Ultrasonic transducer driver circuit reverse-engineering

[Axk]

A Chinese "beauty skin therapy" thing https://www.ebay.com/itm/Ultrasound-Facial-Body-Skin-Massager-Pain-Therapy-Ultrasonic-Machine-beauty-Age/142544969567?hash=item2130579f5f:g:0vYAAOSwd0BVyGue

The inside looks line this

The driver board with the mains transformer:



The driver board closeup (top):



The driver board closeup (bottom):



The uC board (bottom):



My schematic of the main driver part of the driver board (the power supply part and the relay control part are not included, the transducer part is a Butterworth-Van Dyke equivalent circuit):



Oscilloscope captures of the operation:

CH1 - base of Q1 (input from the uC board)
CH2 - collector of Q3 (output to the transducer)
CH3 - base of Q3


The stated resonant frequency of the transducer is 1.1Mhz (~1.05 MHz measured)

The output is PWMed to control output power:



A closeup of the measurement (CH1 is about 3V max):



My questions about the circuit at the moment are:

1. How the "voltage doubling" on CH2 is achieved from the input 39V to the output 92V?
2. What is the purpose of Q2 and the circuitry around it?
3. Why the base of Q1 is out of phase with the base of Q3?
4. What is the purpose of L2 and C5? (L2 is the big vertical heat-shrunk inductor  on the top right on the photo, reading 68uH on the board)

There's probably an explanation of a similar circuit somewhere (a patent or something), but I've not found it yet.

[Paul Price]

I don't see a capacitor "right under" the 68-uH inductor, only a large electroytic cap.

Take more care to clearly label every part on the PCB you wish to have explanation for.

You have not labeled the components on the top of the PCB so one could follow your circuit description.

Q1 inverts the signal to the base of Q3, Q1 is a common-emitter inverting amplifier/gate/switch/trigger.

Your schematic is not drawn correctly showing the base and collector of Q2 circuitry? You should find at least one resistor to the +V  on the collector and/or Q2's base.

FWICS, Q2 is connected as  cascode (grounded-base) to enable level shifting to drive/trigger/gate  the HV Q3 osc. circuitry.

The signal into Q1 is level shifted by Q2 to trigger an oscillation for each input pulse/sinewave on it's base to to drive Q3.

The waveform height on the collector output is apparently doubled by resonant effect and the phase shift/oscillation in this circuit inverts the expected waveform polarity.

[Axk]

Thanks!

I've updated the schematic and the question, the capacitor is C5.

As to Q2, I've re-checked on the board and it's not connected to V+

Will look up and read on grounded-base cascode.

[Paul Price]

You're right, the design does work without any additional resistor on the base or emitter of Q2. The real function of Q2 is to create a driver signal, to pump the resonant circuit created by C2/L1 that is the drive signal to the base of Q3. The idea is to keep C2/L1 oscillating for several cycles and also present a bias that could start/gate oscillation with Q3 but at the same time would not be so high as to bias Q3 to a point that would cause clipping/waveform distortion of the transducer driving waveform. Clever circuit to create an ultrasonic "chirp".

L2 and C5 form a resonant "tank" circuit to set the oscillation at the transducer's optimum operating resonant frequency.

[Axk]

Now I think I'm starting to get it how the whole thing works

1. Can one say that Q2 in this configuration is impedance matching from a low input impedance at the emitter side to a high output impedance at the collector side?

2. So there are no oscillations coming in from the uC board to the base of Q1? The uC board just turns on Q1 and then it's just positive feedback from the collector of Q3 to the base of Q3 and the resonant frequency of the feedback path?

[Paul Price]

The driving signal into Q1 must be at the desired transducer frequency.

The drive circuit is giving a  few cycles of resonant "kick" to a damped "flywheel" and the flywheel is the resonant driver circuit using Q3.

The flywheel is damped because energy is lost in the form of ultrasonic power delivered to the air/nails. C2/L1 form a tank that oscillates for a few cycles to create the chirp. For this to work, the signal into the base of Q1 as on/off pulses must match/sync at the resonant frequency desired.

Q2 operating in a grounded-base configuration which creates a high-impedance at the collector to allow C2/L1 to form an oscillating resonant circuit that is kicked into oscillation by Q2 but doesn't load it once it begins to ring and this signal drives Q3 to create a HV drive for the resonant transducer to create powerful ultrasound chirps.

[Axk]

Measured the base of Q1 without the transducer connected when there's no oscillation and it appears to be just on for the whole length of the oscillation portion, not on/off at the frequency of the oscillation.

Looked up the values of the components and created a spice model, now trying to figure what is a proper equivalent circuit for the transducer to make the model oscillate as the real thing.

Attached is the LT spice model.

[Axk]

Found suitable values for the transducer's equivalent circuit. Now it resonates and the output looks similar to the real thing.
But the input (Q1s base) is the ideal square signal, trying to figure out how to model the circuitry that drives the input at Q1s base so that the simulation is similar to the real distorted signal affected by the output.

Now thinking if measured it correctly, need to re-measure.

Attached is the updated LTSpice model.