I have this output driving stage, reverse engineered from a generic Chinese mist generator:


The goal is to drive the piezo transducer with a given frequency (more or less 1.7MHz or 2.4MHz depending on the transducer) to put it in resonance. It will then do its work and turn water in vapor (more or less, details are not important).

In my understanding it works in this way: from the left come the gate drive wave (at 1.7MHz or whatever, square wave), the MOSFET will charge the coil like a storage choke in a flyback converter and on gate release the inductor will dump the stored energy on the load, down right.

The upper right components are a standard RCD snubber to keep the coil in reset and quench transients, the two resistors are current shunts to help finding resonance (the frequency varies slightly with load and environment).

The 'load' is the thing which baffles me: the transducer in resonance is equivalent to about 1.7nF (body parasitic) in parallel with a couple of ohm (the 'transducing' resistance). Why the two capacitors? Couldn't they simply dump the power right on the transducer or is that some kind of impedance matching network?

At first it could seem that the 22n is here to resonate and cancel with the storage choke (unknown inductance, for now) but in a pulse transfer situation like this what is the 'source impedance' of the discharging choke?

EDIT: in theory, at least when doing power conversion, the storage inductor is modeled as a current source with the ESR in series, however a) current sources have the internal impedance in parallel and b) does matching work with current sources too?

As for the 220p I think it could be a protection to avoid overstressing in case the transducer gets accidentally detached or maybe some help in shaping the pulse (to let pass thru all the higher harmonics in the square wave).

  • \$\begingroup\$ Well, I added a schematic that was far more readable but, you didn't like that change and you reversed it back to what I see as an unreadable schematic. You can take a horse to water etc.. I also fixed your spelling so, if you are happy with your spelling mistakes then so be it. Did I mention a horse? \$\endgroup\$
    – Andy aka
    Mar 12, 2021 at 9:51
  • \$\begingroup\$ uh? didn't see a thing, I only added one note in the middle… maybe simultaneous edit issues??? \$\endgroup\$ Mar 12, 2021 at 9:54
  • \$\begingroup\$ That's probably it. What an awkward situation. \$\endgroup\$
    – Unknown123
    Mar 12, 2021 at 10:02
  • \$\begingroup\$ Yup. At least S.E. keeps the revisions in its belly. Anyway, back to work. Measured the choke, it's about 22µH (XL=234 at 1.7MHz). The whole thingamagick on the botton amounts at about 1.75nF (XC=-53) so unless I messed up it isn't done for resonance or matching. It could be to lower the voltage on the transduces (as a capacitive divider) but it's actually working lower than the nominal voltage. I'm starting to doubt that the people reused values calculated for another kind of transducer \$\endgroup\$ Mar 12, 2021 at 11:31
  • \$\begingroup\$ Would the 22nF be for DC blocking, so the piezo doesn't see 24V DC across it? \$\endgroup\$
    – Simon B
    Mar 12, 2021 at 11:59


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.