The Strobe Light Effect (Levitating Water)




Wouldn’t it be nice to see the water actually levitating in mid air with naked eyes? It is actually possible:



In my previous video I showed how matching the frame rate of the camera and vibration frequency of water can cause the camera to record the water levitating. But that wasn’t visible to the eye. But using a strobe light it is possible to create the illusion that the water is levitating, visible to the eye. A proper strobe light like LED can light up the water for a very short period of time every period. The concept of this levitation is similar to what I explained in my previous post.

To make this work, the strobe light should be on for only a very short period of time every period. The shorter the on-time is, the sharper and less fuzzy the water drops will look like. Because the light would sample a smaller period of time in which the water moves less. In my design, the on-time of the light is less than 1mS (around 0.7mS) every period.
I have already told you how to make a vibrator using a speaker and how to drive it in a previous post. So now all you need to know is how to make and drive a strobe light. Below is the schematic of the circuit.

This driver provides a fixed width pulse.

This driver provides a fixed width pulse.

And here’s the part list of the components:

  • LED: Everlight Elec., 334-15/X1C5-1QSA, Or any white LED, the brighter the better
  • Comparator: Any >5V supply comparator with push-pull output would do
  • Transistor: Fairchild, FQP30N06, or similar power FET capable of >10A
  • Regulator: MCC, 78L05BP or similar
  • All other parts are generic

The circuit uses an input generated from a signal generator, or in my case the audio jack of a smart phone with a signal generator app similar to what I explained in my previous post.

There are a few stages in the design. The first one including U1A is a simple comparator. C1 gets rid of the DC component of the signal coming from the signal generator and then that signal, riding on 2.5V (due to R1, R4 divider), will be compared to a DC 2.5V (from R2, R3 divider) and the output of U1A will be a 0 to 5V square wave. This is done do make a uniform square wave signal because the phone may not be able to generate a good and consistent square wave.



The next step is the R-C circuit of R5, C2 that is a high pass filter. The intention of this filter is to make the signal width much thinner, so that it is only high for a very short period of time. Adjusting the R-C values would affect the width of the pulse. R6 helps boost the high side current drive of U1A. Now to make this signal shaped as a proper digital wave, it passes through a second comparator stage. Below is how the signals look like passing through the stages.

Waveforms of every stage

Waveforms of every stage

Now the last stage is a power FET transistor. A high at the FET’s gate will turn it on like a switch. This FET needs to be strong enough to be able to handle the high current running through the LEDs. This driving circuit is not very power efficient. There are much more efficient circuits, but this should do the trick for this experiment. Here are some notes in designing the circuit.

  • If the light is on for 1mS and the maximum frequency provided is 100Hz, then the period is 10mS and the light is on 10% of the time.
  • This means that we can actually drive the LED 10 times its nominal current rating. (make sure to read the datasheet for recommended maximum pulse current).
  • The higher current will provide more light. We need it as our LED is only on 10% of the time at best.
  • The white LEDs usually have 3V to 4V across them when on.
  • The series resistor for every LED is calculated to provide maximum current required. The average current will be 10% the maximum at 100Hz. The lower the frequency, there will be less current draw as the amount of time the light is on is fixed while the period increases.
  • The FET transistor must be rated to be able to easily handle this average current. In my case the average current would be around 1A max (~10A pulse currents).

Here you have it. Let me see if you can make a fancy water fixture. There is one problem with this fixture that is the room needs to be dark, or at least the water needs to be illuminated only by the strobe light and otherwise the rest of the water that should normally be dark and invisible, would light up like I showed in the video. But even is the room is not totally dark, it might be interesting to see a constant flow of water, within which some water drops are seemingly traveling differently.

And… don’t pour water on your electronics, especially if you decide to run it from 110V or 220V rather than the drill batteries!

[EDIT] Additional Notes:

There might be a little confusion here I like to clarify. VBAT-12V is not -12V, the “-” is just a spacer between VBAT and 12V. It is actually a +12V and the other side of battery is ground. But then the battery has a center contact at +6V. Naming the voltage of the circuit nodes is arbitrary. The fact is that there are three nodes, one with the lowest voltage, the other one 6V above the lowest and the other one 12V above the lowest. Now you can call them 0V, 6V and 12V, or like my other video -6V, 0V and +6V. It is how you design with them that gives them a meaning.

The regulator is a device that takes an input voltage, I called VCC here, which in my case is 5V. The regulator can take an input voltage range, for example 7V to 16V and generate a fixed output of 5V.  I haven’t shown the regulator properly in the diagram. It is in general a 3 pin device with an input, an output and a ground connection. So the input in this case connects to VBAT-12V and generates a fixed 5V called VCC. A fixed voltage is used to ensure consistent behavior of the circuit, as well as making sure that the circuit wouldn’t see an unwanted high voltage.

VIN can be the positive or negative of the audio jack. Since there is no initial ground reference between the two devices (cell phone and the circuit) and there is no DC voltage between the output lines of the cell phone, the two lines can be connected anyway makes you happier. But normally to make it rational, the negative output is connected to ground and the positive is connected to VIN.



106 thoughts on “The Strobe Light Effect (Levitating Water)

  1. So I am a magician and I plan to use this on a stage performance. I need it to be big enough but I will need some advice on how big and how to make the water spouts. I am gonna have some help with this from my family but I need some ideas on some things. Please help me out and reply on how to make the water spout and how much it will aproxametly cost.

    Thanks so much

  2. Pingback: Phone advertisements – Animation Studio Research

    • sorry, it’s a question
      Are there another ways to sync frequency of water and strobe light without speakers(audio freq) ?.

  3. Dear Mehdi S.
    i really want to do this for decorate
    i can control sound frequency
    but i dont know anything about curcuit
    is there any ready-to-use LED strobe light that we can adjust frequency about 40-100 hz i can buy ?

  4. i would like to get your advice. I am doing the same project as you using arduino that control the speed of the strobing light but i am facing problem. For the water, i use a 12V water pump the pump the water but i couldn’t match the frequency of water with the frequency of the strobe light. Here is my question, can I match these two frequency using the following method?
    -I will tie the water pipe to a speaker playing sound of 50Hz. Then I set the LED to strobe at 20ms ( which is 1/(20ms)= 50Hz.
    if i use the method above can i get the result as you get in the video or do you have any suggestion to me? please explain in more shallow words as i am not used to words which are more to technical terms.

  5. cual es la frecuencia que entra a VIN de la luz estroboscopica… y que frecuencia de sonido para la bosnia

  6. Hi,

    I have some enquires i would like to consult you on.

    1. As for the strobe light you are using, are you using some special form of light that is invisible to normal water? As shown in the video the stream of water is continuous and when you shone the normal light the stream lit up. Do we have to buy a special form of light source for the strobe? I am currently trying to do this using arduino to control the strobe light frequency. I tried using a tap as a water source and let the water droplets drip at a constant rate. I then change the frequency of the LEDs strobing to try to achieve the same effect. However, no matter how i changed the frequency, i could still see the whole stream of water. The room itself was dark and only the strobe lights were active.
    2. Is it important to change the frequency of water? Is the frequency the vibration of water molecules or is it the rate of water being dispensed through the tube. This is because i always thought the water flow must be droplets, as the illusion is done when the drop currently being illuminated is replaced by the incoming drop above it, and this happens during the strobing effect where we cant see as it is dark during that period of time. Resulting in levitating water.

    Would appreciate if you could give me advise on how to improve it! Thank you very much.

  7. Hey that’s a really interesting project you did up there.I wanted to apply this logic of stroboscopy to similar fast moving objects like a turbine or a rotor or anything that rotates at a very high frequency to inspect any faults or impaired movements.I know this is possible but wanted to ask you what modifications will i have to perform in my project and is it necessary to have an array of 99 LED’s,can we reduce the number of LED’s and still observe the effect efficiently.

    • In fact we have this strobe light at work that we use to measure PWM of a rotating motor. 99 LEDs is just for brightness, you can definitely go less.

      • I have designed an alternate circuit to get the same result using a microcontroller,in this case an ATMEGA32 which will work on higher frequency.But honestly i’m struggling with my coding for this IC as i’m not too sure what clock rate should i give to those LEDs to attain the desired output(eg:a stationary fan or a levitating water).And about the LEDs,we are using an array of 8 LEDs to achieve this.Is this amount enough or u would suggest to add some more to it?Your inputs or suggession will really help us in our project and i’ll be obliged.

  8. – Just wondering, can it be made up to 6 meter tall and still work as beautifull as you show it on your video??? Pls advise

  9. Hey I am inspired by your
    creati-witty I m newbie to all this but this whole project involves strobe light and the frequency generator right?? How do I sync them together the circuit shown above is for the strobe light ??
    Please don’t mind my stupid questions

    • I am using two phones as my signal generators, one to strobe the light and one to shake the water. So to sync them all you need is to set the frequencies the same.

  10. Hi, I was just wondering, isn’t a regular light bulb plugged into 220v 50Hz, turn on and off 50 times in every second? So it is possible to use it as the light source, isn’t it?

    • Hi, no you can’t use that. Two reasons, regular incandescent lights don’t turn fully off. Because it takes a while for them to cool down, which is how they turn off. For example if you turn your light off, you’ll notice the time it takes for them to turn off. So at 50hz, their brightness just oscillates at some level close to on level.
      The other reason is that in my other post about levitating water, the on period needs to be very short, in orders of faster than 1/500th of a second. This is to capture a more precise snapshot of water falling. If the period is too long, then the water drop travels a longer distance and makes the water blurry, exactly like capturing a photo, where a moving object is much sharper when the shutter speed is faster.

  11. Hey Mehdi!

    first off: great work and entertainment that you are performing here! that’s what the world need: kind people that know what they do and educate others at the same time. I am in the phase of planning that whole thing and I have one question: I plan on buying a cheap MP3 player for the speaker-signal. is it also possible to run the stroboscope with such a MP3 player? Or do I have to use a controller like this: http://www.circuitstoday.com/audio-oscillator-circuit-2 ?

    • one thing I forgot: you use a speaker with 6 ohm impedance. is it possible to use say 4 or 8 ohm instead with your amplifier circuit?

      • Hi LesSV,

        For the strobe lights you need a pulse width of 1mS to momentary turn the lights on, which is like 1kHz. And to have a proper pulse you should have 10 times that frequency, which is 10kHz. Audio players should go up to 20kHz and so I think you should be able to create that waveform. Although you can’t really tune the frequency with it, it is a pre-programed frequency.
        You can also make your own generator like the circuit you showed, and that gives you ability to tune the frequency.
        The amplifier can work with larger impedance easily. But with smaller ones, you have to reduce the output voltage so that you won’t over heat the transistors and burn them.
        Have fun!

  12. Very very cool project.
    You said make a fancy water fixture? How about…. well, what I can best describe as a floating vortex of spiraling water!
    I am trying to super size the idea, and install it as my everyday shower.
    Everything is absolutely working great, except for you are completely correct about that brightness problem when the duty cycle of the LED is only at around 10%.(I built an adjustable frequency oscillator from a 555timer

    My question is, how do I drive the LED at 10 times its nominal current rating? I don’t have a clue where to even start. You are speaking of the amperage, right?

    • The only way to drive an LED or any other component for that matter way above its rating, is to PWM the power so that the average current remains the same as rated or below. If your LED is rated for some current, the datasheet usually provides data on maximum pulse current, which is usually much higher than the rated current as it is a momentary pulse for a certain period. So it means you can load your LED above your rating for a limited time. If it is rated for 0.1A for example, then you can drive 1A through it if you turn it on only for 10% of the period (average 0.1A). Also that 10% time should be too long to over stress the component. It is true that it is on for 10% of the time, but if the 10% length is an hour, all hope is lost!

      • Thank you so much for the quick reply.
        I’m not exactly sure where to find the data sheet for THIS led, but it sure gives me another path to chase at my coolest shower in the world.

        Thanks again

  13. Hey you really had me wondering a lot with this how can you change the frequincy of the water how does that work.
    The circuit looks easy but is there anything i need to know about the water running maybe?

    • I am changing the frequency using my cellphone as a signal generator like I said. That is simple enough. Otherwise you need to make your own signal generator.

  14. Hi, I really enjoyed your video. My question is that will the levitating water effect change if you were to put a poly carbonate or acrylic shield in front of the lights? Thank you.

  15. goodnight:
    Mehdi fasino really levitate me this water project but interested in doing a simulation project. in which software you recommend me

  16. Dear mehdi

    I have seen in your youtube video that you are using a mobile app to control the frequency of the strobe light,
    whould you please let us know how that is possible

    • It’s simple. The audio jack can be used as a function generator signal output. The app basically makes audio signals from this jack, which I take and use in my circuit. If you search the app store you will find many function or signal generator apps.

      • thank you for your reply,
        I have downloaded the app. I also have connected the jack. but the problem is how I can connect the output to LED,s (I mean how to power LEDs and at the same time connect the the jack to them to function as a strobe light). Sorry if I ask silly questions, I am not much familiar with electronics, but this article was absolutely interesting to me, So I am trying to make it myself. I would appreciate if you could help me.

        • You will have to make the circuit I showed above in the article, that is if you want to do what I did. Unless you want to do something else. Vin would be the output from the phone jack.

  17. Hi Mehdi,
    when i simulate my circuit with TLV2372 instead of LM393 and remove resistor R8 from output of second comparator everything is looking good but problem is i cant get TLV2372 and i have to use one of LM….
    When i am using LM393 i get 2 pulses when sine goes down and up on tlv2372 i get only when sinewave goes acros axis from up. So if i have right i get frequency twice bigget than sinewave freq. ?
    Please help to fix it. I attatcht added screens of circuit and osciloscope.

    And sorry for my english 😛

    http://imageshack.us/photo/my-images/31/m9e5.png/

    http://imageshack.us/photo/my-images/17/mygx.jpg/

  18. Hi, one question. Can I use color (red/green) LEDs instead of white? Cause I thought it would be a nice effect to have “colored” levitating water ;).

  19. Thanks for very cool project. I’ll start to collect component for the project from tomorrow. My question is can I use regulated power supply combine with
    5-volts and 12-volts, or use separate power supplies. Waiting for your quick reply. Thanks

        • Could you please mention any part no of 5 volt Comparator?? Shop keeper doesn’t be familiar with by Comparator. And sorry, I have 5 volt (60 Ampere) chopper type power supply specially made for LED’s sign boards my son dealing with. This supply has multiple 5 volts out puts from same supply??
          Secondly, such a high Ampere does any damage to the circuit??
          Thanks,

          • I don’t have a part number in mind, but you can also use any rail to rail OpAmp as well. Doesn’t need to be a comparator. Something like TLV2372IP would work. If your 5V supply is 5V, it is ok to use.

  20. i want to do this experiment for a school science fair .. but i am only in 9th grade .. i can get the speakers and the rest done from the video link below but i want to do it in the dark so i have to use strobe light
    watch?v=uENITui5_jU
    and here is the problem .. i have to find an alternative to strobe light – because i cannot find the particular materials for it .. so could you suggest an alternative light source?

    let me phrase my question properly
    can i do it in a classroom with the lights on ? and will it show the same results?
    (dont worry about the water source)

  21. You forgot a Schottky diode parallel to R5. Your opamp will be dead soon due to negative voltage on its pin.
    Also you can put two LEDs in series to reduce power consumption in half. You may have to use smaller resistors in series to them.

    • The OpAmps (and ICs in general) usually have internal protecting diode and the negative pulse is very low current, so I am not much worried about the OpAmp, although it is a very good point for higher power pulses. Also it is a very good point to put LEDs in series, saves lots of power and reduces the number of required resistors at least to half. I didn’t think of that. Thanks for the comment.

  22. All went Well Mehdi, I did it the water Literally was Levitating only on 16 LED’s I did a test run only thing is if I want to keep this as a attraction I can not use a Mobile phone to Generate a Signal, Can I use anything else ? Any reference to make it standalone both the speaker and the LED’s and do I need the audio only for a sine wave ? I want to make it variable like a rotar switch so you can see the water rise and fall! Please suggest something!

      • Instead can I make a Mp3 of the sound wave and play it on a recorder chip, Now a days people make chips (Circuit board in which they can play the music in a loop… will that work instead I can get the same sound in the signal generator ? Still I will try to make the circuit you ‘v shown me, it nothing of your business of what i am doing but you ‘v just made me interested in electronics suddenly and I just bought lots n lots of things from the market to just try this ( Purchased things over $100 in just 3-4 days) like electronic Project Kits a drill a multimeter, soldering iron a glue gun!!! – I like this .. The only thing i don’t know how to build a circuit like you do… I can just copy!!! – Thanks!

        • You can use that MP3 thingy too, if you know how to make it. That sounds more involved.

          Take a few courses in electronics and you will a guru! When you like what you do, you learn fast.

          Have fun!

      • You can try this circuit, it might have less jitter, but in general the circuits that use resistor and capacitor as timing components are more jittery. Maybe this chip has very accurate internals to help with the jitter. For more accurate waves you would need a resonator or a crystal in your circuit. But then it will make it harder to change the output frequency. A simple solution, like you mentioned, could be getting some cheap MP3 players and play back a constant frequency generated by computer for say 30 minutes in a loop. You can have a few different audio files for different frequencies and select between them.

  23. Dear Mehdi, I did it THANKS A LOT LOT LOT……, my 50% work is done now for the 100 LED Bread Board (25%) and the tap and Speaker Configuration(25%) is left will let you know HOW it goes , But the Major Issue to make a LED strobe is a success because of you Thank man Again

  24. ok I saw your reply just now …. Will my above Configuration work ? I am waiting for your reply! (I think Op-07 Might have problem)

      • your transistor looks fine. OP07 was an non-rail-to-rail OpAmp and wouldn’t work. LM393 is a good comparator, but is open collector output. It’s not a big deal. What it means is that in addition to the schematic above, you need to add a 1k Ohm resistor to the output of the second comparator on the transistor gate. It should work now if your connections are good.

  25. I am so Excited to do this I could not wait I tried to Find a LMV7219 but could not Find it so I bought OP07 instead (Comparator) and for the transistor I also tried to get 30N06 but still could not get it so got a 55N06 instead will this work… ? I am going to do an R&D till I get a reply from you! Please Reply! Thank Again for your HELP Thank you Very much!

  26. Thank you so much Mehdi….. It helps a lot I have one confusion the transistor you said I can use any >10A so the one I have is up to 60A as I can not find a 30n06 can you suggest any other one and also a comparator too as I am not a electronics person I know some of the things but I am only a Pinch of the knowledge you have and in the diagram the (Q1 NMOS) you connected the ground(Negative) with the middle PIN will this not Bypass the Transistor ? ( I don’t Know More Just wanted to clarify) as my LED gets full Bright if I do so but there is no Strobe Effect as I use a Camera to see if there is a strobe effect ( I think the normal eye can not see the strobe Effect)….

    • How about STP36NF06 for transistor and
      AP393NL-U, or LM2903NG, or any thing similar for comparator?
      My transistor has three legs, Gate (the one connected to comparator output) Source (connected to ground) and Drain (the one connected to LEDs). Some transistor have a pad which is also drain. Please search and see the datasheet of the part to make sure which pin is which. The center pin you see in the part symbol is not available on the actual part, and is not the center pin of the device. Every device may have a different pinout depending on the package, which doesn’t have anything to do with the schematic symbol. Good luck!

  27. Oh and Also the regulator is 7805 I also did not find 78L05BP its output is 5 V check on the Multi meter will this work ?

  28. Hi Mehdi, I made the circuit today I have used MOSFET 25n120 as the 30n06 was not available and the Op-AMP I used is LM741 and the capacitor of 1u is not ceramic one but a box type capacitor I checked the value on a multi meter and its 1u and for 10u I use a 10u electrolyte capacitor as that also of ceramic was not available for audio in (VIN) I gave one connection of my mobile jack out to VIN and other to the ground (Negative terminal of the battery) I use the signal genarator software as you said on my android (NOW THE PROBLEM PART) the LED lights up very low the voltage out put is 2.372V from the middle pin of the transistor the input (Output from the LM741) is 4.49V on the First Pin and the third Pin is Ground(Negative) Please help if you allow me to send email to you and if you want I can send the Picture of the Circuit I made on the bread board, PLease PLease Please HELP!

    • A couple of issues, the transistor you are using is IGBT, with a VGE threshold voltage from 3.5 to 7.5, your comparator output provides 5V max so it is possible that it is not turning your transistor on. Use an N-Channel MOSFET with a gate source threshold voltage below 3.5V that can take lots of current.
      You are using an OpAmp, the design needs a comparator, mainly because the OpAmp outputs are often not rail to rail, but comparators are always rail to rail. Like LM741, the output will always be one or two volts away from rail. So you might actually be creating a voltage from 2v to 3V only. Change it to a general purpose comparator. Try these out and make sure your connections are good and the battery power lines are adequately thick. Try to scope your signals and make sure they are similar to my drawing. It shall work!

    • Actually I understand the importance of D2 more than D1. To charge the coil, the OpAmp output has to go high in the circuit you provided so that D1 can conduct. It means that current will flow through D1 into the inductor to ground. When you turn the OpAmp off, the inductor will switch to a generator, continuing to draw current in the same way. So the polarity of the inductor will switch and the polarity of the side at D1 will switch to negative. Now D2 will be forward biased feeding the inductor while it is discharging. So D2 becomes important, but D1 purpose is unknown to me.
      I looked at the datasheet of the OpAmp, thinking that it might instead be a comparator and switches high, but it is just an OpAmp. So it is not like having a high side switch on the inductor. The OpAmp output doesn’t just open, but it will drive the inductor voltage down slower than a switch. So the undershoot would be very small to none. I doubt the purpose of both diodes in that design.

  29. Is there anyway you could go into more depth with the diagram? I have not made a circuit before and am confused about the regulator, VCC, the -12V, and also where the 5v comes. Doesn’t VCC have to be +5v for the transition to output 5v? But there is no +5v terminal on a drill battery. Also how do you get -12v? Arnt the terminals +6v, 0v and -6v? Perhaps a photo of each side of the circuit will help. Thank you so much, awesome project! 🙂

    • A valve! I basically attached the hose to toilet sink water pipes and use those valves to adjust it. The water flow plays an important role in the shape and stability of the water drops you get.

    • 😀 If you are not joking, then I should say the liquid is not actually moving up. The strobe light creates that illusion while the water is actually still falling.

      • I understand SuperUnknown confusion as there are acoustic fluid levitation videos that are buzzing right now. They are very similar to yours, but theirs are not illusions but the real deal…

        Plus I don’t want to be disrespectful but your posts are quite difficult to read… come on “visible to the eye”, “the water levitating”! Anyway I don’t want to belittle your project, it’s quite cool… just that English is not your strongest suit.

        • 🙂 English is not my strongest suit for sure, but it doesn’t mean I should keep quiet, does it? And when I open my mouth, I’m sure I’ll generate a bunch of gibberish once in a while!! It takes reader’s imagination to understand what I’m saying!

          • Your English is great. Anyone that doesn’t realise that the water isn’t actually being levitated should go do high school physics again.

Leave a Reply

Your email address will not be published.