# Electromagnetic signal from spark gap

I am a power electrical engineer college student, and I have a project to make an antenna for magnetic direction finding.

My main topic is to detect lightning strikes, but its to hard due to randomness of the lightning occurence. So my lecturer suggest me to model the lightning strike into a spark gap from impulse voltage generator. I learned that lightning spark radiates electromagnetic waves on radio frequency

My question:

1. How can I know on which frequency I need to make my receiver antenna resonate?
2. Is it true that I can model lightning strike into spark gap from impulse voltage generator? (I can't just write on my paper that my lecturer told me to.)
• Hello Ramanda, and welcome to ham.stackexchange.com! Dec 7 '20 at 23:48
• Yes, welcome! We understand you, and no need to apologize; so I'll edit that out. :-) Dec 13 '20 at 22:54

Don't worry, there have been many researchers before you, and you can borrow some of their work. Your English is fine.

1. If you do a web search for "lightning spectrum", you will find a graph showing lightning intensity vs. frequency. The range of frequencies emitted by lightning discharges is very broad. From amateur radio experience I can tell you that in the US, the 160m, 80m, 60m, and 40m bands (wavelengths) are difficult to use in summer because of noise from lightning.

2. Try it and see! Build a high-voltage source, discharge it through a spark gap, and listen with a radio, such as an AM radio or shortwave radio receiver. I think that you will hear a noise.

• Woah.. one more thing sir. so, on a single spark/discharge, the electromagnetic signal is radiated on multiple/wideband frequency? I thought a electromagnetic wave emitted by lightning discharge or spark grap only has a single frequency. So I need to really know on what specific frequency I have to make my antenna resonates. Thank you very much! I hope I can meet you and shake your hand personally. Dec 8 '20 at 18:21
• You're welcome! If you like my answer, please consider upvoting it (the up arrow). Once sufficient time has passed, if you feel that one of the answers (assuming there are more than one) has answered your question, please consider "accepting" it by clicking the check mark. I recommend giving some time before accepting, in the hope of attracting as many answers as possible. Dec 8 '20 at 18:28
• Okay sir.. I am new in this forum. Thanks for your advice! I will learn how to do it. Umm last thing, can you confirm my last question on my previous comment? I ask about is there correct that you were saying that the single lightning discharge emmits wideband frequency? Is that mean I have many options for the operating frequency on my antenna design? Because the electromagnetic signal is not radiated on a single frequency? I am sorry if that a silly question since I am not very good on my course, but I really want to finish college. Dec 8 '20 at 18:37
• No worries sir! A static discharge, whether through your spark gap or by lightning, sends out radio waves at a very wide range of frequencies simultaneously. You can pick a frequency that is convenient to you inside that range, and listen there. Dec 8 '20 at 19:42
• To explain why, if you recall a square wave can be regarded as the sum of an infinite series of odd-integer harmonics. Lightning is like a square wave, because in one instant there is no current flowing, and then in the next instant there is a large current. If you look at a true square wave with a spectrum analyzer you will see many harmonics at many frequencies. Lightning radiates at a wide range of frequencies simultaneously because it is like a square wave when it goes from 0 to 1, if you know what I mean. Dec 8 '20 at 19:50

A bolt of lightning has an average center frequency of about 16 kHz, which varies with the height of the bolt.

This is a distant stroke that I captured on one of my SDR lightning receivers, using three 4-turn 1-meter dia. vertical loop antennas, spaced 120° apart:

(Ignore the apparent 50 mV offset)

• Note how the energy on this particular stroke is concentrated below 20 kHz, in the LF region (very typical). The two higher frequency, broad areas are almost certainly due to ionospheric Doppler shifting.

• The harmonics extend up into to the VHF region. However, since this FFT spectrogram's vertical axis is based on power —and the filters roll off the unneeded higher frequencies— the much weaker harmonics do not display.

Having said that, occasionally I see many significant ionospherically-Doppler-shifted "bounces" up to 80 or 90 kHz. They appear as equally-spaced bell curves that decrease in height with each higher frequency-shifted bounce.

The 16 kHz figure is based on many observations of mine over a long period of time.

This signal was sent to the blitzortung.org and lightningmaps.org servers in Europe, and was displayed on the map at https://lightningmaps.org as a line from my rural location in SW Missouri to the distant (>1300 km) stroke.

A short spark gap on a laboratory tabletop will tend to produce much higher frequencies, unless you incorporate a suitable tuned circuit in your design. (Googling spark transmitter will return many designs and information from the late 18th and 19th century up to 1927.) Heinrich Hertz' first spark experiment was centered somewhere in the VHF region, IIRC.

The antenna does not necessarily have to be resonant. No one on the Blitzortung network uses resonant antennas. You might get more ideas if you also ask this on https://forum.blitzortung.org.

Some radio amateurs that operate on 137 kHz and 630 kHz use loop antennas with many turns of wire for receiving.

Others report success with short, high vertical antennas with a high-gain, high-impedance preamp like this one from PA0RDT.

• Hello, sir! Thank you very much for your asnwer! I really appreciate it.. may I ask for more? How did you choose on what frequency that you want for your antenna design to operate? I want to make an antenna to detect a spark gap from lightning voltage generator and do not know how determine the dimension for my loop antenna design. Thank you sir! Dec 13 '20 at 23:36
• I mean impuls voltage generator.. Dec 14 '20 at 0:09
• @RamandaNvl I simply made the antennas according to what was recommended by the developers. It is not tuned and is very broadband. The receiver itself is tuned to the desired frequency range. Dec 14 '20 at 3:12
• Sir, does the antenna for detect lightning can also work for detect spark gap from lightning impulse generator? My goal is to test out Magnetic direction finding (MDF) method using loop antenna to locating magnetic source. What do you think sir? Dec 14 '20 at 19:50
• The reason why I test it from spark gap is because it's hard to predict when the natural lightning occurred.. and I am limited by time due the deadline from my college. Dec 14 '20 at 19:51
1. The spark gap spectrum is highly determined by the frequency transfer of the (to the spark-gap source) connected antenna.
2. Near EM-field from a spark gap radio spectrum generator differs essentially from a remote lightning discharge.
3. Both E and H field sense antennas can be used for lightning detection. A combination is very powerfull (direction finding).
4. Approach to find the location of a spark gap (radio) source in the near field (within half a mile of that spark gap!) is diffuse: the magnetic field is the result from the connected wires of the spark gap circuit. The spark gap is a kind of switch in this circuit. The magnetic field is rather homogeneous, NOT stronger near the gap. The E-field is transmitted by the connected wires: the wires function as antenna and mainly the spectral content varies when coming closer to the gap.

Conclusion: the spark gap generator is diffusing the location of the spark and is less usable for the development or investigation of lightning detection equipment.

• What about a spark gap in the center of a vertical dipole, or a vertically-polarized loop? The power supply leads would be filtered right where they connect to the gap [EDIT:], in a manner where any radiation from them would be insignificant. This might include shielding those leads. Dec 17 '20 at 0:32
• Good idea: almost a point source. Spectrum is determined mainly by the dipole antenna. So in total not a real lightning radio spectrum. And also: this is a near field solution. Detection on larger distance requires extreme energy in the spark. Dec 17 '20 at 15:35
• Agreed, nothing is a substitute for a genuine bolt of lightning. :-) Suppose, thought, the antenna could be tuned to ~16 khz? Might it be useful to @Ramanda? Please note my edit here. Dec 17 '20 at 18:21
• Blitzortung.org is the best source. A dipole as transmitter and a loop as receiver in the near field is a poor combination in my opinion. To test the receiver chain a (tuned) loop as transmitter is probably the best approach when the receiver is also based on magnet field reception: wideband loop. Dec 17 '20 at 19:53
• True. But receiving with a loop and transmitting with a dipole in the near field (...wavelength separation at 16 kHz...) is not a good idea. Dec 18 '20 at 9:47