few turns of wire around it and single transistor are enough in most of Europe.


I've been hacking LW time signal receivers for at least a couple of decades, always liked DCF77 because of the additional 512 count sequence (both it and MSF are great intro's to RF hacking in the UK/EU). I'd still got recordings of LW time signals from MSF, HBG and DCF for the midnight rollover from 1999 to 2000 (they didn't all get it right :) ).

It's not well known, but quite a few LW transmissions also carry timecode. TDF Allouis in France and Radio 4 in the UK (R4 is pretty cool because it also carries certain 'additional data'). A couple of source that went off air relatively recently were HBG75[[0] and Loran [1].

[0] https://en.wikipedia.org/wiki/HBG_(time_signal)

[1] https://en.wikipedia.org/wiki/Loran-C


> 512 count sequence

I guess the proper term is "chip", as in DSSS CDMA.


It's funny that it broadcasts the local time, with adjustment for daylight savings. This makes the receivers extremely easy to build, which mattered a lot in the times that the signal was first broadcasted. And a lot of receivers have been manufactured over the decases, probably tens or even hundreds of millions.

Today even the most basic microcontroller could easily adjust the UTC time to the user's timezone. If this was built today people would laugh at the idea of broadcasting anything else that UTC.


I love that it took a 35 horse power transmitter (with what, 20 making it out as EM radiation) to do it. That kind of power in a radio signal is just wild to me, especially because of what kind of components it takes to build the transmitter.



At least it also tells you if DST is active, so you can easily get back to UTC today. It's interesting what other stuff they came up with to include in the signal over time.


I love DCF77. All my clocks in my home synchronize with it and are always spot on, without requiring a network connection. Never have to check whether they are still accurate and no hassle switching between summer/winter time!

I even used to have a wrist watch which was able to synchronize with it.


Same here, the kitchen wall clock and all my alarm clocks use it. It just works, much more efficiently than NTP over a whole network stack, so an AA battery easily lasts a year.


I tried to have the same, but I'm barely out of the signal's range unfortunately.


Try moving the clock to face a different direction. Also the reception varies a lot throughout the day (bad near sunset and sunrise); most radio clocks rely on gettings the occasional good reception and running in the conventional way the rest of the time. So it can be a case of waiting a day or two for the clock to get a good time reception.


Yep, that's what I've done and why I think I'm barely out of range. It's a bit tedious with wall clocks.


Ha, i remember having a parallel port printer cable with an integrated DCF77 receiver (connected on one side to my 486DX33 and on the other to my Epson LQ-850)...

What i can't recall tho is if it really worked to set the date/time on my computer :)


There's software and cheap receivers that allow this even today.

I personally couldn't get my decoder working with new Linux though, the lack of sysfs GPIO broke the only decoder I could find.

I think it kinda indicates how DCF77's popularity has dwindled.


This is not the only time signal emitted from a terrestrial tower. For example France uses a similar signal for their legal time https://en.wikipedia.org/wiki/TDF_time_signal , and as mentioned in https://en.wikipedia.org/wiki/Radio_clock there are many others. Have also a look at LORAN-C / eLoran, which recent releases enable a receiver accuracy close to GNSS as far as I understand while there is no way to properly compensate for various propagation effects in DCF77 and alikes (i.e. you'll get millisecond-ish accuracy, but not microsecond even though the original source clock is very stable. Microsecond accuracy may seem overkill but this is actually required in several industrial applications e.g. 4G/5G TDD base stations are often connected to GNSS in order to be phase/time-synchronised).

Last but not least : I should mention that this researcher J.M. Friedt has an amazing amount of wonderful material about clocks, oscillators, SDR and so on... http://jmfriedt.free.fr And more specifically on DCF77, I could find http://jmfriedt.free.fr/dcf77.pdf and http://jmfriedt.free.fr/dcf77.mp4 (among other things)


There's an interesting project I found, that spoofs this time signal using a soundcard and speaker -

https://bastianborn.de/radio-clock-hack


It's odd that my BMW doesn't use it to set the clock. It's a pita changing the clock every spring and autumn!


Any car with a built in gps or cell phone is picking up time signals from those transmissions, and at least using them internally in some capacity.

Which doesn't necessarily mean one of them is used for setting the actual clock display.


Even more odd if you know that also plain RDS has the time in it.

Imagine my surprise, sitting in my old, beat up Opel Corsa after installing a high-end Kenwood radio which, shortly after turning it on, automagically set the time right.

At first i was like "what the... is going on" and after digging in a little bit, i never understood why other manufacturers and OEMs won't do the same - they rather have you manually change the clock twice a year.

Why?


Same with my 2018 VW. At least it has a DST toggle in the menu, so the most annoying part of changing the DST is left out of this process.


A reason could be that the clock would show the wrong time if you took your car to Greece, Portugal, the UK or the Ukraine.

That’s solvable by “jump one hour forward” and “jump one hour backwards” buttons, though.

If they want to sell the same car worldwide, it would be a bigger problem.


>the clock would show the wrong time if you took your car to Greece, Portugal, the UK or the Ukraine

But why do you think that it'd show the wrong time?

In my mind, if left as is, it'd show the wrong time, no?

What's the first thing you check after you get up from your chair after your flight has landed at its destination? I know i'm looking a) at my smartphone b) for the (local) time and if it has c) changed automatically yet.

Imagine you're making an appointment with a local restaurant and have to do calculations for the right time first, because your smartphone shows the time for your hometown?


That signal broadcasts German local time. So, this would show the wrong everywhere local time doesn’t match that.


True, looking again at the map on Wikipedia, it shows the signal is broadcast far farther than just Germany, something i forgot to take into account...


It mentions the program getting an eight year renewal in 2021, is there any word on the future of the program?



So, how is this single point of failure protected against terrorist attacks? Do they have backup hardware off site? Do they trust all users to be able to handle downtime well?


It’s been operating for decades so don’t worry about it


That’s a lousy argument. The Fukushima power plant had been operating for decades, as had that Miami apartment building that collapsed last week.

If you limit it to terrorist attacks, the twin towers had been operating for decades.


And all of these had a high reward for terrorist attacks (high impact, psychological, financial or sanitary).

Which is obviously not the case of a freaking radio antenna.


If you live in Europe and have a Citizen or Casio watch, it is likely that your watch checks the signal from DCF77 every night to sync up with the atomic clock.


And if you live outside reception distance of JJY or WWVB or DCF77 there's a raspberry pi local transmitter model and it runs off ntp.

Using it to sync my citizen watch in Australia.

https://github.com/hzeller/txtempus


You are right! Here is a manual for one of Citizen watches explicitly mentioning DCF77 (with 900Km radius range): https://www.citizenwatch-global.com/support/pdf/a414/e.pdf




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: