"Sferics" (also spelled "spherics") is short for atmospherics and refers to the crackling static in a radio produced by lightning. This receiver is tuned to an unused portion of the LF band and provides enough sensitivity to detect lightning activity for hundreds or even thousands of miles. Best performance is achieved with an external antenna, but just a few yards of insulated wire is sufficient. A short whip mounted on the unit will work fine for picking up local storms.
Much explanation is needed for the various sections. Starting to the left, a trap or two may be added by resonating a 1 mH choke with a selected capacitor to eliminate strong AM stations, when using an outdoor antenna, especially an amplified antenna. The traps aren't necessary with a short antenna. (I like faint station bleed-through, simply to let me know it's all working properly.) The 100 ohm resistor on the TL592 may be changed to change the gain, higher values giving less gain. A short may be used, but values much below 20 ohms will start to spread the bandwidth, and might cause stability problems, due to all that gain in one box! For more sensitivity try a 33 ohm and increase the 4.7 nF on pin 3 of the LM386 to 15 nF if instability is noticed. You can also add a resistor in series with the 10 uF from pin 1 to 8 on the LM386 since there will be excess of available audio gain that can cause instability when the volume is turned up high.
The 3.6 nF capacitor is an odd value that I happen to have in mica, and isn't critical, giving a receive frequency of around 80 kHz. Any value from one to ten nF should be fine, but change the value if a VLF signal is interfering. I just changed mine to dodge some buzzing interference. Use a good RF capacitor; this value also "shorts out" higher frequency stations. That's the purpose of the 330 ohm resistor, to give the tank capacitor something to work against when rolling off higher frequency stations, especially strong FM stations.
The 1N5711 detector is a biased detector with a twist. The detector has a long discharge time-constant that gives the output a ramp-like response to lightning pulses, with the length of the ramp being proportional to the size of the pulse. That causes the first LM358, acting as a comparator, to output a pulse with a width that is proportional to the received pulse height. The detector's amplitude-to-pulse width response also gives more powerful strikes a deeper sound, making it easy to tell what's happening simply by listening. Those pulses drive a lamp flasher and another op-amp averaging circuit for driving a meter. The tendency of the lamp to flash is easily modified by changing the 1k to ground on pin 6, with a lower value making the light more sensitive. The 22k resistor determines how much each pulse affects the meter reading, and the 10 megohm determines the window of time that is averaged, lower values giving a faster response.
The schematic shows a current meter, but a voltmeter would also work fine. Expect about 3 volts at pin 1 when the lightning is "pegging" the meter circuit.
The LM386 needs care to prevent oscillation, because it has gain in the frequency range of the receiver. First of all, note the 4.7 nF at pin 3 is a little larger than normally seen, and the addition of the series 4.7k makes sure there is roll-off even when the volume control is all the way up. Also, the output stabilization network is modified to block RF from getting into the speaker wires. I simply wound 30 turns of enamel wire around the body of a 1/2 watt, 47 ohm resistor, but one could also wrap four or five turns through a ferrite bead and connect that in parallel with a 47 ohm resistor. It would be a good idea to keep the speaker away from the input circuit and to twist the speaker wires together. I didn't do either but mine works. If fact mine got really crowded!
The experimenter might consider enclosing the front-end circuitry, including the MPSA18, in a box near the antenna, and enclosing the rest of the circuit in an interior case. That would greatly reduce potential stability problems.
Rajkumar from Nepal made a much neater version:
The PCB was made using the toner transfer method from a layout made using ExpressPCB. The results are quite professional!