Note:
The techniques described below should work - with only minor adaptation - for any "Longwave" time signal used by these radio-controlled (non-GPS) clocks - not only WWVB, but DCF77, MSF, BPC and both JJY signals as well.
* * * * *
Last year I moved a bunch of SDRs (KiwiSDRs, RTL-SDR) and a bunch of network gear to a new shelf in my shack, but this placed them much closer to the wall on which I'd previously mounted the two "Atomic" (e.g. radio-controlled) clocks which had been there - and operating - for years. Since then, they hadn't been able to reliably synchronize to the 60 kHz WWVB signal out of Fort Collins, Colorado.
 |
| Figure 1: The two radio clocks surrounded by the four-turn loop, near a number of pieces of "noisy"equipment. The top clock is set to UTC and the bottom for local time. Click for a larger version. |
While annoying, I wasn't terribly surprised. There are several switch-mode power supplies involved in the aformentioned gear and it's not uncommon for them to operate in the 30-60kHz range, offering the potential of "jamming" the receivers. As both the location of these clocks - and the nearby gear - is convenient, I wasn't too inclined to move them again and initial efforts to "filter" the switching power supplies didn't really help - but I wasn't surprised about this, either, since it's likely direct coupling of their magnetic fields that is the culprit rather than any electrostatic field as the clocks themselves use ferrite loopstick antennas sensitive to just the H-field.
A solution
Many years ago a friend came to me to solve a similar problem in a downtown Salt Lake office building where the WWVB clocks in a conference room never synchronized and I constructed a remote loop and amplifier/coupling system, described here:
- Getting "Atomic" (WWVB) clocks to work indoors and in weak signal areas - LINK
In short, a rooftop loop antenna amplified the signal and it was conveyed into the room with the clocks where it was further amplified and then, using inductive loops placed in the proximity of the clocks, and this is how the WWVB signal was coupled to them. To my knowledge, this system worked for many years (well over a decade) and for all I know, it may still be in use.
Why revisit?
I've tackled this type of problem before - but I decided to revisit it as the circumstances are slightly different: I already had a signal source as noted below plus I wanted to see if I could do this with more commonly-available components in a simpler manner.
While I don't have a WWVB loop on my roof, I do have a dedicated LF E-field whip antenna - a 40 year old LF Engineering LF-400B with integrated low-pass filter. This antenna has been on the roof wherever I have lived almost continuously since I purchased it in the mid-late 1980s and with a few repairs over the years, it still works well, having been on the roof of my current house for several decades. Its use for LF reception as described on the following page:
- A (semi)-typical suburban E-field whip receive system for the 630 and 2200 meter amateur bands - LINK
The fact that I already had an LF/VLF receive antenna system meant that I had a "clean" source for WWVB, and other devices that receive signals below 500 kHz (e.g. LF receivers for 630 and 2200 meter operation and my Blitzortung "Blue" receiver) and I decided to add one more to the list.
Other types of outdoor antennas
Note that the circuit described here should work well with other types of active antennas including E-field types such as the PA0RDT "Mini-Whip" and the DX Engineering ARAV3 to name but two as well as an amplified loop like the Wellbrook and similar - provided that it is not oriented such that the desired time signal is not in its nulls.
Buffer/Amplifier
Through back-of-the-envelope calculations I figured that the already-amplified signal from the active whip needed another 15dB or so of boost and it could then be applied to a loop of wire around the WWVB clocks on my wall. One thing that helps greatly is that the WWVB signal is extremely strong here in northern Utah - on the order of 5mV/meter or so - and connecting an oscilloscope to my LF-400B whip's signal output showed that the amplitude-modulated time code of the 60 kHz signal from WWVB was visible among the many others.
What I needed to do was to tap off the signal (e.g. "bridge" the connection) from the existing coaxial cable without affecting was was being sent to the other devices using it, amplify it. and apply it to the loop - and I did this "tap" using a BNC "Tee" connector on my antenna feed.
The circuit diagram below gives more details:
 |
Figure 2: |
Circuit description
Of high importance is L1, a common-mode choke, liberated from a failed switch-mode power supply somewhere. This particular unit has an inductance of about 1mH per winding meaning that it has about 377 Ohms of impedance at 60 kHz and helps to prevent a ground loop and the coupling noise from the mains. If you replicate this circuit I would strongly suggest that whatever you use for L1 have at least a similar amount of inductance. On either side of L1 are electrolytic capacitors (C1, C2 - preferably of low ESR types) to offer low impedance and a degree of reinforcement of common-mode rejection through L1 while C2 and C3 provide RF bypassing for the circuit itself.
A buffer amplifier consisting of Q1 - with a high-impedance input, but no actual gain - couples the signal from the existing antenna: Having several k-Ohm of input impedance, it is unlikely to appreciably load the existing antenna system. On the feed from the E-field whip, I simply installed a coaxial "T" connector to allow me to bridge across the signal feed rather than split the signal, which would have been complicated owing to the fact that the DC power for the whip was also being carried on that same cable. The connection to the amplifier in Figure 3 was made using a very short piece of coaxial cable (about 2 feet long - less than a meter) and since this whip is not used for reception above about 500 kHz, neither its presence or that of the added amplifier had any discernible effect on the other received signals.
Coupling from the existing antenna are series components L2 and C4, selected to resonate at about 60 kHz: The resonance is extremely broad, so the fact that the "ideal" value of C4 - according to the formula below actually calculating as 0.007uF (7000 pF) - is unimportant. This series resonant circuit is probably not essential and a simple coupling capacitor of 0.01uF (10000 pF) could be used (omitting L2 entirely) but I chose built it with L2 to broadly filter off-frequency signals - something that might be important if your E-field whip antenna doesn't have a low-pass filter to remove AM (Mediumwave) signals as mine does as well as to block any stray coupling of HF signals when I transmit.
 |
| Figure 3: The circuit of Figure 2 built on a piece of prototyping board in the case. Bifilar choke L1 is on the right with the BNC connector (J1, input) and output to the loop (J2) on the left. Click on the image for a larger version. |
As the signal from WWVB is quite strong at this location, there is only one stage of amplification shown in Figure 2, but if I lived more distant, greater overall system gain might be required. Replicating the circuit involving Q2 (R4-R8, C5-C6) and cascading it with the existing amplifier would add yet another block of gain to boost the absolute signal level - but this would presume that whatever active antenna you were using outdoors to pick up the WWVB signal was working well, providing a "clean" signal and that the deficit was just in signal strength at the clocks rather than than signal-noise ratio.
Due to the smallness of the project box that I chose I couldn't mount the bifilar choke "through" the PC board so it was mounted on the edge to minimize height. To hold it in place I used UV cured resin along the edge to prevent it from breaking the pin connections mechanically: UV cured resin is very handy as it's about a strong as epoxy, but it is cured almost instantly meaning that it's able to be handled immediately. For the BNC connector, the one that I found didn't have its matching mounting nut, but more UV-cured epoxy did the job for that, too! As can be seen in Figure 3, I didn't bother "mounting" the board in the box, letting it hang about on its own wires.
Indoor Coupling loop
The "coupling loop" - visible in Figure 1 and shown on the schematic - is just a loop of wire - and it is used to inductively couple the signals from the outside antenna to the clocks. In my case, I measured a rectangle that would encompass both of the wall clocks and found a cardboard box with similar dimensions and on it I wound four turns of 22AWG hookup wire. Connecting this loop to the amplifier, I used some shielded microphone cable: Coaxial cable would have been fine as would just some single-pair speaker wire as this frequency is not all that much higher than audio!
Neatly forming the individual conductors, I used small "zip" ties to hold them together and with four screws, attached it to the wall, placing the clocks inside the loop of wire. Within the loop, signals from the amplifier would be strongly coupled into the ferrite loopsticks in the clocks themselves - but being very small in terms of the 60kHz wavelength, this loop is unlikely to radiate more than a few feet/meter outside it.
To improve efficiency of the coupling loop I wanted to series-resonate it at around 60 kHz as this would increase the amount of energy transferred to the loop from the amplifier somewhat, effectively providing "free" signal gain. Measuring the inductance of the loop I found that it happened to be about 22uH and using this simple formula, I calculated the value of C7 - the resonating capacitor in Figure 3:
LC = 25330/(FMHz)2
Where:
LC is the product of the inductance and capacitance (e.g. Capacitance in pF * Inductance in uH)
FMHz is the desired resonant frequency in MHz (e.g. kHz/1000)
Knowing that we have 22uH of inductance in the coupling loop and a frequency of 60 kHz (0.06MHz) we end up with "LC" being equal to 7036111. Dividing this value by the known inductance of our coupling loop (22uH) we get the capacitance, as in (7036111/22) = 319823pF, or 0.319uF.
 |
| Figure 4: The finished amplifier in its box, hanging out below the loop - connected, and in service. (It's just visible in the bottom of Figure 1) Click on the image for a larger version. |
As 0.33uF (330000 pF) is the closest common capacitor value, I used that for C7. Again, as with C4 and L2, the resonance is very broad and precision isn't too important. The article linked near the top of this page goes into more detail on how one would construct and resonate a coupling loop. Based on this formula, if I wanted to resonate the same loop for use with DCF77 at 77.5 kHz I would have picked a 0.18 or 0.2uF (180000 or 200000 pf) capacitor, instead. Similar changes could be made to accommodate longwave time signals on other frequencies (e.g. 40 kHz, 50 kHz, 68 kHz).
The formula above can also be used to calculate the value of C4 with the 1mH (1000uH) L2 inductor: If your interest was for another frequency, such as DCF77 at 77.5 kHz, C4 would be 0.0047uF (4700 pf), instead.
It need not be said that this loop should not be placed very close to whatever outdoor receive antenna you are using - but more than about 10-15 feet (3-5 meters) should suffice: If they are too close to each other, feedback (oscillation) could occur - but as this loop is only around 0.01% of a wavelength in circumference it does not radiate efficiently at all - and since it's inductive, its signals won't efficiently couple to an E-field antenna, anyway.
In the diagram, C7, the resonating capacitor for the coupling loop, is shown at the amplifier - but it could have been placed at the loop itself.
Power supply
First off, do not use a switching power supply for this device!
As noted, common-mode choke L1 was used to "decouple" the power supply from the amplifier - and also from the coaxial cable of the LF antenna. To power this loop amplifier I would strongly recommend using ONLY a transformer-type DC power supply and NOT any type of switching power supply for the simple reason that the switching power supply will be comparatively noisy, and it - its harmonic - will likely operate at/near the frequency of WWVB or whatever time signal you are trying to receive.
This power supply does not need to be regulated: Simple capacitor filtering with low-ish ripple (a few hundred millivolts) will suffice and any voltage between about 11 and 16 volts will work which means that about any old "wall wart" in that voltage range - regulated or not - would be fine.
Conclusion
Having had the parts on hand it took only a bit more than an hour to piece this together and almost as long to put it in the box seen in Figure 4.
When I forced both clocks to re-acquire WWVB's signal for syncing they immediately set themselves to the correct time and date - and since it had been the start of daylight saving time the night before but had not been able to synchronize prior to this - they "knew" the new, correct time, too!
* * * * *
This page stolen from ka7oei.blogspot.com
[END]
