h-bridge proportional valve
Hello phaedrus,
1. I actually imagined a partly different circuit, but the suggested circuit is profound and I didn't see a reason to contradict. A voltage divider with a resistor in parallel to the 2N2222 (Q4) allows to adjust the gate voltage without a zener diode. The dynamic losses at Q1 are somewhat increased. Both variants are possible, I think. If the supply voltage is varying, the zener diode will achieve stable conditions.
The R4 series resistor is a good thing, generally.
2. There are complementary transistors in all discussed circuits, I think
3. Basically yes.
Your questions aren't off topic, but go beyond the original question I think. Some consideration aren't necessary for the said application, but meaningful when discussing more critical (higher power and frequency) applications.
The Zetex application mentions a possible reduced turn-on speed. I think, for general applications, a symmetrical gate waveform is most common, e. g. with SMPS circuits. A special reason to reduce the turn-on speed is to limit the diode reverse recovery current, which is possibly the strongest EMI source and may seriously reduce the breakdown voltage margin of the involved devices.
The turn-on series resistor can be increased by trial, until gentle current and voltage waveforms are achieved.
For the rather low PWM frequencies of a proportional valve driver, even the push-pull gate driver could be probably omitted.