You can win any argument about the time when you have a radio controlled watch. Or, at least, you can if there’s any signal. [Henner Zeller] lives in a place where there is no reception of the DCF77 signal that his European wristwatch expects to receive. Consequently, he decided to make his own tiny transmitter, which emulates the DCF77 signal and allows the watch to synchronise.
A Raspberry Pi Zero W is the heart of the transmitter, and [Henner] manages to coax it into generating 77500.003Hz on a GPIO pin – close enough to the 77.5kHz carrier that DCF77 uses. The signal is AM, and transmits one bit/s, repeating every minute. A second GPIO performs the required attenuation, and a few loops of wire are sufficient for an antenna which only needs to work over a few inches. The Raspberry Pi syncs with NTP Stratum 1 servers, which gives the system time an accuracy of about ±50ms. The whole thing sits in a slick 3D printed case, which provides a stand for the watch to rest on at night; this means that every morning it’s synchronised and ready to go.
[Henner] also kindly took the time to implement the protocols for WWVB (US), MSF (UK) and JJY (Japan). This might be just as well, given that we recently wrote about the possibility of WWVB being switched off. Be sure to check the rules in your area before giving this a try.
We’ve seen WWVB emulators before, like this ATtiny45 build, but we love that this solution is an easy command line tool which supports many geographical locations.
I was fascinated by this hack although I don’t own any of those watches so I did some googling and have come across this site:
http://jrcomputing.com.au/Set_Watch/Set_Watch_Manual.html
He just uses javascript on a PC and earphones for this hack!
What a beautiful solution it is.
This is awesome! Tnx for the link!
Are these units legal? As someone who occasionally uses the WWV stations I would be quite unhappy to find out my neighbor was transmitting at the ‘same’ frequency and screwing up my measurements. As far as I can tell they are operating under the guise of ‘as long as no one notices I can transmit at any frequency I want’
I’d say that given the antenna length and signal power (as I see no RF amp) the range can’t be more than a few feet.. so it would probably fall in the range of “acceptable” RF, I would think
Exactly my thoughts. This sounds like something that would get you in trouble. Not cool. Does the watch lack knobs for you to set the time manually?
Possibly the only thing cool about it really.
If this would get you in trouble you should turn light on/off with aluminum foil on your head. Really. This device maximum power is 500 times lower than allowed by authorities. 2 order of magnitude Carl!
Makes me think of Chuck on Better Call Saul
If you’re subject to the US FCC, you’re welcome (nay, *required*) to study 47CFR part 15, which details in glorious detail exactly what you may do with specific frequencies. It’s fascinating reading.
However, interestingly, Part §15.205 “Restricted bands of operation.” makes no mention at all about operation at 60 kHz.
And §15.5 “General conditions of operation.” says: “(b) Operation of an intentional, unintentional, or incidental radiator is subject to the conditions that no harmful interference is caused”
Which implies that if you keep it quiet, it won’t be an issue. §15.107 and 109 give limits on how loud you can be, but only at frequencies above 150 kHz.
But §15.5 also says “(c) The operator of a radio frequency device shall be required to cease operating the device upon notification by a Commission representative that the device is causing harmful interference. “, and you REALLY do NOT want to get one of those letters, especially if you currently hold any kind of radio licence (ham, general operator, boat, aviation, commercial…)
what kind of harm can transmitting on a a watch setting frequency cause?
it isnt like you can fowl up everyone’s clocks as that frequency is probably reserved just for the watch and if the government is going to turn off the signal it is to save money by no longer supporting obsolete technology?
however 77 khz is in maritime mobile and radio location and fixed bands so you could cause problems with ship based sat or cell phone or radios and navigation beacons
77KHz is barely radio, I’d be amazed if anyone sent speech over it. A quick search tells me some fish finders use 77KHz audio as ultrasound sonar, which gives a good idea of how low a frequency it is. Satellite comms are usually up in the GHz. EPIRB beacons are 406MHz. All miles away from this.
Low frequency bands like this are useful cos they travel a long way. But you’re very limited to how much data you can send. 1 bit / second is a fair bit below the limit, but I suppose they use that because it needs to be reliably received by low-power wristwatch receivers with simple circuitry, and because the correct time isn’t a huge lot of data to send.
If this gadget really does cause interference, he can stick a resistor in to reduce the power, but I really doubt it’s necessary. A Pi’s GPIO can’t supply much power.
The transmitting frequency is coming from a computer so it can be extremely stable. And around that sort of frequency, there isn’t much else transmitting because there’s almost no room on the spectrum. Almost no spectrum!
Presumably he’s only going to switch it on every now and then, to sync his watch. And in fact, the most important thing, it says in the article that there’s no DCF77 signal where Henner lives. So there’s nothing to interfere with! There’s a small chance he might actually be doing a fellow watch-owning neighbour a favour.
in any case of RF interference, you have to be careful with unintended or spurious emissions. Harmonics can be generated by poor design, excessive power…. a multitude of reasons. So, the laws are written to put the onus on the offender. YOU are required to take responsibility for any interference you might cause. PERIOD.
Actually 15.209 Radiated emission limits; general requirements. covers this frequency. Range there is 9-90 kHz. And section (a) Except as provided elsewhere in this subpart, the emissions from an intentional radiator shall not exceed the field strength levels specified in the following table:
In that table you can see that you need to measure level of intentional radiation from distance of 300m. I doubt there is a detector exists on this planet that will detect level of signal sent buy PI gpio on this distance. I am not even saying that level would be order of magnitude close to the limit. No way.
It depends on the country. It is possible the far-field signal generated by this unit is so low that it fall under regulation for unintentional radiators. The main method of coupling will be directly inductive between the transmitter and receiver coil. There probably is almost no radiated far-field energy.
If the range is shorter than the width of room, who cares if it’s legal? How will they find you, a crystal ball?
You may find it funny to set your watch to the time you want, but your neighbours heating control may not.
Without professional measurement equipment, you can’t judge what the range of the transmission really is.
Your watch might not pick it up in the next room, but the outdoor ferrite rod antenna of the alarm system from the bank across the street might.
Should they shut down the transmitter some day, and the frequency isn’t assigned to any other service, then probably nobody will care.
But as long as it still runs, you shouldn’t interfere with it.
Even if YOU can’t get a signal in your area, doesn’t mean that anybody else can’t.
A bit strange to have a neighbor in San Francisco with a heating control that uses a German radio to set the time.
Not really — a controller designed for multiple locations may well attempt to find signals using any of the standard services and use whichever it picks up first.
#jh3141
Ok, show me a heating controller commonly sold in SF which supports DCF77 and WWVB, without a dedicated switch, v-ref switch, RC lock, time-zone lock, or just specified home city which pretty much every multi-band clock has.
Odd indeed. Even if such a device does exist it would be a pretty poor design to just accept random data that just happened to be on a carrier of the same frequency. Without using framing or some sort of checksum and letting bits stack up until there are enough to form a command who knows what would happen.
But why? My Casio Waveceptor watch can be set by hand. If it can’t receive a time signal, it’s like any other watch. I assume it keeps reasonable time, but after ten years there’s only been a few days, never in a group, when it hasn’t sync’d up.
Unless someone bas a great time source, you might as well let the clock or watch run by itself.
And yes, unless it is really accurate, I don’t want the neighbor controlling my clock or watch.
One might just try a big tuned loop, like in the old days, and place the clock or watch within its field, to boost reception in weak locations.
Michael
This isn’t a case of being out of range for a short period of time — he doesn’t live in the same country the watch is designed to work in.
NTP is probably going to be more reliable than the watch’s own quartz crystal, over any sort of time period. And this is nice cos it’s automatic and clever! Since the range is probably only a couple of feet he’s presumably only going to turn it on when the watch is on the stand. A bit of shell script could have it only come on once an hour, if it were actually ever a problem to anyone. Maybe even fit a button to it to activate.
For a magnetic loop like this, with the intended target very close inverse square law means there will be negligible field at any significant distance if the power is adjusted correctly. At 77KHz you’d need a huge antenna to get any serious range
Signal is 3.3v going over a resistance of roughly 5k, so ~2mW. You’d need something seriously sensitive if you wanted to detect it outside the room it’s operated in. For reference, DCF77 is a 50kW transmitter and has a 2,000km detectable range. Based on those stats, this transmitter should have a range <1m.
…and that assumes 100% radiation efficiency. The “radiation resistance” of a coil like that at 77 kHz is probably milli-ohms or less.
First, as many ham will tell you there are often, even if rare propagation modes that will allow low power signals to go much further than one would expect. If you look at hams and their qrp work you will see what I am talking about. According to wikipedia the current record for a QRP connection is 1 µW for 2640 kilometers (1650 miles) on 10-meter band (28-29.7 MHz). If other people depend on the signal, please just set your watch by hand.
There may be acceptable low power devices that the authorities would condone, but my guess is compliance testing would cost many orders of magnitude more than the device itself. I doubt you would get caught using an uncertified device, but IMHO doing so is it just rude.
Sorry, but if you do not understand difference of transitting signal over 28Mhz and 60kHz you should not discuss this topic. Otherwise you need to start to worry of emission you electric wall switch sends out when you turn your light on or off.
Compliance testing is required for commercial devices that you want to sell. For personal devices you just need to obey the law and it is authorities obligation to prove that you violated it. As level of emission on this frequencies measured from 300m good luck even detect this signal from noise on that distance, despite measure its level and prove it is too high.
Better idea: use NTP (or GPS) to get the accurate time, then rebroadcast it (with appropriate delay compensation) on low power 60kHz.
I would like something done like this, gps time and wwvb emulator on some hardware. I use the dcf77 app on android but I would prefer some hardware.
I wonder if something similar would drive a Seiko Message Watch, my dad had one back in the day but they shut the service off in the late 90s/early 00s (Although it was still active in Taiwan, whenever he’d go to Taiwan it’d re-set it self)
There is a paper out there written by the guys that designed the Seiko system, search for “High Speed Subcarrier Data System (HSDS) Gary Gaskill and Ken Gray”. The paper has more info including data structure info but the short answer for how the seiko watch worked is “The HSDS data rate is 19,000 bits per second in a bandwidth of 19kHz, centered at 66.5kHz. The HSDS signal is modulated as a subcarrier ranging from 5% to 20% injection, but typically at 10% on a commercial FM radio station’s carriers in the frequency range of 87.5 to 108kHz.”
Hm, “generating 77500.003Hz on a GPIO pin – close enough to the 77.5kHz carrier that DCF77 uses.”
I would say that THAT IS the exact frequency. You can’t get any closer, you will always be a bit off. This is actually even better then some cheap VCOs output.
This hack is really great. But does it actually work for MSF? I have a MSF driven casio watch. As I live at 1200 km from MSF emitter, I experience some difficulties to get my watch synchronized, especially in winter because of VLF propagation problems during that season. I tried the hack with 2 kinds of antennas : one in a 60KHz resonant circuit (with a capacitor in parallel), one bigger with only a 10K resistor in series. I tried close to the watch and at 1 feet distance but I never succeeded to sync the watch. At the beginning, the watch shows a full level reception symbol but after a while the signal seems to vanish an the watch shows “ERR”. I wonder if someting is wrong in the data frame (maybe the watch is waiting for a bit witch never comes ? ). Did you ever try it with MSF signal ? My Raspberry is a PI 2 model B and the signal seems to be OK on a scope.
I have the same problem with yours except that mine is an Acctim Stratus MSF Radio Control LCD wall clock 7457SL. It seems it can pickup the 60kHz signal emanating from the loop antenna I have created with its Antenna Icon blinking on the LCD Display. However, the said clock won’t update or sync. I suspected that my Raspberry Pi WH was Made from China to be buggy. So, I bought the latest Raspberry Pi 3B+ from Argon40 as it is the only reseller available in the Philippines to be selling Made in UK RBPi’s. My problem was solved then.
The design uses an inductively coupled “magnetic field” loop antenna. The signal won’t propagate as there’s negligible associated electric-field. At a distance beyond a few diameters of the coil, the amplitude of H-field from a loop antenna decreases with 1/d-cubed (and the power with 1/d to the power 6).
So while some authorities may frown on it, in reality with a coil of just a few centimeters diameter, it’s not going to cause any issues in practice.
see also: http://hyperphysics.phy-astr.gsu.edu/hbase/magnetic/curloo.html
Just a little explanation on the antenna issues of the DCF77 emulator:
Loop antennas with a small circumference compared to the wavelength (which is absolutely the case here) radiates the signal in the plane of the loop/coil, polarized in the same plane.
The DCF77 signal is vertically polarized, which is why most clocks have the internal antenna rod placed horizontal, and placing the clock perpendicular to the direction to the DCF77 sender will normally give the best receiving conditions.
Then, to emulate the DCF77 sender with txtempus and a loop antenna you should place the coil in an upright position, and orient the clock with the front or back towards the edge of the coil – not towards the broad side of the coil.
The upright positioning and sending from the edge of the coil may seem a little contra-intuitive – one place you can read more about loop antennas is at ’https://en.wikipedia.org/wiki/Loop_antenna’ especially in the section ‘Radiation pattern and polarization’
My own coil is approx. 7 cm diameter with 6 windings, this easily controls my clocks in up to 30-40 cm distance.