Would it be feasible to put some TIMER ONE code in your software to create a 10 bit PWM signal? Then you'd have 1024 steps of resolution instead of 255?
Hmmmm... that's a good idea! I'm not sure how it would affect my code, because I'm not a super wiz at the Arduino. It'll take some research, or somebody who knows better to tell me. I know the IR library can run on timer one if I move the IR LED to pin 11. Other than that I'd have to dig in.
That many steps would take the 5v model to the 2.6 mL/min resolution range, which I don't think is good enough (at least for smaller tanks). However, if you could under drive the 20v model with 5v, and effectively only use the bottom 1/4 of it's range, then you'd have 0.66 mL/min resolution, which should work.
I know you can drive it with PWM because the data sheet says "The valve may be controlled using DC current, open- or closed-loop control, and even PWM (pulse width modulation) to cover a broad range of applications." It's just a question of under driving the 20v model with 5v.
The other thing that I just thought of is that flow rate is dependent on working pressure. If we used the 5v model (ET-P-05-0950) but only supplied it with 25 PSI working pressure instead of it's 100 PSI max rating, would that knock it down to 1/4 the flow? That would accomplish the same thing, and combined with 1024 steps, you'd have .66 mL/min resolution.
I've submitted all of my questions to Clippard. We'll see what they say.