Turns out you’re right, @o_lampe - I’ve enabled (a very slow) sinusoidal PWM, and I get much higher gate voltages (~17V or so) and full 10V across the phases. So that explains one mystery! Thanks a lot, now to the opamp…
I used to watercool load resistors, when they weren’t up to the amps. Used to use demineralized water in a ceramic bowl and longer wires to keep it away from electronics.