Electrical – Blocking DC when driving a buzzer with a PWM signal

Looks as if it's not intended as an amplifier, but a maker of loud noises (square waves only) such as for alarm signals.

I suppose you could drive it from a PWM output and it would sort-of work; the only place you could put any filtering would be in series with the speaker itself.

Because the other three transistors are driven by Q2, they cannot switch until after Q2 has switched, and that asymmetrical delay may do bad things at PWM switching frequencies. A little thought will show that Q1 cannot turn off until AFTER Q2 has turned on, for example; creating a momentary short circuit across the PSU...

So my verdict would be: it's fine for alarms, but needs more analysis or testing before serious use in PWM applications.

They're passive, you have to create the signal yourself. "Without circuit" is the most obvious way to tell, but there are also the frequency response curves. If the device would have an internal oscillator the frequency would be fixed, and the graph be unnecessary.

Connecting the piezo directly to an I/O pin is the best way to get a loud sound, but piezo elements are highly capacitive, and microcontrollers don't like capacitive loads. A series resistor will protect the I/O pin, but also reduce the loudness. The best thing is to use a transistor to drive the buzzer.

For the loudest sound, look in the datasheet at the frequency response curve, and pick the frequency where it peaks the highest.