A few years ago at Dayton I was visiting the FAR Circuits tent and ran across a little circuit by Larry Coyle, K1QW, for converting the LF and VLF band to 4 MHz. This was done to allow users to experience the perceived improvement in sensitivity of many receivers in the 4 MHz band as compared to the often poor performance down in the LF band. A very well-written two part article accompanied this circuit and can be found in the November 2008 QST, “A Modular Receiver for Exploring the LF/VLF Bands”. Part one covered the circuit I put together, the converter. Part 2, which I will probably not explore (at least not anytime soon) was a software defined receiver. It is described as a good way to get introduced to software defined radio and might be worthy of consideration.
The converter is another SA602-based creation, incorporating a switchable 12 dB attenuator to knock down out-of-band signals which may overload the mixer chip, and a “5-pole elliptic low-pass filter with a cutoff around 550 kHz” (to reject MW broadcast band signals). The parts to build most of this came out of my junkbox and are commonly available. I did have to order the transistor he used for the local oscillator input into the mixer, the 4 MHz crystal, a TL317CLP voltage regulator and the Schottky diode used in the power supply.
I was going to post the circuit but I will refrain as I do not see it in the public domain.
The basic functionality of the circuit is straightforward. The signal from the antenna is low-pass filtered and up-converted to 4 MHz. A transformer off the 602 chip converts impedance to around 50 ohms for export to an external receiver.
The description of this circuit implied that the full spectrum from DC up to about 500 kHz should be received fairly well with a loop antenna and no tuning should be required. However, this was not the experience I had with this circuit.
Antenna for LF
Originally I wound a ferrite rod style antenna, with an inductance of about 1 mH. This was deaf as a fencepost and I could not receive any signals whatsoever either using the converter box or via direct input (mismatch and all) into my Kenwood. I injected signals into the converter box from my signal generator just to verify the circuit was working, and sure enough I was getting output in the 4 MHz band. But no beacons and not even any AM broadcasts were coming through at the low-end of the MW band.
Reluctantly, I wound a larger loop antenna on a wood frame, the wire forming a square of about 2 feet on a side. I had one of these as a teenager and had kept it for a few years but I found traveling with it to be cumbersome and tore it down. This time around I built it to be easily taken apart and re-constructed at some later time. Or at least that was the plan.
Normally these antennas are constructed such that you make some grooves at the ends of the support beams and wind the coil around the frame to sit within those grooves. But I wanted the wires to be contained better than that, and thought I’d instead drill holes at the ends of each beam and just thread the wire through the holes. Bad move! It took HOURS to wind the loop this way! And there really is no simple method to thread these holes with large lengths of wire. I ended up doing 20 painstaking turns in this manner, yielding about 700 uH.
The pain of winding yielded results though: signals! I could receive some NDB beacons in my local area (two total). And at least the local broadcasters could be heard.
But signals were VERY weak. I found that the beacon was almost impossible to hear, and I could actually hear it stronger off of my end fed longwire tuned to 40m on my TS-2000. Something seemed way off to me.
It could be that because my loop was a little smaller than the author’s the impedance was too mismatched (his was 30 inches on a side to my 24 inches). But the entire band was really dead. I did not hear any lightning “whistlers” even as the Virginia summer thunderstorm season was raging on around me. That was kind of a disappointment.
Tuning the loop
I decided to do a little engineering on the circuit and investigate the impedance. I put the antenna on my TAPR VNA and soon learned that the impedance was very high and inductive. Not surprising I suppose for a big coil. The input impedance of the SA602 chip is also rather high, around 1.5 kohms. So it might not be that badly matched, but for such high impedances my VNA tool is a little hard to interpret (the values seem to swim all over the Smith chart). I’m not really sure if I’m seeing a kohm or 100 kohms.
I wanted to see if I could bring the impedance down closer to 50 ohms at around 200 kHz and just feed my Kenwood with it directly, without the converter, and see if I could at least make out the beacon. In order to do this in a practical manner I needed to use a balun, then tune the hot side with a series capacitance (using the balun means I don’t need to try to simultaneously tune two lines). This actually worked very well! I have a really nice air variable in my junkbox that will tune from a few pF up to about 1200 pF, and I pressed that into service. I found someone else on the internet (see his interesting little writeup at lf loops) who had posted a design for their loop antenna and said that they found a good match by using 2200 pF capacitors in series just before the balun. Having difficulty initially finding a match, I decided to copy his layout and amazingly I got very good matches from about 200 kHz up to 600 kHz! Wow. I felt like I struck gold.
Plugging this into the Kenwood, signals began coming in much better. The level of the beacon (at around 270 kHz) was stronger now than off of un-tuned wire antennas and the local broadcasters were booming in. Progress.
But I still wanted to make use of my converter box! After all, I had built the circuit into a nice metal box with connectors, switches and even a nice little green LED! I even painted the box yellow to signify VLF lightning signals.
Since I knew how to tune this to 50 ohms I decided the easiest thing to do would be to just bump the impedance back up to some arbitrary high value to match what the converter box would be expecting. Admittedly by this time I was losing interest in the project and kind of wanted to wrap it up, so making this match perfect was less important to me than just getting it to work reasonably well.
I ended up rolling two more toroidal transformers, one a 1:4 balun (to bump up impedance and convert back to a balanced transmission line) and one a simple 1:2 transformer. With the two transformations I should get an output impedance of about 800 ohms. I also found that if I switched in another capacitor in parallel with my big variable I could get it to tune down to about 170 kHz. The details and a picture are shown below.
The completed system worked very well. Signals and even just the basic noise floor are elevated up to reasonable levels throughout the LF band and a big improvement from the tuned loop by itself. It seems that my Kenwood’s sensitivity in the 4 MHz band is indeed superior to its LF performance. Below 170 kHz though performance rapidly deteriorates and the system is pretty deaf. I could not find values of capacitance or inductance to match down below 170 kHz, although I’m sure such a system could be constructed. Increasing the difficulty is the fact that my VNA does not tune below 200 kHz so I would need to use other methods to design a match.
On the first night after finishing this system I actually heard a broadcast station from Cuba on 570 kHz! It was “Radio Reloj” (https://en.wikipedia.org/wiki/Radio_Reloj) I was surprised to say the least. Apparently when some other stations on the frequency are off-air it is common to hear this Cuban clock station up and down the east coast of the US.
The rest of the LF band was pretty disappointing. I did pick up one more NDB beacon. But there were no pirates or experimental stations heard. And no lightning whistlers! I’ve since learned that there are other design strategies for listening to lightning, employing audio amplifiers. I will either try one of those or contact the author of this article for his thoughts.
I believe I will eventually play with some of the audio-amp style lightning circuits. And I’m also a little curious if anyone else is hearing the megawatt station NAA in Maine that is communicating with submarines in the VLF region (below 100 kHz somewhere). I couldn’t hear NAA, WWVB or the LORAN system at all. So, while I always enjoy tuning things and making baluns and seeing them work, it would also be nice to hear interesting signals!
I’m running out of time before my schedule kills most of my ham radio play time once again. But I’m hoping to look at some Arduino stuff and a Raspberry Pi I picked up at Dayton. Stay tuned and happy soldering.