I am making a water feature that will involve a stream fed by the basin of a fountain with a pump at the end sending water back into the fountain. To control the water flow rate through the stream and prevent flooding, I intend to use this valve controlled by an Arduino connected to a water level sensor. The nice thing about motorized ball valves is that they have very little head loss and can be left in a partially open position to throttle fluid flow.

For safety, I would like the valve to be able to fully shut from fully open in the event of a power loss. US Solid makes a variant of this valve with exactly this capability, unfortunately it does not have the ability to remain half open. Their documentation says the auto-return valves store energy in a capacitor that is used to shut the valve when power is cut off, but they don't give specifics on the size. I'd like to do my own custom set-up with the 3-wire version and my own capacitor, and I'm figuring out what size I need.

Their specs say it can operate in the 9 to 24 volt range, takes 3-5 seconds to close, and can draw up to 2 watts. Taking the worst case, 5 seconds at 2 watts, I need to be able to deliver 10 joules of energy between 9 and 24 volts. Taking stored energy as 1/2CV^2, I calculate that if I'm going from 24 volts to 9 volts, I need 1.3F of capacitance, or if going from 12V to 9V, I need 6.7F. A supercapacitor might work for this, but the closest I have been able to find with anything close to these specs that is not a complete beast is something like this one from Kemet. That one sells for $62, and given that the auto-return valve itself is only $45, I doubt that's what they're using.

Can anyone give me some guidance on selecting the right capacitor for this valve? Alternatively, I suppose I could use a rechargeable 9V battery as the backup power source, but then I would need to build a battery charger into my circuit, and that adds a whole other layer of complication as well.

  • \$\begingroup\$ Have you considered a second on-off ball valve with a spring return? It would act as a fail-safe and should be cheap. \$\endgroup\$
    – Transistor
    Mar 30, 2020 at 20:06
  • \$\begingroup\$ Thank you @Transistor, that's actually a really good suggestion. I'm worried about adding extra head loss with two valves thought, as this is a gravity feed system. I can definitely do that if I can't figure out the capacitor though. \$\endgroup\$ Mar 30, 2020 at 20:21

1 Answer 1


I'm not exactly sure what I was doing wrong with my math before, but my calculations were way off. On re-calculating, assuming the worst case of drawing 2 watts for 5 seconds, the valve would need about 300 mF. Having actually gotten the valve and measured the current draw through it, it was actually drawing 30-40mA at 13V for about 5 seconds to close, and could continue to operate down to about 5V. I selected a 15V power supply, and assuming it could go down to 9V and would draw 40mA at 13V for 5 seconds, I calculated needing a 36 mF capacitor.

To be on the safe side, I ordered a 68,000 μF 16V electrolytic capacitor. This, when charged to 12.5V, could take the valve from fully open to fully closed, back to fully open again, and about 90% of the way to fully closed before quitting.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.