Ultasonic humidifier Tranducer model
I followed through on measuring an ultrasonic humidifier (fog machine) resonator disc.
If you have little or no back ground in crystal resonators, most of what follows will be jiberish. Please note that I do not consider buying a crystal slapping it on a IC oscillator and adding 2 caps expereince, no offense. I highly recomend reading
https://www.electro-tech-online.com/custompdfs/2009/12/AN-CU_GG_e.pdf for a fundimental understanding of crystal resonators.
I measured the disc as a series 2-port on a 50 Ohm network analyzer and got a suprizingly strong resonance without adding shunt capacitance to each port of the resonator. This is only suprizing becuase I have never measured this type of resonator before and have more experience with crystal and SAW resonators.
I have the small disc from
Ultrasonic Water Fogger-The Mist Maker
The disc has many other smaller lower Q resonances, which I ignored, so the model I generate will only be applicable for the primary resonance.
The disc is series resonant at 1.612839 MHz and -0.4 to -0.7 dB
The disc is parallel resonant at 1.83593 and about -20 dB
The 1 dB BW of the series resonance is flow=1.5211MHz, Fhigh=1.679138MHz
I measure the Co of the disc at 10MHz to be about 1.0nF (another surprize, huge!)
From this data I calculate the circuit model using NDK's equations. Please note that NDK uses the component names: L1, C1, R1, Co. I strongly prefer Lm, Cm, Rm and Co, as there should be no confusion on which capacitor is which!
For the small Disc equasions and some minor manual optimazation, I get:
Lm=35.196uH
Cm=276.67pF
Rm=4 Ohms
Co=1.0nF
Note that the Lm and Cm have high precision due to the accuracy of the frequency measurments, don't truncate! Rm was optomized from the series attenuation measurment using PSpice and stray shunt capacitance of 0.5pF on each end of the resonator. Try simulating this with PSpice using a 50 source and load resistance.
I then optomixed the resonator Circuit for the peak amplitude response and got 6.3nF shunt at each disc terminal (Cshunt, NOT Cload!, Cload is approximately Cshunt/2) +3.8dB. The positive amplitude response for a pasive device is probably due to inaccuracy in the broad band calibration I used on the network analyzer and the Q oc the resonator, but is still useful for the peak!
The peak Q is of no practicle interest to the circuit designer, it happens at infinite Cshunt meaning a termination impedance of very close to zero ohms. It is impossible to drive a voltage into a short circuit it always looks like very small voltage and is useless! Around the peak amplitude response of the resonator and circuit is the only practical operating point especialy for a power Amp design!
More updates when I reverse engineer the Circuit picture. Right now my circuit topology guesses are looking pritty good!
Regards,
Pete White