That´s a very old topic but I´ll reply anyhow, just so that the info exists.
Current is not really a problem to use the L298, it´s much unneeded.
So you simply use a in-series RLC circuit, to get some serious voltage kicking in.
Also, it would be very good to calculate the Q factor, cause you don't want to drive them with huge voltages.
And when I say huge, I mean it... An in-series RLC can reach as much as 60V across the cap and the coil, even if you use an H-bridge at 5V...
you mean something like attached image? how to calculate R?
From what I remember I used a series circuit. It´s must be more than a year that I started the project and
to tell the truth never ended it, cause I needed to find better transducers and some other things also.
You put the LC in parallel, that creates a band cut filter, you want a band pass filter. The very reason you are using
a coil is to avoid having greater voltage rails in the circuit. The LC circuit itself can create large voltages, it depends
on the quality factor of the oscillation. Since you have fixed capacitance, you also are limited by the inductance selection.
Still the quality factor should be good enough to produce voltages up to 40 - 50V. Of course that would be catastrophic
for the piezoelectric element. So you place a resistor to lower the quality factor of the oscillation, to a voltage around
20V which can be handled by the transducer. I can´t remember the exact formulas to calculate this, but a Google search
should do it, so no worries. Basically what you do is to dump energy into the resistor. But that also means that you´ll need
more time to reach full oscillation amplitude. It a trade off. Of course you could temper a little with the inductor value, but
since of the nature of the inductors having about 10% - 20% tolerances, ehhh, I don´t like the solution. Of course, it
would be almost impossible to find an inductor at the exact value you want, so you may start thinking making your own.
There are coil calculators in the internet, free and good onces also, so no worries. You just need to find the core material,
ferrite preferably. All the other usually used materials are not good to these frequencies. (I can´t remember for iron powder but
iron cores are really NO go). Be careful of that.
Also if you are going to use the LC circuit you don´t really need an H-Bridge. You can use one but you can do it with only a
typical push-pull circuit with the other end of the circuit tied to ground. That saves you from other troubles also. There are
many things to consider generally...
But do you want a dirt cheap solution that works and doesn´t need a higher power rail, just 5V
MAX232.... problem solved. It has a charge pump, it can drive the load with no problem, it´s not the best of the best
but it works, and much much easier than the coil solution. And it´s much easier to do other things with the transducers,
with little effort like FM and stuff like that... It´s a dirt cheap solution, I know, but it works great.