how to pilot a piezo transducer?

[glias]

Hello,
I would want to control a piezo transducer. I use a mosfet with a PWM from a µC but my supply voltage is 5V instead of 3Vpp speficied by the datasheet of the transducer.

https://www.farnell.com/datasheets/97521.pdf

I read that it's not recommended to use a serial resistor to reduce the voltage (or a zener), can I use it simply with a nMOS and direct connection between the drain and the 5V supply ? can I consider that the transducer is simply a resistor ? (like specified in the datasheet, 3V - 3mA => 1kohms impedance ?)
Is there any recommendation that I should consider ?

regards

 

[keith1200rs]

At the top of page 2 it shows the maximum voltage as 25V so it will be fine with 5V. At 3V it will take 3mA maximum.

Keith

 

[FvM]

can I consider that the transducer is simply a resistor ?

No, the impedance is capacitive (14 nF outside the resonance frequency). The transducer need to be driven push-pull, so a single MOSFET won't work as a driver.

 

[glias]

Hello FvM and Keith,
Thanks a lot for your replies.
I'm not familiar with tranducer, you said that the impedance is capacitive and the value is 14nF outside the resonance frequncy; what does it mean ? how can I calculate the resistor to limit the current with such load ? if I consider that it is capacitive, it will be vary with the frequency of my pwm command ? well, it's not very clear for me, if you could give me more exaplnations.

regards

**broken link removed**

VCC=5V and input of MOSFET too.

 

[kak111]

Driving transducer ( impedance of normal piezo element is high when out of resonance)

Frequency should be 4kHz. ( Element resonance frequency see data sheet )

Check is there any polarity marking in piezo. ( I did not see any in data sheet)

R4 1k R3 10k Q1 2N2222 or equiv. Usupply can be 5V.
you need only these , connect 4kHz pulse output to R3

Regards KAK

 

[minhlam]

Sorry my English not well so I don't know to much your question .But if you want make more volume ( big sound) you can use it with a transformer.(2n2222 or 2n3904 transistor + small transformer.)You can ask me more.

 

[glias]

I'm not sure to understand better... The transducer couldn't oscillate with another frequency ? only 4kHz is possible ? in fact I don't know how the piezo works ...

 

[kak111]

Look at data sheet there is graph: Typical frequency curve.

You can see that sound pressure max. points are 3.5 kHz and 5.2 kHz
there is a drop in 4.8 kHz.

So if you want maximum sound pressure the frequency should be 3.3.....4.5 kHz

Anyway you can use this component in other frequencies too ,
but effiency is poor.

In frequency-area 0.3 ... 20kHz you can safetly use maximum 12 Vpp signal supply for
this piezo transducer.

KAK

 

[FvM]

In my opinion, you should start with a single ended driver (e.g. 1/2 of your circuit from post #4) and try different frequencies. A load resistor of 800 ohm or 1k sounds like a good starting point.

 

[glias]

Hello all, thanks a lot for your replies,
OK, so I don't need a push pull and try to do it with only one MOSFET stage.
I will feedback you the results after my PCB will be finished.
Thanks

 

[glias]

Hello all, I just have received my board and tested the piezo and... it doesn't work well : I hear the frequency @ a very low level , my stage doesn't seem to be the good way to pilot it or I made a mistake in polarization of the MOS.
I have tested it with a square waveform (3V peak to peak with a 1k resistor) directly connected to a generator, it works fine, I can hear it clearly.
So my transistor stage seem to be the wrong.

See my schematic (I tried to simulate my stage with pspice but I'm not sure that the piezo could be modeled by a single cap)
The current in the load does't seem to be enough, I suppose that the current is dependent of the switch time of the MOS and it should be controlled by the RC filter is that right ?

If someone could help me
thanks a lot
**broken link removed**

 

[FvM]

I tried to simulate my stage with pspice but I'm not sure that the piezo could be modeled by a single cap

The model is good enough to explain, why you can't get any output from the piezo! Because it blocks the current through the transistor. As a simple modification to make it work: Connect th piezo in parallel to the resistor instead in series. I'm however not sure, if the 3V specification from the datasheet has to be kept strictly. If so, the transducer would be overloaded.

P.S.: I wonder why you are reducing Vgs to 3V by a voltage divider?

 

[glias]

thanks for your reply, yes I saw it that the DC current can't pass, but I thought that current pics produced by the commutation by the Mosfet could be enough to provide a sound...
for the MOS, the 2N7002 doesn't need more than 3V... and it permits to improve noise immunity by reducing input impedance (another traces of the PCB are very near the command line of the MOS)

 

[FvM]

for the MOS, the 2N7002 doesn't need more than 3V...

All characteristic curves are given for typical Vth. With wost case (maximum) Vth, you won't get much Id with Vgs of 3V. It's possibly sufficient for the piezo, but raises doubts, why you are using a MOSFET at all.

but I thought that current pics produced by the commutation by the Mosfet could be enough to provide a sound...

Vds drops to zero and stays there. A small AC voltage amount is coupled to the output by Cgd, however. If it's sufficient, why you are using a MOSFET?

 

[glias]

because I want to isolate IO of the µC... I only need 3mA ! the mos function perfectly with this Vgs voltage.

 

[FvM]

because I want to isolate IO of the µC... I only need 3mA !

Yes I know. I don't opt against MOSFET drivers generally.

But it seems like a single series resistor between I/O port and piezo would achieve the same in this case.