Simple LowFER Transmitter

by Lyle Koehler, KØLR

One of the neatest things about 1750 meters is that it's a home-brewer's paradise. Transmitter circuits can be very simple and inexpensive, you don't have to worry about hazardous voltages (either DC or RF), and almost any construction technique will work. An entire transmitter can be built on a small solderless plug-in protoboard.
The circuit in Figure 1 uses a complementary-pair final that I've had good success with over the years. Although my present "LEK" transmitter has a home-brew frequency synthesizer that drives the final, a 74HC4060 oscillator/divider circuit works just as well and is about the simplest crystal-controlled "exciter" you can build. Transistors with the "A" suffix seem to work somewhat better in the final than plain 2N2222s and 2N2907s. The cheap plastic PN2222A and PN2907A transistors from Mouser, Tech America or other suppliers are fine. I've tried complementary-pair FETs, but their drive levels are a bit more critical, so I've stuck with bipolar transistors. When driving low-resistance loads, less than 20 ohms, higher-power transistors like the NTE129MCP matched pair or an NTE186/NTE187 combination may give slightly better efficiency. However, the difference will not be noticeable at the receiving end, and it's comforting to know that if the final gets zapped by lightning I can replace both of the transistors for under 20 cents. Besides that, using low-power transistors keeps you from cranking up the power to 5 or 10 watts and spoiling the fun!
In Figure 1, the crystal oscillator runs at 32 times the desired output frequency. For operation at 160 to 190 kHz, the crystal frequency must be in the range of 5.120 to 6.080 MHz. Until this fall (1998) it was advisable to stay above 176 kHz to avoid competition from the Ground Wave Emergency Network (GWEN). However, GWEN has been decommissioned and the bottom 15 kHz of the band is now a pretty good place to operate. The 74HC4060 has outputs at other division ratios which let you use different crystal frequencies. For example, you can use crystals in the 3-MHz range and the ÷16 output (pin 7) or crystals in the 12-MHz range with the ÷64 output (pin 4). Higher division ratios are available, but crystals above 20 MHz are often overtone-type crystals which will not work properly in this circuit. The trimmer capacitor provides a limited range of adjustment of the oscillator frequency, and can be replaced with a fixed value of about 27 pF if you aren't worried about hitting a precise frequency. Pin 9 is the output of the crystal oscillator section. There should be a signal on this pin whenever power is applied to the circuit. The outputs of the various divider stages are only present when the keying line (pin 12, the reset line for the divider chain) is grounded.
The output of the complementary-pair final is a square wave, which contains lots of harmonics. However, it may actually be "cleaner" than other high-efficiency finals. There is theoretically no energy on even harmonics, and the third harmonic is almost 10 dB down from the fundamental. A typical LowFER antenna/loading coil combination acts as a narrowband filter. You can feed the antenna directly with a square wave and the harmonic levels are likely to be well below the -20 dB level required by the FCC Part 15 rules without any low-pass filtering. But to keep peace with the neighbors, I recommend enclosing the transmitter in a metal box and using at least the ferrite bead with the 560-pF capacitor across the output jack. Neither the type of ferrite bead nor the value of the capacitor is critical to the operation of the transmitter; they are just included to eliminate TV and FM interference. In my "LEK" transmitter, I also use the low-pass filter shown enclosed within the dotted lines in Figure 1 just to make sure that the output is clean no matter what kind of antenna is connected.
Values of L1, L2 and C1 shown in Figure 1 are for an antenna impedance of 50 ohms. L1 and L2 can be made with 60 turns of #26 wire on a T80-3 (grey) toroid form, or 40 turns of #22 wire on a 1.9 inch diameter PVC pipe. A plastic pill bottle or the cardboard core from a toilet paper roll are possible substitutes if you don't have any pieces of PVC pipe lying around. L1 and L2 are separate coils, and should be mounted at right angles to each other to minimize inductive coupling if you use the 1.9 inch coil forms. Toroid cores are more or less self-shielding, so there isn't much coupling between them unless they are very close together. Capacitor C1 should have a rating of at least 100 volts. It may be hard to find a .018 uF capacitor, but .015 uF and 2700 pF connected in parallel will be close enough. In fact, you may want to experiment with different capacitance values to get the maximum output across a 50-ohm load resistor with the actual values of inductance in your circuit.
A good LowFER antenna will not necessarily present a 50 ohm load to the transmitter. Actually it should have an impedance of less than 20 ohms at resonance if the loading coil and ground losses are low. You could scale the component values in the filter to "match" the impedance, but the antenna impedance will vary depending on whether the ground is frozen, leaves are on nearby trees, etc. A single external shunt reactance at the base of the antenna (in other words, across the transmitter's output connector) can be used to bring the impedance up to 50 ohms. This technique is often used on mobile antennas and is described in the ARRL Antenna Book. When you insert the shunt reactance, the value of the loading coil inductance must be adjusted slightly to bring the antenna system back to resonance. Either a shunt capacitor or inductor will raise the feed-point impedance. A shunt inductor has the advantage that it provides a DC path to ground to drain off static charges from snow, dust, or lightning. However, a capacitor helps reduce harmonic radiation, and it's easier to get good quality capacitors in small sizes. I use a shunt inductor in summer because it provides better lightning immunity, and a shunt capacitor in winter when I'm trying to squeeze out every last milliwatt. The value of shunt reactance is not very critical. A 30 uH inductor or 0.02 uF capacitor is about right if the antenna impedance is between 15 and 20 ohms.
My LF transmitter also uses a power limiting circuit so that the DC power input can never exceed the FCC limit. The circuit consists simply of a power supply with a DC voltage, V, that is twice as high as the final amplifier needs, and a series resistor equal to V(squared)/2. My LF final gives the best efficiency with about 13 volts at 77 mA, so I'm using a 26-volt supply and a 169-ohm series resistor. This is done more for convenience rather than for fear that the transmitter might be a few milliwatts above the FCC limit. It makes it possible to tune for maximum antenna current (or field strength) without worrying about the DC power input. The output also stays more constant when the antenna is detuned by rain or ice than it would with a "stiff" power supply. If you don't know in advance what voltage gives the best efficiency, you can start with a best guess, measure the voltage the final has "chosen" after the antenna current is maximized, and use this value to modify the supply circuit. One or two iterations are adequate if the first guess was anywhere near the target.
The best LowFER antennas are verticals with big top hats to improve the current distribution and minimize the amount of loading coil inductance required. If the antenna is not properly tuned, your range is going to be measured in feet rather than in miles! The loading coil can be inserted anywhere in the vertical portion of the antenna if a large top hat is used. Best efficiency will usually be obtained when the loading coil is near the middle of the vertical section. However, this may present construction problems, and the required inductance will be greater than it would be if the coil is near the base. Forget about using a tuner or some wimpy little loading coil like the ones used on HF mobile antennas. It typically takes 2500 uH or more for a base-loaded antenna with a big top hat, and about 5,000 uH for a simple base-loaded 15-meter vertical with a 1.5" mast. I've seen published suggestions for LowFER loading coil designs that are totally out of the ballpark in Q, inductance or both. Your best bet is to design your own. To get reasonable Q in a loading coil for these frequencies, use #20 or larger wire, with a winding pitch (ratio of center-to-center spacing to the wire diameter) of about 2, and a diameter-to-length ratio between 1:1 and 2.5:1. Litz wire is great if you can find it in large enough sizes. Avoid lossy coil form materials like PVC and Nylon if you want high-Q coils. Air is the best insulator, but styrofoam comes close. Other suitable materials are Teflon, polyethylene and polystyrene. Included in Reg Edwards' excellent collection of free software on his G4FGQ web site are programs to design transmitting and receiving loops, coils and vertical antennas. His VERTLOAD program will calculate all you need to know about vertical antennas (without top hats), and SOLENOID provides accurate calculations of coil inductance and Q. In Reg's programs, wire diameters and coil dimensions are given in millimeters. I've written a program called AWG-COIL that will convert AWG wire gauges to millimeters, and will also calculate the dimensions of a loading coil if you input the desired inductance, wire size and winding pitch.
For more information on LowFER antennas, my LOWDOWN article on "Getting the Most out of LowFER Transmitting Antennas" is available on-line via the File Libraries section of the Longwave Home Page.