Notes on Small Ferrite-Core Antennas
for 3496 Hz Radiolocation

         I recently completed 2 ferrite beacon antennas and 1 receiving antenna for my 3496 Hz radio. We will soon be drilling a 35 meter well hole into a 1 mtr wide cave passage, which will be enlarged to make a new vertical entrance to a large cave system.  I have already done the radiolocation. I am assuming that we may miss the passage simply from drill-rig error if from nothing else. We must at least open up the bottom of the hole to provide a water drain for the digging project.  Note:  The drill came out close to the center of the passage!  This was confirmed with a camera.

Low-power Beacon Loop

    A landowner needed a water source on land next to a cliff.  The only water was in a small cave below theproperty with an entrance in the cliff   I radiolocated a pool at the rear of the cave.  Unfortunately, the landowner drilled about 6 ft (2 m) to one side to avoid having the well pipe in the middle of the passage.  This put the well shaft in the wall of the pool room.   
I built a submersible beacon loop that was lowered into the 6 inch "well" to generate a signal that was used to find  the location of the hole behind the cave wall. The electronics stayed on the surface. I used a high-u 5/8 inch dia by 12 inch rod with a single layer winding of 22 Gauge wire. I haven't properly measured Q yet. I used my standard beacon circuit except the tuning capacitor is at the loop. The feedline is 150 feet (45 meters) of 2-conductor #18 solid wire with a tough jacket used in houses for wiring thermostats and intercoms. This line does not degrade the beacon signal.  I marked depths directly on the feedline with a fine-point Magic Marker.  The battery drain is only .027 amps because I was only interested in short range.  The feedline also conducted 27 MHZ CB signals into the cave for 2-way voice communications using handheld AM CB radios.  I used this strange beacon with the miniature receive loop described next.

Miniature Receive Loop

        I built a tiny receive loop to allow me to take my beacon receiver into the cave. The high-u rod is 1" dia by 3" long with 1000 turns of fine litz wire.  Sensitivity is about 2.8 nanoamps/mtr in a 1 Hz bandwidth, limited by loop thermal noise. This is about 11 dB worse sensitivity than my normal 22" diameter receive loop (2 lbs of #28 wire). Equivalent E-field effective height of the ferrite loop is 32.5 mm, giving signal levels about 29 dB below the full-size loop.
        Using the above 2 antennas, I got "100" on the receiver's digital meter at 370 ft (112 m) HMD range. This is ~40 dB s/n in 1 Hz BW. This is far greater range than I really needed.
    Using the CB's we had the surface crew lower the beacon loop until it was about 3 ft (1 m) from the floor of the pool room in the cave.  I was then able to use triangulation to locate and mark the beacon's position, then determine that it was about 6 ft (2 m) back in the wall.  It was then up to a mining crew to tunnel the short distance in to it.

High-Power Beacon Loop

        This inspired my to build a powerful ferrite beacon loop to see what was possible. I used a 5/8" dia by 12" long rod and set the tap for 1/2 amp DC from the beacon battery. I believe this ferrite is Stackpole 24B, a low freq ferrite with torroidal u~2000. Q is >100 with a magnetic moment of 7.5 A-T-m squared. This is only 4.5 dB less than the 22 inch loop I used at Wakulla, which also draws 1/2 amp! If only I had known then! I have built this antenna into a submersible pipe container with a cable at one end. Hung from its cable, on any support, it is self-leveling. It could be used for diver tracking by hanging it from a scooter (probably with a water current deflector).
        The beacon circuit is not optimized for this high-Q loop. I recommend using the new Class-E beacon design (Jan 2008) for this and all new beacon designs.  Also read the following caution: Note: The permeability of ferrite rods used for transmitting will gradually increase as the RMS current through the loop winding increases. This happens far below the level that will saturate the ferrite.  This will cause enough increase in inductance to de-tune the loop.  The result is that the final loop tuning must be done while transmitting at the power level that the beacon will actually use.  The final capacitance value will always be less that the value calculated using the small-signal inductance value such as that obtained with an LCR meter.  Another effect of high power is that losses increase, i. e. the Q drops from the small-signal value.  What limits the maximum power level is the ability of the circuit to "re-start" after it has been shut off.  If the loop has been re-tuned too far at high power, it will present such a high impedance at low power (i.e. when the circuit is again turned on) that not enough loop current will flow to "snap" the lindutance back to its high-power value!  In effect, the circuit becomes bi-stable.  It can be tuned for a nice high output power (high loop current), but when shut off then turned on again, it operates at a much lower power level and stays that way.  I found the maximum power level by trial and error once I understood the problem.
There is a new (2008) ferrite beacon antenna design, for the new beacon circiut, here.
    In June 2000 I used this ferrite Beacon to do Radiolocations in NY, PA, and W VA, at depths in excess of 200 feet (60m) with signals stronger than necessary. The underground crews really like the small size and instant setup and leveling.
    In July 2000 I tried varying the duty cycle and driving at 1/2 frequency with no increase in efficiency!  I tried changing the zener snubber from 30v to 40v, which actually decreased efficiency slightly!  This inspired me to try the lowest voltage snubber possible.  The actual signal on the drain of the MOSFET Q1 (Loop Tap) is a 6 volt p-p sinewave riding on +12v. This is not pure class-E which would be close to a 12v p-p sinewave.  What I did was to wire a 5v zener in series with a 1N4001 power diode (the diodes connected anode-anode to prevent "normal diode" conduction in the zener). The cathode of the zener went to the MOSFET drain and the cathode of the 4001 went to +12v. The AC loop voltage went from 122v to 138v and the DC battery current dropped from .50A to .42A. Loop tuning was checked before and after (it did not change). This is a significant improvement in efficiency. Z1 (the 30v zener) no longer does anything, but is still a good precaution if the loop antenna can be unplugged.