logo elektroda
logo elektroda
X
logo elektroda
Advanced Search
logo elektroda

A simple Tesla music coil

patrykkurzejago  32 10809 Cool? (+24)
📢 Listen (AI):

TL;DR

  • A simple musical Tesla coil was rebuilt from a single-transistor design so it could play tunes instead of only lighting fluorescent lamps.
  • The primary was changed to 5 turns, the secondary to about 1000 turns on a 5 cm PVC pipe, and a yellow LED plus a second transistor improved robustness.
  • The coil ran at 30 V and about 30 W using a LRS-100-36 supply reduced from 36 V.
  • An ESP8266 generated square waves to drive a 2N7000, shorting the transistor bases to ground and making the coil play simple tunes.
  • The circuit was only temporarily soldered on a universal board, and the author notes the schematic still has shortcomings.
Generated by the language model.

Hello
I've always wanted to build a musical tesla coil, but I was overwhelmed by the complexity of the design schematics found on the internet. Some time ago I found a schematic of a single transistor based coil and decided to improve it so that it could play sound. I decided to share the project so that others could make such a simple coil.

Here is the diagram I found:


The effect of her work was poor, no spark, only the fluorescent lamps were getting closer to light. I increased the number of windings to 5 on the primary (I made them from a 1.5 mm? cable ;) and around 1000 on the primary. There are 25 cm of windings with 0.25 mm diameter wire on a 5 cm diameter PVC pipe. After this procedure, the spark was immediately bigger, but a new problem appeared, the transistor did not last a long time and soon it fell. I experimented with many transistors and got a good result using the BD239. The transistor burnout problem was rare, but I wanted to eliminate it completely. It was helpful to replace the semiconductor diode connected in reverse to the base of the transistor with the LED - and attention, specifically the yellow one - and the insertion of a second transistor. Due to the fact that these transistors have more power, I increased the voltage of the system to 30 V, increasing the power of the coil to about 30 W (value read from the service power supply). I suspect that the tension can be increased a little more.
The next step was to make the coil play. I used ESP8266 (it was handy, and I had code for arduino on the disk that plays simple tunes). ESP generates a square wave with a given frequency, which controls the unipolar transistor (2N7000), short-circuiting the bases of the transistors controlling the coil to ground. In this way, we managed to get a nice and loud sound from the coil.

Finally, I made a casing from some IKEA box that happened to be at home. As a power supply I used a 36 V impulse power supply (LRS-100-36, I had it from some electro-scrap that I got as a "gift" for parts), I reduced its voltage to 30 V with the potentiometer on it.

I know that the scheme has some shortcomings, it was just created as much as my knowledge allowed, there is definitely something to be improved in it. It is as follows:


I soldered everything temporarily on a universal board, there are plans to make a printed circuit board.
Unfortunately, I didn't take photos during the creation, but I put in those with the coil ready.





1st movie:



Program code for ESP8266 (written in Arduino IDE) I attach to the topic. ESP provides a page on the local network where you can select a track to play. I add a second program that works immediately after startup (it does not connect via Wi-Fi, only plays one track in a circle).
Regards

PS.
It should be noted that the Tesla transformer generates a high voltage of several kilovolts, so be especially careful when working!
Attachments:
  • program.zip (12.82 KB) You must be logged in to download this attachment.
  • program_bez_wifi.zip (4.1 KB) You must be logged in to download this attachment.
About Author
patrykkurzejago wrote 8 posts with rating 33 . Been with us since 2022 year.

Comments

bambus94 15 Aug 2022 11:55

Have you checked the frequency of work? As it goes in hundreds of kHz or even MHz, you could try something similar to a class D amplifier and then even normal mp3 would play. Kind of like my HFVTTC link... [Read more]

patrykkurzejago 15 Aug 2022 12:06

Unfortunately, I do not have an oscilloscope or even a multimeter with frequency measurement. I also do not know how to check or calculate this frequency in another way :( Does such a coil emit radio frequency... [Read more]

bambus94 15 Aug 2022 12:11

Yes, the Tesla coil is essentially a radio transmitter. A well-tuned one should generate radio waves with a length equal to the length of the wire on the secondary winding. A less tuned one will transmit... [Read more]

patrykkurzejago 15 Aug 2022 12:19

Then let me look for that frequency in a moment. I have an SDR with the R820T2 chip and it reaches up to 1864 MHz. [Read more]

bambus94 15 Aug 2022 12:23

Rather, you have to look at the lower range, that is, a single MHz. Even if you catch something, you cannot be 100% sure of this frequency, because it may be one of the next harmonics. [Read more]

patrykkurzejago 15 Aug 2022 12:50

I rummaged through the manual and it turned out that my SDR ranges from 0.1 MHz to 1.7 GHz (it was enough to change the settings). The coil emits a strong wave at 1.037 MHz, if it's playing a song,... [Read more]

Janusz_kk 15 Aug 2022 23:28

Not true, the resonant circuit is inductance and capacitance, not the length of the wire. [Read more]

bambus94 15 Aug 2022 23:32

Yes, but a well-calculated Tesla coil has a resonant frequency equal to that of a wire-length wavelength on the secondary winding. This is why toruses are used to vary the resonant frequency to tune this... [Read more]

Janusz_kk 16 Aug 2022 08:30

equal to inductance and inter-winding capacitance. that someone also wanted to look for such a relationship :) normally no one is concerned with such dependencies. Anyway, the author has too high... [Read more]

bambus94 16 Aug 2022 08:55

Well, someone found it and everyone who wants to have good results applies to this dependence. [Read more]

zulugula 16 Aug 2022 10:03

Every radio amateur will tell you that the best amplifier is the antenna, and by choosing its dimensions and the length of the wire, you tune it. And wasn't the length of the wire 1/4 or 3/4 of the... [Read more]

bambus94 16 Aug 2022 10:08

From what I remember it was 1 or 1/2 wavelength. I have to look for an email from a man who did something about 40 cm of arc on VTTV on GU81. [Read more]

patrykkurzejago 16 Aug 2022 11:15

If someone has a pacemaker, a heart defect, a disturbed potassium-sodium balance, or some other heart conductivity problem, such a coil is very dangerous. At best, a healthy person can burn, but I still... [Read more]

Janusz_kk 16 Aug 2022 18:59

Epidermis, read how much it is at 1 MHz. But the coil is not an antenna. [Read more]

rosomak19 16 Aug 2022 22:16

I would be more tempted to use the term Tesla's generator or transformer, not "coil"; Second thing, the Tesla generator is tuned with the primary winding, not the secondary one. The third thing, the... [Read more]

patrykkurzejago 16 Aug 2022 22:53

What about the electromagnetic field? It seems to me that the starter is shielded in some way and the risk is minimal, but I wonder if anyone knows anything more about it. I even remember in high school,... [Read more]

zulugula 17 Aug 2022 07:35

As @ rosomak19 wrote, it can be considered a quarter-wave plate, but for proper resonance it needs to be tuned with a suitable capacitor, e.g. a torus. Finally, tune the resonance in the primary circuit. ... [Read more]

rosomak19 17 Aug 2022 12:08

I do not know about your musical Tesla, but with my classic one, the field was so strong that the fluorescent lamp in my hand shone from a long distance. I do not know about the starter, but the phone... [Read more]

bambus94 17 Aug 2022 12:12

I would be more afraid of nitrogen oxides than ozone. It was one of the first industrial methods of producing nitrogen oxides and then nitric acid. An electric arc in the air, later refined by the addition... [Read more]

FAQ

TL;DR: A 30 W, 1.037 MHz musical Tesla coil proves that “Tesla coil is essentially a radio transmitter” [Elektroda, bambus94, post #20144574] while remaining buildable with two BD239 transistors [Elektroda, patrykkurzejago, post #20144206] Safe handling and proper tuning are critical. Why it matters: Even small coils can burn skin, jam Wi-Fi and broadcast unintended RF.

Quick Facts

• Resonant frequency: 1.037 MHz detected with RTL-SDR [Elektroda, patrykkurzejago, post #20144646] • Supply: 30 V DC from trimmed 36 V SMPS [Elektroda, patrykkurzejago, post #20144206] • Output power: ≈30 W at 1 A [Elektroda, patrykkurzejago, post #20144206] • Windings: Primary 5 turns / 1.5 mm; Secondary ≈1000 turns / 0.25 mm [Elektroda, patrykkurzejago, post #20144206] • SDR coverage 0.1 MHz–1.7 GHz with R820T2 tuner [Elektroda, patrykkurzejago, post #20144595]

1. What schematic was the starting point for this build?

The author began with a single-transistor SSTC diagram found online, then redrew it to use two BD239 NPN transistors, an LED across the bases and an ESP8266 audio gate [Elektroda, patrykkurzejago, post #20144206]

2. Which modifications increased spark length?

Increasing primary turns from 3 → 5, raising supply to 30 V and adding ~1000 secondary turns boosted spark visibility and lamp excitation [Elektroda, patrykkurzejago, post #20144206]

4. How does the coil play music?

An ESP8266 outputs a square-wave melody; a 2N7000 FET grounds the transistor bases rhythmically, amplitude-modulating the RF and producing audible plasma tones [Elektroda, patrykkurzejago, post #20144206]

5. How can I measure resonance without an oscilloscope?

Use an RTL-SDR dongle: the coil radiates; tune from 0.5–5 MHz until you see a strong peak (1.037 MHz in this project) [Elektroda, patrykkurzejago, post #20144646] Alternatively, compute f ≈ 1 / (2π√LC) with online calculators when L and C are known.

7. What effects on nearby electronics were observed?

Google Nest speakers in the next room randomly changed volume; fluorescent lamps lit from ≥30 cm; smart-band and phone survived close passes [Elektroda, patrykkurzejago, post #20146027]

8. Can I raise the supply voltage for bigger arcs?

Yes, but transistor heating rises with frequency. Users report safe operation up to 30 V; beyond that, core losses and transistor SOA limit apply [Elektroda, patrykkurzejago, post #20144206]

9. How do I tune a Tesla coil efficiently?

  1. Wind secondary first and calculate its self-resonance. 2. Start with extra primary turns. 3. Slide an alligator clip to vary inductance until spark length maximises and transistor heat minimises [Elektroda, rosomak19, post #20147034]

10. What is the wavelength-equals-wire-length debate?

Some hobbyists aim for secondary wire length ≈ ½ λ; others focus solely on LC resonance. Experts note performance depends on L and C, not wire length alone [Elektroda, Janusz_kk, post #20145799]

11. Which gases form around the arc?

Ozone and nitrogen oxides form; litmus paper turned red near the arc, confirming NOx presence [Elektroda, patrykkurzejago, post #20147860] Prolonged exposure above 0.1 ppm ozone irritates lungs [EPA, 2023].

12. What primary and secondary wires work best?

Use 1–2 mm² copper for the primary to handle currents; 0.2–0.3 mm enamelled magnet wire suits a 25 cm secondary, giving ≈1000 turns [Elektroda, patrykkurzejago, post #20144206]

13. How can I reduce transistor heating?

Lower resonance by adding a toroid or increasing secondary height; reduced frequency cuts switching losses. Ensure a heat-sink keeps case <70 °C [Elektroda, Janusz_kk, post #20145799]

14. What myths surround the skin effect?

High-frequency current still penetrates tissues because the body is ionic and inhomogeneous; skin effect does not guarantee safety, especially in DRSSTC class where peak discharge current can reach amps [Elektroda, KJ, post #20162812]

15. 3-step: Measure frequency with RTL-SDR

  1. Power the coil a short distance from the antenna.
  2. Open SDR software, set range 0.5–5 MHz, gain low.
  3. Identify the strongest peak (e.g., 1.037 MHz) and note harmonics. [Elektroda, patrykkurzejago, post #20144646]
Generated by the language model.
%}