[Lowfer] OT: Atomic Clock circuits

[C. Turner]

I've always found that you can locate a logic line on WWVB clocks that 
contain the serial time code - although it sometimes takes a bit of 
poking around with a 'scope to do so.

One trick might be to make sure that the receiver is active, however:  
Most clocks only turn on the receiver once a day (usually at night) to 
try to acquire the signal, retrying 2-5 times if they are unable to do 
so the first time.

A while ago, I did some Googling on this topic and found a site at which 
someone got one of those cheap analog clock movements (such as may be 
obtained from www.klockit.com - but there are no doubt others) and not 
only picked off the time code signal, but managed to find out how to 
force the receiver to be powered up at will.  This time code was then 
fed to whatever it was that he was building...

As for getting a 60kHz signal, 1kHz or 1pps, that's a bit trickier.  
There's no intrinsic signal that relates easily to 1 kHz, but the 1pps 
could be derived from the time code itself.  Note that the uncertainty 
of this pulse is likely to be in the units to 10's of microseconds owing 
to the relatively low detection bandwidth of the signals, variability in 
the slicing level and deviations due to the S/N itself.  While long-term 
averaging may help to make this 1pps more useful, changes in effective 
path length further introduce uncertainty.

These clocks all seem to use 60 kHz "tuning fork" mode crystals as their 
bandpass element:  The chips for which I have seen data sheets usually 
call for two such crystals for optimal performance, but I've seen a 
number of clocks that have just one, and these seem to do just fine.  
Tapping this signal from an existing receiver may be problematic owing 
to their very high impedance, however, but it may be do-able with a JFET 
source-follower - but it may just be easier to roll your own TRF:  The 
60kHz crystals are cheap and readily available from DigiKey - one just 
needs to keep in mind their high series resistance and design the RF 
stages appropriately.

***

One of the big "gotchas" for using WWVB as a frequency reference - in 
addition to the diurnal phase disturbances - is the fact that there's a 
45 degree phase shift to "ID" the carrier.  Years ago (in the mid 80's) 
I homebrewed a WWVB receiver that locked a crystal oscillator to it - 
and it worked very well, until the phase shift occurred.  Had I 
completed the project I would have probably multiplied the carrier 
(starting either at 60 kHz, or a heterodyned-down version) by 8 to 
remove the phase shift to keep everything happy without having had to 
worry about knowing exactly when to introduce my own, local 45 degree 
phase shift.

***

As far as WWVB QRM goes, I've noted that many of my compact fluorescent 
lights seem to have switchers that operate at around 30 kHz, so the 2nd 
harmonic and its sidebands tend to murder nearby WWVB clocks.

Years ago, a friend of mine, who works for a local city government, came 
to me with a problem:  They'd installed WWVB clocks in a conference 
room, but they never acquired a signal - and being in the middle of a 
multi-story office building didn't help.  The solution was to make an 
active re-radiator.

Several turns of wire were run through a piece of EMT conduit with a 
break in the middle (at the top) to form a shielded loop and these turns 
were resonated to 60 kHz with a mylar capacitor.  A simple amplifier was 
then constructed using a several of 2N3904's (providing 40-50dB of gain, 
IIRC) with all of this housed in an aluminum outdoor utility/outlet box 
along with watertight conduit fittings.  An F-connector and RG-6 cable 
were used to feed power and to conduct the signal to where it was needed.

At the other end of RG-6 was a power inserter - just some inductors and 
capacitors to feed DC to the rooftop unit and extract the RF, with power 
being supplied by a 12 volt "wall wart."  This then went to a resistive 
2-way splitter (6dB loss or so) and each of these runs of RG-6 fed a 
coil of wire, 12"-18" diameter, that were also broadly resonated with a 
mylar cap at 60 kHz.  The intent was to drop these loops - which were 
just insulated hookup wire covered with some electrical tape - inside 
the wall behind the clock, but in initial testing, it they were just 
left in the drop ceiling above the clocks and found to provide plenty of 
signal to make the clocks happy.  That was about 10 years ago, and that 
is how they still are, today.

The only problems with the system were near the beginning.  Originally, 
my friend, unfamiliar with the patterns of loop antennas installed the 
loop, which was located on the roof, oriented 90 degrees out:  Rotating 
it 1/4 turn fixed it instantly.  A few weeks later he called to say that 
it had quit working.  I told him to bring it over again - but before he 
did so, he called again and explained that once he opened the cover and 
drained the water out, the amplifier worked again:  The liberal 
application of silicone sealant and the drilling of a drain hole or two 
prevented this from happening again.

I mention the above because it may be of interest to those who have WWVB 
clocks that are being QRMmed - either by local switchmode power supplies 
or, possibly by their own LF transmitter.

73,

Clint
KA7OEI


Mike Staines - WM1KE wrote:
> Slightly off topic, but related to a LF project:
>
> Is anyone aware of a source of circuits or pre-built boards for these
> "Atomic Clocks"? I have looked, and even visited my local clock
> repairman but have not come up with anything.
>
> I hope that there may be some pick-off points where I could get stable
> frequencies like 60 KHZ, 1000 Hz or 1 pps that are locked to NIST. The
> actualy frequencies are unimportant but stability to WWVB standards is
> what I am looking for.
>
> On another subject: I am going into "Disability Mode" here. While that
> means that the income will be lowered it also means that I will have an
> additional 40 hours a week to play radio. So expect TMO to start it's
> wandering ways soon.
>
> Mike
> WM1KE
>
>