Weekend Projects: My Ongoing Plasma Speaker (singing Arc) Project.




This instructable will show you how to make a simple plasma speaker. A plasma speaker (or singing arc) is a device that generates a high voltage electrical discharge when we modulate to emit audible music. All the sound comes from the arc!


The advantages of using an arc to create audio is that it makes it possible (with the correct driving circuitry) to create really high fidelity tweeters. Conventional speakers use directional diaphragms to push and pull air. However, when they need to move back and forth as much as 22 thousand times per second, it can become quite challenging. They have quite a bit of mass and it is difficult to get them to oscillate that fast. This leads to large power consumption, distorted audio, and overall poor performance. The arc generated by this circuit does not rely on moving massive things to create audio, and instead varies the "thickness" and size of the arc, which displaces air. We can similarly control the amount of plasma being generated to control the temperature of the air, therefore controlling the expansion and contraction of said air, and then these oscillations can ripple out and away to your ears as acoustic waves! Also, it is really cool! Why not?? ;)

How It Works:

In the 555 version, we will use the magic of PWM to create audible sparks. An Astable 555 timer circuit is the heart of this design. It makes a high frequency square wave of approximately 48,000 Hz and the pulse-width is controlled by the audio source. This signal that controls a high power transistor which causes the current in the primary of the transformer to change rapidly.The extremely fast switching of the MOSFET is what drives the flyback, and the flyback output spark will essentially get thicker and thinner as a result.

We can also accomplish audio modulation by injecting an audio signal into the power supply rail, causing the output power of the flyback to change relative to the music! This method is used in the ZVS version flyback driver.

So what is PWM?

Well, imagine it sort of like flipping a light switch "ON" and "OFF". (only 50,000 times a second!) If you leave the switch "ON" longer than you leave it "OFF", the light will be brighter than halfway. and if it is left "OFF" for longer than it is "ON", the light would be dimmer. That's PWM in a nutshell. How we just need a way to control how long each pulse is "ON" in proportion to "OFF" based on the signal from a audio source, and now your light will flicker to the music! Now obviously, we cannot use a physical switch, it is too slow, and it needs to be controlled electronically!

Why 2 versions of this instructable?

Well, it is a really, REALLY long story. My original 555 circuit using cheap MOSFETs never really worked. It worked well enough for me to publish it, but it would always fail. Eventually I gave up on the unreliable circuit and built a much easier ZVS circuit, which did work a LOT more reliably, but the audio modulation was done poorly, by simply injecting an audio signal into the series ballast of the flyback transformer. This worked well enough but I eventually returned to the 555 version, and this time with better transistors, more knowledge of electronics, and a tank capacitor to smooth out reactive power from the coil, I got HUGE performance! For the longest time I had 2 separate copies of the instructable that was the same except one was a copy of the other but only referred to the ZVS circuit. I merged the 2 together into this one long instructable. Sorry for the confusion.

Which driver is better?

This is really a matter of opinion. The ZVS version is very robust, almost never fails, and is very efficient. The arcs produced, however, are relatively short and very hot compared to the input voltage, and likewise the sound is very quiet. You would not be able to hear it in a noise environment. The sound quality though can be very good depending on the amplifier used to modulate the power supply, and the amount of ripple/noise on the power supply.

On the other hand, the 555 can produce much longer arcs, but the since it is only a 555, we cannot achieve a true PWM modulation without also affecting frequency, and the sound quality is comparable to a telephone music, because the frequency of PWM is within audible range. The PWM frequency is comparable to a sampling rate, and a 15KHz sampling rate is really bad. You can of course increase the frequency of the driver, but you will lose performance as it becomes further and further out of resonance with the flyback transformer. It is a tradeoff. This circuit is better for impressing people with long, epic arcs. However, although sound quality is not the best in this driver, it can get a bit louder than the ZVS version and does not put nearly as much stress of the amplifier as the ZVS version!

Step 1: Disclaimers and Dangers!

If you have not messed around with high voltages before, Please don't attempt to make this circuit. These drivers I built are easily capable of delivering 25,000 volts (25 KV) or more depending on the input voltage, at 5-50 mA of current. This output current can cause serious problems such as loss of muscular control, involuntary spasms, heart fibrillation, cardiac arrest, death, and perhaps worst of all, involuntary self-urination. You definitely would not want any of that!

Step 2: Electrical Parts You'll Need:


  1. power supply (SLA batteries work, or a good beefy 12-24V linear power-supply)
  2. LARGE heatsink (your MOSFET will let out the magic smoke without it.)
  3. computer fan (to provide active cooling in case the heatsink is not enough.)

~~~~ELECTRONIC COMPONENTS (for 555 version):~~~~~

  1. flyback transformer (you can get them from old TV's or online.)
  2. Many high power MOSFETs (buy these online in bulk or you can salvage them)
  3. Large film/MKP capacitor (value is experimental, about ~0.1µF ish.)
  4. 10 or more 555 timers (Plan on buying these in bulk. They are easily killed.)
  5. 8-pin IC holder (if you plan to make the circuit permanent, makes it easy to replace the 555.)
  6. 2x 10K resistors (these worked well for me. but experimentation may be needed.)
  7. 1nF capacitor (I also used in the video 2 of these in series, to get a total of 500pf.)
  8. 4x 1N4007 or 1N4148 clamping diodes (protects the 555 from excessive voltage spikes.)
  9. Large value electrolytic capacitor (around 1µF --- 470µF is fine.)
  10. BIG high value >1000µF electrolytic filtering capacitor (bigger & lower ESR = better.)
  11. Good, thick copper wire (>14 AWG)
  12. indicator light (optional)
  13. breadboard/protoboard

~~~~ELECTRONIC COMPONENTS (for ZVS driver version):~~~~~

  1. flyback transformer(you can get them from old TV's or online.)
  2. At least 2 high power MOSFETs (order spares though. I recommend at least 6)
  3. Large film/MKP capacitor (value is experimental, about ~0.1µF ish.)
  4. Fast switching diodes (I am using cheap 1N4007's)
  5. 12-15V zener diodes (I am using 1N4742A's)
  6. BIG high value (>1000µF) electrolytic filtering capacitor (The bigger, and lower the ESR, the better.)
  7. Computer fan (recommended for cooling the tank capacitor and flyback)
  8. Thick copper wire (>14 AWG)
  9. Indicator lamp (optional)
  10. breadboard/protoboard

Step 3: Optional (Recommended) Case and Materials:

  • custom PCB
  • duct tape and/or electrical tape (required)
  • epoxy, hot glue, rubber cement, or similar.
  • a housing for everything (I am making a custom wooden box as shown later in the instructable)
    • quarter cut white oak half inch thick, 5 inches wide, and at least 3 feet (with some extra)
    • small brass wood screws
    • small screws with matching nut
    • wood finish
    • wood-working equipment
    • wood glue
    • belt sander, random orbit sander, or good old handheld sandpaper
    • jointer
    • plainer
    • router with 3/8ths inch rabbiting bit
    • table saw
    • box joint jig
    • drill w/ small drill but for the 8 wood screws and metal
    • sheet of metal (10 X 5 inch)
    • metalworking equipment
    • Something to cut it to size. I used a rotary tool but i did not get a good straight cut, so I cut a bit bigger and ground it down with a large stationary belt sander
    • Drill press with various bits ranging from 1/16 inch to over 1/2 inch
    • Sand-blaster
    • Clear-coat

Step 4: Tools Required

  • Soldering tools (if you plan on making the circuit on protoboard):
    • Soldering iron & Solder
    • Desoldering pump or solder wick (for fixing mistakes)
    • Sponge (for cleaning the tip of the iron)
    • Flux solder paste
  • Wire strippers
  • Needle nose pliers
  • diagonal cutting pliers
  • Helping hands
  • Screw-driver
  • Tape
  • Tape measure

Step 5: 555 Circuit Diagram and Prototyping / Troubleshooting

The circuitry is very simple. it is just your standard astable 555 with an audio input connected to pin 5 as you see in the schematic. The 555 drives a high power MOSFET, which drives a resonant tank circuit and the primary of the flyback. Gather all your parts and build your circuit according to the schematic on a breadboard first.


If you are experiencing difficulty with making the circuit work, check and recheck your connections. Often, components will burn out, but not always lead to magic smoke. Here is a list of what to check:

  • Thou Shalt check voltages! Test your power supply output with a multimeter while the circuit is turned ON, and while it is OFF or disconnected. Ideally, have another meter measure current simultaneously to determine if it is a shorting the power to ground.
  • Check the 555! There are many ways to do this, the easiest and most effective way to do this is to make sure the wiring of it is correct (it is easy to mess it up) and with pin 3 disconnected from the gate of the MOSFET, check the signal with an oscilloscope. If you do not have one, I recommend you get one, you can find cheap used one's on craigslist and eBay. Alternatively, you can build a 555 buzzer circuit separately and insert the 555 into it to verify it is functioning.
  • Check the MOSFET! It is quite common to have them die. It is just too easy to exceed the maximum Vgs and Vds ratings. Often, the failure mode is a internal short circuit (and shorted gate.) From my experience, if it dies in this failure mode, it will stress your power supply or batteries, and the MOSFET may then catastrophically fail, possibly exploding! Also, when this fails, the shorted gate will also take the 555 down with it. To test this, replace the flyback transformer with a lamp, and connect gate to +Vcc, and the lamp should light up full brightness. When the gate is grounded, the lamp should go fully out. If not, replace the MOSFET.

Step 6: ZVS Circuit Diagram and Prototyping / Troubleshooting

One of the most popular circuit for flyback drivers is the ZVS flyback driver. This circuit is essentially a Royer oscillator, that uses two MOSFETS to drive an LC tank circuit (C1 and L1). This circuit relies on resonance rather than the flyback mode of operation. Think of resonance like pushing some fat kid on a swing, and the transistors like 2 huge bullies on either side pushing that poor fatso kid back and forth real hard at just the right time, and as they do so he gains momentum (like the voltage). causing the voltage to rise almost 4 times the input voltage, unloaded!


DO NOT build this circuit on a breadboard, YOU WILL MELT IT!!!


I'm also going to attempt to etch my own PCB, with nice, thick traces to handle the current. I have heard people built this driver on a PCB say they had to use external jumpers because they blew the traces off. Also, the PCB is fairly compact and should easily fit into my box I made. I did try to make a custom PCB my hand before, but failed.

UPDATE to creating my own PCBs: I have not managed to make my own, the muriatic acid does not dissolve the copper fast enough, and eventually the toner starts to peel off too. I have since learned that you are supposed to add some hydrogen peroxide which works as a catalyst that speeds of the reaction with the copper. I have given up on making my own PCB and just wired it dead-bug style. If you want to get a custom PCB, I recommend sending off these things I recently learned about called germer files, which are the files used to make these PCB cad drawing properly. I used MSpaint which is not well suited for that at all.

Step 7: Planning to Make This Permanent?

1) Get your PCB ready and place all the parts on, solder them in place.

2) Flip the board over and connect all the parts together.(tip: I go about this by bending over the leads i want connected towards each other so they touch and soldering them in place) remember, when you flip the board around, you are looking at a flipped view of the board so be double sure you are looking at the right lead on the underside.

Step 8: Testing the Finished 555 Circuit With a Small Power Supply

My original 555 circuit (deleted due to its unreliability) With a 0.8A 12V wall wart a get about 1 cm ark and can gets very thin streamers as long as 0.8 inch. My updated 555 circuit achieves MUCH better output at just 11.8V! You may find that a preamp is required to make the output louder, but be careful. It is easy to kill the 555 by driving too much power into pin 5. I added some protection clamping diodes to prevent voltage levels from getting too high or too low.

My ZVS circuit works simalarly well,

Step 9: Building the Wood Case

lay out the scheme for the 1/2 inch wood. the box when finished should be 10 x 5.5 x 4.5 inches. There are slits 1/4 deep and 1/4 inch wide to fit the control panel and backside.

The first thing that needs to be done is polishing up the quartercut board. To do this, first run the board though a jointer on all 4 sides. then, run it through the planer. Also, get the board down to 4 1/2 inch thickness. Then cut it to size with a saw. (Two 10 inch long pieces and two 5 1/2 inch long pieces.)

Now make box joints as shown in the first graphic below. (note: I screwed up on the 10 by 4 1/2 piece, there should be box joints on both ends, not just one.)

After I finished all the pieces, glue them together. The pieces may need to be hammered together. After it has finished gluing, fill in all the holes and spots left over with wood putty. Then start the sanding, First with a belt sander, then a random orbit sander, then with handheld sandpaper. You should end up with something that looks like picture 6.

then use a router to cut the slit or notch (whatever you want to call it) into the open sides as shown in picture 7, 8 and 9.

Step 10: The Front Plate and Back Cover

For the front control plate we will need a sheet of steel to make a front control plate and a back cover plate. So by the appropriate sized sheet. It may be good to buy one that is larger than what you need for a backup in case you messed up on the first one, but I went cheap and bought just enough for 2. Try to pick up one with minimal scratches and dings on it. Spots are OK if they can be wiped off.

I used a dremel to cut out the plates, then used a stationary belt sander to finish off the edges. Then figure out how you want your layout. I show my layout in the first picture. After you made the measurements and you are sure of where everything is going, mark the spots and drill them out.Then finish off with filing the leftover trash.

Step 11:



    • Sensors Contest

      Sensors Contest
    • 1 Hour Challenge

      1 Hour Challenge
    • Pets Challenge

      Pets Challenge

    42 Discussions


    3 years ago

    Hey, do you think using a CFL driver instead of some components will work? How would you go about injecting the audio signal?

    1 reply

    Reply 3 years ago

    I don't know enough about CFL drivers to give you any advice on using that as a flyback driver. I have seen other instructables doing that, you may want to ask them about audio modulation.

    I would reverse engineer the oscillator used in the CFL driver, improve it by replacing the cheap crappy components with better ones in my junk drawers, and see if I can find a convenient node to inject a audio signal to modify either the frequency or amplitude of the oscillator that ultimately drives the flyback.

    If what you mean is to use components from the CFL driver to build this circuit, then my advice is to not do that. 90% of the time CFLs fail because the ballast is poorly and cheaply designed (with cheap parts) and overheats.


    Reply 6 years ago on Introduction

    it has the same danger level as walking down a ladder.

    if you don't know what you are doing, [not experienced with basic electricity and HV] it is dangerous and can seriously hurt you. however, if you use your common sense, play with van-de-graph generators and just get familiar w/ high voltage, you should be fine.

    just attempting to walk on a ladder for the first time and triping and falling, breaking a bone or your neck. in the same way, this can hurt you by: Ventricular fibrillation, cardiac arrest, RF burns, shocks, boo-boos, etc read http://en.wikipedia.org/wiki/High_voltage


    Reply 6 years ago on Introduction

    Boo-boos are the worst! :D
    Nice project. Gonna have to build me one.
    How would you protect the setup from cats (or rather the cats from it)? Would a grounded metal cage do it?


    Reply 6 years ago on Introduction

    I guess you could add a kill switch. And not have cats in the workshop. Other than that I don't know


    Reply 6 years ago on Introduction

    I live in a tiny apartment with my wife and four cats (and soon toddler). I can only dream of having a workshop.
    My reasoning for a metal cage, as you've probably guessed, is a kind of a Faraday cage. I wouldn't wanna touch it and test it, but it *should* keep you out of harms way.


    Reply 3 years ago

    know this is old but you could use a glass dome with a few holes in the top to allow air and sound to exchange while also keeping people and animals from touching it.


    Reply 6 years ago on Introduction

    I would just keep it out of reach when your not using it. remember, this is more of a toy rather than something to replace computer speakers. (I wouldn't trust the circuit, the kickback might go far enough backwards to fry whatever you have it plugged into, such as a sony walk-man or computer.) (you wouldn't want to fry the computers sound card). or a Faraday cage, I would use a wire grid bent into a cylinder shape, glued over the top of the output, and cap it off with a circle cut out of the same metal grid.


    Reply 6 years ago on Introduction

    Oh ok I see your situation. I dont know. there are other designs online for more pro-looking versions. I would just keep it out of reach without anyone around.


    Instead of grounding it to mains ground hook it up to the flybacks HV negative. To find the HV negative I recommend this instructable: How to find the primary and the secondary coil of a Flyback transformer

    PS keep it away from cats and babies :)

    Pete ChristianR

    3 years ago

    Hey how do I make my arcs longer?

    I used 9v to power 555 and 12v to power the coil

    8 replies
    -max-Pete ChristianR

    Reply 3 years ago

    First, try reversing the polarity of the primary coil of the flyback transformer. Next, try increasing or decreasing the capacitor in series with the flyback primary, next, try different frequencies on the 555, if you are still not satisfied, upgrade to better MOSFETs, add more filter capacitors on the power rails, and/or increase the voltage. Really unless you are aiming to get better results than me, increasing the voltage is not necessary. Just make sure your power supply can deliver enough current and is low impedance (which just means the voltage does not drop when drawing lots of current.)

    Pete ChristianR-max-

    Reply 3 years ago

    I already tried interchanging the coil from the transformer. but still only less an inch arch. I am sing the irfp260(wanted to build you slayer exciter first but didn't find ds2006 so I decided to switch to this).

    What capacitor do you mean? the film capacitor? or the electrolytic one?

    I am using a 12v power from my routers power adapter. and 9v from my piano adapter.

    My aim is to just get at least a 1.5 inch arc. that's enough.

    thanks for the fast reply btw. I do really need this now though it's our ITEL project and it's is our final day now.

    Thanks :)

    -max-Pete ChristianR

    Reply 3 years ago

    Ok great! I think your problem is really just the power supplies you are using. Switch mode power supplies are crap. They have lots of "safety" features and stuff to keep us from doing fun stuff with them, like foldback current limiting. (when the power supply voltage & current drop to almost nothing when even a small overload is detected) You need to get some good, huge, beefy unregulated power supplies, or hack the switchmode power supplys.

    Pete ChristianR-max-

    Reply 3 years ago

    Oh ok thanks.

    Sadly we broke it.

    I tried changing the electrolytic capacitor to a 1000uf one. I think the problem is with the power supply but when i got home i tried testing it back with my router and it worked.

    btw i forgot to say that im using 50v 3300uf capacitor last time and now becuase that's the only one we have and for the film capacitor I don't really know what we used because we just used the one that we found on an old tv sadly we only found one. it's slightly bigger than the one you used and it not round like yours.

    any suggestion on how to fix it?

    thanks for the help by the way.

    -max-Pete ChristianR

    Reply 3 years ago

    The value of that film capacitor is pretty critical, but needs experimentation to find the perfect value. I would try to remove it for now and see what happens. Most likely you might have popped the MOSFET. That's why I recommend buying them in packs of 10 or more.

    Pete ChristianR-max-

    Reply 3 years ago

    sadly i only bought one. I bought it from the next city 3 hours from mine. I did not hear any popping sound with the circuit though.

    do you know how to test the irfp260 with multimeter?

    I didn't know that my mosfet could break with 12v power. I Hook it up in an heat sink and when I try to test it in the past(when it was still working) I plug it off not longer than 1-2min and I always touch my heat sink if it gets really hot but it never did.

    irfp260 is kinda pricy and at that time I dont know which one to buy most except for the 555 . I tried touching the positive wire with the poisitve end of my capacitor like in your vid. It sparks sometimes.

    -max-Pete ChristianR

    Reply 3 years ago

    You cannot really test MOSFETS with a multimeter, but you can verify that it is dead if you think you killed it. The resistance between the gate and the source or drain should be nearly infinite. To test that, set your multimeter to the highest resistance range (or just the resistance range if there is only one) and test the resistance between the gate and the other 2 pins. There should be no movement on the abalig meter and a digital meter should either read infinite or overload or open circuit.

    Pete ChristianR-max-

    Reply 3 years ago

    By the way does using a 12 volt in power from an computers power supply be better?

    I have an atx 500w

    it has some 12 V and 0.8A.

    will this work better?