ELF Receiver Design Avoids Inductors

I have an interest in extremely low-frequency (ELF) radio — transmission frequencies below 300kHz. Of particular interest to me is WWVB, the official government time station operated by the National Institute of Standards and Technology (NIST). It transmits on 60kHz. The NIST also operates WWV, which transmits on 2.5, 5, 10, 15, and 20MHz, but since those are HF shortwave transmissions, I don't pay much attention to them. All that RF stuff is too tricky.

But 60kHz — that's practically audio. So here's what I'm wondering. Could I make a WWVB receiver that is not a conventional superhetrodyne receiver? Instead, I think the classic tuned RF (TRF) receiver might be practical. This type of receiver was popular from 1915 to 1930 and preceded Armstrong's invention of the superhet. The TRF was not as selective as the superhet. Further, tuning multiple, cascaded stages was an iterative (and therefore tedious) process.

In spite of these issues, it would be an interesting design to investigate. Here are the details. We should need only one inductor: the loop antenna. This loop (aperture facing toward Fort Collins, Colo.) would have a parallel tuning capacitor selected to resonate with the loop's inductance at 60kHz. This tank circuit feeds a gain-and-tuning stage built around an op-amp configured as a VCVS bandpass filter (BPF) or possibly a three-op-amp biquad BPF.

I would design the filter to have as high Q as is practical without breaking into oscillation. The NIST would not appreciate someone splattering all over its transmissions. If more gain were needed, I could add a second stage just like the first. If I used the biquad topology, I could tune it by simultaneously tweaking two resistor values per filter. This is a good application for a dual (or quad) digitally controlled pot; I believe Maxim has these in its portfolio.

After the gain stage, I'd install the detector — probably an ideal diode stage built around another op-amp. After the detector, I'd add the decoder circuitry to extract the time and date information. Since that's all digital, and since that's not my forte, I'd leave that portion to someone far wiser than I am.

Would this work? Is it practical? Let me know.

10 comments on “ELF Receiver Design Avoids Inductors

  1. Michael Dunn
    January 21, 2013

    A DSP could handle all the filtering and decoding quite nicely. Switched-cap filter chips might be worth considering too. Not that there's anything wrong with a standard analog active filter.

    AFAIK, so-called atomic watches and clocks tune to WWVB, so the watches at least manage without inductors!

  2. EMCgenius
    January 23, 2013

    If your loop is stationary you can orient it for maximum signal strength.  For portable use, you have to accept that sometimes you will be in the null of the receive pattern and have to move.  In that instance a signal strength indication would be very helpful.

    Resonating the loop antenna as your only tuned circuit is problematic because LF ambient noise is high.  You really need narrow bandwidth for long distance reception.  Try a 60 KHz PLL for synchronous detection.  If you live near the transmitter in Ft. Collins, maybe SNR is not a problem.

    Without a loop, the alternative is a capacitive antenna (AM car radio antennas are of this type).  That's probably what the watches use, with the human as the reference.  I'm guessing a lot of signal averaging is used to extract the signal from the noise.

  3. BPaddock_#2
    January 23, 2013

    Take a look at:

    B Roehrig, “Most Accurate Frequency Standard, Parts 1,2,3”, 73 magazine, Jan/Feb/Mar 1994.
    Circuit boards available from FAR Circuits This is a WWVB frequency standard and needs additional circuitry to demodulate and decode the time signals. 

    I've always wanted to try one of NASA's Black Hole Antennas to see if it could be made to work at 60 kHz. 

    “A Broadband Active Antenna for ELF Magnetic Fields” by John F. Sutton and G. Craig Spaniol” in Physics Essays March 1993, Vol 6, #1, 1993. See:

  4. Michael Dunn
    January 23, 2013

    >a capacitive antenna (AM car radio antennas are of this type).

    Please expand on this.

    All the AM antennas I've ever seen have been ferrite rods, which, as I understand it, work only with the magnetic field. EM waves…still magic to me.

  5. EMCgenius
    January 23, 2013

    The whip antenna on an automobile is a small fraction of a wavelength (300 meters at 1 MHz versus a 1 meter whip) and therefore is basically a capacitor plate open to the air.  It and the capacitance of the coaxial cable that connects it are resonated with an inductor in the AM receiver (older designs) or folded into a bandpass filter (synthesizer based designs).  The receiver is in a shielded enclosure so the capacitive whip antenna is the sole source of reception.

  6. Bruce Bailey
    January 23, 2013

    BPaddock_2 –

    Good call on the articles by Roehrig in “73 Magazine”.  Content shows all that is needed – resonant, B-field antenna (lower noise than E-field?), Hi-Z JFET buffer, diode AGC (so you can haul it around), band-pass amplifiers to reduce harmonics generated in the diode AGC.  I think Brad can implement a TRF without inductors.

    Bruce Bailey, Ft Collins, CO

    A ferrite bar antenna wound to resonate with a fixed value cap into the Hi-Z JFET stage receives the signal in my basement – quite well!

  7. WKetel
    January 24, 2013

    Brad, The system as you describe it would probably work to some degree, but there are a few challenges. ELF is a very noisy part of the spectrum, and some of the noise is quite strong. Where that becomes a problem is in dynamic range and intermodulation product generation. The challenge being to have enough gain prior to the filter to overcome filter noise, while avoiding overloads due to unwanted noise. The solution is greater dynamic range, but that is not quite so easy.

    One common trick is to shield the tuned loop so that it does not pick up any electrostatic signals, which would not be so readily rejected by the single tuned loop. That is why many of the TRF receivers had several tuned circuits, each with an intermediate amplifier. Another option would be a direct conversion receiver, which would allow you to recover both the I and Q values of the signal. Then a DSP section would allow the recovery of the modulation signal.

  8. Brad Albing
    March 28, 2013

    Hmm… looks like s/o else already thought of my clever idea. I'm bummed.

  9. Jean-Luc.Suchail
    December 3, 2013

    I remember I had built an heterodyne receiver for ELF with a TL071 as RF stage, followed by a J-FET as mixer, gate driven with a 74C04 as local oscillator, frquency adjusted with a potentiometer, two IF bandpass filters also built with TL071 OP amps and a zero threshold detector with a diode in the feedback of another TL071.

    This was in 1982, perfectly working for DCF77 time signals. I had improved the selectivity with a manual variable capacitor and -what an horror- a coil wound in a ferrite RM type core.


  10. Brad_Albing
    December 3, 2013

    @F1GFK – So, someone else thought of my clever idea before me. Hmm… looks like I'll need to come up with some other clever ideas.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.