Surplus electronic parts :
Stock and Crypto AI Prediction :

A look at a very common little atomiser module sold on eBay. This style of circuit is often found in the USB powered desktop humidifiers that use a wick and piezo disk to put a visual stream of moisture into the air.
This is a very minimalist circuit, which makes it very interesting. There's no feedback, so the microcontroller is just banging out a rough 110kHz to the 16mm piezo disk.
I really went to town on reverse engineering this module, including desoldering and measuring all the capacitors, and unwinding the little transformer to count the windings and measure the wire diameter.
To find these on eBay search for - usb humidifier module
It's often cheaper to buy a multipack, plus that gives you spares.
If you enjoy these videos you can help support the channel with a dollar for coffee, cookies and random gadgets for disassembly at:-
This also keeps the channel independent of YouTube's advertising algorithms allowing it to be a bit more dangerous and naughty.

Let's take a look at an ebay, ultrasonic, atomizer module, i bought three, it tends to be quite cheap and you buy three. It was about roughly three pounds, each or less and what you get in the packet. Is you get a wick for soaking the water up, because these are kind of designed for humidifying? You get your little 16 millimeter diameter, piezoelectric disc, for atomizing, the water and you get the control module, which has micro, usb connector and the processor for running the circuitry and a little output connector from the step-up transformer to the piezo disk, and the disk itself is Uh i'll zoom down this uh, the disc itself is a little stainless steel disc. Oh it's got a number on it.

I didn't spot the number etched on it all right, i'm going to have to take a look at that. What's it say it says: k35d5a. Okay um, but this is a disc with a tiny little holes through it and uh. It's got a ring of piezoelectric material in the back with another conductive ring and one connection goes onto the stainless steel.

The other connection goes onto the conductive ring and when a sort of highest voltage at high frequency is applied about, it's essentially listing 110 kilohertz, then it causes this uh piezoelectric material to distort at very high frequency and that smashes the water and causes it to be Atomized through the holes, it's very odd distinctive pattern in those holes. I wonder if it's etched there's a lot of holes in that strange. I know they do that laser or etching here's one. I've set up on a butchered, earwick mist to show your operating.

So i'm going to turn it on and uh you'll see the led lights and then the mist comes out in force. It draws about 150 milliamps per way. This is spraying everywhere and if you press the button again, it goes into burst mode where it pauses momentarily. It goes for about 12 pulses and then it stops and then starts again after about four or five, so roughly uh sort of like 75 on 25 percent off mode.

I'm not really sure advantage of that. Presumably it times out after a while. I'm not really sure. I haven't tested that i'm going to turn this off, because this is an aroma bottle and it's absolutely stinking here now so you're, probably wanting to see oscilloscope trace, as i did so.

I've taken the oscilloscope trace from the little microcontroller. I've got a couple of connections here: let's take a look at the oscilloscope trace and for those wanting the oscilloscope waveform here it is so i'm just going to lock that a bit jittery, but it shows this is the output from the chip. The haze wafting across is the aroma. It's stinking uh.

It is showing 4.5 sections here at 2 microseconds, which is 9 microseconds 1 divided by 9. Microseconds is approximately 111 kilohertz, so it's very close to what they actually state in the listing. Let's take a look at the pcb here is the usb connector coming on. There are three capacitors in the vicinity of that two 10 micro farad and one 100 nanofarad all in parallel.

There's also the option for electrolytic which they've not used, presumably just because they've got a couple of 10 microfarad ones and it provides a decent level of capacitance. There is an led with its own little resistor here, 510 ohms and there's a push button signaling straight over to the tiny little six pin microcontroller of which five pins are used. There is a 100 nanofarads capacitor in series with the gate of the mosfet here, which is an x01v to give it its full number. It's an ap3400mi and there's a 100k pulldown resistor in that, then that drives the primary 28 turns in a transformer.

165 turns in the secondary of the sim wire and that then drives the piezoelectric disk. So if you want to marvel at this side of the circuit board there, it is, if you wish, to take a look at the back of the circuit board there. It is notable huge sort of ground plane with lots and lots of pleated through holes for current handling right schematic time. Let me bring on the schematic here is the schematic here's a usb coming.

We've got their 210 mega fire capacitors on the 100 nano farad capacitor in the vicinity of each other. It's a very, very bare circuit board. We have the led and its 510 ohm resistor. We have the measly little six pin microcontroller anonymous, nothing marked on it and a push button going to the zero vote, rail just to actually activate the modes initially, when you turn it on it's off press it once, and it goes continuously press the second time it Goes into that sort of 75 on 25 off type mode.

Here is the slightly unusual approach of driving the mosfet. The mosfet is a pull-down resistor, quite high value, one there it's got a hundred nanofarad capacitor in a. I guess that this might be in case the processor crashes, so if this suddenly locks up high instead of causing problems by just powering this all the time, it will only power it until that capacitor is fully charged up and then because this isn't changing state, it Will just be uh, it will just stop driving the mosfet and it will naturally pull down to the zero volt rail and turn off the little inductor, which i i unwound this little inductor here. The wire is 0.17 millimeter diameter and it starts off at one.

Pin. Has 28 turns goes to the mosfet and then continues on with the 165 turns to step that voltage up to the piezo so, technically speaking, that's actually quite a decent step-up ratio. What's uh, let me try and calculate this 165 turns if it was 100 efficient, divided by the 28 turns equals a step up ratio about, say 5.89 times, roughly 5 volts uh goes up to about 30 volts theoretically in ideal circumstances, but i'm not really sure what It would actually be in real life. I didn't want to scope that side in case it turned out to be quite spiky and peaky and wrecked my little scope, but that is it um.

That's all there is to it. So it's quite a neat little circuit. I really decimated this this time to get these values. I de-soldered all the capacitors and i de-soldered the um connector, so i could see the traces under it and the little um transformer here.

Just purely so, i could unwind it turn by turn just to see how it was wound, and it literally is it's just one winding starting the positive doing. Those 28 turns going on to the middle pin and then continuing the 165 turns. So it means they just turn it with a single wire, but that just gives effectively an isolated uh secondary with that sort of higher ratio for the higher voltage. So there we go.

That's what's inside those little piezo electric modules, they work quite well, but it's notable that this is what i thought the uh. I thought that the earwick missed unit was going to be super evolved and this was going to be the circuitry unit and it turned out to be quite complicated. This is about the most minimalist minimalist approach and they're, relying on the accuracy of the microcontroller's clock. Without a crystal or anything to drive this a very specific frequency, it's worth mentioning that you know if this was running at say one megahertz, or so it's quite easy just to an internal clock frequency, it's quite easy, just to divide it down to create.

That's a very simple crude square wave at that frequency plus still potentially allow another clock to be another counter to be clocked just to provide timing functions or that cycling on and off uh. Very basic, very simple. But it works quite an interesting little circuit and it was quite fun to reverse engineer.

14 thoughts on “Ebay atomiser module test with schematic”
  1. Avataaar/Circle Created with python_avatars Levi Campos says:

    It would be interesting to know how to make a custom piezoelectric disc or plate of a determined size? What would the fundamental requirements to fulfill and how can one be made that meets local requirements.

  2. Avataaar/Circle Created with python_avatars MrShwaggins says:

    I have one of those cool vapor humidifiers and it was on but the little atomizer disc that makes the vapor had a collection of bubbles on it. I stuck my finger on it to agitate the bubble off and boy thats a sensation I don't think I want to feel again. I wonder just how dangerous that was? Could it cause fluid in my finger to make captivate bubbles? It sort of felt like a shock you would get from those electric muscle stimulators. Sometimes I can still feel the tingles from it but I have a feeling its just my mind remembering the sensation as I'm writing this.

  3. Avataaar/Circle Created with python_avatars Bunny Love says:

    hey Clive l have a board that does a similar thing (its for my Titanic model) but the atomizer bit does not work? any ideas how to make it work

  4. Avataaar/Circle Created with python_avatars Mark D. says:

    "Atomiser" is such a fun word – but "Discombobulator" or perhaps "Illudium Q-36 Space Modulator" would have so much more sparkle and pizzazz.

  5. Avataaar/Circle Created with python_avatars James Barclay says:

    That thing looks like it would feed the MOSFET gate +/- 2.5V. Maybe the peaks are higher if it uses a narrower duty-cycle. I’m surprised it’s not a resonant circuit, but I guess the components are constant enough that it was found experimentally to be best at 110kHz so they stuck with that.

  6. Avataaar/Circle Created with python_avatars Richard Halliday says:

    I have used these for a couple of years now they work quite well, unfortunately they tend to self destruct necessitating in a repair or replacement ( usually the piezo) as I run them with a few drops of Zaflora added to humidify the air with a fragrance but the liquid tends to contaminate and bung up the piezo. Cheap enough to replace though.

  7. Avataaar/Circle Created with python_avatars MrBun9l3 says:

    Interesting device. It's a tapped inductor as opposed to a transformer. The output is probably much greater than 30V but will depend on the load.

  8. Avataaar/Circle Created with python_avatars 何馬士 says:

    My atomiser failed to work recently due to controller malfunction. The signal pin to switch the FET didn't work. I used arduino nano to generate the signal instead.

  9. Avataaar/Circle Created with python_avatars Lion McLionhead says:

    The lion kingdom made an air conditioner out of 3 of those, automating the button presses. It ended up creating more body odor than cooling during the day, but during the night, it could cool the room down more than it would naturally.

  10. Avataaar/Circle Created with python_avatars Alberto Gregory says:

    Hello Clive, literally working on a project involving these rn. If you don't mind, how would I wire a few of their switches together to make all of them go off at once?

  11. Avataaar/Circle Created with python_avatars Quick Fix says:

    A while back I wondered if it would be possible to atomize fuel, like Diesel or kerosine, using these modules to make it flamable like a whoofler. Never tried it though.

  12. Avataaar/Circle Created with python_avatars Dan Coulson says:

    Is there an issue is these ever run dry?
    Considering it's just a piezo just making a high frequency, I would guess not?

  13. Avataaar/Circle Created with python_avatars Silverfoxwolfen says:

    Funnily enough I've ordered this exact same one, the one sold on the Canadian site worked out to £2.54 and 71p postage. Not a bad deal all in all. Very nice to see the video and all the details.

  14. Avataaar/Circle Created with python_avatars zh84 says:

    In my 1970s childhood if you had a model railway and you wanted the engines to produce scale "steam" there was an unsatisfactory system for having them emit electrically produced smoke instead by using a heating element in oil. You can see it in early episodes of the Thomas the Tank Engine TV series. These clouds of microdroplets look a lot more like steam, and would be 100% safe, non-toxic and easy to refill. I wonder if anyone has fitted them to a model engine?

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.