I'm interested in developing a PID controlled proportional valve system for fairly precise control of water pressure and/or flow rate, probably based around an arduino and a 24vdc water valve (plus pwm ic, power supplies, transistors, diodes, etc.)
The thing I can't figure out is how to determine the ideal pwm frequency for a particular valve. The existing commercial solutions include valves that list preferred pwm frequencies, but those valves are crazy expensive and I'd like to experiment with cheaper components, at least as a starting point while I'm also grappling with sensors, feedback, programming, etc.
Are there good ways to test a valve and determine how best to drive it with pwm? Is it directly related to inductance? Could I use something like the ring tester discussed recently here, maybe measuring the frequency of the rings and deriving something from that?
I've tried a lot of Google searches and all I can find are really expensive valves that provide the required spec, or hobbyists like me who seem to all just guess and experiment through trial and error. I'd rather not go with trial and error. Seems like this should be testable.
Any ideas?
The thing I can't figure out is how to determine the ideal pwm frequency for a particular valve. The existing commercial solutions include valves that list preferred pwm frequencies, but those valves are crazy expensive and I'd like to experiment with cheaper components, at least as a starting point while I'm also grappling with sensors, feedback, programming, etc.
Are there good ways to test a valve and determine how best to drive it with pwm? Is it directly related to inductance? Could I use something like the ring tester discussed recently here, maybe measuring the frequency of the rings and deriving something from that?
I've tried a lot of Google searches and all I can find are really expensive valves that provide the required spec, or hobbyists like me who seem to all just guess and experiment through trial and error. I'd rather not go with trial and error. Seems like this should be testable.
Any ideas?