New Lightning Detector

(Also see Magnetic Antenna Version and an experimental op-amp version)

Built to make the wiring easy to follow. Read the text below!

This new lightning detector circuit uses a single inductor tuned circuit to receive static pulses from lightning at a frequency near 200 kHz. The need for a tapped tuned circuit is eliminated by employing a very high input impedance RF amplifier, using a darlington transistor. The amplifier is micro-power, and the whole circuit draws only about 200 uA from two alkaline D cells, hardly denting the shelf life. The flasher portion of the circuit is similar to the earlier versions, only the polarity of the transistors is reversed. As a result, the output pulses momentarily go to ground from a normally-high state.

The schematic and close-up photos were made with the beginner in mind. Questions about the schematic can be answered by close-up examination of the photos. Try a magnifier utility for an even closer look.

The 150k resistor in the base of the 2N4401 may be replaced with a series combination of a 100k resistor and a 100k (or greater) potentiometer to add sensitivity control. Simply adjust the potentiometer until flashing just stops for maximum sensitivity. The fixed resistor is plenty sensitive for most users. But, the receiver will only work well outside or near a window. A few yards of wire could be used to run an antenna out a window, but connect a 47 pF capacitor in series with the wire at the detector end. Also, connecting a good ground will improve the sensitivity. Keep the unit away from electronic or electrical appliances and wiring for best performance.

A common mistake is to select a 10 uH choke instead of a 10 mH. The large green choke in the photo is typical of the size and markings to expect. Color bands are brown, black, orange. A wide silver band often indicates the beginning of the color code.

Some of the other parts in the prototype were chosen to make it easy for the beginner to see the values. For example, the large .1 uF "orange drop" can be a tiny ceramic type. Parts may be in odd positions, too. The orange 10 uF is twisted to make the markings visible and other parts are oriented to make viewing easy. 

The 100 uF shown across the battery is the blue part right below the MPSA63 in the photo below. Notice how it isn't right by the battery connections. Sometimes the position of a component relative to others is important, but that information is often not on a schematic, unless specifically stated. Schematics aren't usually physical representations; they just show which legs of the various parts are electrically connected. Schematics are usually accompanied by assembly drawings as part of a complete design, especially when layout is critical. The point is to not take schematics to be mechanical layout instructions, although they often are a fairly good start. On the other hand, schematics generally aren't intentionally misleading about a good location for the part, either, but a little electronics experience helps a great deal when there's only a schematic to go by.

The parts are not particularly critical. The darlington PNP transistor may be a different number, as may the other transistors. The diodes can be any silicon switching diode, like the 1N4448 (or probably any orange one you have). An amber LED with a 22 ohm resistor draws about 30 mA with fresh batteries. Vary the resistor to achieve the desired LED current for other colors (connect the LED and resistor to the battery through a current meter to select the value). The output transistor can sink several hundred mA, so several LEDs or other loads may be connected in parallel. The head from a cheap LED flahslight that uses two cells will be really bright! The pulse is pretty short to save power and for quick response, but for longer pulses, increase the 10 uF capacitor value. You might want a longer pulse to flash a 3 volt incandescent bulb (my favorite). Actually, I'm using a 1.5 volt bulb and it's really bright (not shown in the photos), even with this short pulse! I have lots of them and I'm sure I'll need to replace it fairly often.

Below is a "photoshopped" close-up without the sensitivity adjustment option. The construction technique was chosen to make it easy to follow the wiring. Normally, the leads would be stuck through the holes and wired on the other side of the perf board. I used the perf board simply as a guide for copper nails driven into a pine board - strictly to make it easy to see all the connections.  The perf board serves no purpose beyond visibility, so one could also use nails and wood for the assembly. The copper nails from the home improvement store really solder easily, by the way.

I made several image edits for clarity and to correct mistakes; they're not perfect but I think the connections are clear.

The next photo shows the optional potentiometer and resistor for adjusting the sensitivity. Notice that the 150k resistor has been removed. The odd-looking color bands on the resistors were painted on to make them more visible after cleaning faded the original bands. The violet band looks a little dark on the 27k resistor between the two diodes; that should be red, violet, orange.


Marcos in Brazil made a video of his detector.

Magnetic Antenna Version

With minor modifications it's possible to replace the telescoping antenna with a solenoid type inductor. One could use a ferrite loopstick operating at radio frequency but I've decided to tune this antenna to about 5 kHz. The resonant tank rolls off high-frequency noise and the circuit rejects lower frequency line-related hum. The lightning impulses ring this tank like a bell, easily seen on the collector of the darlington with an oscilloscope.

The 1 Henry inductor and 1 nF capacitor resonate near 5 kHz. Choose a value of capacitor that resonated your inductor around 5 kHz (not particularly critical). The inductor needs to be a straight solenoid type with no magnetic shielding on the ends. My prototype's inductor is from the high-voltage generator in an old CDV-715 radiation survey meter. The inductance should be between about 50 mH and 1 H. Typical inductors are shown below:

Bobbins from power transformers may also work but the core must be removed and optionally replaced by a ferrite rod or audio transformer laminated iron (only a straight "solenoid" core so that the external field gets in). The black coil above is from an industrial solenoid and the large #33 Amadon core brings the inductance up to nearly 1 Henry. I'm using the tiny one on the bottom right with good results. The pretty beige coil (an eBay find) measures about 50 mH. Use a good quality film or NPO ceramic capacitor to resonate your inductor. You can't use toroids, pot cores, or other cores that close the magnetic path back around the coil. You can probably use a relay coil if you remove all the metal pieces. Hook your prospective coil to a high-gain audio amplifier and if you hear distant lightning crackles you probably have a good candidate (assuming a storm is within a hundred miles or so). When indoors loud AC hum is another indication that the coil might be suitable.

Hook the inductor and capacitor in parallel across your scope and apply a signal through a 1 megohm resistor with an audio generator to see the resonance. Or use the lightning simulator and observe the frequency of the ringing on the collector of the darlington (practical with a digital oscilloscope). I just looked at the ringing caused by nearby lightning.

Preliminary Op-Amp Version

Here's a preliminary schematic using a micro-power op-amp:

The coil and ferrite rod are from eBay and measure 2.5 Henry shunted by 90 pF. Adding the 330 pF brings the resonant frequency down to 5 kHz, a great frequency for lightning energy but also just about as fast as this op-amp can go. The micropower op-amp draws about 20 uA so this can run on 3 C-cell batteries for quite a long time - except - it's so sensitive that it flashes quite often. I'd turn down the sensitivity, frankly. Right now it's flashing in response to a storm in Louisiana and I'm in Austin, Texas. That's just too much sensitivity. This prototype draws 24 uA when not flashing.

Or run it on a molded power supply. I prefer the power hungry incandescent bulb; they just look better in this application, at least to me. The bottom two diodes provide a little positive voltage to bias the op-amps even though these particular ones can look below ground a little. I was a little worried about large signals exceeding the negative voltage input rating and I want plenty of swing to turn on the detector diode. It's tempting to use a micro-power voltage reference there. The top diode in the string generates a threshold voltage for the second op-amp that acts like a comparator. The horizontal diode rectifies the 5 kHz ringing signal from the tuned circuit and the .01u and 10 megohm stretch the detected pulse for a bright flash. The time-constant is plenty short to see lots of quick flashes in succession. Strong pulses light the lamp for longer. When using an LED the time-constant could be shortened by reducing the .01 uF capacitor. Some of the other caps may seem small for the job but I am also trying to kill sensitivity at 60 Hz.

I stuck it in a peanut butter jar. A large standoff holds three D-cells taped together at the bottom of the jar and the circuitry is in the lid. The lamp screws into a Leviton socket so I can unscrew the bulb to save the battery during busy storms. One could replace the 2N2219 with a VMOS transistor (with very low turn-on voltage) and unscrewing the lamp would result in extremely low current even during a busy storm since there would be no base current (about 3 mA) during flashes. Even so, these batteries will last for years in my experience so I just soldered them in series, no battery holder and no power switch. It immediately started flashing and, sure enough, a storm is approaching from the northwest.